1 \input texinfo @c -*-texinfo-*-
2 @comment %**start of header
3 @setfilename ../../info/eintr
4 @c setfilename emacs-lisp-intro.info
5 @c sethtmlfilename emacs-lisp-intro.html
6 @settitle Programming in Emacs Lisp
12 @c <<<< For hard copy printing, this file is now
13 @c set for smallbook, which works for all sizes
14 @c of paper, and with PostScript figures >>>>
20 @set print-postscript-figures
22 @c clear print-postscript-figures
25 @comment %**end of header
27 @c per rms and peterb, use 10pt fonts for the main text, mostly to
28 @c save on paper cost.
29 @c Do this inside @tex for now, so current makeinfo does not complain.
35 \global\hbadness=6666 % don't worry about not-too-underfull boxes
38 @set edition-number 3.10
39 @set update-date 28 October 2009
42 ## Summary of shell commands to create various output formats:
44 pushd /usr/local/src/emacs/lispintro/
48 makeinfo --paragraph-indent=0 --verbose emacs-lisp-intro.texi
50 ## ;; (progn (when (bufferp (get-buffer "*info*")) (kill-buffer "*info*")) (info "/usr/local/src/emacs/info/eintr"))
53 texi2dvi emacs-lisp-intro.texi
55 ## xdvi -margins 24pt -topmargin 4pt -offsets 24pt -geometry 760x1140 -s 5 -useTeXpages -mousemode 1 emacs-lisp-intro.dvi &
58 makeinfo --html --no-split --verbose emacs-lisp-intro.texi
60 ## galeon emacs-lisp-intro.html
63 makeinfo --fill-column=70 --no-split --paragraph-indent=0 \
64 --verbose --no-headers --output=emacs-lisp-intro.txt emacs-lisp-intro.texi
70 mtusb # mount -v -t ext3 /dev/sda /mnt
71 cp -v /u/intro/emacs-lisp-intro.texi /mnt/backup/intro/emacs-lisp-intro.texi
72 umtusb # umount -v /mnt
75 ## Other shell commands
77 pushd /usr/local/src/emacs/lispintro/
81 texi2dvi --pdf emacs-lisp-intro.texi
82 # xpdf emacs-lisp-intro.pdf &
84 ## DocBook -- note file extension
85 makeinfo --docbook --no-split --paragraph-indent=0 \
86 --verbose --output=emacs-lisp-intro.docbook emacs-lisp-intro.texi
88 ## XML with a Texinfo DTD -- note file extension
89 makeinfo --xml --no-split --paragraph-indent=0 \
90 --verbose --output=emacs-lisp-intro.texinfoxml emacs-lisp-intro.texi
92 ## PostScript (needs DVI)
93 # gv emacs-lisp-intro.ps &
94 # Create DVI if we lack it
95 # texi2dvi emacs-lisp-intro.texi
96 dvips emacs-lisp-intro.dvi -o emacs-lisp-intro.ps
99 # Use OpenOffice to view RTF
100 # Create HTML if we lack it
101 # makeinfo --no-split --html emacs-lisp-intro.texi
102 /usr/local/src/html2rtf.pl emacs-lisp-intro.html
105 /usr/bin/rtf2latex emacs-lisp-intro.rtf
111 @c ================ Included Figures ================
113 @c Set print-postscript-figures if you print PostScript figures.
114 @c If you clear this, the ten figures will be printed as ASCII diagrams.
115 @c (This is not relevant to Info, since Info only handles ASCII.)
116 @c Your site may require editing changes to print PostScript; in this
117 @c case, search for `print-postscript-figures' and make appropriate changes.
119 @c ================ How to Create an Info file ================
121 @c If you have `makeinfo' installed, run the following command
123 @c makeinfo emacs-lisp-intro.texi
125 @c or, if you want a single, large Info file, and no paragraph indents:
126 @c makeinfo --no-split --paragraph-indent=0 --verbose emacs-lisp-intro.texi
128 @c After creating the Info file, edit your Info `dir' file, if the
129 @c `dircategory' section below does not enable your system to
130 @c install the manual automatically.
131 @c (The `dir' file is often in the `/usr/local/share/info/' directory.)
133 @c ================ How to Create an HTML file ================
135 @c To convert to HTML format
136 @c makeinfo --html --no-split --verbose emacs-lisp-intro.texi
138 @c ================ How to Print a Book in Various Sizes ================
140 @c This book can be printed in any of three different sizes.
141 @c In the above header, set @-commands appropriately.
151 @c European A4 size paper:
156 @c ================ How to Typeset and Print ================
158 @c If you do not include PostScript figures, run either of the
159 @c following command sequences, or similar commands suited to your
162 @c texi2dvi emacs-lisp-intro.texi
163 @c lpr -d emacs-lisp-intro.dvi
167 @c tex emacs-lisp-intro.texi
168 @c texindex emacs-lisp-intro.??
169 @c tex emacs-lisp-intro.texi
170 @c lpr -d emacs-lisp-intro.dvi
172 @c If you include the PostScript figures, and you have old software,
173 @c you may need to convert the .dvi file to a .ps file before
174 @c printing. Run either of the following command sequences, or one
177 @c dvips -f < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
181 @c postscript -p < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
184 @c (Note: if you edit the book so as to change the length of the
185 @c table of contents, you may have to change the value of `pageno' below.)
187 @c ================ End of Formatting Sections ================
189 @c For next or subsequent edition:
190 @c create function using with-output-to-temp-buffer
191 @c create a major mode, with keymaps
192 @c run an asynchronous process, like grep or diff
194 @c For 8.5 by 11 inch format: do not use such a small amount of
195 @c whitespace between paragraphs as smallbook format
198 \global\parskip 6pt plus 1pt
202 @c For all sized formats: print within-book cross
203 @c reference with ``...'' rather than [...]
205 @c This works with the texinfo.tex file, version 2003-05-04.08,
206 @c in the Texinfo version 4.6 of the 2003 Jun 13 distribution.
209 \if \xrefprintnodename
210 \global\def\xrefprintnodename#1{\unskip, ``#1''}
212 \global\def\xrefprintnodename#1{ ``#1''}
214 % \global\def\xrefprintnodename#1{, ``#1''}
217 @c ----------------------------------------------------
219 @dircategory GNU Emacs Lisp
221 * Emacs Lisp Intro: (eintr).
222 A simple introduction to Emacs Lisp programming.
226 This is an @cite{Introduction to Programming in Emacs Lisp}, for
227 people who are not programmers.
229 Edition @value{edition-number}, @value{update-date}
231 Copyright @copyright{} 1990-1995, 1997, 2001-2012 Free Software Foundation, Inc.
237 GNU Press, @hfill @uref{http://www.gnupress.org}@*
238 a division of the @hfill General: @email{press@@gnu.org}@*
239 Free Software Foundation, Inc. @hfill Orders:@w{ } @email{sales@@gnu.org}@*
240 51 Franklin Street, Fifth Floor @hfill Tel: +1 (617) 542-5942@*
241 Boston, MA 02110-1301 USA @hfill Fax: +1 (617) 542-2652@*
248 GNU Press, Website: http://www.gnupress.org
249 a division of the General: press@@gnu.org
250 Free Software Foundation, Inc. Orders: sales@@gnu.org
251 51 Franklin Street, Fifth Floor Tel: +1 (617) 542-5942
252 Boston, MA 02110-1301 USA Fax: +1 (617) 542-2652
257 @c Printed copies are available for $30 each.@*
260 Permission is granted to copy, distribute and/or modify this document
261 under the terms of the GNU Free Documentation License, Version 1.3 or
262 any later version published by the Free Software Foundation; there
263 being no Invariant Section, with the Front-Cover Texts being ``A GNU
264 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
265 the license is included in the section entitled ``GNU Free
266 Documentation License''.
268 (a) The FSF's Back-Cover Text is: ``You have the freedom to
269 copy and modify this GNU manual. Buying copies from the FSF
270 supports it in developing GNU and promoting software freedom.''
273 @c half title; two lines here, so do not use `shorttitlepage'
276 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
278 {\begingroup\hbox{}\vskip 0.25in \chaprm%
279 \centerline{Programming in Emacs Lisp}%
280 \endgroup\page\hbox{}\page}
285 @center @titlefont{An Introduction to}
287 @center @titlefont{Programming in Emacs Lisp}
289 @center Revised Third Edition
291 @center by Robert J. Chassell
294 @vskip 0pt plus 1filll
300 @evenheading @thispage @| @| @thischapter
301 @oddheading @thissection @| @| @thispage
305 @c Keep T.O.C. short by tightening up for largebook
308 \global\parskip 2pt plus 1pt
309 \global\advance\baselineskip by -1pt
318 @node Top, Preface, (dir), (dir)
319 @top An Introduction to Programming in Emacs Lisp
323 This master menu first lists each chapter and index; then it lists
324 every node in every chapter.
327 @c >>>> Set pageno appropriately <<<<
329 @c The first page of the Preface is a roman numeral; it is the first
330 @c right handed page after the Table of Contents; hence the following
331 @c setting must be for an odd negative number.
334 @c global@pageno = -11
337 @set COUNT-WORDS count-words-example
338 @c Length of variable name chosen so that things still line up when expanded.
341 * Preface:: What to look for.
342 * List Processing:: What is Lisp?
343 * Practicing Evaluation:: Running several programs.
344 * Writing Defuns:: How to write function definitions.
345 * Buffer Walk Through:: Exploring a few buffer-related functions.
346 * More Complex:: A few, even more complex functions.
347 * Narrowing & Widening:: Restricting your and Emacs attention to
349 * car cdr & cons:: Fundamental functions in Lisp.
350 * Cutting & Storing Text:: Removing text and saving it.
351 * List Implementation:: How lists are implemented in the computer.
352 * Yanking:: Pasting stored text.
353 * Loops & Recursion:: How to repeat a process.
354 * Regexp Search:: Regular expression searches.
355 * Counting Words:: A review of repetition and regexps.
356 * Words in a defun:: Counting words in a @code{defun}.
357 * Readying a Graph:: A prototype graph printing function.
358 * Emacs Initialization:: How to write a @file{.emacs} file.
359 * Debugging:: How to run the Emacs Lisp debuggers.
360 * Conclusion:: Now you have the basics.
361 * the-the:: An appendix: how to find reduplicated words.
362 * Kill Ring:: An appendix: how the kill ring works.
363 * Full Graph:: How to create a graph with labeled axes.
364 * Free Software and Free Manuals::
365 * GNU Free Documentation License::
370 --- The Detailed Node Listing ---
374 * Why:: Why learn Emacs Lisp?
375 * On Reading this Text:: Read, gain familiarity, pick up habits....
376 * Who You Are:: For whom this is written.
378 * Note for Novices:: You can read this as a novice.
383 * Lisp Lists:: What are lists?
384 * Run a Program:: Any list in Lisp is a program ready to run.
385 * Making Errors:: Generating an error message.
386 * Names & Definitions:: Names of symbols and function definitions.
387 * Lisp Interpreter:: What the Lisp interpreter does.
388 * Evaluation:: Running a program.
389 * Variables:: Returning a value from a variable.
390 * Arguments:: Passing information to a function.
391 * set & setq:: Setting the value of a variable.
392 * Summary:: The major points.
393 * Error Message Exercises::
397 * Numbers Lists:: List have numbers, other lists, in them.
398 * Lisp Atoms:: Elemental entities.
399 * Whitespace in Lists:: Formatting lists to be readable.
400 * Typing Lists:: How GNU Emacs helps you type lists.
404 * Complications:: Variables, Special forms, Lists within.
405 * Byte Compiling:: Specially processing code for speed.
409 * How the Interpreter Acts:: Returns and Side Effects...
410 * Evaluating Inner Lists:: Lists within lists...
414 * fill-column Example::
415 * Void Function:: The error message for a symbol
417 * Void Variable:: The error message for a symbol without a value.
421 * Data types:: Types of data passed to a function.
422 * Args as Variable or List:: An argument can be the value
423 of a variable or list.
424 * Variable Number of Arguments:: Some functions may take a
425 variable number of arguments.
426 * Wrong Type of Argument:: Passing an argument of the wrong type
428 * message:: A useful function for sending messages.
430 Setting the Value of a Variable
432 * Using set:: Setting values.
433 * Using setq:: Setting a quoted value.
434 * Counting:: Using @code{setq} to count.
436 Practicing Evaluation
438 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
440 * Buffer Names:: Buffers and files are different.
441 * Getting Buffers:: Getting a buffer itself, not merely its name.
442 * Switching Buffers:: How to change to another buffer.
443 * Buffer Size & Locations:: Where point is located and the size of
445 * Evaluation Exercise::
447 How To Write Function Definitions
449 * Primitive Functions::
450 * defun:: The @code{defun} special form.
451 * Install:: Install a function definition.
452 * Interactive:: Making a function interactive.
453 * Interactive Options:: Different options for @code{interactive}.
454 * Permanent Installation:: Installing code permanently.
455 * let:: Creating and initializing local variables.
457 * else:: If--then--else expressions.
458 * Truth & Falsehood:: What Lisp considers false and true.
459 * save-excursion:: Keeping track of point, mark, and buffer.
463 Install a Function Definition
465 * Effect of installation::
466 * Change a defun:: How to change a function definition.
468 Make a Function Interactive
470 * Interactive multiply-by-seven:: An overview.
471 * multiply-by-seven in detail:: The interactive version.
475 * Prevent confusion::
476 * Parts of let Expression::
477 * Sample let Expression::
478 * Uninitialized let Variables::
480 The @code{if} Special Form
482 * if in more detail::
483 * type-of-animal in detail:: An example of an @code{if} expression.
485 Truth and Falsehood in Emacs Lisp
487 * nil explained:: @code{nil} has two meanings.
489 @code{save-excursion}
491 * Point and mark:: A review of various locations.
492 * Template for save-excursion::
494 A Few Buffer--Related Functions
496 * Finding More:: How to find more information.
497 * simplified-beginning-of-buffer:: Shows @code{goto-char},
498 @code{point-min}, and @code{push-mark}.
499 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
500 * append-to-buffer:: Uses @code{save-excursion} and
501 @code{insert-buffer-substring}.
502 * Buffer Related Review:: Review.
505 The Definition of @code{mark-whole-buffer}
507 * mark-whole-buffer overview::
508 * Body of mark-whole-buffer:: Only three lines of code.
510 The Definition of @code{append-to-buffer}
512 * append-to-buffer overview::
513 * append interactive:: A two part interactive expression.
514 * append-to-buffer body:: Incorporates a @code{let} expression.
515 * append save-excursion:: How the @code{save-excursion} works.
517 A Few More Complex Functions
519 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
520 * insert-buffer:: Read-only, and with @code{or}.
521 * beginning-of-buffer:: Shows @code{goto-char},
522 @code{point-min}, and @code{push-mark}.
523 * Second Buffer Related Review::
524 * optional Exercise::
526 The Definition of @code{insert-buffer}
528 * insert-buffer code::
529 * insert-buffer interactive:: When you can read, but not write.
530 * insert-buffer body:: The body has an @code{or} and a @code{let}.
531 * if & or:: Using an @code{if} instead of an @code{or}.
532 * Insert or:: How the @code{or} expression works.
533 * Insert let:: Two @code{save-excursion} expressions.
534 * New insert-buffer::
536 The Interactive Expression in @code{insert-buffer}
538 * Read-only buffer:: When a buffer cannot be modified.
539 * b for interactive:: An existing buffer or else its name.
541 Complete Definition of @code{beginning-of-buffer}
543 * Optional Arguments::
544 * beginning-of-buffer opt arg:: Example with optional argument.
545 * beginning-of-buffer complete::
547 @code{beginning-of-buffer} with an Argument
549 * Disentangle beginning-of-buffer::
550 * Large buffer case::
551 * Small buffer case::
553 Narrowing and Widening
555 * Narrowing advantages:: The advantages of narrowing
556 * save-restriction:: The @code{save-restriction} special form.
557 * what-line:: The number of the line that point is on.
560 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
562 * Strange Names:: An historical aside: why the strange names?
563 * car & cdr:: Functions for extracting part of a list.
564 * cons:: Constructing a list.
565 * nthcdr:: Calling @code{cdr} repeatedly.
567 * setcar:: Changing the first element of a list.
568 * setcdr:: Changing the rest of a list.
574 * length:: How to find the length of a list.
576 Cutting and Storing Text
578 * Storing Text:: Text is stored in a list.
579 * zap-to-char:: Cutting out text up to a character.
580 * kill-region:: Cutting text out of a region.
581 * copy-region-as-kill:: A definition for copying text.
582 * Digression into C:: Minor note on C programming language macros.
583 * defvar:: How to give a variable an initial value.
584 * cons & search-fwd Review::
589 * Complete zap-to-char:: The complete implementation.
590 * zap-to-char interactive:: A three part interactive expression.
591 * zap-to-char body:: A short overview.
592 * search-forward:: How to search for a string.
593 * progn:: The @code{progn} special form.
594 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
598 * Complete kill-region:: The function definition.
599 * condition-case:: Dealing with a problem.
602 @code{copy-region-as-kill}
604 * Complete copy-region-as-kill:: The complete function definition.
605 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
607 The Body of @code{copy-region-as-kill}
609 * last-command & this-command::
610 * kill-append function::
611 * kill-new function::
613 Initializing a Variable with @code{defvar}
615 * See variable current value::
616 * defvar and asterisk::
618 How Lists are Implemented
621 * Symbols as Chest:: Exploring a powerful metaphor.
626 * Kill Ring Overview::
627 * kill-ring-yank-pointer:: The kill ring is a list.
628 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
632 * while:: Causing a stretch of code to repeat.
634 * Recursion:: Causing a function to call itself.
639 * Looping with while:: Repeat so long as test returns true.
640 * Loop Example:: A @code{while} loop that uses a list.
641 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
642 * Incrementing Loop:: A loop with an incrementing counter.
643 * Incrementing Loop Details::
644 * Decrementing Loop:: A loop with a decrementing counter.
646 Details of an Incrementing Loop
648 * Incrementing Example:: Counting pebbles in a triangle.
649 * Inc Example parts:: The parts of the function definition.
650 * Inc Example altogether:: Putting the function definition together.
652 Loop with a Decrementing Counter
654 * Decrementing Example:: More pebbles on the beach.
655 * Dec Example parts:: The parts of the function definition.
656 * Dec Example altogether:: Putting the function definition together.
658 Save your time: @code{dolist} and @code{dotimes}
665 * Building Robots:: Same model, different serial number ...
666 * Recursive Definition Parts:: Walk until you stop ...
667 * Recursion with list:: Using a list as the test whether to recurse.
668 * Recursive triangle function::
669 * Recursion with cond::
670 * Recursive Patterns:: Often used templates.
671 * No Deferment:: Don't store up work ...
672 * No deferment solution::
674 Recursion in Place of a Counter
676 * Recursive Example arg of 1 or 2::
677 * Recursive Example arg of 3 or 4::
685 Regular Expression Searches
687 * sentence-end:: The regular expression for @code{sentence-end}.
688 * re-search-forward:: Very similar to @code{search-forward}.
689 * forward-sentence:: A straightforward example of regexp search.
690 * forward-paragraph:: A somewhat complex example.
691 * etags:: How to create your own @file{TAGS} table.
693 * re-search Exercises::
695 @code{forward-sentence}
697 * Complete forward-sentence::
698 * fwd-sentence while loops:: Two @code{while} loops.
699 * fwd-sentence re-search:: A regular expression search.
701 @code{forward-paragraph}: a Goldmine of Functions
703 * forward-paragraph in brief:: Key parts of the function definition.
704 * fwd-para let:: The @code{let*} expression.
705 * fwd-para while:: The forward motion @code{while} loop.
707 Counting: Repetition and Regexps
710 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
711 * recursive-count-words:: Start with case of no words in region.
712 * Counting Exercise::
714 The @code{@value{COUNT-WORDS}} Function
716 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
717 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
719 Counting Words in a @code{defun}
721 * Divide and Conquer::
722 * Words and Symbols:: What to count?
723 * Syntax:: What constitutes a word or symbol?
724 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
725 * Several defuns:: Counting several defuns in a file.
726 * Find a File:: Do you want to look at a file?
727 * lengths-list-file:: A list of the lengths of many definitions.
728 * Several files:: Counting in definitions in different files.
729 * Several files recursively:: Recursively counting in different files.
730 * Prepare the data:: Prepare the data for display in a graph.
732 Count Words in @code{defuns} in Different Files
734 * lengths-list-many-files:: Return a list of the lengths of defuns.
735 * append:: Attach one list to another.
737 Prepare the Data for Display in a Graph
739 * Data for Display in Detail::
740 * Sorting:: Sorting lists.
741 * Files List:: Making a list of files.
742 * Counting function definitions::
746 * Columns of a graph::
747 * graph-body-print:: How to print the body of a graph.
748 * recursive-graph-body-print::
750 * Line Graph Exercise::
752 Your @file{.emacs} File
754 * Default Configuration::
755 * Site-wide Init:: You can write site-wide init files.
756 * defcustom:: Emacs will write code for you.
757 * Beginning a .emacs File:: How to write a @code{.emacs file}.
758 * Text and Auto-fill:: Automatically wrap lines.
759 * Mail Aliases:: Use abbreviations for email addresses.
760 * Indent Tabs Mode:: Don't use tabs with @TeX{}
761 * Keybindings:: Create some personal keybindings.
762 * Keymaps:: More about key binding.
763 * Loading Files:: Load (i.e., evaluate) files automatically.
764 * Autoload:: Make functions available.
765 * Simple Extension:: Define a function; bind it to a key.
766 * X11 Colors:: Colors in X.
768 * Mode Line:: How to customize your mode line.
772 * debug:: How to use the built-in debugger.
773 * debug-on-entry:: Start debugging when you call a function.
774 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
775 * edebug:: How to use Edebug, a source level debugger.
776 * Debugging Exercises::
778 Handling the Kill Ring
780 * What the Kill Ring Does::
782 * yank:: Paste a copy of a clipped element.
783 * yank-pop:: Insert element pointed to.
786 The @code{current-kill} Function
788 * Code for current-kill::
789 * Understanding current-kill::
791 @code{current-kill} in Outline
793 * Body of current-kill::
794 * Digression concerning error:: How to mislead humans, but not computers.
795 * Determining the Element::
797 A Graph with Labeled Axes
800 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
801 * print-Y-axis:: Print a label for the vertical axis.
802 * print-X-axis:: Print a horizontal label.
803 * Print Whole Graph:: The function to print a complete graph.
805 The @code{print-Y-axis} Function
807 * print-Y-axis in Detail::
808 * Height of label:: What height for the Y axis?
809 * Compute a Remainder:: How to compute the remainder of a division.
810 * Y Axis Element:: Construct a line for the Y axis.
811 * Y-axis-column:: Generate a list of Y axis labels.
812 * print-Y-axis Penultimate:: A not quite final version.
814 The @code{print-X-axis} Function
816 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
817 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
819 Printing the Whole Graph
821 * The final version:: A few changes.
822 * Test print-graph:: Run a short test.
823 * Graphing words in defuns:: Executing the final code.
824 * lambda:: How to write an anonymous function.
825 * mapcar:: Apply a function to elements of a list.
826 * Another Bug:: Yet another bug @dots{} most insidious.
827 * Final printed graph:: The graph itself!
832 @node Preface, List Processing, Top, Top
833 @comment node-name, next, previous, up
836 Most of the GNU Emacs integrated environment is written in the programming
837 language called Emacs Lisp. The code written in this programming
838 language is the software---the sets of instructions---that tell the
839 computer what to do when you give it commands. Emacs is designed so
840 that you can write new code in Emacs Lisp and easily install it as an
841 extension to the editor.
843 (GNU Emacs is sometimes called an ``extensible editor'', but it does
844 much more than provide editing capabilities. It is better to refer to
845 Emacs as an ``extensible computing environment''. However, that
846 phrase is quite a mouthful. It is easier to refer to Emacs simply as
847 an editor. Moreover, everything you do in Emacs---find the Mayan date
848 and phases of the moon, simplify polynomials, debug code, manage
849 files, read letters, write books---all these activities are kinds of
850 editing in the most general sense of the word.)
853 * Why:: Why learn Emacs Lisp?
854 * On Reading this Text:: Read, gain familiarity, pick up habits....
855 * Who You Are:: For whom this is written.
857 * Note for Novices:: You can read this as a novice.
861 @node Why, On Reading this Text, Preface, Preface
863 @unnumberedsec Why Study Emacs Lisp?
866 Although Emacs Lisp is usually thought of in association only with Emacs,
867 it is a full computer programming language. You can use Emacs Lisp as
868 you would any other programming language.
870 Perhaps you want to understand programming; perhaps you want to extend
871 Emacs; or perhaps you want to become a programmer. This introduction to
872 Emacs Lisp is designed to get you started: to guide you in learning the
873 fundamentals of programming, and more importantly, to show you how you
874 can teach yourself to go further.
876 @node On Reading this Text, Who You Are, Why, Preface
877 @comment node-name, next, previous, up
878 @unnumberedsec On Reading this Text
880 All through this document, you will see little sample programs you can
881 run inside of Emacs. If you read this document in Info inside of GNU
882 Emacs, you can run the programs as they appear. (This is easy to do and
883 is explained when the examples are presented.) Alternatively, you can
884 read this introduction as a printed book while sitting beside a computer
885 running Emacs. (This is what I like to do; I like printed books.) If
886 you don't have a running Emacs beside you, you can still read this book,
887 but in this case, it is best to treat it as a novel or as a travel guide
888 to a country not yet visited: interesting, but not the same as being
891 Much of this introduction is dedicated to walkthroughs or guided tours
892 of code used in GNU Emacs. These tours are designed for two purposes:
893 first, to give you familiarity with real, working code (code you use
894 every day); and, second, to give you familiarity with the way Emacs
895 works. It is interesting to see how a working environment is
898 hope that you will pick up the habit of browsing through source code.
899 You can learn from it and mine it for ideas. Having GNU Emacs is like
900 having a dragon's cave of treasures.
902 In addition to learning about Emacs as an editor and Emacs Lisp as a
903 programming language, the examples and guided tours will give you an
904 opportunity to get acquainted with Emacs as a Lisp programming
905 environment. GNU Emacs supports programming and provides tools that
906 you will want to become comfortable using, such as @kbd{M-.} (the key
907 which invokes the @code{find-tag} command). You will also learn about
908 buffers and other objects that are part of the environment.
909 Learning about these features of Emacs is like learning new routes
910 around your home town.
913 In addition, I have written several programs as extended examples.
914 Although these are examples, the programs are real. I use them.
915 Other people use them. You may use them. Beyond the fragments of
916 programs used for illustrations, there is very little in here that is
917 `just for teaching purposes'; what you see is used. This is a great
918 advantage of Emacs Lisp: it is easy to learn to use it for work.
921 Finally, I hope to convey some of the skills for using Emacs to
922 learn aspects of programming that you don't know. You can often use
923 Emacs to help you understand what puzzles you or to find out how to do
924 something new. This self-reliance is not only a pleasure, but an
927 @node Who You Are, Lisp History, On Reading this Text, Preface
928 @comment node-name, next, previous, up
929 @unnumberedsec For Whom This is Written
931 This text is written as an elementary introduction for people who are
932 not programmers. If you are a programmer, you may not be satisfied with
933 this primer. The reason is that you may have become expert at reading
934 reference manuals and be put off by the way this text is organized.
936 An expert programmer who reviewed this text said to me:
939 @i{I prefer to learn from reference manuals. I ``dive into'' each
940 paragraph, and ``come up for air'' between paragraphs.}
942 @i{When I get to the end of a paragraph, I assume that that subject is
943 done, finished, that I know everything I need (with the
944 possible exception of the case when the next paragraph starts talking
945 about it in more detail). I expect that a well written reference manual
946 will not have a lot of redundancy, and that it will have excellent
947 pointers to the (one) place where the information I want is.}
950 This introduction is not written for this person!
952 Firstly, I try to say everything at least three times: first, to
953 introduce it; second, to show it in context; and third, to show it in a
954 different context, or to review it.
956 Secondly, I hardly ever put all the information about a subject in one
957 place, much less in one paragraph. To my way of thinking, that imposes
958 too heavy a burden on the reader. Instead I try to explain only what
959 you need to know at the time. (Sometimes I include a little extra
960 information so you won't be surprised later when the additional
961 information is formally introduced.)
963 When you read this text, you are not expected to learn everything the
964 first time. Frequently, you need only make, as it were, a `nodding
965 acquaintance' with some of the items mentioned. My hope is that I have
966 structured the text and given you enough hints that you will be alert to
967 what is important, and concentrate on it.
969 You will need to ``dive into'' some paragraphs; there is no other way
970 to read them. But I have tried to keep down the number of such
971 paragraphs. This book is intended as an approachable hill, rather than
972 as a daunting mountain.
974 This introduction to @cite{Programming in Emacs Lisp} has a companion
977 @cite{The GNU Emacs Lisp Reference Manual}.
980 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
981 Emacs Lisp Reference Manual}.
983 The reference manual has more detail than this introduction. In the
984 reference manual, all the information about one topic is concentrated
985 in one place. You should turn to it if you are like the programmer
986 quoted above. And, of course, after you have read this
987 @cite{Introduction}, you will find the @cite{Reference Manual} useful
988 when you are writing your own programs.
990 @node Lisp History, Note for Novices, Who You Are, Preface
991 @unnumberedsec Lisp History
994 Lisp was first developed in the late 1950s at the Massachusetts
995 Institute of Technology for research in artificial intelligence. The
996 great power of the Lisp language makes it superior for other purposes as
997 well, such as writing editor commands and integrated environments.
1001 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
1002 in the 1960s. It is somewhat inspired by Common Lisp, which became a
1003 standard in the 1980s. However, Emacs Lisp is much simpler than Common
1004 Lisp. (The standard Emacs distribution contains an optional extensions
1005 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
1007 @node Note for Novices, Thank You, Lisp History, Preface
1008 @comment node-name, next, previous, up
1009 @unnumberedsec A Note for Novices
1011 If you don't know GNU Emacs, you can still read this document
1012 profitably. However, I recommend you learn Emacs, if only to learn to
1013 move around your computer screen. You can teach yourself how to use
1014 Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This
1015 means you press and release the @key{CTRL} key and the @kbd{h} at the
1016 same time, and then press and release @kbd{t}.)
1018 Also, I often refer to one of Emacs's standard commands by listing the
1019 keys which you press to invoke the command and then giving the name of
1020 the command in parentheses, like this: @kbd{M-C-\}
1021 (@code{indent-region}). What this means is that the
1022 @code{indent-region} command is customarily invoked by typing
1023 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
1024 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
1025 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
1026 @key{META} key, @key{CTRL} key and @key{\} key all at the same time.
1027 (On many modern keyboards the @key{META} key is labeled
1029 Sometimes a combination like this is called a keychord, since it is
1030 similar to the way you play a chord on a piano. If your keyboard does
1031 not have a @key{META} key, the @key{ESC} key prefix is used in place
1032 of it. In this case, @kbd{M-C-\} means that you press and release your
1033 @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at
1034 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
1035 along with the key that is labeled @key{ALT} and, at the same time,
1036 press the @key{\} key.
1038 In addition to typing a lone keychord, you can prefix what you type
1039 with @kbd{C-u}, which is called the `universal argument'. The
1040 @kbd{C-u} keychord passes an argument to the subsequent command.
1041 Thus, to indent a region of plain text by 6 spaces, mark the region,
1042 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
1043 Emacs either passes the number 4 to the command or otherwise runs the
1044 command differently than it would otherwise.) @xref{Arguments, ,
1045 Numeric Arguments, emacs, The GNU Emacs Manual}.
1047 If you are reading this in Info using GNU Emacs, you can read through
1048 this whole document just by pressing the space bar, @key{SPC}.
1049 (To learn about Info, type @kbd{C-h i} and then select Info.)
1051 A note on terminology: when I use the word Lisp alone, I often am
1052 referring to the various dialects of Lisp in general, but when I speak
1053 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
1055 @node Thank You, , Note for Novices, Preface
1056 @comment node-name, next, previous, up
1057 @unnumberedsec Thank You
1059 My thanks to all who helped me with this book. My especial thanks to
1060 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
1061 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.@:
1062 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
1063 @w{Philip Johnson} and @w{David Stampe} for their patient
1064 encouragement. My mistakes are my own.
1068 @email{bob@@gnu.org}
1071 @c ================ Beginning of main text ================
1073 @c Start main text on right-hand (verso) page
1076 \par\vfill\supereject
1079 \par\vfill\supereject
1081 \par\vfill\supereject
1083 \par\vfill\supereject
1087 @c Note: this resetting of the page number back to 1 causes TeX to gripe
1088 @c about already having seen page numbers 1-4 before (in the preface):
1089 @c pdfTeX warning (ext4): destination with the same identifier (name{1})
1090 @c has been already used, duplicate ignored
1091 @c I guess that is harmless (what happens if a later part of the text
1092 @c makes a link to something in the first 4 pages though?).
1093 @c Note that eg the Emacs manual has a preface, but does not bother
1094 @c resetting the page numbers back to 1 after that.
1097 @evenheading @thispage @| @| @thischapter
1098 @oddheading @thissection @| @| @thispage
1102 @node List Processing, Practicing Evaluation, Preface, Top
1103 @comment node-name, next, previous, up
1104 @chapter List Processing
1106 To the untutored eye, Lisp is a strange programming language. In Lisp
1107 code there are parentheses everywhere. Some people even claim that
1108 the name stands for `Lots of Isolated Silly Parentheses'. But the
1109 claim is unwarranted. Lisp stands for LISt Processing, and the
1110 programming language handles @emph{lists} (and lists of lists) by
1111 putting them between parentheses. The parentheses mark the boundaries
1112 of the list. Sometimes a list is preceded by a single apostrophe or
1113 quotation mark, @samp{'}@footnote{The single apostrophe or quotation
1114 mark is an abbreviation for the function @code{quote}; you need not
1115 think about functions now; functions are defined in @ref{Making
1116 Errors, , Generate an Error Message}.} Lists are the basis of Lisp.
1119 * Lisp Lists:: What are lists?
1120 * Run a Program:: Any list in Lisp is a program ready to run.
1121 * Making Errors:: Generating an error message.
1122 * Names & Definitions:: Names of symbols and function definitions.
1123 * Lisp Interpreter:: What the Lisp interpreter does.
1124 * Evaluation:: Running a program.
1125 * Variables:: Returning a value from a variable.
1126 * Arguments:: Passing information to a function.
1127 * set & setq:: Setting the value of a variable.
1128 * Summary:: The major points.
1129 * Error Message Exercises::
1132 @node Lisp Lists, Run a Program, List Processing, List Processing
1133 @comment node-name, next, previous, up
1137 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1138 This list is preceded by a single apostrophe. It could just as well be
1139 written as follows, which looks more like the kind of list you are likely
1140 to be familiar with:
1152 The elements of this list are the names of the four different flowers,
1153 separated from each other by whitespace and surrounded by parentheses,
1154 like flowers in a field with a stone wall around them.
1155 @cindex Flowers in a field
1158 * Numbers Lists:: List have numbers, other lists, in them.
1159 * Lisp Atoms:: Elemental entities.
1160 * Whitespace in Lists:: Formatting lists to be readable.
1161 * Typing Lists:: How GNU Emacs helps you type lists.
1164 @node Numbers Lists, Lisp Atoms, Lisp Lists, Lisp Lists
1166 @unnumberedsubsec Numbers, Lists inside of Lists
1169 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1170 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1171 separated by whitespace.
1173 In Lisp, both data and programs are represented the same way; that is,
1174 they are both lists of words, numbers, or other lists, separated by
1175 whitespace and surrounded by parentheses. (Since a program looks like
1176 data, one program may easily serve as data for another; this is a very
1177 powerful feature of Lisp.) (Incidentally, these two parenthetical
1178 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1179 @samp{.} as punctuation marks.)
1182 Here is another list, this time with a list inside of it:
1185 '(this list has (a list inside of it))
1188 The components of this list are the words @samp{this}, @samp{list},
1189 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1190 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1191 @samp{of}, @samp{it}.
1193 @node Lisp Atoms, Whitespace in Lists, Numbers Lists, Lisp Lists
1194 @comment node-name, next, previous, up
1195 @subsection Lisp Atoms
1198 In Lisp, what we have been calling words are called @dfn{atoms}. This
1199 term comes from the historical meaning of the word atom, which means
1200 `indivisible'. As far as Lisp is concerned, the words we have been
1201 using in the lists cannot be divided into any smaller parts and still
1202 mean the same thing as part of a program; likewise with numbers and
1203 single character symbols like @samp{+}. On the other hand, unlike an
1204 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1205 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1207 In a list, atoms are separated from each other by whitespace. They can be
1208 right next to a parenthesis.
1210 @cindex @samp{empty list} defined
1211 Technically speaking, a list in Lisp consists of parentheses surrounding
1212 atoms separated by whitespace or surrounding other lists or surrounding
1213 both atoms and other lists. A list can have just one atom in it or
1214 have nothing in it at all. A list with nothing in it looks like this:
1215 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1216 empty list is considered both an atom and a list at the same time.
1218 @cindex Symbolic expressions, introduced
1219 @cindex @samp{expression} defined
1220 @cindex @samp{form} defined
1221 The printed representation of both atoms and lists are called
1222 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1223 The word @dfn{expression} by itself can refer to either the printed
1224 representation, or to the atom or list as it is held internally in the
1225 computer. Often, people use the term @dfn{expression}
1226 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1227 as a synonym for expression.)
1229 Incidentally, the atoms that make up our universe were named such when
1230 they were thought to be indivisible; but it has been found that physical
1231 atoms are not indivisible. Parts can split off an atom or it can
1232 fission into two parts of roughly equal size. Physical atoms were named
1233 prematurely, before their truer nature was found. In Lisp, certain
1234 kinds of atom, such as an array, can be separated into parts; but the
1235 mechanism for doing this is different from the mechanism for splitting a
1236 list. As far as list operations are concerned, the atoms of a list are
1239 As in English, the meanings of the component letters of a Lisp atom
1240 are different from the meaning the letters make as a word. For
1241 example, the word for the South American sloth, the @samp{ai}, is
1242 completely different from the two words, @samp{a}, and @samp{i}.
1244 There are many kinds of atom in nature but only a few in Lisp: for
1245 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1246 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1247 listed in the examples above are all symbols. In everyday Lisp
1248 conversation, the word ``atom'' is not often used, because programmers
1249 usually try to be more specific about what kind of atom they are dealing
1250 with. Lisp programming is mostly about symbols (and sometimes numbers)
1251 within lists. (Incidentally, the preceding three word parenthetical
1252 remark is a proper list in Lisp, since it consists of atoms, which in
1253 this case are symbols, separated by whitespace and enclosed by
1254 parentheses, without any non-Lisp punctuation.)
1257 Text between double quotation marks---even sentences or
1258 paragraphs---is also an atom. Here is an example:
1259 @cindex Text between double quotation marks
1262 '(this list includes "text between quotation marks.")
1265 @cindex @samp{string} defined
1267 In Lisp, all of the quoted text including the punctuation mark and the
1268 blank spaces is a single atom. This kind of atom is called a
1269 @dfn{string} (for `string of characters') and is the sort of thing that
1270 is used for messages that a computer can print for a human to read.
1271 Strings are a different kind of atom than numbers or symbols and are
1274 @node Whitespace in Lists, Typing Lists, Lisp Atoms, Lisp Lists
1275 @comment node-name, next, previous, up
1276 @subsection Whitespace in Lists
1277 @cindex Whitespace in lists
1280 The amount of whitespace in a list does not matter. From the point of view
1281 of the Lisp language,
1292 is exactly the same as this:
1295 '(this list looks like this)
1298 Both examples show what to Lisp is the same list, the list made up of
1299 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1300 @samp{this} in that order.
1302 Extra whitespace and newlines are designed to make a list more readable
1303 by humans. When Lisp reads the expression, it gets rid of all the extra
1304 whitespace (but it needs to have at least one space between atoms in
1305 order to tell them apart.)
1307 Odd as it seems, the examples we have seen cover almost all of what Lisp
1308 lists look like! Every other list in Lisp looks more or less like one
1309 of these examples, except that the list may be longer and more complex.
1310 In brief, a list is between parentheses, a string is between quotation
1311 marks, a symbol looks like a word, and a number looks like a number.
1312 (For certain situations, square brackets, dots and a few other special
1313 characters may be used; however, we will go quite far without them.)
1315 @node Typing Lists, , Whitespace in Lists, Lisp Lists
1316 @comment node-name, next, previous, up
1317 @subsection GNU Emacs Helps You Type Lists
1318 @cindex Help typing lists
1319 @cindex Formatting help
1321 When you type a Lisp expression in GNU Emacs using either Lisp
1322 Interaction mode or Emacs Lisp mode, you have available to you several
1323 commands to format the Lisp expression so it is easy to read. For
1324 example, pressing the @key{TAB} key automatically indents the line the
1325 cursor is on by the right amount. A command to properly indent the
1326 code in a region is customarily bound to @kbd{M-C-\}. Indentation is
1327 designed so that you can see which elements of a list belong to which
1328 list---elements of a sub-list are indented more than the elements of
1331 In addition, when you type a closing parenthesis, Emacs momentarily
1332 jumps the cursor back to the matching opening parenthesis, so you can
1333 see which one it is. This is very useful, since every list you type
1334 in Lisp must have its closing parenthesis match its opening
1335 parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs
1336 Manual}, for more information about Emacs's modes.)
1338 @node Run a Program, Making Errors, Lisp Lists, List Processing
1339 @comment node-name, next, previous, up
1340 @section Run a Program
1341 @cindex Run a program
1342 @cindex Program, running one
1344 @cindex @samp{evaluate} defined
1345 A list in Lisp---any list---is a program ready to run. If you run it
1346 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1347 of three things: do nothing except return to you the list itself; send
1348 you an error message; or, treat the first symbol in the list as a
1349 command to do something. (Usually, of course, it is the last of these
1350 three things that you really want!)
1352 @c use code for the single apostrophe, not samp.
1353 The single apostrophe, @code{'}, that I put in front of some of the
1354 example lists in preceding sections is called a @dfn{quote}; when it
1355 precedes a list, it tells Lisp to do nothing with the list, other than
1356 take it as it is written. But if there is no quote preceding a list,
1357 the first item of the list is special: it is a command for the computer
1358 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1359 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1360 understands that the @code{+} is an instruction to do something with the
1361 rest of the list: add the numbers that follow.
1364 If you are reading this inside of GNU Emacs in Info, here is how you can
1365 evaluate such a list: place your cursor immediately after the right
1366 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1372 @c use code for the number four, not samp.
1374 You will see the number @code{4} appear in the echo area. (In the
1375 jargon, what you have just done is ``evaluate the list.'' The echo area
1376 is the line at the bottom of the screen that displays or ``echoes''
1377 text.) Now try the same thing with a quoted list: place the cursor
1378 right after the following list and type @kbd{C-x C-e}:
1381 '(this is a quoted list)
1385 You will see @code{(this is a quoted list)} appear in the echo area.
1387 @cindex Lisp interpreter, explained
1388 @cindex Interpreter, Lisp, explained
1389 In both cases, what you are doing is giving a command to the program
1390 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1391 interpreter a command to evaluate the expression. The name of the Lisp
1392 interpreter comes from the word for the task done by a human who comes
1393 up with the meaning of an expression---who ``interprets'' it.
1395 You can also evaluate an atom that is not part of a list---one that is
1396 not surrounded by parentheses; again, the Lisp interpreter translates
1397 from the humanly readable expression to the language of the computer.
1398 But before discussing this (@pxref{Variables}), we will discuss what the
1399 Lisp interpreter does when you make an error.
1401 @node Making Errors, Names & Definitions, Run a Program, List Processing
1402 @comment node-name, next, previous, up
1403 @section Generate an Error Message
1404 @cindex Generate an error message
1405 @cindex Error message generation
1407 Partly so you won't worry if you do it accidentally, we will now give
1408 a command to the Lisp interpreter that generates an error message.
1409 This is a harmless activity; and indeed, we will often try to generate
1410 error messages intentionally. Once you understand the jargon, error
1411 messages can be informative. Instead of being called ``error''
1412 messages, they should be called ``help'' messages. They are like
1413 signposts to a traveler in a strange country; deciphering them can be
1414 hard, but once understood, they can point the way.
1416 The error message is generated by a built-in GNU Emacs debugger. We
1417 will `enter the debugger'. You get out of the debugger by typing @code{q}.
1419 What we will do is evaluate a list that is not quoted and does not
1420 have a meaningful command as its first element. Here is a list almost
1421 exactly the same as the one we just used, but without the single-quote
1422 in front of it. Position the cursor right after it and type @kbd{C-x
1426 (this is an unquoted list)
1430 What you see depends on which version of Emacs you are running. GNU
1431 Emacs version 22 provides more information than version 20 and before.
1432 First, the more recent result of generating an error; then the
1433 earlier, version 20 result.
1437 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1438 you will see the following in it:
1442 ---------- Buffer: *Backtrace* ----------
1443 Debugger entered--Lisp error: (void-function this)
1444 (this is an unquoted list)
1445 eval((this is an unquoted list))
1446 eval-last-sexp-1(nil)
1448 call-interactively(eval-last-sexp)
1449 ---------- Buffer: *Backtrace* ----------
1455 Your cursor will be in this window (you may have to wait a few seconds
1456 before it becomes visible). To quit the debugger and make the
1457 debugger window go away, type:
1464 Please type @kbd{q} right now, so you become confident that you can
1465 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1468 @cindex @samp{function} defined
1469 Based on what we already know, we can almost read this error message.
1471 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1472 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1473 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1474 an abbreviation for `evaluate' and @code{sexp} is an abbreviation for
1475 `symbolic expression'. The command means `evaluate last symbolic
1476 expression', which is the expression just before your cursor.
1478 Each line above tells you what the Lisp interpreter evaluated next.
1479 The most recent action is at the top. The buffer is called the
1480 @file{*Backtrace*} buffer because it enables you to track Emacs
1484 At the top of the @file{*Backtrace*} buffer, you see the line:
1487 Debugger entered--Lisp error: (void-function this)
1491 The Lisp interpreter tried to evaluate the first atom of the list, the
1492 word @samp{this}. It is this action that generated the error message
1493 @samp{void-function this}.
1495 The message contains the words @samp{void-function} and @samp{this}.
1497 @cindex @samp{function} defined
1498 The word @samp{function} was mentioned once before. It is a very
1499 important word. For our purposes, we can define it by saying that a
1500 @dfn{function} is a set of instructions to the computer that tell the
1501 computer to do something.
1503 Now we can begin to understand the error message: @samp{void-function
1504 this}. The function (that is, the word @samp{this}) does not have a
1505 definition of any set of instructions for the computer to carry out.
1507 The slightly odd word, @samp{void-function}, is designed to cover the
1508 way Emacs Lisp is implemented, which is that when a symbol does not
1509 have a function definition attached to it, the place that should
1510 contain the instructions is `void'.
1512 On the other hand, since we were able to add 2 plus 2 successfully, by
1513 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1514 have a set of instructions for the computer to obey and those
1515 instructions must be to add the numbers that follow the @code{+}.
1518 In GNU Emacs version 20, and in earlier versions, you will see only
1519 one line of error message; it will appear in the echo area and look
1523 Symbol's function definition is void:@: this
1527 (Also, your terminal may beep at you---some do, some don't; and others
1528 blink. This is just a device to get your attention.) The message goes
1529 away as soon as you type another key, even just to move the cursor.
1531 We know the meaning of the word @samp{Symbol}. It refers to the first
1532 atom of the list, the word @samp{this}. The word @samp{function}
1533 refers to the instructions that tell the computer what to do.
1534 (Technically, the symbol tells the computer where to find the
1535 instructions, but this is a complication we can ignore for the
1538 The error message can be understood: @samp{Symbol's function
1539 definition is void:@: this}. The symbol (that is, the word
1540 @samp{this}) lacks instructions for the computer to carry out.
1542 @node Names & Definitions, Lisp Interpreter, Making Errors, List Processing
1543 @comment node-name, next, previous, up
1544 @section Symbol Names and Function Definitions
1545 @cindex Symbol names
1547 We can articulate another characteristic of Lisp based on what we have
1548 discussed so far---an important characteristic: a symbol, like
1549 @code{+}, is not itself the set of instructions for the computer to
1550 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1551 of locating the definition or set of instructions. What we see is the
1552 name through which the instructions can be found. Names of people
1553 work the same way. I can be referred to as @samp{Bob}; however, I am
1554 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1555 consciousness consistently associated with a particular life-form.
1556 The name is not me, but it can be used to refer to me.
1558 In Lisp, one set of instructions can be attached to several names.
1559 For example, the computer instructions for adding numbers can be
1560 linked to the symbol @code{plus} as well as to the symbol @code{+}
1561 (and are in some dialects of Lisp). Among humans, I can be referred
1562 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1564 On the other hand, a symbol can have only one function definition
1565 attached to it at a time. Otherwise, the computer would be confused as
1566 to which definition to use. If this were the case among people, only
1567 one person in the world could be named @samp{Bob}. However, the function
1568 definition to which the name refers can be changed readily.
1569 (@xref{Install, , Install a Function Definition}.)
1571 Since Emacs Lisp is large, it is customary to name symbols in a way
1572 that identifies the part of Emacs to which the function belongs.
1573 Thus, all the names for functions that deal with Texinfo start with
1574 @samp{texinfo-} and those for functions that deal with reading mail
1575 start with @samp{rmail-}.
1577 @node Lisp Interpreter, Evaluation, Names & Definitions, List Processing
1578 @comment node-name, next, previous, up
1579 @section The Lisp Interpreter
1580 @cindex Lisp interpreter, what it does
1581 @cindex Interpreter, what it does
1583 Based on what we have seen, we can now start to figure out what the
1584 Lisp interpreter does when we command it to evaluate a list.
1585 First, it looks to see whether there is a quote before the list; if
1586 there is, the interpreter just gives us the list. On the other
1587 hand, if there is no quote, the interpreter looks at the first element
1588 in the list and sees whether it has a function definition. If it does,
1589 the interpreter carries out the instructions in the function definition.
1590 Otherwise, the interpreter prints an error message.
1592 This is how Lisp works. Simple. There are added complications which we
1593 will get to in a minute, but these are the fundamentals. Of course, to
1594 write Lisp programs, you need to know how to write function definitions
1595 and attach them to names, and how to do this without confusing either
1596 yourself or the computer.
1599 * Complications:: Variables, Special forms, Lists within.
1600 * Byte Compiling:: Specially processing code for speed.
1603 @node Complications, Byte Compiling, Lisp Interpreter, Lisp Interpreter
1605 @unnumberedsubsec Complications
1608 Now, for the first complication. In addition to lists, the Lisp
1609 interpreter can evaluate a symbol that is not quoted and does not have
1610 parentheses around it. The Lisp interpreter will attempt to determine
1611 the symbol's value as a @dfn{variable}. This situation is described
1612 in the section on variables. (@xref{Variables}.)
1614 @cindex Special form
1615 The second complication occurs because some functions are unusual and do
1616 not work in the usual manner. Those that don't are called @dfn{special
1617 forms}. They are used for special jobs, like defining a function, and
1618 there are not many of them. In the next few chapters, you will be
1619 introduced to several of the more important special forms.
1621 The third and final complication is this: if the function that the
1622 Lisp interpreter is looking at is not a special form, and if it is part
1623 of a list, the Lisp interpreter looks to see whether the list has a list
1624 inside of it. If there is an inner list, the Lisp interpreter first
1625 figures out what it should do with the inside list, and then it works on
1626 the outside list. If there is yet another list embedded inside the
1627 inner list, it works on that one first, and so on. It always works on
1628 the innermost list first. The interpreter works on the innermost list
1629 first, to evaluate the result of that list. The result may be
1630 used by the enclosing expression.
1632 Otherwise, the interpreter works left to right, from one expression to
1635 @node Byte Compiling, , Complications, Lisp Interpreter
1636 @subsection Byte Compiling
1637 @cindex Byte compiling
1639 One other aspect of interpreting: the Lisp interpreter is able to
1640 interpret two kinds of entity: humanly readable code, on which we will
1641 focus exclusively, and specially processed code, called @dfn{byte
1642 compiled} code, which is not humanly readable. Byte compiled code
1643 runs faster than humanly readable code.
1645 You can transform humanly readable code into byte compiled code by
1646 running one of the compile commands such as @code{byte-compile-file}.
1647 Byte compiled code is usually stored in a file that ends with a
1648 @file{.elc} extension rather than a @file{.el} extension. You will
1649 see both kinds of file in the @file{emacs/lisp} directory; the files
1650 to read are those with @file{.el} extensions.
1652 As a practical matter, for most things you might do to customize or
1653 extend Emacs, you do not need to byte compile; and I will not discuss
1654 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1655 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1658 @node Evaluation, Variables, Lisp Interpreter, List Processing
1659 @comment node-name, next, previous, up
1663 When the Lisp interpreter works on an expression, the term for the
1664 activity is called @dfn{evaluation}. We say that the interpreter
1665 `evaluates the expression'. I've used this term several times before.
1666 The word comes from its use in everyday language, `to ascertain the
1667 value or amount of; to appraise', according to @cite{Webster's New
1668 Collegiate Dictionary}.
1671 * How the Interpreter Acts:: Returns and Side Effects...
1672 * Evaluating Inner Lists:: Lists within lists...
1675 @node How the Interpreter Acts, Evaluating Inner Lists, Evaluation, Evaluation
1677 @unnumberedsubsec How the Lisp Interpreter Acts
1680 @cindex @samp{returned value} explained
1681 After evaluating an expression, the Lisp interpreter will most likely
1682 @dfn{return} the value that the computer produces by carrying out the
1683 instructions it found in the function definition, or perhaps it will
1684 give up on that function and produce an error message. (The interpreter
1685 may also find itself tossed, so to speak, to a different function or it
1686 may attempt to repeat continually what it is doing for ever and ever in
1687 what is called an `infinite loop'. These actions are less common; and
1688 we can ignore them.) Most frequently, the interpreter returns a value.
1690 @cindex @samp{side effect} defined
1691 At the same time the interpreter returns a value, it may do something
1692 else as well, such as move a cursor or copy a file; this other kind of
1693 action is called a @dfn{side effect}. Actions that we humans think are
1694 important, such as printing results, are often ``side effects'' to the
1695 Lisp interpreter. The jargon can sound peculiar, but it turns out that
1696 it is fairly easy to learn to use side effects.
1698 In summary, evaluating a symbolic expression most commonly causes the
1699 Lisp interpreter to return a value and perhaps carry out a side effect;
1700 or else produce an error.
1702 @node Evaluating Inner Lists, , How the Interpreter Acts, Evaluation
1703 @comment node-name, next, previous, up
1704 @subsection Evaluating Inner Lists
1705 @cindex Inner list evaluation
1706 @cindex Evaluating inner lists
1708 If evaluation applies to a list that is inside another list, the outer
1709 list may use the value returned by the first evaluation as information
1710 when the outer list is evaluated. This explains why inner expressions
1711 are evaluated first: the values they return are used by the outer
1715 We can investigate this process by evaluating another addition example.
1716 Place your cursor after the following expression and type @kbd{C-x C-e}:
1723 The number 8 will appear in the echo area.
1725 What happens is that the Lisp interpreter first evaluates the inner
1726 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1727 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1728 returns the value 8. Since there are no more enclosing expressions to
1729 evaluate, the interpreter prints that value in the echo area.
1731 Now it is easy to understand the name of the command invoked by the
1732 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1733 letters @code{sexp} are an abbreviation for `symbolic expression', and
1734 @code{eval} is an abbreviation for `evaluate'. The command means
1735 `evaluate last symbolic expression'.
1737 As an experiment, you can try evaluating the expression by putting the
1738 cursor at the beginning of the next line immediately following the
1739 expression, or inside the expression.
1742 Here is another copy of the expression:
1749 If you place the cursor at the beginning of the blank line that
1750 immediately follows the expression and type @kbd{C-x C-e}, you will
1751 still get the value 8 printed in the echo area. Now try putting the
1752 cursor inside the expression. If you put it right after the next to
1753 last parenthesis (so it appears to sit on top of the last parenthesis),
1754 you will get a 6 printed in the echo area! This is because the command
1755 evaluates the expression @code{(+ 3 3)}.
1757 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1758 you will get the number itself. In Lisp, if you evaluate a number, you
1759 get the number itself---this is how numbers differ from symbols. If you
1760 evaluate a list starting with a symbol like @code{+}, you will get a
1761 value returned that is the result of the computer carrying out the
1762 instructions in the function definition attached to that name. If a
1763 symbol by itself is evaluated, something different happens, as we will
1764 see in the next section.
1766 @node Variables, Arguments, Evaluation, List Processing
1767 @comment node-name, next, previous, up
1771 In Emacs Lisp, a symbol can have a value attached to it just as it can
1772 have a function definition attached to it. The two are different.
1773 The function definition is a set of instructions that a computer will
1774 obey. A value, on the other hand, is something, such as number or a
1775 name, that can vary (which is why such a symbol is called a variable).
1776 The value of a symbol can be any expression in Lisp, such as a symbol,
1777 number, list, or string. A symbol that has a value is often called a
1780 A symbol can have both a function definition and a value attached to
1781 it at the same time. Or it can have just one or the other.
1782 The two are separate. This is somewhat similar
1783 to the way the name Cambridge can refer to the city in Massachusetts
1784 and have some information attached to the name as well, such as
1785 ``great programming center''.
1788 (Incidentally, in Emacs Lisp, a symbol can have two
1789 other things attached to it, too: a property list and a documentation
1790 string; these are discussed later.)
1793 Another way to think about this is to imagine a symbol as being a chest
1794 of drawers. The function definition is put in one drawer, the value in
1795 another, and so on. What is put in the drawer holding the value can be
1796 changed without affecting the contents of the drawer holding the
1797 function definition, and vice-verse.
1800 * fill-column Example::
1801 * Void Function:: The error message for a symbol
1803 * Void Variable:: The error message for a symbol without a value.
1806 @node fill-column Example, Void Function, Variables, Variables
1808 @unnumberedsubsec @code{fill-column}, an Example Variable
1811 @findex fill-column, @r{an example variable}
1812 @cindex Example variable, @code{fill-column}
1813 @cindex Variable, example of, @code{fill-column}
1814 The variable @code{fill-column} illustrates a symbol with a value
1815 attached to it: in every GNU Emacs buffer, this symbol is set to some
1816 value, usually 72 or 70, but sometimes to some other value. To find the
1817 value of this symbol, evaluate it by itself. If you are reading this in
1818 Info inside of GNU Emacs, you can do this by putting the cursor after
1819 the symbol and typing @kbd{C-x C-e}:
1826 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1827 area. This is the value for which @code{fill-column} is set for me as I
1828 write this. It may be different for you in your Info buffer. Notice
1829 that the value returned as a variable is printed in exactly the same way
1830 as the value returned by a function carrying out its instructions. From
1831 the point of view of the Lisp interpreter, a value returned is a value
1832 returned. What kind of expression it came from ceases to matter once
1835 A symbol can have any value attached to it or, to use the jargon, we can
1836 @dfn{bind} the variable to a value: to a number, such as 72; to a
1837 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1838 oak)}; we can even bind a variable to a function definition.
1840 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1841 Setting the Value of a Variable}, for information about one way to do
1844 @node Void Function, Void Variable, fill-column Example, Variables
1845 @comment node-name, next, previous, up
1846 @subsection Error Message for a Symbol Without a Function
1847 @cindex Symbol without function error
1848 @cindex Error for symbol without function
1850 When we evaluated @code{fill-column} to find its value as a variable,
1851 we did not place parentheses around the word. This is because we did
1852 not intend to use it as a function name.
1854 If @code{fill-column} were the first or only element of a list, the
1855 Lisp interpreter would attempt to find the function definition
1856 attached to it. But @code{fill-column} has no function definition.
1857 Try evaluating this:
1865 In GNU Emacs version 22, you will create a @file{*Backtrace*} buffer
1870 ---------- Buffer: *Backtrace* ----------
1871 Debugger entered--Lisp error: (void-function fill-column)
1874 eval-last-sexp-1(nil)
1876 call-interactively(eval-last-sexp)
1877 ---------- Buffer: *Backtrace* ----------
1882 (Remember, to quit the debugger and make the debugger window go away,
1883 type @kbd{q} in the @file{*Backtrace*} buffer.)
1887 In GNU Emacs 20 and before, you will produce an error message that says:
1890 Symbol's function definition is void:@: fill-column
1894 (The message will go away as soon as you move the cursor or type
1898 @node Void Variable, , Void Function, Variables
1899 @comment node-name, next, previous, up
1900 @subsection Error Message for a Symbol Without a Value
1901 @cindex Symbol without value error
1902 @cindex Error for symbol without value
1904 If you attempt to evaluate a symbol that does not have a value bound to
1905 it, you will receive an error message. You can see this by
1906 experimenting with our 2 plus 2 addition. In the following expression,
1907 put your cursor right after the @code{+}, before the first number 2,
1916 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1921 ---------- Buffer: *Backtrace* ----------
1922 Debugger entered--Lisp error: (void-variable +)
1924 eval-last-sexp-1(nil)
1926 call-interactively(eval-last-sexp)
1927 ---------- Buffer: *Backtrace* ----------
1932 (As with the other times we entered the debugger, you can quit by
1933 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1935 This backtrace is different from the very first error message we saw,
1936 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1937 In this case, the function does not have a value as a variable; while
1938 in the other error message, the function (the word `this') did not
1941 In this experiment with the @code{+}, what we did was cause the Lisp
1942 interpreter to evaluate the @code{+} and look for the value of the
1943 variable instead of the function definition. We did this by placing the
1944 cursor right after the symbol rather than after the parenthesis of the
1945 enclosing list as we did before. As a consequence, the Lisp interpreter
1946 evaluated the preceding s-expression, which in this case was the
1949 Since @code{+} does not have a value bound to it, just the function
1950 definition, the error message reported that the symbol's value as a
1955 In GNU Emacs version 20 and before, your error message will say:
1958 Symbol's value as variable is void:@: +
1962 The meaning is the same as in GNU Emacs 22.
1965 @node Arguments, set & setq, Variables, List Processing
1966 @comment node-name, next, previous, up
1969 @cindex Passing information to functions
1971 To see how information is passed to functions, let's look again at
1972 our old standby, the addition of two plus two. In Lisp, this is written
1979 If you evaluate this expression, the number 4 will appear in your echo
1980 area. What the Lisp interpreter does is add the numbers that follow
1983 @cindex @samp{argument} defined
1984 The numbers added by @code{+} are called the @dfn{arguments} of the
1985 function @code{+}. These numbers are the information that is given to
1986 or @dfn{passed} to the function.
1988 The word `argument' comes from the way it is used in mathematics and
1989 does not refer to a disputation between two people; instead it refers to
1990 the information presented to the function, in this case, to the
1991 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1992 that follow the function. The values returned by the evaluation of
1993 these atoms or lists are passed to the function. Different functions
1994 require different numbers of arguments; some functions require none at
1995 all.@footnote{It is curious to track the path by which the word `argument'
1996 came to have two different meanings, one in mathematics and the other in
1997 everyday English. According to the @cite{Oxford English Dictionary},
1998 the word derives from the Latin for @samp{to make clear, prove}; thus it
1999 came to mean, by one thread of derivation, `the evidence offered as
2000 proof', which is to say, `the information offered', which led to its
2001 meaning in Lisp. But in the other thread of derivation, it came to mean
2002 `to assert in a manner against which others may make counter
2003 assertions', which led to the meaning of the word as a disputation.
2004 (Note here that the English word has two different definitions attached
2005 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
2006 have two different function definitions at the same time.)}
2009 * Data types:: Types of data passed to a function.
2010 * Args as Variable or List:: An argument can be the value
2011 of a variable or list.
2012 * Variable Number of Arguments:: Some functions may take a
2013 variable number of arguments.
2014 * Wrong Type of Argument:: Passing an argument of the wrong type
2016 * message:: A useful function for sending messages.
2019 @node Data types, Args as Variable or List, Arguments, Arguments
2020 @comment node-name, next, previous, up
2021 @subsection Arguments' Data Types
2023 @cindex Types of data
2024 @cindex Arguments' data types
2026 The type of data that should be passed to a function depends on what
2027 kind of information it uses. The arguments to a function such as
2028 @code{+} must have values that are numbers, since @code{+} adds numbers.
2029 Other functions use different kinds of data for their arguments.
2033 For example, the @code{concat} function links together or unites two or
2034 more strings of text to produce a string. The arguments are strings.
2035 Concatenating the two character strings @code{abc}, @code{def} produces
2036 the single string @code{abcdef}. This can be seen by evaluating the
2040 (concat "abc" "def")
2044 The value produced by evaluating this expression is @code{"abcdef"}.
2046 A function such as @code{substring} uses both a string and numbers as
2047 arguments. The function returns a part of the string, a substring of
2048 the first argument. This function takes three arguments. Its first
2049 argument is the string of characters, the second and third arguments are
2050 numbers that indicate the beginning and end of the substring. The
2051 numbers are a count of the number of characters (including spaces and
2052 punctuation) from the beginning of the string.
2055 For example, if you evaluate the following:
2058 (substring "The quick brown fox jumped." 16 19)
2062 you will see @code{"fox"} appear in the echo area. The arguments are the
2063 string and the two numbers.
2065 Note that the string passed to @code{substring} is a single atom even
2066 though it is made up of several words separated by spaces. Lisp counts
2067 everything between the two quotation marks as part of the string,
2068 including the spaces. You can think of the @code{substring} function as
2069 a kind of `atom smasher' since it takes an otherwise indivisible atom
2070 and extracts a part. However, @code{substring} is only able to extract
2071 a substring from an argument that is a string, not from another type of
2072 atom such as a number or symbol.
2074 @node Args as Variable or List, Variable Number of Arguments, Data types, Arguments
2075 @comment node-name, next, previous, up
2076 @subsection An Argument as the Value of a Variable or List
2078 An argument can be a symbol that returns a value when it is evaluated.
2079 For example, when the symbol @code{fill-column} by itself is evaluated,
2080 it returns a number. This number can be used in an addition.
2083 Position the cursor after the following expression and type @kbd{C-x
2091 The value will be a number two more than what you get by evaluating
2092 @code{fill-column} alone. For me, this is 74, because my value of
2093 @code{fill-column} is 72.
2095 As we have just seen, an argument can be a symbol that returns a value
2096 when evaluated. In addition, an argument can be a list that returns a
2097 value when it is evaluated. For example, in the following expression,
2098 the arguments to the function @code{concat} are the strings
2099 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2100 @code{(number-to-string (+ 2 fill-column))}.
2102 @c For GNU Emacs 22, need number-to-string
2104 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2108 If you evaluate this expression---and if, as with my Emacs,
2109 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2110 appear in the echo area. (Note that you must put spaces after the
2111 word @samp{The} and before the word @samp{red} so they will appear in
2112 the final string. The function @code{number-to-string} converts the
2113 integer that the addition function returns to a string.
2114 @code{number-to-string} is also known as @code{int-to-string}.)
2116 @node Variable Number of Arguments, Wrong Type of Argument, Args as Variable or List, Arguments
2117 @comment node-name, next, previous, up
2118 @subsection Variable Number of Arguments
2119 @cindex Variable number of arguments
2120 @cindex Arguments, variable number of
2122 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2123 number of arguments. (The @code{*} is the symbol for multiplication.)
2124 This can be seen by evaluating each of the following expressions in
2125 the usual way. What you will see in the echo area is printed in this
2126 text after @samp{@result{}}, which you may read as `evaluates to'.
2129 In the first set, the functions have no arguments:
2140 In this set, the functions have one argument each:
2151 In this set, the functions have three arguments each:
2155 (+ 3 4 5) @result{} 12
2157 (* 3 4 5) @result{} 60
2161 @node Wrong Type of Argument, message, Variable Number of Arguments, Arguments
2162 @comment node-name, next, previous, up
2163 @subsection Using the Wrong Type Object as an Argument
2164 @cindex Wrong type of argument
2165 @cindex Argument, wrong type of
2167 When a function is passed an argument of the wrong type, the Lisp
2168 interpreter produces an error message. For example, the @code{+}
2169 function expects the values of its arguments to be numbers. As an
2170 experiment we can pass it the quoted symbol @code{hello} instead of a
2171 number. Position the cursor after the following expression and type
2179 When you do this you will generate an error message. What has happened
2180 is that @code{+} has tried to add the 2 to the value returned by
2181 @code{'hello}, but the value returned by @code{'hello} is the symbol
2182 @code{hello}, not a number. Only numbers can be added. So @code{+}
2183 could not carry out its addition.
2186 In GNU Emacs version 22, you will create and enter a
2187 @file{*Backtrace*} buffer that says:
2192 ---------- Buffer: *Backtrace* ----------
2193 Debugger entered--Lisp error:
2194 (wrong-type-argument number-or-marker-p hello)
2196 eval((+ 2 (quote hello)))
2197 eval-last-sexp-1(nil)
2199 call-interactively(eval-last-sexp)
2200 ---------- Buffer: *Backtrace* ----------
2205 As usual, the error message tries to be helpful and makes sense after you
2206 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2207 the abbreviation @code{'hello}.}
2209 The first part of the error message is straightforward; it says
2210 @samp{wrong type argument}. Next comes the mysterious jargon word
2211 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2212 kind of argument the @code{+} expected.
2214 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2215 trying to determine whether the information presented it (the value of
2216 the argument) is a number or a marker (a special object representing a
2217 buffer position). What it does is test to see whether the @code{+} is
2218 being given numbers to add. It also tests to see whether the
2219 argument is something called a marker, which is a specific feature of
2220 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2221 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2222 its position is kept as a marker. The mark can be considered a
2223 number---the number of characters the location is from the beginning
2224 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2225 numeric value of marker positions as numbers.
2227 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2228 practice started in the early days of Lisp programming. The @samp{p}
2229 stands for `predicate'. In the jargon used by the early Lisp
2230 researchers, a predicate refers to a function to determine whether some
2231 property is true or false. So the @samp{p} tells us that
2232 @code{number-or-marker-p} is the name of a function that determines
2233 whether it is true or false that the argument supplied is a number or
2234 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2235 a function that tests whether its argument has the value of zero, and
2236 @code{listp}, a function that tests whether its argument is a list.
2238 Finally, the last part of the error message is the symbol @code{hello}.
2239 This is the value of the argument that was passed to @code{+}. If the
2240 addition had been passed the correct type of object, the value passed
2241 would have been a number, such as 37, rather than a symbol like
2242 @code{hello}. But then you would not have got the error message.
2246 In GNU Emacs version 20 and before, the echo area displays an error
2250 Wrong type argument:@: number-or-marker-p, hello
2253 This says, in different words, the same as the top line of the
2254 @file{*Backtrace*} buffer.
2257 @node message, , Wrong Type of Argument, Arguments
2258 @comment node-name, next, previous, up
2259 @subsection The @code{message} Function
2262 Like @code{+}, the @code{message} function takes a variable number of
2263 arguments. It is used to send messages to the user and is so useful
2264 that we will describe it here.
2267 A message is printed in the echo area. For example, you can print a
2268 message in your echo area by evaluating the following list:
2271 (message "This message appears in the echo area!")
2274 The whole string between double quotation marks is a single argument
2275 and is printed @i{in toto}. (Note that in this example, the message
2276 itself will appear in the echo area within double quotes; that is
2277 because you see the value returned by the @code{message} function. In
2278 most uses of @code{message} in programs that you write, the text will
2279 be printed in the echo area as a side-effect, without the quotes.
2280 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2281 detail}, for an example of this.)
2283 However, if there is a @samp{%s} in the quoted string of characters, the
2284 @code{message} function does not print the @samp{%s} as such, but looks
2285 to the argument that follows the string. It evaluates the second
2286 argument and prints the value at the location in the string where the
2290 You can see this by positioning the cursor after the following
2291 expression and typing @kbd{C-x C-e}:
2294 (message "The name of this buffer is: %s." (buffer-name))
2298 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2299 echo area. The function @code{buffer-name} returns the name of the
2300 buffer as a string, which the @code{message} function inserts in place
2303 To print a value as an integer, use @samp{%d} in the same way as
2304 @samp{%s}. For example, to print a message in the echo area that
2305 states the value of the @code{fill-column}, evaluate the following:
2308 (message "The value of fill-column is %d." fill-column)
2312 On my system, when I evaluate this list, @code{"The value of
2313 fill-column is 72."} appears in my echo area@footnote{Actually, you
2314 can use @code{%s} to print a number. It is non-specific. @code{%d}
2315 prints only the part of a number left of a decimal point, and not
2316 anything that is not a number.}.
2318 If there is more than one @samp{%s} in the quoted string, the value of
2319 the first argument following the quoted string is printed at the
2320 location of the first @samp{%s} and the value of the second argument is
2321 printed at the location of the second @samp{%s}, and so on.
2324 For example, if you evaluate the following,
2328 (message "There are %d %s in the office!"
2329 (- fill-column 14) "pink elephants")
2334 a rather whimsical message will appear in your echo area. On my system
2335 it says, @code{"There are 58 pink elephants in the office!"}.
2337 The expression @code{(- fill-column 14)} is evaluated and the resulting
2338 number is inserted in place of the @samp{%d}; and the string in double
2339 quotes, @code{"pink elephants"}, is treated as a single argument and
2340 inserted in place of the @samp{%s}. (That is to say, a string between
2341 double quotes evaluates to itself, like a number.)
2343 Finally, here is a somewhat complex example that not only illustrates
2344 the computation of a number, but also shows how you can use an
2345 expression within an expression to generate the text that is substituted
2350 (message "He saw %d %s"
2354 "The quick brown foxes jumped." 16 21)
2359 In this example, @code{message} has three arguments: the string,
2360 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2361 the expression beginning with the function @code{concat}. The value
2362 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2363 in place of the @samp{%d}; and the value returned by the expression
2364 beginning with @code{concat} is inserted in place of the @samp{%s}.
2366 When your fill column is 70 and you evaluate the expression, the
2367 message @code{"He saw 38 red foxes leaping."} appears in your echo
2370 @node set & setq, Summary, Arguments, List Processing
2371 @comment node-name, next, previous, up
2372 @section Setting the Value of a Variable
2373 @cindex Variable, setting value
2374 @cindex Setting value of variable
2376 @cindex @samp{bind} defined
2377 There are several ways by which a variable can be given a value. One of
2378 the ways is to use either the function @code{set} or the function
2379 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2380 jargon for this process is to @dfn{bind} a variable to a value.)
2382 The following sections not only describe how @code{set} and @code{setq}
2383 work but also illustrate how arguments are passed.
2386 * Using set:: Setting values.
2387 * Using setq:: Setting a quoted value.
2388 * Counting:: Using @code{setq} to count.
2391 @node Using set, Using setq, set & setq, set & setq
2392 @comment node-name, next, previous, up
2393 @subsection Using @code{set}
2396 To set the value of the symbol @code{flowers} to the list @code{'(rose
2397 violet daisy buttercup)}, evaluate the following expression by
2398 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2401 (set 'flowers '(rose violet daisy buttercup))
2405 The list @code{(rose violet daisy buttercup)} will appear in the echo
2406 area. This is what is @emph{returned} by the @code{set} function. As a
2407 side effect, the symbol @code{flowers} is bound to the list; that is,
2408 the symbol @code{flowers}, which can be viewed as a variable, is given
2409 the list as its value. (This process, by the way, illustrates how a
2410 side effect to the Lisp interpreter, setting the value, can be the
2411 primary effect that we humans are interested in. This is because every
2412 Lisp function must return a value if it does not get an error, but it
2413 will only have a side effect if it is designed to have one.)
2415 After evaluating the @code{set} expression, you can evaluate the symbol
2416 @code{flowers} and it will return the value you just set. Here is the
2417 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2424 When you evaluate @code{flowers}, the list
2425 @code{(rose violet daisy buttercup)} appears in the echo area.
2427 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2428 in front of it, what you will see in the echo area is the symbol itself,
2429 @code{flowers}. Here is the quoted symbol, so you can try this:
2435 Note also, that when you use @code{set}, you need to quote both
2436 arguments to @code{set}, unless you want them evaluated. Since we do
2437 not want either argument evaluated, neither the variable
2438 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2439 are quoted. (When you use @code{set} without quoting its first
2440 argument, the first argument is evaluated before anything else is
2441 done. If you did this and @code{flowers} did not have a value
2442 already, you would get an error message that the @samp{Symbol's value
2443 as variable is void}; on the other hand, if @code{flowers} did return
2444 a value after it was evaluated, the @code{set} would attempt to set
2445 the value that was returned. There are situations where this is the
2446 right thing for the function to do; but such situations are rare.)
2448 @node Using setq, Counting, Using set, set & setq
2449 @comment node-name, next, previous, up
2450 @subsection Using @code{setq}
2453 As a practical matter, you almost always quote the first argument to
2454 @code{set}. The combination of @code{set} and a quoted first argument
2455 is so common that it has its own name: the special form @code{setq}.
2456 This special form is just like @code{set} except that the first argument
2457 is quoted automatically, so you don't need to type the quote mark
2458 yourself. Also, as an added convenience, @code{setq} permits you to set
2459 several different variables to different values, all in one expression.
2461 To set the value of the variable @code{carnivores} to the list
2462 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2466 (setq carnivores '(lion tiger leopard))
2470 This is exactly the same as using @code{set} except the first argument
2471 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2472 means @code{quote}.)
2475 With @code{set}, the expression would look like this:
2478 (set 'carnivores '(lion tiger leopard))
2481 Also, @code{setq} can be used to assign different values to
2482 different variables. The first argument is bound to the value
2483 of the second argument, the third argument is bound to the value of the
2484 fourth argument, and so on. For example, you could use the following to
2485 assign a list of trees to the symbol @code{trees} and a list of herbivores
2486 to the symbol @code{herbivores}:
2490 (setq trees '(pine fir oak maple)
2491 herbivores '(gazelle antelope zebra))
2496 (The expression could just as well have been on one line, but it might
2497 not have fit on a page; and humans find it easier to read nicely
2500 Although I have been using the term `assign', there is another way of
2501 thinking about the workings of @code{set} and @code{setq}; and that is to
2502 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2503 list. This latter way of thinking is very common and in forthcoming
2504 chapters we shall come upon at least one symbol that has `pointer' as
2505 part of its name. The name is chosen because the symbol has a value,
2506 specifically a list, attached to it; or, expressed another way,
2507 the symbol is set to ``point'' to the list.
2509 @node Counting, , Using setq, set & setq
2510 @comment node-name, next, previous, up
2511 @subsection Counting
2514 Here is an example that shows how to use @code{setq} in a counter. You
2515 might use this to count how many times a part of your program repeats
2516 itself. First set a variable to zero; then add one to the number each
2517 time the program repeats itself. To do this, you need a variable that
2518 serves as a counter, and two expressions: an initial @code{setq}
2519 expression that sets the counter variable to zero; and a second
2520 @code{setq} expression that increments the counter each time it is
2525 (setq counter 0) ; @r{Let's call this the initializer.}
2527 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2529 counter ; @r{This is the counter.}
2534 (The text following the @samp{;} are comments. @xref{Change a
2535 defun, , Change a Function Definition}.)
2537 If you evaluate the first of these expressions, the initializer,
2538 @code{(setq counter 0)}, and then evaluate the third expression,
2539 @code{counter}, the number @code{0} will appear in the echo area. If
2540 you then evaluate the second expression, the incrementer, @code{(setq
2541 counter (+ counter 1))}, the counter will get the value 1. So if you
2542 again evaluate @code{counter}, the number @code{1} will appear in the
2543 echo area. Each time you evaluate the second expression, the value of
2544 the counter will be incremented.
2546 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2547 the Lisp interpreter first evaluates the innermost list; this is the
2548 addition. In order to evaluate this list, it must evaluate the variable
2549 @code{counter} and the number @code{1}. When it evaluates the variable
2550 @code{counter}, it receives its current value. It passes this value and
2551 the number @code{1} to the @code{+} which adds them together. The sum
2552 is then returned as the value of the inner list and passed to the
2553 @code{setq} which sets the variable @code{counter} to this new value.
2554 Thus, the value of the variable, @code{counter}, is changed.
2556 @node Summary, Error Message Exercises, set & setq, List Processing
2557 @comment node-name, next, previous, up
2560 Learning Lisp is like climbing a hill in which the first part is the
2561 steepest. You have now climbed the most difficult part; what remains
2562 becomes easier as you progress onwards.
2570 Lisp programs are made up of expressions, which are lists or single atoms.
2573 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2574 surrounded by parentheses. A list can be empty.
2577 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2578 character symbols like @code{+}, strings of characters between double
2579 quotation marks, or numbers.
2582 A number evaluates to itself.
2585 A string between double quotes also evaluates to itself.
2588 When you evaluate a symbol by itself, its value is returned.
2591 When you evaluate a list, the Lisp interpreter looks at the first symbol
2592 in the list and then at the function definition bound to that symbol.
2593 Then the instructions in the function definition are carried out.
2596 A single quotation mark,
2603 , tells the Lisp interpreter that it should
2604 return the following expression as written, and not evaluate it as it
2605 would if the quote were not there.
2608 Arguments are the information passed to a function. The arguments to a
2609 function are computed by evaluating the rest of the elements of the list
2610 of which the function is the first element.
2613 A function always returns a value when it is evaluated (unless it gets
2614 an error); in addition, it may also carry out some action called a
2615 ``side effect''. In many cases, a function's primary purpose is to
2616 create a side effect.
2619 @node Error Message Exercises, , Summary, List Processing
2620 @comment node-name, next, previous, up
2623 A few simple exercises:
2627 Generate an error message by evaluating an appropriate symbol that is
2628 not within parentheses.
2631 Generate an error message by evaluating an appropriate symbol that is
2632 between parentheses.
2635 Create a counter that increments by two rather than one.
2638 Write an expression that prints a message in the echo area when
2642 @node Practicing Evaluation, Writing Defuns, List Processing, Top
2643 @comment node-name, next, previous, up
2644 @chapter Practicing Evaluation
2645 @cindex Practicing evaluation
2646 @cindex Evaluation practice
2648 Before learning how to write a function definition in Emacs Lisp, it is
2649 useful to spend a little time evaluating various expressions that have
2650 already been written. These expressions will be lists with the
2651 functions as their first (and often only) element. Since some of the
2652 functions associated with buffers are both simple and interesting, we
2653 will start with those. In this section, we will evaluate a few of
2654 these. In another section, we will study the code of several other
2655 buffer-related functions, to see how they were written.
2658 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2660 * Buffer Names:: Buffers and files are different.
2661 * Getting Buffers:: Getting a buffer itself, not merely its name.
2662 * Switching Buffers:: How to change to another buffer.
2663 * Buffer Size & Locations:: Where point is located and the size of
2665 * Evaluation Exercise::
2668 @node How to Evaluate, Buffer Names, Practicing Evaluation, Practicing Evaluation
2670 @unnumberedsec How to Evaluate
2673 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2674 command to move the cursor or to scroll the screen, @i{you are evaluating
2675 an expression,} the first element of which is a function. @i{This is
2678 @cindex @samp{interactive function} defined
2679 @cindex @samp{command} defined
2680 When you type keys, you cause the Lisp interpreter to evaluate an
2681 expression and that is how you get your results. Even typing plain text
2682 involves evaluating an Emacs Lisp function, in this case, one that uses
2683 @code{self-insert-command}, which simply inserts the character you
2684 typed. The functions you evaluate by typing keystrokes are called
2685 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2686 interactive will be illustrated in the chapter on how to write function
2687 definitions. @xref{Interactive, , Making a Function Interactive}.
2689 In addition to typing keyboard commands, we have seen a second way to
2690 evaluate an expression: by positioning the cursor after a list and
2691 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2692 section. There are other ways to evaluate an expression as well; these
2693 will be described as we come to them.
2695 Besides being used for practicing evaluation, the functions shown in the
2696 next few sections are important in their own right. A study of these
2697 functions makes clear the distinction between buffers and files, how to
2698 switch to a buffer, and how to determine a location within it.
2700 @node Buffer Names, Getting Buffers, How to Evaluate, Practicing Evaluation
2701 @comment node-name, next, previous, up
2702 @section Buffer Names
2704 @findex buffer-file-name
2706 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2707 the difference between a file and a buffer. When you evaluate the
2708 following expression, @code{(buffer-name)}, the name of the buffer
2709 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2710 the name of the file to which the buffer refers appears in the echo
2711 area. Usually, the name returned by @code{(buffer-name)} is the same as
2712 the name of the file to which it refers, and the name returned by
2713 @code{(buffer-file-name)} is the full path-name of the file.
2715 A file and a buffer are two different entities. A file is information
2716 recorded permanently in the computer (unless you delete it). A buffer,
2717 on the other hand, is information inside of Emacs that will vanish at
2718 the end of the editing session (or when you kill the buffer). Usually,
2719 a buffer contains information that you have copied from a file; we say
2720 the buffer is @dfn{visiting} that file. This copy is what you work on
2721 and modify. Changes to the buffer do not change the file, until you
2722 save the buffer. When you save the buffer, the buffer is copied to the file
2723 and is thus saved permanently.
2726 If you are reading this in Info inside of GNU Emacs, you can evaluate
2727 each of the following expressions by positioning the cursor after it and
2728 typing @kbd{C-x C-e}.
2739 When I do this in Info, the value returned by evaluating
2740 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2741 evaluating @code{(buffer-file-name)} is @file{nil}.
2743 On the other hand, while I am writing this document, the value
2744 returned by evaluating @code{(buffer-name)} is
2745 @file{"introduction.texinfo"}, and the value returned by evaluating
2746 @code{(buffer-file-name)} is
2747 @file{"/gnu/work/intro/introduction.texinfo"}.
2749 @cindex @code{nil}, history of word
2750 The former is the name of the buffer and the latter is the name of the
2751 file. In Info, the buffer name is @file{"*info*"}. Info does not
2752 point to any file, so the result of evaluating
2753 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2754 from the Latin word for `nothing'; in this case, it means that the
2755 buffer is not associated with any file. (In Lisp, @code{nil} is also
2756 used to mean `false' and is a synonym for the empty list, @code{()}.)
2758 When I am writing, the name of my buffer is
2759 @file{"introduction.texinfo"}. The name of the file to which it
2760 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2762 (In the expressions, the parentheses tell the Lisp interpreter to
2763 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2764 functions; without the parentheses, the interpreter would attempt to
2765 evaluate the symbols as variables. @xref{Variables}.)
2767 In spite of the distinction between files and buffers, you will often
2768 find that people refer to a file when they mean a buffer and vice-verse.
2769 Indeed, most people say, ``I am editing a file,'' rather than saying,
2770 ``I am editing a buffer which I will soon save to a file.'' It is
2771 almost always clear from context what people mean. When dealing with
2772 computer programs, however, it is important to keep the distinction in mind,
2773 since the computer is not as smart as a person.
2775 @cindex Buffer, history of word
2776 The word `buffer', by the way, comes from the meaning of the word as a
2777 cushion that deadens the force of a collision. In early computers, a
2778 buffer cushioned the interaction between files and the computer's
2779 central processing unit. The drums or tapes that held a file and the
2780 central processing unit were pieces of equipment that were very
2781 different from each other, working at their own speeds, in spurts. The
2782 buffer made it possible for them to work together effectively.
2783 Eventually, the buffer grew from being an intermediary, a temporary
2784 holding place, to being the place where work is done. This
2785 transformation is rather like that of a small seaport that grew into a
2786 great city: once it was merely the place where cargo was warehoused
2787 temporarily before being loaded onto ships; then it became a business
2788 and cultural center in its own right.
2790 Not all buffers are associated with files. For example, a
2791 @file{*scratch*} buffer does not visit any file. Similarly, a
2792 @file{*Help*} buffer is not associated with any file.
2794 In the old days, when you lacked a @file{~/.emacs} file and started an
2795 Emacs session by typing the command @code{emacs} alone, without naming
2796 any files, Emacs started with the @file{*scratch*} buffer visible.
2797 Nowadays, you will see a splash screen. You can follow one of the
2798 commands suggested on the splash screen, visit a file, or press the
2799 spacebar to reach the @file{*scratch*} buffer.
2801 If you switch to the @file{*scratch*} buffer, type
2802 @code{(buffer-name)}, position the cursor after it, and then type
2803 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2804 will be returned and will appear in the echo area. @code{"*scratch*"}
2805 is the name of the buffer. When you type @code{(buffer-file-name)} in
2806 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2807 in the echo area, just as it does when you evaluate
2808 @code{(buffer-file-name)} in Info.
2810 Incidentally, if you are in the @file{*scratch*} buffer and want the
2811 value returned by an expression to appear in the @file{*scratch*}
2812 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2813 instead of @kbd{C-x C-e}. This causes the value returned to appear
2814 after the expression. The buffer will look like this:
2817 (buffer-name)"*scratch*"
2821 You cannot do this in Info since Info is read-only and it will not allow
2822 you to change the contents of the buffer. But you can do this in any
2823 buffer you can edit; and when you write code or documentation (such as
2824 this book), this feature is very useful.
2826 @node Getting Buffers, Switching Buffers, Buffer Names, Practicing Evaluation
2827 @comment node-name, next, previous, up
2828 @section Getting Buffers
2829 @findex current-buffer
2830 @findex other-buffer
2831 @cindex Getting a buffer
2833 The @code{buffer-name} function returns the @emph{name} of the buffer;
2834 to get the buffer @emph{itself}, a different function is needed: the
2835 @code{current-buffer} function. If you use this function in code, what
2836 you get is the buffer itself.
2838 A name and the object or entity to which the name refers are different
2839 from each other. You are not your name. You are a person to whom
2840 others refer by name. If you ask to speak to George and someone hands you
2841 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2842 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2843 not be satisfied. You do not want to speak to the name, but to the
2844 person to whom the name refers. A buffer is similar: the name of the
2845 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2846 get a buffer itself, you need to use a function such as
2847 @code{current-buffer}.
2849 However, there is a slight complication: if you evaluate
2850 @code{current-buffer} in an expression on its own, as we will do here,
2851 what you see is a printed representation of the name of the buffer
2852 without the contents of the buffer. Emacs works this way for two
2853 reasons: the buffer may be thousands of lines long---too long to be
2854 conveniently displayed; and, another buffer may have the same contents
2855 but a different name, and it is important to distinguish between them.
2858 Here is an expression containing the function:
2865 If you evaluate this expression in Info in Emacs in the usual way,
2866 @file{#<buffer *info*>} will appear in the echo area. The special
2867 format indicates that the buffer itself is being returned, rather than
2870 Incidentally, while you can type a number or symbol into a program, you
2871 cannot do that with the printed representation of a buffer: the only way
2872 to get a buffer itself is with a function such as @code{current-buffer}.
2874 A related function is @code{other-buffer}. This returns the most
2875 recently selected buffer other than the one you are in currently, not
2876 a printed representation of its name. If you have recently switched
2877 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2878 will return that buffer.
2881 You can see this by evaluating the expression:
2888 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2889 the name of whatever other buffer you switched back from most
2890 recently@footnote{Actually, by default, if the buffer from which you
2891 just switched is visible to you in another window, @code{other-buffer}
2892 will choose the most recent buffer that you cannot see; this is a
2893 subtlety that I often forget.}.
2895 @node Switching Buffers, Buffer Size & Locations, Getting Buffers, Practicing Evaluation
2896 @comment node-name, next, previous, up
2897 @section Switching Buffers
2898 @findex switch-to-buffer
2900 @cindex Switching to a buffer
2902 The @code{other-buffer} function actually provides a buffer when it is
2903 used as an argument to a function that requires one. We can see this
2904 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2907 But first, a brief introduction to the @code{switch-to-buffer}
2908 function. When you switched back and forth from Info to the
2909 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2910 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2911 rather, to save typing, you probably only typed @kbd{RET} if the
2912 default buffer was @file{*scratch*}, or if it was different, then you
2913 typed just part of the name, such as @code{*sc}, pressed your
2914 @kbd{TAB} key to cause it to expand to the full name, and then typed
2915 your @kbd{RET} key.} when prompted in the minibuffer for the name of
2916 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2917 b}, cause the Lisp interpreter to evaluate the interactive function
2918 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2919 different keystrokes call or run different functions. For example,
2920 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2921 @code{forward-sentence}, and so on.
2923 By writing @code{switch-to-buffer} in an expression, and giving it a
2924 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2928 Here is the Lisp expression:
2931 (switch-to-buffer (other-buffer))
2935 The symbol @code{switch-to-buffer} is the first element of the list,
2936 so the Lisp interpreter will treat it as a function and carry out the
2937 instructions that are attached to it. But before doing that, the
2938 interpreter will note that @code{other-buffer} is inside parentheses
2939 and work on that symbol first. @code{other-buffer} is the first (and
2940 in this case, the only) element of this list, so the Lisp interpreter
2941 calls or runs the function. It returns another buffer. Next, the
2942 interpreter runs @code{switch-to-buffer}, passing to it, as an
2943 argument, the other buffer, which is what Emacs will switch to. If
2944 you are reading this in Info, try this now. Evaluate the expression.
2945 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2946 expression will move you to your most recent other buffer that you
2947 cannot see. If you really want to go to your most recently selected
2948 buffer, even if you can still see it, you need to evaluate the
2949 following more complex expression:
2952 (switch-to-buffer (other-buffer (current-buffer) t))
2956 In this case, the first argument to @code{other-buffer} tells it which
2957 buffer to skip---the current one---and the second argument tells
2958 @code{other-buffer} it is OK to switch to a visible buffer.
2959 In regular use, @code{switch-to-buffer} takes you to an invisible
2960 window since you would most likely use @kbd{C-x o} (@code{other-window})
2961 to go to another visible buffer.}
2963 In the programming examples in later sections of this document, you will
2964 see the function @code{set-buffer} more often than
2965 @code{switch-to-buffer}. This is because of a difference between
2966 computer programs and humans: humans have eyes and expect to see the
2967 buffer on which they are working on their computer terminals. This is
2968 so obvious, it almost goes without saying. However, programs do not
2969 have eyes. When a computer program works on a buffer, that buffer does
2970 not need to be visible on the screen.
2972 @code{switch-to-buffer} is designed for humans and does two different
2973 things: it switches the buffer to which Emacs's attention is directed; and
2974 it switches the buffer displayed in the window to the new buffer.
2975 @code{set-buffer}, on the other hand, does only one thing: it switches
2976 the attention of the computer program to a different buffer. The buffer
2977 on the screen remains unchanged (of course, normally nothing happens
2978 there until the command finishes running).
2980 @cindex @samp{call} defined
2981 Also, we have just introduced another jargon term, the word @dfn{call}.
2982 When you evaluate a list in which the first symbol is a function, you
2983 are calling that function. The use of the term comes from the notion of
2984 the function as an entity that can do something for you if you `call'
2985 it---just as a plumber is an entity who can fix a leak if you call him
2988 @node Buffer Size & Locations, Evaluation Exercise, Switching Buffers, Practicing Evaluation
2989 @comment node-name, next, previous, up
2990 @section Buffer Size and the Location of Point
2991 @cindex Size of buffer
2993 @cindex Point location
2994 @cindex Location of point
2996 Finally, let's look at several rather simple functions,
2997 @code{buffer-size}, @code{point}, @code{point-min}, and
2998 @code{point-max}. These give information about the size of a buffer and
2999 the location of point within it.
3001 The function @code{buffer-size} tells you the size of the current
3002 buffer; that is, the function returns a count of the number of
3003 characters in the buffer.
3010 You can evaluate this in the usual way, by positioning the
3011 cursor after the expression and typing @kbd{C-x C-e}.
3013 @cindex @samp{point} defined
3014 In Emacs, the current position of the cursor is called @dfn{point}.
3015 The expression @code{(point)} returns a number that tells you where the
3016 cursor is located as a count of the number of characters from the
3017 beginning of the buffer up to point.
3020 You can see the character count for point in this buffer by evaluating
3021 the following expression in the usual way:
3028 As I write this, the value of @code{point} is 65724. The @code{point}
3029 function is frequently used in some of the examples later in this
3033 The value of point depends, of course, on its location within the
3034 buffer. If you evaluate point in this spot, the number will be larger:
3041 For me, the value of point in this location is 66043, which means that
3042 there are 319 characters (including spaces) between the two
3043 expressions. (Doubtless, you will see different numbers, since I will
3044 have edited this since I first evaluated point.)
3046 @cindex @samp{narrowing} defined
3047 The function @code{point-min} is somewhat similar to @code{point}, but
3048 it returns the value of the minimum permissible value of point in the
3049 current buffer. This is the number 1 unless @dfn{narrowing} is in
3050 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3051 or a program, to operations on just a part of a buffer.
3052 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3053 function @code{point-max} returns the value of the maximum permissible
3054 value of point in the current buffer.
3056 @node Evaluation Exercise, , Buffer Size & Locations, Practicing Evaluation
3059 Find a file with which you are working and move towards its middle.
3060 Find its buffer name, file name, length, and your position in the file.
3062 @node Writing Defuns, Buffer Walk Through, Practicing Evaluation, Top
3063 @comment node-name, next, previous, up
3064 @chapter How To Write Function Definitions
3065 @cindex Definition writing
3066 @cindex Function definition writing
3067 @cindex Writing a function definition
3069 When the Lisp interpreter evaluates a list, it looks to see whether the
3070 first symbol on the list has a function definition attached to it; or,
3071 put another way, whether the symbol points to a function definition. If
3072 it does, the computer carries out the instructions in the definition. A
3073 symbol that has a function definition is called, simply, a function
3074 (although, properly speaking, the definition is the function and the
3075 symbol refers to it.)
3078 * Primitive Functions::
3079 * defun:: The @code{defun} special form.
3080 * Install:: Install a function definition.
3081 * Interactive:: Making a function interactive.
3082 * Interactive Options:: Different options for @code{interactive}.
3083 * Permanent Installation:: Installing code permanently.
3084 * let:: Creating and initializing local variables.
3086 * else:: If--then--else expressions.
3087 * Truth & Falsehood:: What Lisp considers false and true.
3088 * save-excursion:: Keeping track of point, mark, and buffer.
3093 @node Primitive Functions, defun, Writing Defuns, Writing Defuns
3095 @unnumberedsec An Aside about Primitive Functions
3097 @cindex Primitive functions
3098 @cindex Functions, primitive
3100 @cindex C language primitives
3101 @cindex Primitives written in C
3102 All functions are defined in terms of other functions, except for a few
3103 @dfn{primitive} functions that are written in the C programming
3104 language. When you write functions' definitions, you will write them in
3105 Emacs Lisp and use other functions as your building blocks. Some of the
3106 functions you will use will themselves be written in Emacs Lisp (perhaps
3107 by you) and some will be primitives written in C. The primitive
3108 functions are used exactly like those written in Emacs Lisp and behave
3109 like them. They are written in C so we can easily run GNU Emacs on any
3110 computer that has sufficient power and can run C.
3112 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3113 distinguish between the use of functions written in C and the use of
3114 functions written in Emacs Lisp. The difference is irrelevant. I
3115 mention the distinction only because it is interesting to know. Indeed,
3116 unless you investigate, you won't know whether an already-written
3117 function is written in Emacs Lisp or C.
3119 @node defun, Install, Primitive Functions, Writing Defuns
3120 @comment node-name, next, previous, up
3121 @section The @code{defun} Special Form
3123 @cindex Special form of @code{defun}
3125 @cindex @samp{function definition} defined
3126 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3127 it that tells the computer what to do when the function is called.
3128 This code is called the @dfn{function definition} and is created by
3129 evaluating a Lisp expression that starts with the symbol @code{defun}
3130 (which is an abbreviation for @emph{define function}). Because
3131 @code{defun} does not evaluate its arguments in the usual way, it is
3132 called a @dfn{special form}.
3134 In subsequent sections, we will look at function definitions from the
3135 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3136 we will describe a simple function definition so you can see how it
3137 looks. This function definition uses arithmetic because it makes for a
3138 simple example. Some people dislike examples using arithmetic; however,
3139 if you are such a person, do not despair. Hardly any of the code we
3140 will study in the remainder of this introduction involves arithmetic or
3141 mathematics. The examples mostly involve text in one way or another.
3143 A function definition has up to five parts following the word
3148 The name of the symbol to which the function definition should be
3152 A list of the arguments that will be passed to the function. If no
3153 arguments will be passed to the function, this is an empty list,
3157 Documentation describing the function. (Technically optional, but
3158 strongly recommended.)
3161 Optionally, an expression to make the function interactive so you can
3162 use it by typing @kbd{M-x} and then the name of the function; or by
3163 typing an appropriate key or keychord.
3165 @cindex @samp{body} defined
3167 The code that instructs the computer what to do: the @dfn{body} of the
3168 function definition.
3171 It is helpful to think of the five parts of a function definition as
3172 being organized in a template, with slots for each part:
3176 (defun @var{function-name} (@var{arguments}@dots{})
3177 "@var{optional-documentation}@dots{}"
3178 (interactive @var{argument-passing-info}) ; @r{optional}
3183 As an example, here is the code for a function that multiplies its
3184 argument by 7. (This example is not interactive. @xref{Interactive,
3185 , Making a Function Interactive}, for that information.)
3189 (defun multiply-by-seven (number)
3190 "Multiply NUMBER by seven."
3195 This definition begins with a parenthesis and the symbol @code{defun},
3196 followed by the name of the function.
3198 @cindex @samp{argument list} defined
3199 The name of the function is followed by a list that contains the
3200 arguments that will be passed to the function. This list is called
3201 the @dfn{argument list}. In this example, the list has only one
3202 element, the symbol, @code{number}. When the function is used, the
3203 symbol will be bound to the value that is used as the argument to the
3206 Instead of choosing the word @code{number} for the name of the argument,
3207 I could have picked any other name. For example, I could have chosen
3208 the word @code{multiplicand}. I picked the word `number' because it
3209 tells what kind of value is intended for this slot; but I could just as
3210 well have chosen the word `multiplicand' to indicate the role that the
3211 value placed in this slot will play in the workings of the function. I
3212 could have called it @code{foogle}, but that would have been a bad
3213 choice because it would not tell humans what it means. The choice of
3214 name is up to the programmer and should be chosen to make the meaning of
3217 Indeed, you can choose any name you wish for a symbol in an argument
3218 list, even the name of a symbol used in some other function: the name
3219 you use in an argument list is private to that particular definition.
3220 In that definition, the name refers to a different entity than any use
3221 of the same name outside the function definition. Suppose you have a
3222 nick-name `Shorty' in your family; when your family members refer to
3223 `Shorty', they mean you. But outside your family, in a movie, for
3224 example, the name `Shorty' refers to someone else. Because a name in an
3225 argument list is private to the function definition, you can change the
3226 value of such a symbol inside the body of a function without changing
3227 its value outside the function. The effect is similar to that produced
3228 by a @code{let} expression. (@xref{let, , @code{let}}.)
3231 Note also that we discuss the word `number' in two different ways: as a
3232 symbol that appears in the code, and as the name of something that will
3233 be replaced by a something else during the evaluation of the function.
3234 In the first case, @code{number} is a symbol, not a number; it happens
3235 that within the function, it is a variable who value is the number in
3236 question, but our primary interest in it is as a symbol. On the other
3237 hand, when we are talking about the function, our interest is that we
3238 will substitute a number for the word @var{number}. To keep this
3239 distinction clear, we use different typography for the two
3240 circumstances. When we talk about this function, or about how it works,
3241 we refer to this number by writing @var{number}. In the function
3242 itself, we refer to it by writing @code{number}.
3245 The argument list is followed by the documentation string that
3246 describes the function. This is what you see when you type
3247 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3248 write a documentation string like this, you should make the first line
3249 a complete sentence since some commands, such as @code{apropos}, print
3250 only the first line of a multi-line documentation string. Also, you
3251 should not indent the second line of a documentation string, if you
3252 have one, because that looks odd when you use @kbd{C-h f}
3253 (@code{describe-function}). The documentation string is optional, but
3254 it is so useful, it should be included in almost every function you
3257 @findex * @r{(multiplication)}
3258 The third line of the example consists of the body of the function
3259 definition. (Most functions' definitions, of course, are longer than
3260 this.) In this function, the body is the list, @code{(* 7 number)}, which
3261 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3262 @code{*} is the function for multiplication, just as @code{+} is the
3263 function for addition.)
3265 When you use the @code{multiply-by-seven} function, the argument
3266 @code{number} evaluates to the actual number you want used. Here is an
3267 example that shows how @code{multiply-by-seven} is used; but don't try
3268 to evaluate this yet!
3271 (multiply-by-seven 3)
3275 The symbol @code{number}, specified in the function definition in the
3276 next section, is given or ``bound to'' the value 3 in the actual use of
3277 the function. Note that although @code{number} was inside parentheses
3278 in the function definition, the argument passed to the
3279 @code{multiply-by-seven} function is not in parentheses. The
3280 parentheses are written in the function definition so the computer can
3281 figure out where the argument list ends and the rest of the function
3284 If you evaluate this example, you are likely to get an error message.
3285 (Go ahead, try it!) This is because we have written the function
3286 definition, but not yet told the computer about the definition---we have
3287 not yet installed (or `loaded') the function definition in Emacs.
3288 Installing a function is the process that tells the Lisp interpreter the
3289 definition of the function. Installation is described in the next
3292 @node Install, Interactive, defun, Writing Defuns
3293 @comment node-name, next, previous, up
3294 @section Install a Function Definition
3295 @cindex Install a Function Definition
3296 @cindex Definition installation
3297 @cindex Function definition installation
3299 If you are reading this inside of Info in Emacs, you can try out the
3300 @code{multiply-by-seven} function by first evaluating the function
3301 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3302 the function definition follows. Place the cursor after the last
3303 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3304 do this, @code{multiply-by-seven} will appear in the echo area. (What
3305 this means is that when a function definition is evaluated, the value it
3306 returns is the name of the defined function.) At the same time, this
3307 action installs the function definition.
3311 (defun multiply-by-seven (number)
3312 "Multiply NUMBER by seven."
3318 By evaluating this @code{defun}, you have just installed
3319 @code{multiply-by-seven} in Emacs. The function is now just as much a
3320 part of Emacs as @code{forward-word} or any other editing function you
3321 use. (@code{multiply-by-seven} will stay installed until you quit
3322 Emacs. To reload code automatically whenever you start Emacs, see
3323 @ref{Permanent Installation, , Installing Code Permanently}.)
3326 * Effect of installation::
3327 * Change a defun:: How to change a function definition.
3330 @node Effect of installation, Change a defun, Install, Install
3332 @unnumberedsubsec The effect of installation
3335 You can see the effect of installing @code{multiply-by-seven} by
3336 evaluating the following sample. Place the cursor after the following
3337 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3341 (multiply-by-seven 3)
3344 If you wish, you can read the documentation for the function by typing
3345 @kbd{C-h f} (@code{describe-function}) and then the name of the
3346 function, @code{multiply-by-seven}. When you do this, a
3347 @file{*Help*} window will appear on your screen that says:
3351 multiply-by-seven is a Lisp function.
3352 (multiply-by-seven NUMBER)
3354 Multiply NUMBER by seven.
3359 (To return to a single window on your screen, type @kbd{C-x 1}.)
3361 @node Change a defun, , Effect of installation, Install
3362 @comment node-name, next, previous, up
3363 @subsection Change a Function Definition
3364 @cindex Changing a function definition
3365 @cindex Function definition, how to change
3366 @cindex Definition, how to change
3368 If you want to change the code in @code{multiply-by-seven}, just rewrite
3369 it. To install the new version in place of the old one, evaluate the
3370 function definition again. This is how you modify code in Emacs. It is
3373 As an example, you can change the @code{multiply-by-seven} function to
3374 add the number to itself seven times instead of multiplying the number
3375 by seven. It produces the same answer, but by a different path. At
3376 the same time, we will add a comment to the code; a comment is text
3377 that the Lisp interpreter ignores, but that a human reader may find
3378 useful or enlightening. The comment is that this is the ``second
3383 (defun multiply-by-seven (number) ; @r{Second version.}
3384 "Multiply NUMBER by seven."
3385 (+ number number number number number number number))
3389 @cindex Comments in Lisp code
3390 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3391 line that follows a semicolon is a comment. The end of the line is the
3392 end of the comment. To stretch a comment over two or more lines, begin
3393 each line with a semicolon.
3395 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3396 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3397 Reference Manual}, for more about comments.
3399 You can install this version of the @code{multiply-by-seven} function by
3400 evaluating it in the same way you evaluated the first function: place
3401 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3403 In summary, this is how you write code in Emacs Lisp: you write a
3404 function; install it; test it; and then make fixes or enhancements and
3407 @node Interactive, Interactive Options, Install, Writing Defuns
3408 @comment node-name, next, previous, up
3409 @section Make a Function Interactive
3410 @cindex Interactive functions
3413 You make a function interactive by placing a list that begins with
3414 the special form @code{interactive} immediately after the
3415 documentation. A user can invoke an interactive function by typing
3416 @kbd{M-x} and then the name of the function; or by typing the keys to
3417 which it is bound, for example, by typing @kbd{C-n} for
3418 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3420 Interestingly, when you call an interactive function interactively,
3421 the value returned is not automatically displayed in the echo area.
3422 This is because you often call an interactive function for its side
3423 effects, such as moving forward by a word or line, and not for the
3424 value returned. If the returned value were displayed in the echo area
3425 each time you typed a key, it would be very distracting.
3428 * Interactive multiply-by-seven:: An overview.
3429 * multiply-by-seven in detail:: The interactive version.
3432 @node Interactive multiply-by-seven, multiply-by-seven in detail, Interactive, Interactive
3434 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3437 Both the use of the special form @code{interactive} and one way to
3438 display a value in the echo area can be illustrated by creating an
3439 interactive version of @code{multiply-by-seven}.
3446 (defun multiply-by-seven (number) ; @r{Interactive version.}
3447 "Multiply NUMBER by seven."
3449 (message "The result is %d" (* 7 number)))
3454 You can install this code by placing your cursor after it and typing
3455 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3456 Then, you can use this code by typing @kbd{C-u} and a number and then
3457 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3458 @samp{The result is @dots{}} followed by the product will appear in the
3461 Speaking more generally, you invoke a function like this in either of two
3466 By typing a prefix argument that contains the number to be passed, and
3467 then typing @kbd{M-x} and the name of the function, as with
3468 @kbd{C-u 3 M-x forward-sentence}; or,
3471 By typing whatever key or keychord the function is bound to, as with
3476 Both the examples just mentioned work identically to move point forward
3477 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3478 it could not be used as an example of key binding.)
3480 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3483 A prefix argument is passed to an interactive function by typing the
3484 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3485 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3486 type @kbd{C-u} without a number, it defaults to 4).
3488 @node multiply-by-seven in detail, , Interactive multiply-by-seven, Interactive
3489 @comment node-name, next, previous, up
3490 @subsection An Interactive @code{multiply-by-seven}
3492 Let's look at the use of the special form @code{interactive} and then at
3493 the function @code{message} in the interactive version of
3494 @code{multiply-by-seven}. You will recall that the function definition
3499 (defun multiply-by-seven (number) ; @r{Interactive version.}
3500 "Multiply NUMBER by seven."
3502 (message "The result is %d" (* 7 number)))
3506 In this function, the expression, @code{(interactive "p")}, is a list of
3507 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3508 the function and use its value for the argument of the function.
3511 The argument will be a number. This means that the symbol
3512 @code{number} will be bound to a number in the line:
3515 (message "The result is %d" (* 7 number))
3520 For example, if your prefix argument is 5, the Lisp interpreter will
3521 evaluate the line as if it were:
3524 (message "The result is %d" (* 7 5))
3528 (If you are reading this in GNU Emacs, you can evaluate this expression
3529 yourself.) First, the interpreter will evaluate the inner list, which
3530 is @code{(* 7 5)}. This returns a value of 35. Next, it
3531 will evaluate the outer list, passing the values of the second and
3532 subsequent elements of the list to the function @code{message}.
3534 As we have seen, @code{message} is an Emacs Lisp function especially
3535 designed for sending a one line message to a user. (@xref{message, ,
3536 The @code{message} function}.) In summary, the @code{message}
3537 function prints its first argument in the echo area as is, except for
3538 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3539 which we have not mentioned). When it sees a control sequence, the
3540 function looks to the second or subsequent arguments and prints the
3541 value of the argument in the location in the string where the control
3542 sequence is located.
3544 In the interactive @code{multiply-by-seven} function, the control string
3545 is @samp{%d}, which requires a number, and the value returned by
3546 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3547 is printed in place of the @samp{%d} and the message is @samp{The result
3550 (Note that when you call the function @code{multiply-by-seven}, the
3551 message is printed without quotes, but when you call @code{message}, the
3552 text is printed in double quotes. This is because the value returned by
3553 @code{message} is what appears in the echo area when you evaluate an
3554 expression whose first element is @code{message}; but when embedded in a
3555 function, @code{message} prints the text as a side effect without
3558 @node Interactive Options, Permanent Installation, Interactive, Writing Defuns
3559 @comment node-name, next, previous, up
3560 @section Different Options for @code{interactive}
3561 @cindex Options for @code{interactive}
3562 @cindex Interactive options
3564 In the example, @code{multiply-by-seven} used @code{"p"} as the
3565 argument to @code{interactive}. This argument told Emacs to interpret
3566 your typing either @kbd{C-u} followed by a number or @key{META}
3567 followed by a number as a command to pass that number to the function
3568 as its argument. Emacs has more than twenty characters predefined for
3569 use with @code{interactive}. In almost every case, one of these
3570 options will enable you to pass the right information interactively to
3571 a function. (@xref{Interactive Codes, , Code Characters for
3572 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3575 Consider the function @code{zap-to-char}. Its interactive expression
3579 (interactive "p\ncZap to char: ")
3582 The first part of the argument to @code{interactive} is @samp{p}, with
3583 which you are already familiar. This argument tells Emacs to
3584 interpret a `prefix', as a number to be passed to the function. You
3585 can specify a prefix either by typing @kbd{C-u} followed by a number
3586 or by typing @key{META} followed by a number. The prefix is the
3587 number of specified characters. Thus, if your prefix is three and the
3588 specified character is @samp{x}, then you will delete all the text up
3589 to and including the third next @samp{x}. If you do not set a prefix,
3590 then you delete all the text up to and including the specified
3591 character, but no more.
3593 The @samp{c} tells the function the name of the character to which to delete.
3595 More formally, a function with two or more arguments can have
3596 information passed to each argument by adding parts to the string that
3597 follows @code{interactive}. When you do this, the information is
3598 passed to each argument in the same order it is specified in the
3599 @code{interactive} list. In the string, each part is separated from
3600 the next part by a @samp{\n}, which is a newline. For example, you
3601 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3602 This causes Emacs to pass the value of the prefix argument (if there
3603 is one) and the character.
3605 In this case, the function definition looks like the following, where
3606 @code{arg} and @code{char} are the symbols to which @code{interactive}
3607 binds the prefix argument and the specified character:
3611 (defun @var{name-of-function} (arg char)
3612 "@var{documentation}@dots{}"
3613 (interactive "p\ncZap to char: ")
3614 @var{body-of-function}@dots{})
3619 (The space after the colon in the prompt makes it look better when you
3620 are prompted. @xref{copy-to-buffer, , The Definition of
3621 @code{copy-to-buffer}}, for an example.)
3623 When a function does not take arguments, @code{interactive} does not
3624 require any. Such a function contains the simple expression
3625 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3628 Alternatively, if the special letter-codes are not right for your
3629 application, you can pass your own arguments to @code{interactive} as
3632 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3633 for an example. @xref{Using Interactive, , Using @code{Interactive},
3634 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3635 explanation about this technique.
3637 @node Permanent Installation, let, Interactive Options, Writing Defuns
3638 @comment node-name, next, previous, up
3639 @section Install Code Permanently
3640 @cindex Install code permanently
3641 @cindex Permanent code installation
3642 @cindex Code installation
3644 When you install a function definition by evaluating it, it will stay
3645 installed until you quit Emacs. The next time you start a new session
3646 of Emacs, the function will not be installed unless you evaluate the
3647 function definition again.
3649 At some point, you may want to have code installed automatically
3650 whenever you start a new session of Emacs. There are several ways of
3655 If you have code that is just for yourself, you can put the code for the
3656 function definition in your @file{.emacs} initialization file. When you
3657 start Emacs, your @file{.emacs} file is automatically evaluated and all
3658 the function definitions within it are installed.
3659 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3662 Alternatively, you can put the function definitions that you want
3663 installed in one or more files of their own and use the @code{load}
3664 function to cause Emacs to evaluate and thereby install each of the
3665 functions in the files.
3666 @xref{Loading Files, , Loading Files}.
3669 Thirdly, if you have code that your whole site will use, it is usual
3670 to put it in a file called @file{site-init.el} that is loaded when
3671 Emacs is built. This makes the code available to everyone who uses
3672 your machine. (See the @file{INSTALL} file that is part of the Emacs
3676 Finally, if you have code that everyone who uses Emacs may want, you
3677 can post it on a computer network or send a copy to the Free Software
3678 Foundation. (When you do this, please license the code and its
3679 documentation under a license that permits other people to run, copy,
3680 study, modify, and redistribute the code and which protects you from
3681 having your work taken from you.) If you send a copy of your code to
3682 the Free Software Foundation, and properly protect yourself and
3683 others, it may be included in the next release of Emacs. In large
3684 part, this is how Emacs has grown over the past years, by donations.
3686 @node let, if, Permanent Installation, Writing Defuns
3687 @comment node-name, next, previous, up
3691 The @code{let} expression is a special form in Lisp that you will need
3692 to use in most function definitions.
3694 @code{let} is used to attach or bind a symbol to a value in such a way
3695 that the Lisp interpreter will not confuse the variable with a
3696 variable of the same name that is not part of the function.
3698 To understand why the @code{let} special form is necessary, consider
3699 the situation in which you own a home that you generally refer to as
3700 `the house', as in the sentence, ``The house needs painting.'' If you
3701 are visiting a friend and your host refers to `the house', he is
3702 likely to be referring to @emph{his} house, not yours, that is, to a
3705 If your friend is referring to his house and you think he is referring
3706 to your house, you may be in for some confusion. The same thing could
3707 happen in Lisp if a variable that is used inside of one function has
3708 the same name as a variable that is used inside of another function,
3709 and the two are not intended to refer to the same value. The
3710 @code{let} special form prevents this kind of confusion.
3713 * Prevent confusion::
3714 * Parts of let Expression::
3715 * Sample let Expression::
3716 * Uninitialized let Variables::
3719 @node Prevent confusion, Parts of let Expression, let, let
3721 @unnumberedsubsec @code{let} Prevents Confusion
3724 @cindex @samp{local variable} defined
3725 @cindex @samp{variable, local}, defined
3726 The @code{let} special form prevents confusion. @code{let} creates a
3727 name for a @dfn{local variable} that overshadows any use of the same
3728 name outside the @code{let} expression. This is like understanding
3729 that whenever your host refers to `the house', he means his house, not
3730 yours. (Symbols used in argument lists work the same way.
3731 @xref{defun, , The @code{defun} Special Form}.)
3733 Local variables created by a @code{let} expression retain their value
3734 @emph{only} within the @code{let} expression itself (and within
3735 expressions called within the @code{let} expression); the local
3736 variables have no effect outside the @code{let} expression.
3738 Another way to think about @code{let} is that it is like a @code{setq}
3739 that is temporary and local. The values set by @code{let} are
3740 automatically undone when the @code{let} is finished. The setting
3741 only affects expressions that are inside the bounds of the @code{let}
3742 expression. In computer science jargon, we would say ``the binding of
3743 a symbol is visible only in functions called in the @code{let} form;
3744 in Emacs Lisp, scoping is dynamic, not lexical.''
3746 @code{let} can create more than one variable at once. Also,
3747 @code{let} gives each variable it creates an initial value, either a
3748 value specified by you, or @code{nil}. (In the jargon, this is called
3749 `binding the variable to the value'.) After @code{let} has created
3750 and bound the variables, it executes the code in the body of the
3751 @code{let}, and returns the value of the last expression in the body,
3752 as the value of the whole @code{let} expression. (`Execute' is a jargon
3753 term that means to evaluate a list; it comes from the use of the word
3754 meaning `to give practical effect to' (@cite{Oxford English
3755 Dictionary}). Since you evaluate an expression to perform an action,
3756 `execute' has evolved as a synonym to `evaluate'.)
3758 @node Parts of let Expression, Sample let Expression, Prevent confusion, let
3759 @comment node-name, next, previous, up
3760 @subsection The Parts of a @code{let} Expression
3761 @cindex @code{let} expression, parts of
3762 @cindex Parts of @code{let} expression
3764 @cindex @samp{varlist} defined
3765 A @code{let} expression is a list of three parts. The first part is
3766 the symbol @code{let}. The second part is a list, called a
3767 @dfn{varlist}, each element of which is either a symbol by itself or a
3768 two-element list, the first element of which is a symbol. The third
3769 part of the @code{let} expression is the body of the @code{let}. The
3770 body usually consists of one or more lists.
3773 A template for a @code{let} expression looks like this:
3776 (let @var{varlist} @var{body}@dots{})
3780 The symbols in the varlist are the variables that are given initial
3781 values by the @code{let} special form. Symbols by themselves are given
3782 the initial value of @code{nil}; and each symbol that is the first
3783 element of a two-element list is bound to the value that is returned
3784 when the Lisp interpreter evaluates the second element.
3786 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3787 this case, in a @code{let} expression, Emacs binds the symbol
3788 @code{thread} to an initial value of @code{nil}, and binds the symbol
3789 @code{needles} to an initial value of 3.
3791 When you write a @code{let} expression, what you do is put the
3792 appropriate expressions in the slots of the @code{let} expression
3795 If the varlist is composed of two-element lists, as is often the case,
3796 the template for the @code{let} expression looks like this:
3800 (let ((@var{variable} @var{value})
3801 (@var{variable} @var{value})
3807 @node Sample let Expression, Uninitialized let Variables, Parts of let Expression, let
3808 @comment node-name, next, previous, up
3809 @subsection Sample @code{let} Expression
3810 @cindex Sample @code{let} expression
3811 @cindex @code{let} expression sample
3813 The following expression creates and gives initial values
3814 to the two variables @code{zebra} and @code{tiger}. The body of the
3815 @code{let} expression is a list which calls the @code{message} function.
3819 (let ((zebra 'stripes)
3821 (message "One kind of animal has %s and another is %s."
3826 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3828 The two variables are @code{zebra} and @code{tiger}. Each variable is
3829 the first element of a two-element list and each value is the second
3830 element of its two-element list. In the varlist, Emacs binds the
3831 variable @code{zebra} to the value @code{stripes}@footnote{According
3832 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3833 become impossibly dangerous as they grow older'' but the claim here is
3834 that they do not become fierce like a tiger. (1997, W. W. Norton and
3835 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3836 variable @code{tiger} to the value @code{fierce}. In this example,
3837 both values are symbols preceded by a quote. The values could just as
3838 well have been another list or a string. The body of the @code{let}
3839 follows after the list holding the variables. In this example, the
3840 body is a list that uses the @code{message} function to print a string
3844 You may evaluate the example in the usual fashion, by placing the
3845 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3846 this, the following will appear in the echo area:
3849 "One kind of animal has stripes and another is fierce."
3852 As we have seen before, the @code{message} function prints its first
3853 argument, except for @samp{%s}. In this example, the value of the variable
3854 @code{zebra} is printed at the location of the first @samp{%s} and the
3855 value of the variable @code{tiger} is printed at the location of the
3858 @node Uninitialized let Variables, , Sample let Expression, let
3859 @comment node-name, next, previous, up
3860 @subsection Uninitialized Variables in a @code{let} Statement
3861 @cindex Uninitialized @code{let} variables
3862 @cindex @code{let} variables uninitialized
3864 If you do not bind the variables in a @code{let} statement to specific
3865 initial values, they will automatically be bound to an initial value of
3866 @code{nil}, as in the following expression:
3875 "Here are %d variables with %s, %s, and %s value."
3876 birch pine fir oak))
3881 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3884 If you evaluate this expression in the usual way, the following will
3885 appear in your echo area:
3888 "Here are 3 variables with nil, nil, and some value."
3892 In this example, Emacs binds the symbol @code{birch} to the number 3,
3893 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3894 the symbol @code{oak} to the value @code{some}.
3896 Note that in the first part of the @code{let}, the variables @code{pine}
3897 and @code{fir} stand alone as atoms that are not surrounded by
3898 parentheses; this is because they are being bound to @code{nil}, the
3899 empty list. But @code{oak} is bound to @code{some} and so is a part of
3900 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3901 number 3 and so is in a list with that number. (Since a number
3902 evaluates to itself, the number does not need to be quoted. Also, the
3903 number is printed in the message using a @samp{%d} rather than a
3904 @samp{%s}.) The four variables as a group are put into a list to
3905 delimit them from the body of the @code{let}.
3907 @node if, else, let, Writing Defuns
3908 @comment node-name, next, previous, up
3909 @section The @code{if} Special Form
3911 @cindex Conditional with @code{if}
3913 A third special form, in addition to @code{defun} and @code{let}, is the
3914 conditional @code{if}. This form is used to instruct the computer to
3915 make decisions. You can write function definitions without using
3916 @code{if}, but it is used often enough, and is important enough, to be
3917 included here. It is used, for example, in the code for the
3918 function @code{beginning-of-buffer}.
3920 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3921 @emph{then} an expression is evaluated.'' If the test is not true, the
3922 expression is not evaluated. For example, you might make a decision
3923 such as, ``if it is warm and sunny, then go to the beach!''
3926 * if in more detail::
3927 * type-of-animal in detail:: An example of an @code{if} expression.
3930 @node if in more detail, type-of-animal in detail, if, if
3932 @unnumberedsubsec @code{if} in more detail
3935 @cindex @samp{if-part} defined
3936 @cindex @samp{then-part} defined
3937 An @code{if} expression written in Lisp does not use the word `then';
3938 the test and the action are the second and third elements of the list
3939 whose first element is @code{if}. Nonetheless, the test part of an
3940 @code{if} expression is often called the @dfn{if-part} and the second
3941 argument is often called the @dfn{then-part}.
3943 Also, when an @code{if} expression is written, the true-or-false-test
3944 is usually written on the same line as the symbol @code{if}, but the
3945 action to carry out if the test is true, the ``then-part'', is written
3946 on the second and subsequent lines. This makes the @code{if}
3947 expression easier to read.
3951 (if @var{true-or-false-test}
3952 @var{action-to-carry-out-if-test-is-true})
3957 The true-or-false-test will be an expression that
3958 is evaluated by the Lisp interpreter.
3960 Here is an example that you can evaluate in the usual manner. The test
3961 is whether the number 5 is greater than the number 4. Since it is, the
3962 message @samp{5 is greater than 4!} will be printed.
3966 (if (> 5 4) ; @r{if-part}
3967 (message "5 is greater than 4!")) ; @r{then-part}
3972 (The function @code{>} tests whether its first argument is greater than
3973 its second argument and returns true if it is.)
3974 @findex > (greater than)
3976 Of course, in actual use, the test in an @code{if} expression will not
3977 be fixed for all time as it is by the expression @code{(> 5 4)}.
3978 Instead, at least one of the variables used in the test will be bound to
3979 a value that is not known ahead of time. (If the value were known ahead
3980 of time, we would not need to run the test!)
3982 For example, the value may be bound to an argument of a function
3983 definition. In the following function definition, the character of the
3984 animal is a value that is passed to the function. If the value bound to
3985 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3986 tiger!} will be printed; otherwise, @code{nil} will be returned.
3990 (defun type-of-animal (characteristic)
3991 "Print message in echo area depending on CHARACTERISTIC.
3992 If the CHARACTERISTIC is the symbol `fierce',
3993 then warn of a tiger."
3994 (if (equal characteristic 'fierce)
3995 (message "It's a tiger!")))
4001 If you are reading this inside of GNU Emacs, you can evaluate the
4002 function definition in the usual way to install it in Emacs, and then you
4003 can evaluate the following two expressions to see the results:
4007 (type-of-animal 'fierce)
4009 (type-of-animal 'zebra)
4014 @c Following sentences rewritten to prevent overfull hbox.
4016 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4017 following message printed in the echo area: @code{"It's a tiger!"}; and
4018 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
4019 printed in the echo area.
4021 @node type-of-animal in detail, , if in more detail, if
4022 @comment node-name, next, previous, up
4023 @subsection The @code{type-of-animal} Function in Detail
4025 Let's look at the @code{type-of-animal} function in detail.
4027 The function definition for @code{type-of-animal} was written by filling
4028 the slots of two templates, one for a function definition as a whole, and
4029 a second for an @code{if} expression.
4032 The template for every function that is not interactive is:
4036 (defun @var{name-of-function} (@var{argument-list})
4037 "@var{documentation}@dots{}"
4043 The parts of the function that match this template look like this:
4047 (defun type-of-animal (characteristic)
4048 "Print message in echo area depending on CHARACTERISTIC.
4049 If the CHARACTERISTIC is the symbol `fierce',
4050 then warn of a tiger."
4051 @var{body: the} @code{if} @var{expression})
4055 The name of function is @code{type-of-animal}; it is passed the value
4056 of one argument. The argument list is followed by a multi-line
4057 documentation string. The documentation string is included in the
4058 example because it is a good habit to write documentation string for
4059 every function definition. The body of the function definition
4060 consists of the @code{if} expression.
4063 The template for an @code{if} expression looks like this:
4067 (if @var{true-or-false-test}
4068 @var{action-to-carry-out-if-the-test-returns-true})
4073 In the @code{type-of-animal} function, the code for the @code{if}
4078 (if (equal characteristic 'fierce)
4079 (message "It's a tiger!")))
4084 Here, the true-or-false-test is the expression:
4087 (equal characteristic 'fierce)
4091 In Lisp, @code{equal} is a function that determines whether its first
4092 argument is equal to its second argument. The second argument is the
4093 quoted symbol @code{'fierce} and the first argument is the value of the
4094 symbol @code{characteristic}---in other words, the argument passed to
4097 In the first exercise of @code{type-of-animal}, the argument
4098 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4099 is equal to @code{fierce}, the expression, @code{(equal characteristic
4100 'fierce)}, returns a value of true. When this happens, the @code{if}
4101 evaluates the second argument or then-part of the @code{if}:
4102 @code{(message "It's tiger!")}.
4104 On the other hand, in the second exercise of @code{type-of-animal}, the
4105 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4106 is not equal to @code{fierce}, so the then-part is not evaluated and
4107 @code{nil} is returned by the @code{if} expression.
4109 @node else, Truth & Falsehood, if, Writing Defuns
4110 @comment node-name, next, previous, up
4111 @section If--then--else Expressions
4114 An @code{if} expression may have an optional third argument, called
4115 the @dfn{else-part}, for the case when the true-or-false-test returns
4116 false. When this happens, the second argument or then-part of the
4117 overall @code{if} expression is @emph{not} evaluated, but the third or
4118 else-part @emph{is} evaluated. You might think of this as the cloudy
4119 day alternative for the decision ``if it is warm and sunny, then go to
4120 the beach, else read a book!''.
4122 The word ``else'' is not written in the Lisp code; the else-part of an
4123 @code{if} expression comes after the then-part. In the written Lisp, the
4124 else-part is usually written to start on a line of its own and is
4125 indented less than the then-part:
4129 (if @var{true-or-false-test}
4130 @var{action-to-carry-out-if-the-test-returns-true}
4131 @var{action-to-carry-out-if-the-test-returns-false})
4135 For example, the following @code{if} expression prints the message @samp{4
4136 is not greater than 5!} when you evaluate it in the usual way:
4140 (if (> 4 5) ; @r{if-part}
4141 (message "4 falsely greater than 5!") ; @r{then-part}
4142 (message "4 is not greater than 5!")) ; @r{else-part}
4147 Note that the different levels of indentation make it easy to
4148 distinguish the then-part from the else-part. (GNU Emacs has several
4149 commands that automatically indent @code{if} expressions correctly.
4150 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4152 We can extend the @code{type-of-animal} function to include an
4153 else-part by simply incorporating an additional part to the @code{if}
4157 You can see the consequences of doing this if you evaluate the following
4158 version of the @code{type-of-animal} function definition to install it
4159 and then evaluate the two subsequent expressions to pass different
4160 arguments to the function.
4164 (defun type-of-animal (characteristic) ; @r{Second version.}
4165 "Print message in echo area depending on CHARACTERISTIC.
4166 If the CHARACTERISTIC is the symbol `fierce',
4167 then warn of a tiger;
4168 else say it's not fierce."
4169 (if (equal characteristic 'fierce)
4170 (message "It's a tiger!")
4171 (message "It's not fierce!")))
4178 (type-of-animal 'fierce)
4180 (type-of-animal 'zebra)
4185 @c Following sentence rewritten to prevent overfull hbox.
4187 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4188 following message printed in the echo area: @code{"It's a tiger!"}; but
4189 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4190 @code{"It's not fierce!"}.
4192 (Of course, if the @var{characteristic} were @code{ferocious}, the
4193 message @code{"It's not fierce!"} would be printed; and it would be
4194 misleading! When you write code, you need to take into account the
4195 possibility that some such argument will be tested by the @code{if}
4196 and write your program accordingly.)
4198 @node Truth & Falsehood, save-excursion, else, Writing Defuns
4199 @comment node-name, next, previous, up
4200 @section Truth and Falsehood in Emacs Lisp
4201 @cindex Truth and falsehood in Emacs Lisp
4202 @cindex Falsehood and truth in Emacs Lisp
4205 There is an important aspect to the truth test in an @code{if}
4206 expression. So far, we have spoken of `true' and `false' as values of
4207 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4208 `false' is just our old friend @code{nil}. Anything else---anything
4211 The expression that tests for truth is interpreted as @dfn{true}
4212 if the result of evaluating it is a value that is not @code{nil}. In
4213 other words, the result of the test is considered true if the value
4214 returned is a number such as 47, a string such as @code{"hello"}, or a
4215 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4216 long as it is not empty), or even a buffer!
4219 * nil explained:: @code{nil} has two meanings.
4222 @node nil explained, , Truth & Falsehood, Truth & Falsehood
4224 @unnumberedsubsec An explanation of @code{nil}
4227 Before illustrating a test for truth, we need an explanation of @code{nil}.
4229 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4230 empty list. Second, it means false and is the value returned when a
4231 true-or-false-test tests false. @code{nil} can be written as an empty
4232 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4233 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4234 to use @code{nil} for false and @code{()} for the empty list.
4236 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4237 list---is considered true. This means that if an evaluation returns
4238 something that is not an empty list, an @code{if} expression will test
4239 true. For example, if a number is put in the slot for the test, it
4240 will be evaluated and will return itself, since that is what numbers
4241 do when evaluated. In this conditional, the @code{if} expression will
4242 test true. The expression tests false only when @code{nil}, an empty
4243 list, is returned by evaluating the expression.
4245 You can see this by evaluating the two expressions in the following examples.
4247 In the first example, the number 4 is evaluated as the test in the
4248 @code{if} expression and returns itself; consequently, the then-part
4249 of the expression is evaluated and returned: @samp{true} appears in
4250 the echo area. In the second example, the @code{nil} indicates false;
4251 consequently, the else-part of the expression is evaluated and
4252 returned: @samp{false} appears in the echo area.
4269 Incidentally, if some other useful value is not available for a test that
4270 returns true, then the Lisp interpreter will return the symbol @code{t}
4271 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4272 when evaluated, as you can see by evaluating it in the usual way:
4280 On the other hand, this function returns @code{nil} if the test is false.
4286 @node save-excursion, Review, Truth & Falsehood, Writing Defuns
4287 @comment node-name, next, previous, up
4288 @section @code{save-excursion}
4289 @findex save-excursion
4290 @cindex Region, what it is
4291 @cindex Preserving point, mark, and buffer
4292 @cindex Point, mark, buffer preservation
4296 The @code{save-excursion} function is the fourth and final special form
4297 that we will discuss in this chapter.
4299 In Emacs Lisp programs used for editing, the @code{save-excursion}
4300 function is very common. It saves the location of point and mark,
4301 executes the body of the function, and then restores point and mark to
4302 their previous positions if their locations were changed. Its primary
4303 purpose is to keep the user from being surprised and disturbed by
4304 unexpected movement of point or mark.
4307 * Point and mark:: A review of various locations.
4308 * Template for save-excursion::
4311 @node Point and mark, Template for save-excursion, save-excursion, save-excursion
4313 @unnumberedsubsec Point and Mark
4316 Before discussing @code{save-excursion}, however, it may be useful
4317 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4318 the current location of the cursor. Wherever the cursor
4319 is, that is point. More precisely, on terminals where the cursor
4320 appears to be on top of a character, point is immediately before the
4321 character. In Emacs Lisp, point is an integer. The first character in
4322 a buffer is number one, the second is number two, and so on. The
4323 function @code{point} returns the current position of the cursor as a
4324 number. Each buffer has its own value for point.
4326 The @dfn{mark} is another position in the buffer; its value can be set
4327 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4328 a mark has been set, you can use the command @kbd{C-x C-x}
4329 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4330 and set the mark to be the previous position of point. In addition, if
4331 you set another mark, the position of the previous mark is saved in the
4332 mark ring. Many mark positions can be saved this way. You can jump the
4333 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4336 The part of the buffer between point and mark is called @dfn{the
4337 region}. Numerous commands work on the region, including
4338 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4339 @code{print-region}.
4341 The @code{save-excursion} special form saves the locations of point and
4342 mark and restores those positions after the code within the body of the
4343 special form is evaluated by the Lisp interpreter. Thus, if point were
4344 in the beginning of a piece of text and some code moved point to the end
4345 of the buffer, the @code{save-excursion} would put point back to where
4346 it was before, after the expressions in the body of the function were
4349 In Emacs, a function frequently moves point as part of its internal
4350 workings even though a user would not expect this. For example,
4351 @code{count-lines-region} moves point. To prevent the user from being
4352 bothered by jumps that are both unexpected and (from the user's point of
4353 view) unnecessary, @code{save-excursion} is often used to keep point and
4354 mark in the location expected by the user. The use of
4355 @code{save-excursion} is good housekeeping.
4357 To make sure the house stays clean, @code{save-excursion} restores the
4358 values of point and mark even if something goes wrong in the code inside
4359 of it (or, to be more precise and to use the proper jargon, ``in case of
4360 abnormal exit''). This feature is very helpful.
4362 In addition to recording the values of point and mark,
4363 @code{save-excursion} keeps track of the current buffer, and restores
4364 it, too. This means you can write code that will change the buffer and
4365 have @code{save-excursion} switch you back to the original buffer.
4366 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4367 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4369 @node Template for save-excursion, , Point and mark, save-excursion
4370 @comment node-name, next, previous, up
4371 @subsection Template for a @code{save-excursion} Expression
4374 The template for code using @code{save-excursion} is simple:
4384 The body of the function is one or more expressions that will be
4385 evaluated in sequence by the Lisp interpreter. If there is more than
4386 one expression in the body, the value of the last one will be returned
4387 as the value of the @code{save-excursion} function. The other
4388 expressions in the body are evaluated only for their side effects; and
4389 @code{save-excursion} itself is used only for its side effect (which
4390 is restoring the positions of point and mark).
4393 In more detail, the template for a @code{save-excursion} expression
4399 @var{first-expression-in-body}
4400 @var{second-expression-in-body}
4401 @var{third-expression-in-body}
4403 @var{last-expression-in-body})
4408 An expression, of course, may be a symbol on its own or a list.
4410 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4411 within the body of a @code{let} expression. It looks like this:
4421 @node Review, defun Exercises, save-excursion, Writing Defuns
4422 @comment node-name, next, previous, up
4425 In the last few chapters we have introduced a fair number of functions
4426 and special forms. Here they are described in brief, along with a few
4427 similar functions that have not been mentioned yet.
4430 @item eval-last-sexp
4431 Evaluate the last symbolic expression before the current location of
4432 point. The value is printed in the echo area unless the function is
4433 invoked with an argument; in that case, the output is printed in the
4434 current buffer. This command is normally bound to @kbd{C-x C-e}.
4437 Define function. This special form has up to five parts: the name,
4438 a template for the arguments that will be passed to the function,
4439 documentation, an optional interactive declaration, and the body of the
4443 For example, in an early version of Emacs, the function definition was
4444 as follows. (It is slightly more complex now that it seeks the first
4445 non-whitespace character rather than the first visible character.)
4449 (defun back-to-indentation ()
4450 "Move point to first visible character on line."
4452 (beginning-of-line 1)
4453 (skip-chars-forward " \t"))
4460 (defun backward-to-indentation (&optional arg)
4461 "Move backward ARG lines and position at first nonblank character."
4463 (forward-line (- (or arg 1)))
4464 (skip-chars-forward " \t"))
4466 (defun back-to-indentation ()
4467 "Move point to the first non-whitespace character on this line."
4469 (beginning-of-line 1)
4470 (skip-syntax-forward " " (line-end-position))
4471 ;; Move back over chars that have whitespace syntax but have the p flag.
4472 (backward-prefix-chars))
4476 Declare to the interpreter that the function can be used
4477 interactively. This special form may be followed by a string with one
4478 or more parts that pass the information to the arguments of the
4479 function, in sequence. These parts may also tell the interpreter to
4480 prompt for information. Parts of the string are separated by
4481 newlines, @samp{\n}.
4484 Common code characters are:
4488 The name of an existing buffer.
4491 The name of an existing file.
4494 The numeric prefix argument. (Note that this `p' is lower case.)
4497 Point and the mark, as two numeric arguments, smallest first. This
4498 is the only code letter that specifies two successive arguments
4502 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4503 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4507 Declare that a list of variables is for use within the body of the
4508 @code{let} and give them an initial value, either @code{nil} or a
4509 specified value; then evaluate the rest of the expressions in the body
4510 of the @code{let} and return the value of the last one. Inside the
4511 body of the @code{let}, the Lisp interpreter does not see the values of
4512 the variables of the same names that are bound outside of the
4520 (let ((foo (buffer-name))
4521 (bar (buffer-size)))
4523 "This buffer is %s and has %d characters."
4528 @item save-excursion
4529 Record the values of point and mark and the current buffer before
4530 evaluating the body of this special form. Restore the values of point
4531 and mark and buffer afterward.
4538 (message "We are %d characters into this buffer."
4541 (goto-char (point-min)) (point))))
4546 Evaluate the first argument to the function; if it is true, evaluate
4547 the second argument; else evaluate the third argument, if there is one.
4549 The @code{if} special form is called a @dfn{conditional}. There are
4550 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4558 (if (= 22 emacs-major-version)
4559 (message "This is version 22 Emacs")
4560 (message "This is not version 22 Emacs"))
4569 The @code{<} function tests whether its first argument is smaller than
4570 its second argument. A corresponding function, @code{>}, tests whether
4571 the first argument is greater than the second. Likewise, @code{<=}
4572 tests whether the first argument is less than or equal to the second and
4573 @code{>=} tests whether the first argument is greater than or equal to
4574 the second. In all cases, both arguments must be numbers or markers
4575 (markers indicate positions in buffers).
4579 The @code{=} function tests whether two arguments, both numbers or
4585 Test whether two objects are the same. @code{equal} uses one meaning
4586 of the word `same' and @code{eq} uses another: @code{equal} returns
4587 true if the two objects have a similar structure and contents, such as
4588 two copies of the same book. On the other hand, @code{eq}, returns
4589 true if both arguments are actually the same object.
4598 The @code{string-lessp} function tests whether its first argument is
4599 smaller than the second argument. A shorter, alternative name for the
4600 same function (a @code{defalias}) is @code{string<}.
4602 The arguments to @code{string-lessp} must be strings or symbols; the
4603 ordering is lexicographic, so case is significant. The print names of
4604 symbols are used instead of the symbols themselves.
4606 @cindex @samp{empty string} defined
4607 An empty string, @samp{""}, a string with no characters in it, is
4608 smaller than any string of characters.
4610 @code{string-equal} provides the corresponding test for equality. Its
4611 shorter, alternative name is @code{string=}. There are no string test
4612 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4615 Print a message in the echo area. The first argument is a string that
4616 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4617 arguments that follow the string. The argument used by @samp{%s} must
4618 be a string or a symbol; the argument used by @samp{%d} must be a
4619 number. The argument used by @samp{%c} must be an @sc{ascii} code
4620 number; it will be printed as the character with that @sc{ascii} code.
4621 (Various other %-sequences have not been mentioned.)
4625 The @code{setq} function sets the value of its first argument to the
4626 value of the second argument. The first argument is automatically
4627 quoted by @code{setq}. It does the same for succeeding pairs of
4628 arguments. Another function, @code{set}, takes only two arguments and
4629 evaluates both of them before setting the value returned by its first
4630 argument to the value returned by its second argument.
4633 Without an argument, return the name of the buffer, as a string.
4635 @itemx buffer-file-name
4636 Without an argument, return the name of the file the buffer is
4639 @item current-buffer
4640 Return the buffer in which Emacs is active; it may not be
4641 the buffer that is visible on the screen.
4644 Return the most recently selected buffer (other than the buffer passed
4645 to @code{other-buffer} as an argument and other than the current
4648 @item switch-to-buffer
4649 Select a buffer for Emacs to be active in and display it in the current
4650 window so users can look at it. Usually bound to @kbd{C-x b}.
4653 Switch Emacs's attention to a buffer on which programs will run. Don't
4654 alter what the window is showing.
4657 Return the number of characters in the current buffer.
4660 Return the value of the current position of the cursor, as an
4661 integer counting the number of characters from the beginning of the
4665 Return the minimum permissible value of point in
4666 the current buffer. This is 1, unless narrowing is in effect.
4669 Return the value of the maximum permissible value of point in the
4670 current buffer. This is the end of the buffer, unless narrowing is in
4675 @node defun Exercises, , Review, Writing Defuns
4680 Write a non-interactive function that doubles the value of its
4681 argument, a number. Make that function interactive.
4684 Write a function that tests whether the current value of
4685 @code{fill-column} is greater than the argument passed to the function,
4686 and if so, prints an appropriate message.
4689 @node Buffer Walk Through, More Complex, Writing Defuns, Top
4690 @comment node-name, next, previous, up
4691 @chapter A Few Buffer--Related Functions
4693 In this chapter we study in detail several of the functions used in GNU
4694 Emacs. This is called a ``walk-through''. These functions are used as
4695 examples of Lisp code, but are not imaginary examples; with the
4696 exception of the first, simplified function definition, these functions
4697 show the actual code used in GNU Emacs. You can learn a great deal from
4698 these definitions. The functions described here are all related to
4699 buffers. Later, we will study other functions.
4702 * Finding More:: How to find more information.
4703 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4704 @code{point-min}, and @code{push-mark}.
4705 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4706 * append-to-buffer:: Uses @code{save-excursion} and
4707 @code{insert-buffer-substring}.
4708 * Buffer Related Review:: Review.
4709 * Buffer Exercises::
4712 @node Finding More, simplified-beginning-of-buffer, Buffer Walk Through, Buffer Walk Through
4713 @section Finding More Information
4715 @findex describe-function, @r{introduced}
4716 @cindex Find function documentation
4717 In this walk-through, I will describe each new function as we come to
4718 it, sometimes in detail and sometimes briefly. If you are interested,
4719 you can get the full documentation of any Emacs Lisp function at any
4720 time by typing @kbd{C-h f} and then the name of the function (and then
4721 @key{RET}). Similarly, you can get the full documentation for a
4722 variable by typing @kbd{C-h v} and then the name of the variable (and
4725 @cindex Find source of function
4726 @c In version 22, tells location both of C and of Emacs Lisp
4727 Also, @code{describe-function} will tell you the location of the
4728 function definition.
4730 Put point into the name of the file that contains the function and
4731 press the @key{RET} key. In this case, @key{RET} means
4732 @code{push-button} rather than `return' or `enter'. Emacs will take
4733 you directly to the function definition.
4738 If you move point over the file name and press
4739 the @key{RET} key, which in this case means @code{help-follow} rather
4740 than `return' or `enter', Emacs will take you directly to the function
4744 More generally, if you want to see a function in its original source
4745 file, you can use the @code{find-tag} function to jump to it.
4746 @code{find-tag} works with a wide variety of languages, not just
4747 Lisp, and C, and it works with non-programming text as well. For
4748 example, @code{find-tag} will jump to the various nodes in the
4749 Texinfo source file of this document.
4750 The @code{find-tag} function depends on `tags tables' that record
4751 the locations of the functions, variables, and other items to which
4752 @code{find-tag} jumps.
4754 To use the @code{find-tag} command, type @kbd{M-.} (i.e., press the
4755 period key while holding down the @key{META} key, or else type the
4756 @key{ESC} key and then type the period key), and then, at the prompt,
4757 type in the name of the function whose source code you want to see,
4758 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4759 switch buffers and display the source code for the function on your
4760 screen. To switch back to your current buffer, type @kbd{C-x b
4761 @key{RET}}. (On some keyboards, the @key{META} key is labeled
4764 @c !!! 22.1.1 tags table location in this paragraph
4765 @cindex TAGS table, specifying
4767 Depending on how the initial default values of your copy of Emacs are
4768 set, you may also need to specify the location of your `tags table',
4769 which is a file called @file{TAGS}. For example, if you are
4770 interested in Emacs sources, the tags table you will most likely want,
4771 if it has already been created for you, will be in a subdirectory of
4772 the @file{/usr/local/share/emacs/} directory; thus you would use the
4773 @code{M-x visit-tags-table} command and specify a pathname such as
4774 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4775 has not already been created, you will have to create it yourself. It
4776 will be in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4779 To create a @file{TAGS} file in a specific directory, switch to that
4780 directory in Emacs using @kbd{M-x cd} command, or list the directory
4781 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4782 @w{@code{etags *.el}} as the command to execute:
4785 M-x compile RET etags *.el RET
4788 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4790 After you become more familiar with Emacs Lisp, you will find that you will
4791 frequently use @code{find-tag} to navigate your way around source code;
4792 and you will create your own @file{TAGS} tables.
4794 @cindex Library, as term for `file'
4795 Incidentally, the files that contain Lisp code are conventionally
4796 called @dfn{libraries}. The metaphor is derived from that of a
4797 specialized library, such as a law library or an engineering library,
4798 rather than a general library. Each library, or file, contains
4799 functions that relate to a particular topic or activity, such as
4800 @file{abbrev.el} for handling abbreviations and other typing
4801 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4802 libraries provide code for a single activity, as the various
4803 @file{rmail@dots{}} files provide code for reading electronic mail.)
4804 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4805 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4806 by topic keywords.''
4808 @node simplified-beginning-of-buffer, mark-whole-buffer, Finding More, Buffer Walk Through
4809 @comment node-name, next, previous, up
4810 @section A Simplified @code{beginning-of-buffer} Definition
4811 @findex simplified-beginning-of-buffer
4813 The @code{beginning-of-buffer} command is a good function to start with
4814 since you are likely to be familiar with it and it is easy to
4815 understand. Used as an interactive command, @code{beginning-of-buffer}
4816 moves the cursor to the beginning of the buffer, leaving the mark at the
4817 previous position. It is generally bound to @kbd{M-<}.
4819 In this section, we will discuss a shortened version of the function
4820 that shows how it is most frequently used. This shortened function
4821 works as written, but it does not contain the code for a complex option.
4822 In another section, we will describe the entire function.
4823 (@xref{beginning-of-buffer, , Complete Definition of
4824 @code{beginning-of-buffer}}.)
4826 Before looking at the code, let's consider what the function
4827 definition has to contain: it must include an expression that makes
4828 the function interactive so it can be called by typing @kbd{M-x
4829 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4830 must include code to leave a mark at the original position in the
4831 buffer; and it must include code to move the cursor to the beginning
4835 Here is the complete text of the shortened version of the function:
4839 (defun simplified-beginning-of-buffer ()
4840 "Move point to the beginning of the buffer;
4841 leave mark at previous position."
4844 (goto-char (point-min)))
4848 Like all function definitions, this definition has five parts following
4849 the special form @code{defun}:
4853 The name: in this example, @code{simplified-beginning-of-buffer}.
4856 A list of the arguments: in this example, an empty list, @code{()},
4859 The documentation string.
4862 The interactive expression.
4869 In this function definition, the argument list is empty; this means that
4870 this function does not require any arguments. (When we look at the
4871 definition for the complete function, we will see that it may be passed
4872 an optional argument.)
4874 The interactive expression tells Emacs that the function is intended to
4875 be used interactively. In this example, @code{interactive} does not have
4876 an argument because @code{simplified-beginning-of-buffer} does not
4880 The body of the function consists of the two lines:
4885 (goto-char (point-min))
4889 The first of these lines is the expression, @code{(push-mark)}. When
4890 this expression is evaluated by the Lisp interpreter, it sets a mark at
4891 the current position of the cursor, wherever that may be. The position
4892 of this mark is saved in the mark ring.
4894 The next line is @code{(goto-char (point-min))}. This expression
4895 jumps the cursor to the minimum point in the buffer, that is, to the
4896 beginning of the buffer (or to the beginning of the accessible portion
4897 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4898 Narrowing and Widening}.)
4900 The @code{push-mark} command sets a mark at the place where the cursor
4901 was located before it was moved to the beginning of the buffer by the
4902 @code{(goto-char (point-min))} expression. Consequently, you can, if
4903 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4905 That is all there is to the function definition!
4907 @findex describe-function
4908 When you are reading code such as this and come upon an unfamiliar
4909 function, such as @code{goto-char}, you can find out what it does by
4910 using the @code{describe-function} command. To use this command, type
4911 @kbd{C-h f} and then type in the name of the function and press
4912 @key{RET}. The @code{describe-function} command will print the
4913 function's documentation string in a @file{*Help*} window. For
4914 example, the documentation for @code{goto-char} is:
4918 Set point to POSITION, a number or marker.
4919 Beginning of buffer is position (point-min), end is (point-max).
4924 The function's one argument is the desired position.
4927 (The prompt for @code{describe-function} will offer you the symbol
4928 under or preceding the cursor, so you can save typing by positioning
4929 the cursor right over or after the function and then typing @kbd{C-h f
4932 The @code{end-of-buffer} function definition is written in the same way as
4933 the @code{beginning-of-buffer} definition except that the body of the
4934 function contains the expression @code{(goto-char (point-max))} in place
4935 of @code{(goto-char (point-min))}.
4937 @node mark-whole-buffer, append-to-buffer, simplified-beginning-of-buffer, Buffer Walk Through
4938 @comment node-name, next, previous, up
4939 @section The Definition of @code{mark-whole-buffer}
4940 @findex mark-whole-buffer
4942 The @code{mark-whole-buffer} function is no harder to understand than the
4943 @code{simplified-beginning-of-buffer} function. In this case, however,
4944 we will look at the complete function, not a shortened version.
4946 The @code{mark-whole-buffer} function is not as commonly used as the
4947 @code{beginning-of-buffer} function, but is useful nonetheless: it
4948 marks a whole buffer as a region by putting point at the beginning and
4949 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4953 * mark-whole-buffer overview::
4954 * Body of mark-whole-buffer:: Only three lines of code.
4957 @node mark-whole-buffer overview, Body of mark-whole-buffer, mark-whole-buffer, mark-whole-buffer
4959 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4963 In GNU Emacs 22, the code for the complete function looks like this:
4967 (defun mark-whole-buffer ()
4968 "Put point at beginning and mark at end of buffer.
4969 You probably should not use this function in Lisp programs;
4970 it is usually a mistake for a Lisp function to use any subroutine
4971 that uses or sets the mark."
4974 (push-mark (point-max) nil t)
4975 (goto-char (point-min)))
4980 Like all other functions, the @code{mark-whole-buffer} function fits
4981 into the template for a function definition. The template looks like
4986 (defun @var{name-of-function} (@var{argument-list})
4987 "@var{documentation}@dots{}"
4988 (@var{interactive-expression}@dots{})
4993 Here is how the function works: the name of the function is
4994 @code{mark-whole-buffer}; it is followed by an empty argument list,
4995 @samp{()}, which means that the function does not require arguments.
4996 The documentation comes next.
4998 The next line is an @code{(interactive)} expression that tells Emacs
4999 that the function will be used interactively. These details are similar
5000 to the @code{simplified-beginning-of-buffer} function described in the
5004 @node Body of mark-whole-buffer, , mark-whole-buffer overview, mark-whole-buffer
5005 @comment node-name, next, previous, up
5006 @subsection Body of @code{mark-whole-buffer}
5008 The body of the @code{mark-whole-buffer} function consists of three
5015 (push-mark (point-max) nil t)
5016 (goto-char (point-min))
5020 The first of these lines is the expression, @code{(push-mark (point))}.
5022 This line does exactly the same job as the first line of the body of
5023 the @code{simplified-beginning-of-buffer} function, which is written
5024 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
5025 at the current position of the cursor.
5027 I don't know why the expression in @code{mark-whole-buffer} is written
5028 @code{(push-mark (point))} and the expression in
5029 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
5030 whoever wrote the code did not know that the arguments for
5031 @code{push-mark} are optional and that if @code{push-mark} is not
5032 passed an argument, the function automatically sets mark at the
5033 location of point by default. Or perhaps the expression was written
5034 so as to parallel the structure of the next line. In any case, the
5035 line causes Emacs to determine the position of point and set a mark
5038 In earlier versions of GNU Emacs, the next line of
5039 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
5040 expression sets a mark at the point in the buffer that has the highest
5041 number. This will be the end of the buffer (or, if the buffer is
5042 narrowed, the end of the accessible portion of the buffer.
5043 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
5044 narrowing.) After this mark has been set, the previous mark, the one
5045 set at point, is no longer set, but Emacs remembers its position, just
5046 as all other recent marks are always remembered. This means that you
5047 can, if you wish, go back to that position by typing @kbd{C-u
5051 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5055 (push-mark (point-max) nil t)
5059 The expression works nearly the same as before. It sets a mark at the
5060 highest numbered place in the buffer that it can. However, in this
5061 version, @code{push-mark} has two additional arguments. The second
5062 argument to @code{push-mark} is @code{nil}. This tells the function
5063 it @emph{should} display a message that says `Mark set' when it pushes
5064 the mark. The third argument is @code{t}. This tells
5065 @code{push-mark} to activate the mark when Transient Mark mode is
5066 turned on. Transient Mark mode highlights the currently active
5067 region. It is often turned off.
5069 Finally, the last line of the function is @code{(goto-char
5070 (point-min)))}. This is written exactly the same way as it is written
5071 in @code{beginning-of-buffer}. The expression moves the cursor to
5072 the minimum point in the buffer, that is, to the beginning of the buffer
5073 (or to the beginning of the accessible portion of the buffer). As a
5074 result of this, point is placed at the beginning of the buffer and mark
5075 is set at the end of the buffer. The whole buffer is, therefore, the
5078 @node append-to-buffer, Buffer Related Review, mark-whole-buffer, Buffer Walk Through
5079 @comment node-name, next, previous, up
5080 @section The Definition of @code{append-to-buffer}
5081 @findex append-to-buffer
5083 The @code{append-to-buffer} command is more complex than the
5084 @code{mark-whole-buffer} command. What it does is copy the region
5085 (that is, the part of the buffer between point and mark) from the
5086 current buffer to a specified buffer.
5089 * append-to-buffer overview::
5090 * append interactive:: A two part interactive expression.
5091 * append-to-buffer body:: Incorporates a @code{let} expression.
5092 * append save-excursion:: How the @code{save-excursion} works.
5095 @node append-to-buffer overview, append interactive, append-to-buffer, append-to-buffer
5097 @unnumberedsubsec An Overview of @code{append-to-buffer}
5100 @findex insert-buffer-substring
5101 The @code{append-to-buffer} command uses the
5102 @code{insert-buffer-substring} function to copy the region.
5103 @code{insert-buffer-substring} is described by its name: it takes a
5104 string of characters from part of a buffer, a ``substring'', and
5105 inserts them into another buffer.
5107 Most of @code{append-to-buffer} is
5108 concerned with setting up the conditions for
5109 @code{insert-buffer-substring} to work: the code must specify both the
5110 buffer to which the text will go, the window it comes from and goes
5111 to, and the region that will be copied.
5114 Here is the complete text of the function:
5118 (defun append-to-buffer (buffer start end)
5119 "Append to specified buffer the text of the region.
5120 It is inserted into that buffer before its point.
5124 When calling from a program, give three arguments:
5125 BUFFER (or buffer name), START and END.
5126 START and END specify the portion of the current buffer to be copied."
5128 (list (read-buffer "Append to buffer: " (other-buffer
5129 (current-buffer) t))
5130 (region-beginning) (region-end)))
5133 (let ((oldbuf (current-buffer)))
5135 (let* ((append-to (get-buffer-create buffer))
5136 (windows (get-buffer-window-list append-to t t))
5138 (set-buffer append-to)
5139 (setq point (point))
5140 (barf-if-buffer-read-only)
5141 (insert-buffer-substring oldbuf start end)
5142 (dolist (window windows)
5143 (when (= (window-point window) point)
5144 (set-window-point window (point))))))))
5148 The function can be understood by looking at it as a series of
5149 filled-in templates.
5151 The outermost template is for the function definition. In this
5152 function, it looks like this (with several slots filled in):
5156 (defun append-to-buffer (buffer start end)
5157 "@var{documentation}@dots{}"
5158 (interactive @dots{})
5163 The first line of the function includes its name and three arguments.
5164 The arguments are the @code{buffer} to which the text will be copied, and
5165 the @code{start} and @code{end} of the region in the current buffer that
5168 The next part of the function is the documentation, which is clear and
5169 complete. As is conventional, the three arguments are written in
5170 upper case so you will notice them easily. Even better, they are
5171 described in the same order as in the argument list.
5173 Note that the documentation distinguishes between a buffer and its
5174 name. (The function can handle either.)
5176 @node append interactive, append-to-buffer body, append-to-buffer overview, append-to-buffer
5177 @comment node-name, next, previous, up
5178 @subsection The @code{append-to-buffer} Interactive Expression
5180 Since the @code{append-to-buffer} function will be used interactively,
5181 the function must have an @code{interactive} expression. (For a
5182 review of @code{interactive}, see @ref{Interactive, , Making a
5183 Function Interactive}.) The expression reads as follows:
5189 "Append to buffer: "
5190 (other-buffer (current-buffer) t))
5197 This expression is not one with letters standing for parts, as
5198 described earlier. Instead, it starts a list with these parts:
5200 The first part of the list is an expression to read the name of a
5201 buffer and return it as a string. That is @code{read-buffer}. The
5202 function requires a prompt as its first argument, @samp{"Append to
5203 buffer: "}. Its second argument tells the command what value to
5204 provide if you don't specify anything.
5206 In this case that second argument is an expression containing the
5207 function @code{other-buffer}, an exception, and a @samp{t}, standing
5210 The first argument to @code{other-buffer}, the exception, is yet
5211 another function, @code{current-buffer}. That is not going to be
5212 returned. The second argument is the symbol for true, @code{t}. that
5213 tells @code{other-buffer} that it may show visible buffers (except in
5214 this case, it will not show the current buffer, which makes sense).
5217 The expression looks like this:
5220 (other-buffer (current-buffer) t)
5223 The second and third arguments to the @code{list} expression are
5224 @code{(region-beginning)} and @code{(region-end)}. These two
5225 functions specify the beginning and end of the text to be appended.
5228 Originally, the command used the letters @samp{B} and @samp{r}.
5229 The whole @code{interactive} expression looked like this:
5232 (interactive "BAppend to buffer:@: \nr")
5236 But when that was done, the default value of the buffer switched to
5237 was invisible. That was not wanted.
5239 (The prompt was separated from the second argument with a newline,
5240 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5241 two arguments that follow the symbol @code{buffer} in the function's
5242 argument list (that is, @code{start} and @code{end}) to the values of
5243 point and mark. That argument worked fine.)
5245 @node append-to-buffer body, append save-excursion, append interactive, append-to-buffer
5246 @comment node-name, next, previous, up
5247 @subsection The Body of @code{append-to-buffer}
5250 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5252 (defun append-to-buffer (buffer start end)
5253 "Append to specified buffer the text of the region.
5254 It is inserted into that buffer before its point.
5256 When calling from a program, give three arguments:
5257 BUFFER (or buffer name), START and END.
5258 START and END specify the portion of the current buffer to be copied."
5260 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5261 (region-beginning) (region-end)))
5262 (let ((oldbuf (current-buffer)))
5264 (let* ((append-to (get-buffer-create buffer))
5265 (windows (get-buffer-window-list append-to t t))
5267 (set-buffer append-to)
5268 (setq point (point))
5269 (barf-if-buffer-read-only)
5270 (insert-buffer-substring oldbuf start end)
5271 (dolist (window windows)
5272 (when (= (window-point window) point)
5273 (set-window-point window (point))))))))
5276 The body of the @code{append-to-buffer} function begins with @code{let}.
5278 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5279 @code{let} expression is to create and give initial values to one or
5280 more variables that will only be used within the body of the
5281 @code{let}. This means that such a variable will not be confused with
5282 any variable of the same name outside the @code{let} expression.
5284 We can see how the @code{let} expression fits into the function as a
5285 whole by showing a template for @code{append-to-buffer} with the
5286 @code{let} expression in outline:
5290 (defun append-to-buffer (buffer start end)
5291 "@var{documentation}@dots{}"
5292 (interactive @dots{})
5293 (let ((@var{variable} @var{value}))
5298 The @code{let} expression has three elements:
5302 The symbol @code{let};
5305 A varlist containing, in this case, a single two-element list,
5306 @code{(@var{variable} @var{value})};
5309 The body of the @code{let} expression.
5313 In the @code{append-to-buffer} function, the varlist looks like this:
5316 (oldbuf (current-buffer))
5320 In this part of the @code{let} expression, the one variable,
5321 @code{oldbuf}, is bound to the value returned by the
5322 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5323 used to keep track of the buffer in which you are working and from
5324 which you will copy.
5326 The element or elements of a varlist are surrounded by a set of
5327 parentheses so the Lisp interpreter can distinguish the varlist from
5328 the body of the @code{let}. As a consequence, the two-element list
5329 within the varlist is surrounded by a circumscribing set of parentheses.
5330 The line looks like this:
5334 (let ((oldbuf (current-buffer)))
5340 The two parentheses before @code{oldbuf} might surprise you if you did
5341 not realize that the first parenthesis before @code{oldbuf} marks the
5342 boundary of the varlist and the second parenthesis marks the beginning
5343 of the two-element list, @code{(oldbuf (current-buffer))}.
5345 @node append save-excursion, , append-to-buffer body, append-to-buffer
5346 @comment node-name, next, previous, up
5347 @subsection @code{save-excursion} in @code{append-to-buffer}
5349 The body of the @code{let} expression in @code{append-to-buffer}
5350 consists of a @code{save-excursion} expression.
5352 The @code{save-excursion} function saves the locations of point and
5353 mark, and restores them to those positions after the expressions in the
5354 body of the @code{save-excursion} complete execution. In addition,
5355 @code{save-excursion} keeps track of the original buffer, and
5356 restores it. This is how @code{save-excursion} is used in
5357 @code{append-to-buffer}.
5360 @cindex Indentation for formatting
5361 @cindex Formatting convention
5362 Incidentally, it is worth noting here that a Lisp function is normally
5363 formatted so that everything that is enclosed in a multi-line spread is
5364 indented more to the right than the first symbol. In this function
5365 definition, the @code{let} is indented more than the @code{defun}, and
5366 the @code{save-excursion} is indented more than the @code{let}, like
5382 This formatting convention makes it easy to see that the lines in
5383 the body of the @code{save-excursion} are enclosed by the parentheses
5384 associated with @code{save-excursion}, just as the
5385 @code{save-excursion} itself is enclosed by the parentheses associated
5386 with the @code{let}:
5390 (let ((oldbuf (current-buffer)))
5393 (set-buffer @dots{})
5394 (insert-buffer-substring oldbuf start end)
5400 The use of the @code{save-excursion} function can be viewed as a process
5401 of filling in the slots of a template:
5406 @var{first-expression-in-body}
5407 @var{second-expression-in-body}
5409 @var{last-expression-in-body})
5415 In this function, the body of the @code{save-excursion} contains only
5416 one expression, the @code{let*} expression. You know about a
5417 @code{let} function. The @code{let*} function is different. It has a
5418 @samp{*} in its name. It enables Emacs to set each variable in its
5419 varlist in sequence, one after another.
5421 Its critical feature is that variables later in the varlist can make
5422 use of the values to which Emacs set variables earlier in the varlist.
5423 @xref{fwd-para let, , The @code{let*} expression}.
5425 We will skip functions like @code{let*} and focus on two: the
5426 @code{set-buffer} function and the @code{insert-buffer-substring}
5430 In the old days, the @code{set-buffer} expression was simply
5433 (set-buffer (get-buffer-create buffer))
5441 (set-buffer append-to)
5445 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5446 on in the @code{let*} expression. That extra binding would not be
5447 necessary except for that @code{append-to} is used later in the
5448 varlist as an argument to @code{get-buffer-window-list}.
5453 (let ((oldbuf (current-buffer)))
5455 (let* ((append-to (get-buffer-create buffer))
5456 (windows (get-buffer-window-list append-to t t))
5458 (set-buffer append-to)
5459 (setq point (point))
5460 (barf-if-buffer-read-only)
5461 (insert-buffer-substring oldbuf start end)
5462 (dolist (window windows)
5463 (when (= (window-point window) point)
5464 (set-window-point window (point))))))))
5467 The @code{append-to-buffer} function definition inserts text from the
5468 buffer in which you are currently to a named buffer. It happens that
5469 @code{insert-buffer-substring} copies text from another buffer to the
5470 current buffer, just the reverse---that is why the
5471 @code{append-to-buffer} definition starts out with a @code{let} that
5472 binds the local symbol @code{oldbuf} to the value returned by
5473 @code{current-buffer}.
5476 The @code{insert-buffer-substring} expression looks like this:
5479 (insert-buffer-substring oldbuf start end)
5483 The @code{insert-buffer-substring} function copies a string
5484 @emph{from} the buffer specified as its first argument and inserts the
5485 string into the present buffer. In this case, the argument to
5486 @code{insert-buffer-substring} is the value of the variable created
5487 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5488 was the current buffer when you gave the @code{append-to-buffer}
5491 After @code{insert-buffer-substring} has done its work,
5492 @code{save-excursion} will restore the action to the original buffer
5493 and @code{append-to-buffer} will have done its job.
5496 Written in skeletal form, the workings of the body look like this:
5500 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5501 (save-excursion ; @r{Keep track of buffer.}
5503 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5505 @var{change-back-to-original-buffer-when-finished}
5506 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5510 In summary, @code{append-to-buffer} works as follows: it saves the
5511 value of the current buffer in the variable called @code{oldbuf}. It
5512 gets the new buffer (creating one if need be) and switches Emacs's
5513 attention to it. Using the value of @code{oldbuf}, it inserts the
5514 region of text from the old buffer into the new buffer; and then using
5515 @code{save-excursion}, it brings you back to your original buffer.
5517 In looking at @code{append-to-buffer}, you have explored a fairly
5518 complex function. It shows how to use @code{let} and
5519 @code{save-excursion}, and how to change to and come back from another
5520 buffer. Many function definitions use @code{let},
5521 @code{save-excursion}, and @code{set-buffer} this way.
5523 @node Buffer Related Review, Buffer Exercises, append-to-buffer, Buffer Walk Through
5524 @comment node-name, next, previous, up
5527 Here is a brief summary of the various functions discussed in this chapter.
5530 @item describe-function
5531 @itemx describe-variable
5532 Print the documentation for a function or variable.
5533 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5536 Find the file containing the source for a function or variable and
5537 switch buffers to it, positioning point at the beginning of the item.
5538 Conventionally bound to @kbd{M-.} (that's a period following the
5541 @item save-excursion
5542 Save the location of point and mark and restore their values after the
5543 arguments to @code{save-excursion} have been evaluated. Also, remember
5544 the current buffer and return to it.
5547 Set mark at a location and record the value of the previous mark on the
5548 mark ring. The mark is a location in the buffer that will keep its
5549 relative position even if text is added to or removed from the buffer.
5552 Set point to the location specified by the value of the argument, which
5553 can be a number, a marker, or an expression that returns the number of
5554 a position, such as @code{(point-min)}.
5556 @item insert-buffer-substring
5557 Copy a region of text from a buffer that is passed to the function as
5558 an argument and insert the region into the current buffer.
5560 @item mark-whole-buffer
5561 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5564 Switch the attention of Emacs to another buffer, but do not change the
5565 window being displayed. Used when the program rather than a human is
5566 to work on a different buffer.
5568 @item get-buffer-create
5570 Find a named buffer or create one if a buffer of that name does not
5571 exist. The @code{get-buffer} function returns @code{nil} if the named
5572 buffer does not exist.
5576 @node Buffer Exercises, , Buffer Related Review, Buffer Walk Through
5581 Write your own @code{simplified-end-of-buffer} function definition;
5582 then test it to see whether it works.
5585 Use @code{if} and @code{get-buffer} to write a function that prints a
5586 message telling you whether a buffer exists.
5589 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5593 @node More Complex, Narrowing & Widening, Buffer Walk Through, Top
5594 @comment node-name, next, previous, up
5595 @chapter A Few More Complex Functions
5597 In this chapter, we build on what we have learned in previous chapters
5598 by looking at more complex functions. The @code{copy-to-buffer}
5599 function illustrates use of two @code{save-excursion} expressions in
5600 one definition, while the @code{insert-buffer} function illustrates
5601 use of an asterisk in an @code{interactive} expression, use of
5602 @code{or}, and the important distinction between a name and the object
5603 to which the name refers.
5606 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5607 * insert-buffer:: Read-only, and with @code{or}.
5608 * beginning-of-buffer:: Shows @code{goto-char},
5609 @code{point-min}, and @code{push-mark}.
5610 * Second Buffer Related Review::
5611 * optional Exercise::
5614 @node copy-to-buffer, insert-buffer, More Complex, More Complex
5615 @comment node-name, next, previous, up
5616 @section The Definition of @code{copy-to-buffer}
5617 @findex copy-to-buffer
5619 After understanding how @code{append-to-buffer} works, it is easy to
5620 understand @code{copy-to-buffer}. This function copies text into a
5621 buffer, but instead of adding to the second buffer, it replaces all the
5622 previous text in the second buffer.
5625 The body of @code{copy-to-buffer} looks like this,
5630 (interactive "BCopy to buffer: \nr")
5631 (let ((oldbuf (current-buffer)))
5632 (with-current-buffer (get-buffer-create buffer)
5633 (barf-if-buffer-read-only)
5636 (insert-buffer-substring oldbuf start end)))))
5640 The @code{copy-to-buffer} function has a simpler @code{interactive}
5641 expression than @code{append-to-buffer}.
5644 The definition then says
5647 (with-current-buffer (get-buffer-create buffer) @dots{}
5650 First, look at the earliest inner expression; that is evaluated first.
5651 That expression starts with @code{get-buffer-create buffer}. The
5652 function tells the computer to use the buffer with the name specified
5653 as the one to which you are copying, or if such a buffer does not
5654 exist, to create it. Then, the @code{with-current-buffer} function
5655 evaluates its body with that buffer temporarily current.
5657 (This demonstrates another way to shift the computer's attention but
5658 not the user's. The @code{append-to-buffer} function showed how to do
5659 the same with @code{save-excursion} and @code{set-buffer}.
5660 @code{with-current-buffer} is a newer, and arguably easier,
5663 The @code{barf-if-buffer-read-only} function sends you an error
5664 message saying the buffer is read-only if you cannot modify it.
5666 The next line has the @code{erase-buffer} function as its sole
5667 contents. That function erases the buffer.
5669 Finally, the last two lines contain the @code{save-excursion}
5670 expression with @code{insert-buffer-substring} as its body.
5671 The @code{insert-buffer-substring} expression copies the text from
5672 the buffer you are in (and you have not seen the computer shift its
5673 attention, so you don't know that that buffer is now called
5676 Incidentally, this is what is meant by `replacement'. To replace text,
5677 Emacs erases the previous text and then inserts new text.
5680 In outline, the body of @code{copy-to-buffer} looks like this:
5684 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5685 (@var{with-the-buffer-you-are-copying-to}
5686 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5689 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5693 @node insert-buffer, beginning-of-buffer, copy-to-buffer, More Complex
5694 @comment node-name, next, previous, up
5695 @section The Definition of @code{insert-buffer}
5696 @findex insert-buffer
5698 @code{insert-buffer} is yet another buffer-related function. This
5699 command copies another buffer @emph{into} the current buffer. It is the
5700 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5701 copy a region of text @emph{from} the current buffer to another buffer.
5703 Here is a discussion based on the original code. The code was
5704 simplified in 2003 and is harder to understand.
5706 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5707 a discussion of the new body.)
5709 In addition, this code illustrates the use of @code{interactive} with a
5710 buffer that might be @dfn{read-only} and the important distinction
5711 between the name of an object and the object actually referred to.
5714 * insert-buffer code::
5715 * insert-buffer interactive:: When you can read, but not write.
5716 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5717 * if & or:: Using an @code{if} instead of an @code{or}.
5718 * Insert or:: How the @code{or} expression works.
5719 * Insert let:: Two @code{save-excursion} expressions.
5720 * New insert-buffer::
5723 @node insert-buffer code, insert-buffer interactive, insert-buffer, insert-buffer
5725 @unnumberedsubsec The Code for @code{insert-buffer}
5729 Here is the earlier code:
5733 (defun insert-buffer (buffer)
5734 "Insert after point the contents of BUFFER.
5735 Puts mark after the inserted text.
5736 BUFFER may be a buffer or a buffer name."
5737 (interactive "*bInsert buffer:@: ")
5740 (or (bufferp buffer)
5741 (setq buffer (get-buffer buffer)))
5742 (let (start end newmark)
5746 (setq start (point-min) end (point-max)))
5749 (insert-buffer-substring buffer start end)
5750 (setq newmark (point)))
5751 (push-mark newmark)))
5756 As with other function definitions, you can use a template to see an
5757 outline of the function:
5761 (defun insert-buffer (buffer)
5762 "@var{documentation}@dots{}"
5763 (interactive "*bInsert buffer:@: ")
5768 @node insert-buffer interactive, insert-buffer body, insert-buffer code, insert-buffer
5769 @comment node-name, next, previous, up
5770 @subsection The Interactive Expression in @code{insert-buffer}
5771 @findex interactive, @r{example use of}
5773 In @code{insert-buffer}, the argument to the @code{interactive}
5774 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5778 * Read-only buffer:: When a buffer cannot be modified.
5779 * b for interactive:: An existing buffer or else its name.
5782 @node Read-only buffer, b for interactive, insert-buffer interactive, insert-buffer interactive
5783 @comment node-name, next, previous, up
5784 @unnumberedsubsubsec A Read-only Buffer
5785 @cindex Read-only buffer
5786 @cindex Asterisk for read-only buffer
5787 @findex * @r{for read-only buffer}
5789 The asterisk is for the situation when the current buffer is a
5790 read-only buffer---a buffer that cannot be modified. If
5791 @code{insert-buffer} is called when the current buffer is read-only, a
5792 message to this effect is printed in the echo area and the terminal
5793 may beep or blink at you; you will not be permitted to insert anything
5794 into current buffer. The asterisk does not need to be followed by a
5795 newline to separate it from the next argument.
5797 @node b for interactive, , Read-only buffer, insert-buffer interactive
5798 @comment node-name, next, previous, up
5799 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5801 The next argument in the interactive expression starts with a lower
5802 case @samp{b}. (This is different from the code for
5803 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5804 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5805 The lower-case @samp{b} tells the Lisp interpreter that the argument
5806 for @code{insert-buffer} should be an existing buffer or else its
5807 name. (The upper-case @samp{B} option provides for the possibility
5808 that the buffer does not exist.) Emacs will prompt you for the name
5809 of the buffer, offering you a default buffer, with name completion
5810 enabled. If the buffer does not exist, you receive a message that
5811 says ``No match''; your terminal may beep at you as well.
5813 The new and simplified code generates a list for @code{interactive}.
5814 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5815 functions with which we are already familiar and the @code{progn}
5816 special form with which we are not. (It will be described later.)
5818 @node insert-buffer body, if & or, insert-buffer interactive, insert-buffer
5819 @comment node-name, next, previous, up
5820 @subsection The Body of the @code{insert-buffer} Function
5822 The body of the @code{insert-buffer} function has two major parts: an
5823 @code{or} expression and a @code{let} expression. The purpose of the
5824 @code{or} expression is to ensure that the argument @code{buffer} is
5825 bound to a buffer and not just the name of a buffer. The body of the
5826 @code{let} expression contains the code which copies the other buffer
5827 into the current buffer.
5830 In outline, the two expressions fit into the @code{insert-buffer}
5835 (defun insert-buffer (buffer)
5836 "@var{documentation}@dots{}"
5837 (interactive "*bInsert buffer:@: ")
5842 (let (@var{varlist})
5843 @var{body-of-}@code{let}@dots{} )
5847 To understand how the @code{or} expression ensures that the argument
5848 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5849 is first necessary to understand the @code{or} function.
5851 Before doing this, let me rewrite this part of the function using
5852 @code{if} so that you can see what is done in a manner that will be familiar.
5854 @node if & or, Insert or, insert-buffer body, insert-buffer
5855 @comment node-name, next, previous, up
5856 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5858 The job to be done is to make sure the value of @code{buffer} is a
5859 buffer itself and not the name of a buffer. If the value is the name,
5860 then the buffer itself must be got.
5862 You can imagine yourself at a conference where an usher is wandering
5863 around holding a list with your name on it and looking for you: the
5864 usher is ``bound'' to your name, not to you; but when the usher finds
5865 you and takes your arm, the usher becomes ``bound'' to you.
5868 In Lisp, you might describe this situation like this:
5872 (if (not (holding-on-to-guest))
5873 (find-and-take-arm-of-guest))
5877 We want to do the same thing with a buffer---if we do not have the
5878 buffer itself, we want to get it.
5881 Using a predicate called @code{bufferp} that tells us whether we have a
5882 buffer (rather than its name), we can write the code like this:
5886 (if (not (bufferp buffer)) ; @r{if-part}
5887 (setq buffer (get-buffer buffer))) ; @r{then-part}
5892 Here, the true-or-false-test of the @code{if} expression is
5893 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5894 @w{@code{(setq buffer (get-buffer buffer))}}.
5896 In the test, the function @code{bufferp} returns true if its argument is
5897 a buffer---but false if its argument is the name of the buffer. (The
5898 last character of the function name @code{bufferp} is the character
5899 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5900 indicates that the function is a predicate, which is a term that means
5901 that the function will determine whether some property is true or false.
5902 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5906 The function @code{not} precedes the expression @code{(bufferp buffer)},
5907 so the true-or-false-test looks like this:
5910 (not (bufferp buffer))
5914 @code{not} is a function that returns true if its argument is false
5915 and false if its argument is true. So if @code{(bufferp buffer)}
5916 returns true, the @code{not} expression returns false and vice-verse:
5917 what is ``not true'' is false and what is ``not false'' is true.
5919 Using this test, the @code{if} expression works as follows: when the
5920 value of the variable @code{buffer} is actually a buffer rather than
5921 its name, the true-or-false-test returns false and the @code{if}
5922 expression does not evaluate the then-part. This is fine, since we do
5923 not need to do anything to the variable @code{buffer} if it really is
5926 On the other hand, when the value of @code{buffer} is not a buffer
5927 itself, but the name of a buffer, the true-or-false-test returns true
5928 and the then-part of the expression is evaluated. In this case, the
5929 then-part is @code{(setq buffer (get-buffer buffer))}. This
5930 expression uses the @code{get-buffer} function to return an actual
5931 buffer itself, given its name. The @code{setq} then sets the variable
5932 @code{buffer} to the value of the buffer itself, replacing its previous
5933 value (which was the name of the buffer).
5935 @node Insert or, Insert let, if & or, insert-buffer
5936 @comment node-name, next, previous, up
5937 @subsection The @code{or} in the Body
5939 The purpose of the @code{or} expression in the @code{insert-buffer}
5940 function is to ensure that the argument @code{buffer} is bound to a
5941 buffer and not just to the name of a buffer. The previous section shows
5942 how the job could have been done using an @code{if} expression.
5943 However, the @code{insert-buffer} function actually uses @code{or}.
5944 To understand this, it is necessary to understand how @code{or} works.
5947 An @code{or} function can have any number of arguments. It evaluates
5948 each argument in turn and returns the value of the first of its
5949 arguments that is not @code{nil}. Also, and this is a crucial feature
5950 of @code{or}, it does not evaluate any subsequent arguments after
5951 returning the first non-@code{nil} value.
5954 The @code{or} expression looks like this:
5958 (or (bufferp buffer)
5959 (setq buffer (get-buffer buffer)))
5964 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5965 This expression returns true (a non-@code{nil} value) if the buffer is
5966 actually a buffer, and not just the name of a buffer. In the @code{or}
5967 expression, if this is the case, the @code{or} expression returns this
5968 true value and does not evaluate the next expression---and this is fine
5969 with us, since we do not want to do anything to the value of
5970 @code{buffer} if it really is a buffer.
5972 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5973 which it will be if the value of @code{buffer} is the name of a buffer,
5974 the Lisp interpreter evaluates the next element of the @code{or}
5975 expression. This is the expression @code{(setq buffer (get-buffer
5976 buffer))}. This expression returns a non-@code{nil} value, which
5977 is the value to which it sets the variable @code{buffer}---and this
5978 value is a buffer itself, not the name of a buffer.
5980 The result of all this is that the symbol @code{buffer} is always
5981 bound to a buffer itself rather than to the name of a buffer. All
5982 this is necessary because the @code{set-buffer} function in a
5983 following line only works with a buffer itself, not with the name to a
5987 Incidentally, using @code{or}, the situation with the usher would be
5991 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5994 @node Insert let, New insert-buffer, Insert or, insert-buffer
5995 @comment node-name, next, previous, up
5996 @subsection The @code{let} Expression in @code{insert-buffer}
5998 After ensuring that the variable @code{buffer} refers to a buffer itself
5999 and not just to the name of a buffer, the @code{insert-buffer function}
6000 continues with a @code{let} expression. This specifies three local
6001 variables, @code{start}, @code{end}, and @code{newmark} and binds them
6002 to the initial value @code{nil}. These variables are used inside the
6003 remainder of the @code{let} and temporarily hide any other occurrence of
6004 variables of the same name in Emacs until the end of the @code{let}.
6007 The body of the @code{let} contains two @code{save-excursion}
6008 expressions. First, we will look at the inner @code{save-excursion}
6009 expression in detail. The expression looks like this:
6015 (setq start (point-min) end (point-max)))
6020 The expression @code{(set-buffer buffer)} changes Emacs's attention
6021 from the current buffer to the one from which the text will copied.
6022 In that buffer, the variables @code{start} and @code{end} are set to
6023 the beginning and end of the buffer, using the commands
6024 @code{point-min} and @code{point-max}. Note that we have here an
6025 illustration of how @code{setq} is able to set two variables in the
6026 same expression. The first argument of @code{setq} is set to the
6027 value of its second, and its third argument is set to the value of its
6030 After the body of the inner @code{save-excursion} is evaluated, the
6031 @code{save-excursion} restores the original buffer, but @code{start} and
6032 @code{end} remain set to the values of the beginning and end of the
6033 buffer from which the text will be copied.
6036 The outer @code{save-excursion} expression looks like this:
6041 (@var{inner-}@code{save-excursion}@var{-expression}
6042 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
6043 (insert-buffer-substring buffer start end)
6044 (setq newmark (point)))
6049 The @code{insert-buffer-substring} function copies the text
6050 @emph{into} the current buffer @emph{from} the region indicated by
6051 @code{start} and @code{end} in @code{buffer}. Since the whole of the
6052 second buffer lies between @code{start} and @code{end}, the whole of
6053 the second buffer is copied into the buffer you are editing. Next,
6054 the value of point, which will be at the end of the inserted text, is
6055 recorded in the variable @code{newmark}.
6057 After the body of the outer @code{save-excursion} is evaluated, point
6058 and mark are relocated to their original places.
6060 However, it is convenient to locate a mark at the end of the newly
6061 inserted text and locate point at its beginning. The @code{newmark}
6062 variable records the end of the inserted text. In the last line of
6063 the @code{let} expression, the @code{(push-mark newmark)} expression
6064 function sets a mark to this location. (The previous location of the
6065 mark is still accessible; it is recorded on the mark ring and you can
6066 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6067 located at the beginning of the inserted text, which is where it was
6068 before you called the insert function, the position of which was saved
6069 by the first @code{save-excursion}.
6072 The whole @code{let} expression looks like this:
6076 (let (start end newmark)
6080 (setq start (point-min) end (point-max)))
6081 (insert-buffer-substring buffer start end)
6082 (setq newmark (point)))
6083 (push-mark newmark))
6087 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6088 function uses @code{let}, @code{save-excursion}, and
6089 @code{set-buffer}. In addition, the function illustrates one way to
6090 use @code{or}. All these functions are building blocks that we will
6091 find and use again and again.
6093 @node New insert-buffer, , Insert let, insert-buffer
6094 @comment node-name, next, previous, up
6095 @subsection New Body for @code{insert-buffer}
6096 @findex insert-buffer, new version body
6097 @findex new version body for insert-buffer
6099 The body in the GNU Emacs 22 version is more confusing than the original.
6102 It consists of two expressions,
6108 (insert-buffer-substring (get-buffer buffer))
6116 except, and this is what confuses novices, very important work is done
6117 inside the @code{push-mark} expression.
6119 The @code{get-buffer} function returns a buffer with the name
6120 provided. You will note that the function is @emph{not} called
6121 @code{get-buffer-create}; it does not create a buffer if one does not
6122 already exist. The buffer returned by @code{get-buffer}, an existing
6123 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6124 whole of the buffer (since you did not specify anything else).
6126 The location into which the buffer is inserted is recorded by
6127 @code{push-mark}. Then the function returns @code{nil}, the value of
6128 its last command. Put another way, the @code{insert-buffer} function
6129 exists only to produce a side effect, inserting another buffer, not to
6132 @node beginning-of-buffer, Second Buffer Related Review, insert-buffer, More Complex
6133 @comment node-name, next, previous, up
6134 @section Complete Definition of @code{beginning-of-buffer}
6135 @findex beginning-of-buffer
6137 The basic structure of the @code{beginning-of-buffer} function has
6138 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6139 Simplified @code{beginning-of-buffer} Definition}.)
6140 This section describes the complex part of the definition.
6142 As previously described, when invoked without an argument,
6143 @code{beginning-of-buffer} moves the cursor to the beginning of the
6144 buffer (in truth, the beginning of the accessible portion of the
6145 buffer), leaving the mark at the previous position. However, when the
6146 command is invoked with a number between one and ten, the function
6147 considers that number to be a fraction of the length of the buffer,
6148 measured in tenths, and Emacs moves the cursor that fraction of the
6149 way from the beginning of the buffer. Thus, you can either call this
6150 function with the key command @kbd{M-<}, which will move the cursor to
6151 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6152 M-<} which will move the cursor to a point 70% of the way through the
6153 buffer. If a number bigger than ten is used for the argument, it
6154 moves to the end of the buffer.
6156 The @code{beginning-of-buffer} function can be called with or without an
6157 argument. The use of the argument is optional.
6160 * Optional Arguments::
6161 * beginning-of-buffer opt arg:: Example with optional argument.
6162 * beginning-of-buffer complete::
6165 @node Optional Arguments, beginning-of-buffer opt arg, beginning-of-buffer, beginning-of-buffer
6166 @subsection Optional Arguments
6168 Unless told otherwise, Lisp expects that a function with an argument in
6169 its function definition will be called with a value for that argument.
6170 If that does not happen, you get an error and a message that says
6171 @samp{Wrong number of arguments}.
6173 @cindex Optional arguments
6176 However, optional arguments are a feature of Lisp: a particular
6177 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6178 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6179 @samp{optional} is part of the keyword.) In a function definition, if
6180 an argument follows the keyword @code{&optional}, no value need be
6181 passed to that argument when the function is called.
6184 The first line of the function definition of @code{beginning-of-buffer}
6185 therefore looks like this:
6188 (defun beginning-of-buffer (&optional arg)
6192 In outline, the whole function looks like this:
6196 (defun beginning-of-buffer (&optional arg)
6197 "@var{documentation}@dots{}"
6199 (or (@var{is-the-argument-a-cons-cell} arg)
6200 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6202 (let (@var{determine-size-and-set-it})
6204 (@var{if-there-is-an-argument}
6205 @var{figure-out-where-to-go}
6212 The function is similar to the @code{simplified-beginning-of-buffer}
6213 function except that the @code{interactive} expression has @code{"P"}
6214 as an argument and the @code{goto-char} function is followed by an
6215 if-then-else expression that figures out where to put the cursor if
6216 there is an argument that is not a cons cell.
6218 (Since I do not explain a cons cell for many more chapters, please
6219 consider ignoring the function @code{consp}. @xref{List
6220 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6221 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6224 The @code{"P"} in the @code{interactive} expression tells Emacs to
6225 pass a prefix argument, if there is one, to the function in raw form.
6226 A prefix argument is made by typing the @key{META} key followed by a
6227 number, or by typing @kbd{C-u} and then a number. (If you don't type
6228 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6229 @code{"p"} in the @code{interactive} expression causes the function to
6230 convert a prefix arg to a number.)
6232 The true-or-false-test of the @code{if} expression looks complex, but
6233 it is not: it checks whether @code{arg} has a value that is not
6234 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6235 does; it checks whether its argument is a cons cell.) If @code{arg}
6236 has a value that is not @code{nil} (and is not a cons cell), which
6237 will be the case if @code{beginning-of-buffer} is called with a
6238 numeric argument, then this true-or-false-test will return true and
6239 the then-part of the @code{if} expression will be evaluated. On the
6240 other hand, if @code{beginning-of-buffer} is not called with an
6241 argument, the value of @code{arg} will be @code{nil} and the else-part
6242 of the @code{if} expression will be evaluated. The else-part is
6243 simply @code{point-min}, and when this is the outcome, the whole
6244 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6245 is how we saw the @code{beginning-of-buffer} function in its
6248 @node beginning-of-buffer opt arg, beginning-of-buffer complete, Optional Arguments, beginning-of-buffer
6249 @subsection @code{beginning-of-buffer} with an Argument
6251 When @code{beginning-of-buffer} is called with an argument, an
6252 expression is evaluated which calculates what value to pass to
6253 @code{goto-char}. This expression is rather complicated at first sight.
6254 It includes an inner @code{if} expression and much arithmetic. It looks
6259 (if (> (buffer-size) 10000)
6260 ;; @r{Avoid overflow for large buffer sizes!}
6261 (* (prefix-numeric-value arg)
6266 size (prefix-numeric-value arg))) 10)))
6271 * Disentangle beginning-of-buffer::
6272 * Large buffer case::
6273 * Small buffer case::
6276 @node Disentangle beginning-of-buffer, Large buffer case, beginning-of-buffer opt arg, beginning-of-buffer opt arg
6278 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6281 Like other complex-looking expressions, the conditional expression
6282 within @code{beginning-of-buffer} can be disentangled by looking at it
6283 as parts of a template, in this case, the template for an if-then-else
6284 expression. In skeletal form, the expression looks like this:
6288 (if (@var{buffer-is-large}
6289 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6290 @var{else-use-alternate-calculation}
6294 The true-or-false-test of this inner @code{if} expression checks the
6295 size of the buffer. The reason for this is that the old version 18
6296 Emacs used numbers that are no bigger than eight million or so and in
6297 the computation that followed, the programmer feared that Emacs might
6298 try to use over-large numbers if the buffer were large. The term
6299 `overflow', mentioned in the comment, means numbers that are over
6300 large. More recent versions of Emacs use larger numbers, but this
6301 code has not been touched, if only because people now look at buffers
6302 that are far, far larger than ever before.
6304 There are two cases: if the buffer is large and if it is not.
6306 @node Large buffer case, Small buffer case, Disentangle beginning-of-buffer, beginning-of-buffer opt arg
6307 @comment node-name, next, previous, up
6308 @unnumberedsubsubsec What happens in a large buffer
6310 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6311 whether the size of the buffer is greater than 10,000 characters. To do
6312 this, it uses the @code{>} function and the computation of @code{size}
6313 that comes from the let expression.
6315 In the old days, the function @code{buffer-size} was used. Not only
6316 was that function called several times, it gave the size of the whole
6317 buffer, not the accessible part. The computation makes much more
6318 sense when it handles just the accessible part. (@xref{Narrowing &
6319 Widening, , Narrowing and Widening}, for more information on focusing
6320 attention to an `accessible' part.)
6323 The line looks like this:
6331 When the buffer is large, the then-part of the @code{if} expression is
6332 evaluated. It reads like this (after formatting for easy reading):
6337 (prefix-numeric-value arg)
6343 This expression is a multiplication, with two arguments to the function
6346 The first argument is @code{(prefix-numeric-value arg)}. When
6347 @code{"P"} is used as the argument for @code{interactive}, the value
6348 passed to the function as its argument is passed a ``raw prefix
6349 argument'', and not a number. (It is a number in a list.) To perform
6350 the arithmetic, a conversion is necessary, and
6351 @code{prefix-numeric-value} does the job.
6353 @findex / @r{(division)}
6355 The second argument is @code{(/ size 10)}. This expression divides
6356 the numeric value by ten --- the numeric value of the size of the
6357 accessible portion of the buffer. This produces a number that tells
6358 how many characters make up one tenth of the buffer size. (In Lisp,
6359 @code{/} is used for division, just as @code{*} is used for
6363 In the multiplication expression as a whole, this amount is multiplied
6364 by the value of the prefix argument---the multiplication looks like this:
6368 (* @var{numeric-value-of-prefix-arg}
6369 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6374 If, for example, the prefix argument is @samp{7}, the one-tenth value
6375 will be multiplied by 7 to give a position 70% of the way through.
6378 The result of all this is that if the accessible portion of the buffer
6379 is large, the @code{goto-char} expression reads like this:
6383 (goto-char (* (prefix-numeric-value arg)
6388 This puts the cursor where we want it.
6390 @node Small buffer case, , Large buffer case, beginning-of-buffer opt arg
6391 @comment node-name, next, previous, up
6392 @unnumberedsubsubsec What happens in a small buffer
6394 If the buffer contains fewer than 10,000 characters, a slightly
6395 different computation is performed. You might think this is not
6396 necessary, since the first computation could do the job. However, in
6397 a small buffer, the first method may not put the cursor on exactly the
6398 desired line; the second method does a better job.
6401 The code looks like this:
6403 @c Keep this on one line.
6405 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6410 This is code in which you figure out what happens by discovering how the
6411 functions are embedded in parentheses. It is easier to read if you
6412 reformat it with each expression indented more deeply than its
6413 enclosing expression:
6421 (prefix-numeric-value arg)))
6428 Looking at parentheses, we see that the innermost operation is
6429 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6430 a number. In the following expression, this number is multiplied by
6431 the size of the accessible portion of the buffer:
6434 (* size (prefix-numeric-value arg))
6438 This multiplication creates a number that may be larger than the size of
6439 the buffer---seven times larger if the argument is 7, for example. Ten
6440 is then added to this number and finally the large number is divided by
6441 ten to provide a value that is one character larger than the percentage
6442 position in the buffer.
6444 The number that results from all this is passed to @code{goto-char} and
6445 the cursor is moved to that point.
6448 @node beginning-of-buffer complete, , beginning-of-buffer opt arg, beginning-of-buffer
6449 @comment node-name, next, previous, up
6450 @subsection The Complete @code{beginning-of-buffer}
6453 Here is the complete text of the @code{beginning-of-buffer} function:
6459 (defun beginning-of-buffer (&optional arg)
6460 "Move point to the beginning of the buffer;
6461 leave mark at previous position.
6462 With \\[universal-argument] prefix,
6463 do not set mark at previous position.
6465 put point N/10 of the way from the beginning.
6467 If the buffer is narrowed,
6468 this command uses the beginning and size
6469 of the accessible part of the buffer.
6473 Don't use this command in Lisp programs!
6474 \(goto-char (point-min)) is faster
6475 and avoids clobbering the mark."
6478 (and transient-mark-mode mark-active)
6482 (let ((size (- (point-max) (point-min))))
6483 (goto-char (if (and arg (not (consp arg)))
6486 ;; Avoid overflow for large buffer sizes!
6487 (* (prefix-numeric-value arg)
6489 (/ (+ 10 (* size (prefix-numeric-value arg)))
6492 (if arg (forward-line 1)))
6497 From before GNU Emacs 22
6500 (defun beginning-of-buffer (&optional arg)
6501 "Move point to the beginning of the buffer;
6502 leave mark at previous position.
6503 With arg N, put point N/10 of the way
6504 from the true beginning.
6507 Don't use this in Lisp programs!
6508 \(goto-char (point-min)) is faster
6509 and does not set the mark."
6516 (if (> (buffer-size) 10000)
6517 ;; @r{Avoid overflow for large buffer sizes!}
6518 (* (prefix-numeric-value arg)
6519 (/ (buffer-size) 10))
6522 (/ (+ 10 (* (buffer-size)
6523 (prefix-numeric-value arg)))
6526 (if arg (forward-line 1)))
6532 Except for two small points, the previous discussion shows how this
6533 function works. The first point deals with a detail in the
6534 documentation string, and the second point concerns the last line of
6538 In the documentation string, there is reference to an expression:
6541 \\[universal-argument]
6545 A @samp{\\} is used before the first square bracket of this
6546 expression. This @samp{\\} tells the Lisp interpreter to substitute
6547 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6548 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6549 be different. (@xref{Documentation Tips, , Tips for Documentation
6550 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6554 Finally, the last line of the @code{beginning-of-buffer} command says
6555 to move point to the beginning of the next line if the command is
6556 invoked with an argument:
6559 (if arg (forward-line 1)))
6563 This puts the cursor at the beginning of the first line after the
6564 appropriate tenths position in the buffer. This is a flourish that
6565 means that the cursor is always located @emph{at least} the requested
6566 tenths of the way through the buffer, which is a nicety that is,
6567 perhaps, not necessary, but which, if it did not occur, would be sure
6570 On the other hand, it also means that if you specify the command with
6571 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6572 argument' is simply a cons cell, then the command puts you at the
6573 beginning of the second line @dots{} I don't know whether this is
6574 intended or whether no one has dealt with the code to avoid this
6577 @node Second Buffer Related Review, optional Exercise, beginning-of-buffer, More Complex
6578 @comment node-name, next, previous, up
6581 Here is a brief summary of some of the topics covered in this chapter.
6585 Evaluate each argument in sequence, and return the value of the first
6586 argument that is not @code{nil}; if none return a value that is not
6587 @code{nil}, return @code{nil}. In brief, return the first true value
6588 of the arguments; return a true value if one @emph{or} any of the
6592 Evaluate each argument in sequence, and if any are @code{nil}, return
6593 @code{nil}; if none are @code{nil}, return the value of the last
6594 argument. In brief, return a true value only if all the arguments are
6595 true; return a true value if one @emph{and} each of the others is
6599 A keyword used to indicate that an argument to a function definition
6600 is optional; this means that the function can be evaluated without the
6601 argument, if desired.
6603 @item prefix-numeric-value
6604 Convert the `raw prefix argument' produced by @code{(interactive
6605 "P")} to a numeric value.
6608 Move point forward to the beginning of the next line, or if the argument
6609 is greater than one, forward that many lines. If it can't move as far
6610 forward as it is supposed to, @code{forward-line} goes forward as far as
6611 it can and then returns a count of the number of additional lines it was
6612 supposed to move but couldn't.
6615 Delete the entire contents of the current buffer.
6618 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6621 @node optional Exercise, , Second Buffer Related Review, More Complex
6622 @section @code{optional} Argument Exercise
6624 Write an interactive function with an optional argument that tests
6625 whether its argument, a number, is greater than or equal to, or else,
6626 less than the value of @code{fill-column}, and tells you which, in a
6627 message. However, if you do not pass an argument to the function, use
6628 56 as a default value.
6630 @node Narrowing & Widening, car cdr & cons, More Complex, Top
6631 @comment node-name, next, previous, up
6632 @chapter Narrowing and Widening
6633 @cindex Focusing attention (narrowing)
6637 Narrowing is a feature of Emacs that makes it possible for you to focus
6638 on a specific part of a buffer, and work without accidentally changing
6639 other parts. Narrowing is normally disabled since it can confuse
6643 * Narrowing advantages:: The advantages of narrowing
6644 * save-restriction:: The @code{save-restriction} special form.
6645 * what-line:: The number of the line that point is on.
6649 @node Narrowing advantages, save-restriction, Narrowing & Widening, Narrowing & Widening
6651 @unnumberedsec The Advantages of Narrowing
6654 With narrowing, the rest of a buffer is made invisible, as if it weren't
6655 there. This is an advantage if, for example, you want to replace a word
6656 in one part of a buffer but not in another: you narrow to the part you want
6657 and the replacement is carried out only in that section, not in the rest
6658 of the buffer. Searches will only work within a narrowed region, not
6659 outside of one, so if you are fixing a part of a document, you can keep
6660 yourself from accidentally finding parts you do not need to fix by
6661 narrowing just to the region you want.
6662 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6664 However, narrowing does make the rest of the buffer invisible, which
6665 can scare people who inadvertently invoke narrowing and think they
6666 have deleted a part of their file. Moreover, the @code{undo} command
6667 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6668 (nor should it), so people can become quite desperate if they do not
6669 know that they can return the rest of a buffer to visibility with the
6670 @code{widen} command.
6671 (The key binding for @code{widen} is @kbd{C-x n w}.)
6673 Narrowing is just as useful to the Lisp interpreter as to a human.
6674 Often, an Emacs Lisp function is designed to work on just part of a
6675 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6676 buffer that has been narrowed. The @code{what-line} function, for
6677 example, removes the narrowing from a buffer, if it has any narrowing
6678 and when it has finished its job, restores the narrowing to what it was.
6679 On the other hand, the @code{count-lines} function, which is called by
6680 @code{what-line}, uses narrowing to restrict itself to just that portion
6681 of the buffer in which it is interested and then restores the previous
6684 @node save-restriction, what-line, Narrowing advantages, Narrowing & Widening
6685 @comment node-name, next, previous, up
6686 @section The @code{save-restriction} Special Form
6687 @findex save-restriction
6689 In Emacs Lisp, you can use the @code{save-restriction} special form to
6690 keep track of whatever narrowing is in effect, if any. When the Lisp
6691 interpreter meets with @code{save-restriction}, it executes the code
6692 in the body of the @code{save-restriction} expression, and then undoes
6693 any changes to narrowing that the code caused. If, for example, the
6694 buffer is narrowed and the code that follows @code{save-restriction}
6695 gets rid of the narrowing, @code{save-restriction} returns the buffer
6696 to its narrowed region afterwards. In the @code{what-line} command,
6697 any narrowing the buffer may have is undone by the @code{widen}
6698 command that immediately follows the @code{save-restriction} command.
6699 Any original narrowing is restored just before the completion of the
6703 The template for a @code{save-restriction} expression is simple:
6713 The body of the @code{save-restriction} is one or more expressions that
6714 will be evaluated in sequence by the Lisp interpreter.
6716 Finally, a point to note: when you use both @code{save-excursion} and
6717 @code{save-restriction}, one right after the other, you should use
6718 @code{save-excursion} outermost. If you write them in reverse order,
6719 you may fail to record narrowing in the buffer to which Emacs switches
6720 after calling @code{save-excursion}. Thus, when written together,
6721 @code{save-excursion} and @code{save-restriction} should be written
6732 In other circumstances, when not written together, the
6733 @code{save-excursion} and @code{save-restriction} special forms must
6734 be written in the order appropriate to the function.
6750 /usr/local/src/emacs/lisp/simple.el
6753 "Print the current buffer line number and narrowed line number of point."
6755 (let ((start (point-min))
6756 (n (line-number-at-pos)))
6758 (message "Line %d" n)
6762 (message "line %d (narrowed line %d)"
6763 (+ n (line-number-at-pos start) -1) n))))))
6765 (defun line-number-at-pos (&optional pos)
6766 "Return (narrowed) buffer line number at position POS.
6767 If POS is nil, use current buffer location.
6768 Counting starts at (point-min), so the value refers
6769 to the contents of the accessible portion of the buffer."
6770 (let ((opoint (or pos (point))) start)
6772 (goto-char (point-min))
6773 (setq start (point))
6776 (1+ (count-lines start (point))))))
6778 (defun count-lines (start end)
6779 "Return number of lines between START and END.
6780 This is usually the number of newlines between them,
6781 but can be one more if START is not equal to END
6782 and the greater of them is not at the start of a line."
6785 (narrow-to-region start end)
6786 (goto-char (point-min))
6787 (if (eq selective-display t)
6790 (while (re-search-forward "[\n\C-m]" nil t 40)
6791 (setq done (+ 40 done)))
6792 (while (re-search-forward "[\n\C-m]" nil t 1)
6793 (setq done (+ 1 done)))
6794 (goto-char (point-max))
6795 (if (and (/= start end)
6799 (- (buffer-size) (forward-line (buffer-size)))))))
6802 @node what-line, narrow Exercise, save-restriction, Narrowing & Widening
6803 @comment node-name, next, previous, up
6804 @section @code{what-line}
6806 @cindex Widening, example of
6808 The @code{what-line} command tells you the number of the line in which
6809 the cursor is located. The function illustrates the use of the
6810 @code{save-restriction} and @code{save-excursion} commands. Here is the
6811 original text of the function:
6816 "Print the current line number (in the buffer) of point."
6823 (1+ (count-lines 1 (point)))))))
6827 (In recent versions of GNU Emacs, the @code{what-line} function has
6828 been expanded to tell you your line number in a narrowed buffer as
6829 well as your line number in a widened buffer. The recent version is
6830 more complex than the version shown here. If you feel adventurous,
6831 you might want to look at it after figuring out how this version
6832 works. You will probably need to use @kbd{C-h f}
6833 (@code{describe-function}). The newer version uses a conditional to
6834 determine whether the buffer has been narrowed.
6836 (Also, it uses @code{line-number-at-pos}, which among other simple
6837 expressions, such as @code{(goto-char (point-min))}, moves point to
6838 the beginning of the current line with @code{(forward-line 0)} rather
6839 than @code{beginning-of-line}.)
6841 The @code{what-line} function as shown here has a documentation line
6842 and is interactive, as you would expect. The next two lines use the
6843 functions @code{save-restriction} and @code{widen}.
6845 The @code{save-restriction} special form notes whatever narrowing is in
6846 effect, if any, in the current buffer and restores that narrowing after
6847 the code in the body of the @code{save-restriction} has been evaluated.
6849 The @code{save-restriction} special form is followed by @code{widen}.
6850 This function undoes any narrowing the current buffer may have had
6851 when @code{what-line} was called. (The narrowing that was there is
6852 the narrowing that @code{save-restriction} remembers.) This widening
6853 makes it possible for the line counting commands to count from the
6854 beginning of the buffer. Otherwise, they would have been limited to
6855 counting within the accessible region. Any original narrowing is
6856 restored just before the completion of the function by the
6857 @code{save-restriction} special form.
6859 The call to @code{widen} is followed by @code{save-excursion}, which
6860 saves the location of the cursor (i.e., of point) and of the mark, and
6861 restores them after the code in the body of the @code{save-excursion}
6862 uses the @code{beginning-of-line} function to move point.
6864 (Note that the @code{(widen)} expression comes between the
6865 @code{save-restriction} and @code{save-excursion} special forms. When
6866 you write the two @code{save- @dots{}} expressions in sequence, write
6867 @code{save-excursion} outermost.)
6870 The last two lines of the @code{what-line} function are functions to
6871 count the number of lines in the buffer and then print the number in the
6877 (1+ (count-lines 1 (point)))))))
6881 The @code{message} function prints a one-line message at the bottom of
6882 the Emacs screen. The first argument is inside of quotation marks and
6883 is printed as a string of characters. However, it may contain a
6884 @samp{%d} expression to print a following argument. @samp{%d} prints
6885 the argument as a decimal, so the message will say something such as
6889 The number that is printed in place of the @samp{%d} is computed by the
6890 last line of the function:
6893 (1+ (count-lines 1 (point)))
6899 (defun count-lines (start end)
6900 "Return number of lines between START and END.
6901 This is usually the number of newlines between them,
6902 but can be one more if START is not equal to END
6903 and the greater of them is not at the start of a line."
6906 (narrow-to-region start end)
6907 (goto-char (point-min))
6908 (if (eq selective-display t)
6911 (while (re-search-forward "[\n\C-m]" nil t 40)
6912 (setq done (+ 40 done)))
6913 (while (re-search-forward "[\n\C-m]" nil t 1)
6914 (setq done (+ 1 done)))
6915 (goto-char (point-max))
6916 (if (and (/= start end)
6920 (- (buffer-size) (forward-line (buffer-size)))))))
6924 What this does is count the lines from the first position of the
6925 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6926 one to that number. (The @code{1+} function adds one to its
6927 argument.) We add one to it because line 2 has only one line before
6928 it, and @code{count-lines} counts only the lines @emph{before} the
6931 After @code{count-lines} has done its job, and the message has been
6932 printed in the echo area, the @code{save-excursion} restores point and
6933 mark to their original positions; and @code{save-restriction} restores
6934 the original narrowing, if any.
6936 @node narrow Exercise, , what-line, Narrowing & Widening
6937 @section Exercise with Narrowing
6939 Write a function that will display the first 60 characters of the
6940 current buffer, even if you have narrowed the buffer to its latter
6941 half so that the first line is inaccessible. Restore point, mark, and
6942 narrowing. For this exercise, you need to use a whole potpourri of
6943 functions, including @code{save-restriction}, @code{widen},
6944 @code{goto-char}, @code{point-min}, @code{message}, and
6945 @code{buffer-substring}.
6947 @cindex Properties, mention of @code{buffer-substring-no-properties}
6948 (@code{buffer-substring} is a previously unmentioned function you will
6949 have to investigate yourself; or perhaps you will have to use
6950 @code{buffer-substring-no-properties} or
6951 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6952 properties are a feature otherwise not discussed here. @xref{Text
6953 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6956 Additionally, do you really need @code{goto-char} or @code{point-min}?
6957 Or can you write the function without them?
6959 @node car cdr & cons, Cutting & Storing Text, Narrowing & Widening, Top
6960 @comment node-name, next, previous, up
6961 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6962 @findex car, @r{introduced}
6963 @findex cdr, @r{introduced}
6965 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6966 functions. The @code{cons} function is used to construct lists, and
6967 the @code{car} and @code{cdr} functions are used to take them apart.
6969 In the walk through of the @code{copy-region-as-kill} function, we
6970 will see @code{cons} as well as two variants on @code{cdr},
6971 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6974 * Strange Names:: An historical aside: why the strange names?
6975 * car & cdr:: Functions for extracting part of a list.
6976 * cons:: Constructing a list.
6977 * nthcdr:: Calling @code{cdr} repeatedly.
6979 * setcar:: Changing the first element of a list.
6980 * setcdr:: Changing the rest of a list.
6984 @node Strange Names, car & cdr, car cdr & cons, car cdr & cons
6986 @unnumberedsec Strange Names
6989 The name of the @code{cons} function is not unreasonable: it is an
6990 abbreviation of the word `construct'. The origins of the names for
6991 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6992 is an acronym from the phrase `Contents of the Address part of the
6993 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6994 the phrase `Contents of the Decrement part of the Register'. These
6995 phrases refer to specific pieces of hardware on the very early
6996 computer on which the original Lisp was developed. Besides being
6997 obsolete, the phrases have been completely irrelevant for more than 25
6998 years to anyone thinking about Lisp. Nonetheless, although a few
6999 brave scholars have begun to use more reasonable names for these
7000 functions, the old terms are still in use. In particular, since the
7001 terms are used in the Emacs Lisp source code, we will use them in this
7004 @node car & cdr, cons, Strange Names, car cdr & cons
7005 @comment node-name, next, previous, up
7006 @section @code{car} and @code{cdr}
7008 The @sc{car} of a list is, quite simply, the first item in the list.
7009 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
7013 If you are reading this in Info in GNU Emacs, you can see this by
7014 evaluating the following:
7017 (car '(rose violet daisy buttercup))
7021 After evaluating the expression, @code{rose} will appear in the echo
7024 Clearly, a more reasonable name for the @code{car} function would be
7025 @code{first} and this is often suggested.
7027 @code{car} does not remove the first item from the list; it only reports
7028 what it is. After @code{car} has been applied to a list, the list is
7029 still the same as it was. In the jargon, @code{car} is
7030 `non-destructive'. This feature turns out to be important.
7032 The @sc{cdr} of a list is the rest of the list, that is, the
7033 @code{cdr} function returns the part of the list that follows the
7034 first item. Thus, while the @sc{car} of the list @code{'(rose violet
7035 daisy buttercup)} is @code{rose}, the rest of the list, the value
7036 returned by the @code{cdr} function, is @code{(violet daisy
7040 You can see this by evaluating the following in the usual way:
7043 (cdr '(rose violet daisy buttercup))
7047 When you evaluate this, @code{(violet daisy buttercup)} will appear in
7050 Like @code{car}, @code{cdr} does not remove any elements from the
7051 list---it just returns a report of what the second and subsequent
7054 Incidentally, in the example, the list of flowers is quoted. If it were
7055 not, the Lisp interpreter would try to evaluate the list by calling
7056 @code{rose} as a function. In this example, we do not want to do that.
7058 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
7060 (There is a lesson here: when you name new functions, consider very
7061 carefully what you are doing, since you may be stuck with the names
7062 for far longer than you expect. The reason this document perpetuates
7063 these names is that the Emacs Lisp source code uses them, and if I did
7064 not use them, you would have a hard time reading the code; but do,
7065 please, try to avoid using these terms yourself. The people who come
7066 after you will be grateful to you.)
7068 When @code{car} and @code{cdr} are applied to a list made up of symbols,
7069 such as the list @code{(pine fir oak maple)}, the element of the list
7070 returned by the function @code{car} is the symbol @code{pine} without
7071 any parentheses around it. @code{pine} is the first element in the
7072 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7073 oak maple)}, as you can see by evaluating the following expressions in
7078 (car '(pine fir oak maple))
7080 (cdr '(pine fir oak maple))
7084 On the other hand, in a list of lists, the first element is itself a
7085 list. @code{car} returns this first element as a list. For example,
7086 the following list contains three sub-lists, a list of carnivores, a
7087 list of herbivores and a list of sea mammals:
7091 (car '((lion tiger cheetah)
7092 (gazelle antelope zebra)
7093 (whale dolphin seal)))
7098 In this example, the first element or @sc{car} of the list is the list of
7099 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7100 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7104 (cdr '((lion tiger cheetah)
7105 (gazelle antelope zebra)
7106 (whale dolphin seal)))
7110 It is worth saying again that @code{car} and @code{cdr} are
7111 non-destructive---that is, they do not modify or change lists to which
7112 they are applied. This is very important for how they are used.
7114 Also, in the first chapter, in the discussion about atoms, I said that
7115 in Lisp, ``certain kinds of atom, such as an array, can be separated
7116 into parts; but the mechanism for doing this is different from the
7117 mechanism for splitting a list. As far as Lisp is concerned, the
7118 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7119 @code{car} and @code{cdr} functions are used for splitting lists and
7120 are considered fundamental to Lisp. Since they cannot split or gain
7121 access to the parts of an array, an array is considered an atom.
7122 Conversely, the other fundamental function, @code{cons}, can put
7123 together or construct a list, but not an array. (Arrays are handled
7124 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7125 Emacs Lisp Reference Manual}.)
7127 @node cons, nthcdr, car & cdr, car cdr & cons
7128 @comment node-name, next, previous, up
7129 @section @code{cons}
7130 @findex cons, @r{introduced}
7132 The @code{cons} function constructs lists; it is the inverse of
7133 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7134 a four element list from the three element list, @code{(fir oak maple)}:
7137 (cons 'pine '(fir oak maple))
7142 After evaluating this list, you will see
7145 (pine fir oak maple)
7149 appear in the echo area. @code{cons} causes the creation of a new
7150 list in which the element is followed by the elements of the original
7153 We often say that `@code{cons} puts a new element at the beginning of
7154 a list; it attaches or pushes elements onto the list', but this
7155 phrasing can be misleading, since @code{cons} does not change an
7156 existing list, but creates a new one.
7158 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7162 * length:: How to find the length of a list.
7165 @node Build a list, length, cons, cons
7167 @unnumberedsubsec Build a list
7170 @code{cons} must have a list to attach to.@footnote{Actually, you can
7171 @code{cons} an element to an atom to produce a dotted pair. Dotted
7172 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7173 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7174 cannot start from absolutely nothing. If you are building a list, you
7175 need to provide at least an empty list at the beginning. Here is a
7176 series of @code{cons} expressions that build up a list of flowers. If
7177 you are reading this in Info in GNU Emacs, you can evaluate each of
7178 the expressions in the usual way; the value is printed in this text
7179 after @samp{@result{}}, which you may read as `evaluates to'.
7183 (cons 'buttercup ())
7184 @result{} (buttercup)
7188 (cons 'daisy '(buttercup))
7189 @result{} (daisy buttercup)
7193 (cons 'violet '(daisy buttercup))
7194 @result{} (violet daisy buttercup)
7198 (cons 'rose '(violet daisy buttercup))
7199 @result{} (rose violet daisy buttercup)
7204 In the first example, the empty list is shown as @code{()} and a list
7205 made up of @code{buttercup} followed by the empty list is constructed.
7206 As you can see, the empty list is not shown in the list that was
7207 constructed. All that you see is @code{(buttercup)}. The empty list is
7208 not counted as an element of a list because there is nothing in an empty
7209 list. Generally speaking, an empty list is invisible.
7211 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7212 two element list by putting @code{daisy} in front of @code{buttercup};
7213 and the third example constructs a three element list by putting
7214 @code{violet} in front of @code{daisy} and @code{buttercup}.
7216 @node length, , Build a list, cons
7217 @comment node-name, next, previous, up
7218 @subsection Find the Length of a List: @code{length}
7221 You can find out how many elements there are in a list by using the Lisp
7222 function @code{length}, as in the following examples:
7226 (length '(buttercup))
7231 (length '(daisy buttercup))
7236 (length (cons 'violet '(daisy buttercup)))
7242 In the third example, the @code{cons} function is used to construct a
7243 three element list which is then passed to the @code{length} function as
7247 We can also use @code{length} to count the number of elements in an
7258 As you would expect, the number of elements in an empty list is zero.
7260 An interesting experiment is to find out what happens if you try to find
7261 the length of no list at all; that is, if you try to call @code{length}
7262 without giving it an argument, not even an empty list:
7270 What you see, if you evaluate this, is the error message
7273 Lisp error: (wrong-number-of-arguments length 0)
7277 This means that the function receives the wrong number of
7278 arguments, zero, when it expects some other number of arguments. In
7279 this case, one argument is expected, the argument being a list whose
7280 length the function is measuring. (Note that @emph{one} list is
7281 @emph{one} argument, even if the list has many elements inside it.)
7283 The part of the error message that says @samp{length} is the name of
7287 @code{length} is still a subroutine, but you need C-h f to discover that.
7289 In an earlier version:
7290 This is written with a special notation, @samp{#<subr},
7291 that indicates that the function @code{length} is one of the primitive
7292 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7293 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7294 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7298 @node nthcdr, nth, cons, car cdr & cons
7299 @comment node-name, next, previous, up
7300 @section @code{nthcdr}
7303 The @code{nthcdr} function is associated with the @code{cdr} function.
7304 What it does is take the @sc{cdr} of a list repeatedly.
7306 If you take the @sc{cdr} of the list @code{(pine fir
7307 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7308 repeat this on what was returned, you will be returned the list
7309 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7310 list will just give you the original @sc{cdr} since the function does
7311 not change the list. You need to evaluate the @sc{cdr} of the
7312 @sc{cdr} and so on.) If you continue this, eventually you will be
7313 returned an empty list, which in this case, instead of being shown as
7314 @code{()} is shown as @code{nil}.
7317 For review, here is a series of repeated @sc{cdr}s, the text following
7318 the @samp{@result{}} shows what is returned.
7322 (cdr '(pine fir oak maple))
7323 @result{}(fir oak maple)
7327 (cdr '(fir oak maple))
7328 @result{} (oak maple)
7353 You can also do several @sc{cdr}s without printing the values in
7358 (cdr (cdr '(pine fir oak maple)))
7359 @result{} (oak maple)
7364 In this example, the Lisp interpreter evaluates the innermost list first.
7365 The innermost list is quoted, so it just passes the list as it is to the
7366 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7367 second and subsequent elements of the list to the outermost @code{cdr},
7368 which produces a list composed of the third and subsequent elements of
7369 the original list. In this example, the @code{cdr} function is repeated
7370 and returns a list that consists of the original list without its
7373 The @code{nthcdr} function does the same as repeating the call to
7374 @code{cdr}. In the following example, the argument 2 is passed to the
7375 function @code{nthcdr}, along with the list, and the value returned is
7376 the list without its first two items, which is exactly the same
7377 as repeating @code{cdr} twice on the list:
7381 (nthcdr 2 '(pine fir oak maple))
7382 @result{} (oak maple)
7387 Using the original four element list, we can see what happens when
7388 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7393 ;; @r{Leave the list as it was.}
7394 (nthcdr 0 '(pine fir oak maple))
7395 @result{} (pine fir oak maple)
7399 ;; @r{Return a copy without the first element.}
7400 (nthcdr 1 '(pine fir oak maple))
7401 @result{} (fir oak maple)
7405 ;; @r{Return a copy of the list without three elements.}
7406 (nthcdr 3 '(pine fir oak maple))
7411 ;; @r{Return a copy lacking all four elements.}
7412 (nthcdr 4 '(pine fir oak maple))
7417 ;; @r{Return a copy lacking all elements.}
7418 (nthcdr 5 '(pine fir oak maple))
7423 @node nth, setcar, nthcdr, car cdr & cons
7424 @comment node-name, next, previous, up
7428 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7429 The @code{nth} function takes the @sc{car} of the result returned by
7430 @code{nthcdr}. It returns the Nth element of the list.
7433 Thus, if it were not defined in C for speed, the definition of
7434 @code{nth} would be:
7439 "Returns the Nth element of LIST.
7440 N counts from zero. If LIST is not that long, nil is returned."
7441 (car (nthcdr n list)))
7446 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7447 but its definition was redone in C in the 1980s.)
7449 The @code{nth} function returns a single element of a list.
7450 This can be very convenient.
7452 Note that the elements are numbered from zero, not one. That is to
7453 say, the first element of a list, its @sc{car} is the zeroth element.
7454 This is called `zero-based' counting and often bothers people who
7455 are accustomed to the first element in a list being number one, which
7463 (nth 0 '("one" "two" "three"))
7466 (nth 1 '("one" "two" "three"))
7471 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7472 @code{cdr}, does not change the original list---the function is
7473 non-destructive. This is in sharp contrast to the @code{setcar} and
7474 @code{setcdr} functions.
7476 @node setcar, setcdr, nth, car cdr & cons
7477 @comment node-name, next, previous, up
7478 @section @code{setcar}
7481 As you might guess from their names, the @code{setcar} and @code{setcdr}
7482 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7483 They actually change the original list, unlike @code{car} and @code{cdr}
7484 which leave the original list as it was. One way to find out how this
7485 works is to experiment. We will start with the @code{setcar} function.
7488 First, we can make a list and then set the value of a variable to the
7489 list, using the @code{setq} function. Here is a list of animals:
7492 (setq animals '(antelope giraffe lion tiger))
7496 If you are reading this in Info inside of GNU Emacs, you can evaluate
7497 this expression in the usual fashion, by positioning the cursor after
7498 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7499 as I write this. This is one of the advantages of having the
7500 interpreter built into the computing environment. Incidentally, when
7501 there is nothing on the line after the final parentheses, such as a
7502 comment, point can be on the next line. Thus, if your cursor is in
7503 the first column of the next line, you do not need to move it.
7504 Indeed, Emacs permits any amount of white space after the final
7508 When we evaluate the variable @code{animals}, we see that it is bound to
7509 the list @code{(antelope giraffe lion tiger)}:
7514 @result{} (antelope giraffe lion tiger)
7519 Put another way, the variable @code{animals} points to the list
7520 @code{(antelope giraffe lion tiger)}.
7522 Next, evaluate the function @code{setcar} while passing it two
7523 arguments, the variable @code{animals} and the quoted symbol
7524 @code{hippopotamus}; this is done by writing the three element list
7525 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7529 (setcar animals 'hippopotamus)
7534 After evaluating this expression, evaluate the variable @code{animals}
7535 again. You will see that the list of animals has changed:
7540 @result{} (hippopotamus giraffe lion tiger)
7545 The first element on the list, @code{antelope} is replaced by
7546 @code{hippopotamus}.
7548 So we can see that @code{setcar} did not add a new element to the list
7549 as @code{cons} would have; it replaced @code{antelope} with
7550 @code{hippopotamus}; it @emph{changed} the list.
7552 @node setcdr, cons Exercise, setcar, car cdr & cons
7553 @comment node-name, next, previous, up
7554 @section @code{setcdr}
7557 The @code{setcdr} function is similar to the @code{setcar} function,
7558 except that the function replaces the second and subsequent elements of
7559 a list rather than the first element.
7561 (To see how to change the last element of a list, look ahead to
7562 @ref{kill-new function, , The @code{kill-new} function}, which uses
7563 the @code{nthcdr} and @code{setcdr} functions.)
7566 To see how this works, set the value of the variable to a list of
7567 domesticated animals by evaluating the following expression:
7570 (setq domesticated-animals '(horse cow sheep goat))
7575 If you now evaluate the list, you will be returned the list
7576 @code{(horse cow sheep goat)}:
7580 domesticated-animals
7581 @result{} (horse cow sheep goat)
7586 Next, evaluate @code{setcdr} with two arguments, the name of the
7587 variable which has a list as its value, and the list to which the
7588 @sc{cdr} of the first list will be set;
7591 (setcdr domesticated-animals '(cat dog))
7595 If you evaluate this expression, the list @code{(cat dog)} will appear
7596 in the echo area. This is the value returned by the function. The
7597 result we are interested in is the ``side effect'', which we can see by
7598 evaluating the variable @code{domesticated-animals}:
7602 domesticated-animals
7603 @result{} (horse cat dog)
7608 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7609 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7610 @code{(cow sheep goat)} to @code{(cat dog)}.
7612 @node cons Exercise, , setcdr, car cdr & cons
7615 Construct a list of four birds by evaluating several expressions with
7616 @code{cons}. Find out what happens when you @code{cons} a list onto
7617 itself. Replace the first element of the list of four birds with a
7618 fish. Replace the rest of that list with a list of other fish.
7620 @node Cutting & Storing Text, List Implementation, car cdr & cons, Top
7621 @comment node-name, next, previous, up
7622 @chapter Cutting and Storing Text
7623 @cindex Cutting and storing text
7624 @cindex Storing and cutting text
7625 @cindex Killing text
7626 @cindex Clipping text
7627 @cindex Erasing text
7628 @cindex Deleting text
7630 Whenever you cut or clip text out of a buffer with a `kill' command in
7631 GNU Emacs, it is stored in a list and you can bring it back with a
7634 (The use of the word `kill' in Emacs for processes which specifically
7635 @emph{do not} destroy the values of the entities is an unfortunate
7636 historical accident. A much more appropriate word would be `clip' since
7637 that is what the kill commands do; they clip text out of a buffer and
7638 put it into storage from which it can be brought back. I have often
7639 been tempted to replace globally all occurrences of `kill' in the Emacs
7640 sources with `clip' and all occurrences of `killed' with `clipped'.)
7643 * Storing Text:: Text is stored in a list.
7644 * zap-to-char:: Cutting out text up to a character.
7645 * kill-region:: Cutting text out of a region.
7646 * copy-region-as-kill:: A definition for copying text.
7647 * Digression into C:: Minor note on C programming language macros.
7648 * defvar:: How to give a variable an initial value.
7649 * cons & search-fwd Review::
7650 * search Exercises::
7653 @node Storing Text, zap-to-char, Cutting & Storing Text, Cutting & Storing Text
7655 @unnumberedsec Storing Text in a List
7658 When text is cut out of a buffer, it is stored on a list. Successive
7659 pieces of text are stored on the list successively, so the list might
7663 ("a piece of text" "previous piece")
7668 The function @code{cons} can be used to create a new list from a piece
7669 of text (an `atom', to use the jargon) and an existing list, like
7674 (cons "another piece"
7675 '("a piece of text" "previous piece"))
7681 If you evaluate this expression, a list of three elements will appear in
7685 ("another piece" "a piece of text" "previous piece")
7688 With the @code{car} and @code{nthcdr} functions, you can retrieve
7689 whichever piece of text you want. For example, in the following code,
7690 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7691 and the @code{car} returns the first element of that remainder---the
7692 second element of the original list:
7696 (car (nthcdr 1 '("another piece"
7699 @result{} "a piece of text"
7703 The actual functions in Emacs are more complex than this, of course.
7704 The code for cutting and retrieving text has to be written so that
7705 Emacs can figure out which element in the list you want---the first,
7706 second, third, or whatever. In addition, when you get to the end of
7707 the list, Emacs should give you the first element of the list, rather
7708 than nothing at all.
7710 The list that holds the pieces of text is called the @dfn{kill ring}.
7711 This chapter leads up to a description of the kill ring and how it is
7712 used by first tracing how the @code{zap-to-char} function works. This
7713 function uses (or `calls') a function that invokes a function that
7714 manipulates the kill ring. Thus, before reaching the mountains, we
7715 climb the foothills.
7717 A subsequent chapter describes how text that is cut from the buffer is
7718 retrieved. @xref{Yanking, , Yanking Text Back}.
7720 @node zap-to-char, kill-region, Storing Text, Cutting & Storing Text
7721 @comment node-name, next, previous, up
7722 @section @code{zap-to-char}
7725 The @code{zap-to-char} function changed little between GNU Emacs
7726 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7727 calls another function, @code{kill-region}, which enjoyed a major
7730 The @code{kill-region} function in Emacs 19 is complex, but does not
7731 use code that is important at this time. We will skip it.
7733 The @code{kill-region} function in Emacs 22 is easier to read than the
7734 same function in Emacs 19 and introduces a very important concept,
7735 that of error handling. We will walk through the function.
7737 But first, let us look at the interactive @code{zap-to-char} function.
7740 * Complete zap-to-char:: The complete implementation.
7741 * zap-to-char interactive:: A three part interactive expression.
7742 * zap-to-char body:: A short overview.
7743 * search-forward:: How to search for a string.
7744 * progn:: The @code{progn} special form.
7745 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7748 @node Complete zap-to-char, zap-to-char interactive, zap-to-char, zap-to-char
7750 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7753 The @code{zap-to-char} function removes the text in the region between
7754 the location of the cursor (i.e., of point) up to and including the
7755 next occurrence of a specified character. The text that
7756 @code{zap-to-char} removes is put in the kill ring; and it can be
7757 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7758 the command is given an argument, it removes text through that number
7759 of occurrences. Thus, if the cursor were at the beginning of this
7760 sentence and the character were @samp{s}, @samp{Thus} would be
7761 removed. If the argument were two, @samp{Thus, if the curs} would be
7762 removed, up to and including the @samp{s} in @samp{cursor}.
7764 If the specified character is not found, @code{zap-to-char} will say
7765 ``Search failed'', tell you the character you typed, and not remove
7768 In order to determine how much text to remove, @code{zap-to-char} uses
7769 a search function. Searches are used extensively in code that
7770 manipulates text, and we will focus attention on them as well as on the
7774 @c GNU Emacs version 19
7775 (defun zap-to-char (arg char) ; version 19 implementation
7776 "Kill up to and including ARG'th occurrence of CHAR.
7777 Goes backward if ARG is negative; error if CHAR not found."
7778 (interactive "*p\ncZap to char: ")
7779 (kill-region (point)
7782 (char-to-string char) nil nil arg)
7787 Here is the complete text of the version 22 implementation of the function:
7792 (defun zap-to-char (arg char)
7793 "Kill up to and including ARG'th occurrence of CHAR.
7794 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7795 Goes backward if ARG is negative; error if CHAR not found."
7796 (interactive "p\ncZap to char: ")
7797 (if (char-table-p translation-table-for-input)
7798 (setq char (or (aref translation-table-for-input char) char)))
7799 (kill-region (point) (progn
7800 (search-forward (char-to-string char)
7806 The documentation is thorough. You do need to know the jargon meaning
7809 @node zap-to-char interactive, zap-to-char body, Complete zap-to-char, zap-to-char
7810 @comment node-name, next, previous, up
7811 @subsection The @code{interactive} Expression
7814 The interactive expression in the @code{zap-to-char} command looks like
7818 (interactive "p\ncZap to char: ")
7821 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7822 two different things. First, and most simply, is the @samp{p}.
7823 This part is separated from the next part by a newline, @samp{\n}.
7824 The @samp{p} means that the first argument to the function will be
7825 passed the value of a `processed prefix'. The prefix argument is
7826 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7827 the function is called interactively without a prefix, 1 is passed to
7830 The second part of @code{"p\ncZap to char:@: "} is
7831 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7832 indicates that @code{interactive} expects a prompt and that the
7833 argument will be a character. The prompt follows the @samp{c} and is
7834 the string @samp{Zap to char:@: } (with a space after the colon to
7837 What all this does is prepare the arguments to @code{zap-to-char} so they
7838 are of the right type, and give the user a prompt.
7840 In a read-only buffer, the @code{zap-to-char} function copies the text
7841 to the kill ring, but does not remove it. The echo area displays a
7842 message saying that the buffer is read-only. Also, the terminal may
7843 beep or blink at you.
7845 @node zap-to-char body, search-forward, zap-to-char interactive, zap-to-char
7846 @comment node-name, next, previous, up
7847 @subsection The Body of @code{zap-to-char}
7849 The body of the @code{zap-to-char} function contains the code that
7850 kills (that is, removes) the text in the region from the current
7851 position of the cursor up to and including the specified character.
7853 The first part of the code looks like this:
7856 (if (char-table-p translation-table-for-input)
7857 (setq char (or (aref translation-table-for-input char) char)))
7858 (kill-region (point) (progn
7859 (search-forward (char-to-string char) nil nil arg)
7864 @code{char-table-p} is an hitherto unseen function. It determines
7865 whether its argument is a character table. When it is, it sets the
7866 character passed to @code{zap-to-char} to one of them, if that
7867 character exists, or to the character itself. (This becomes important
7868 for certain characters in non-European languages. The @code{aref}
7869 function extracts an element from an array. It is an array-specific
7870 function that is not described in this document. @xref{Arrays, ,
7871 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7874 @code{(point)} is the current position of the cursor.
7876 The next part of the code is an expression using @code{progn}. The body
7877 of the @code{progn} consists of calls to @code{search-forward} and
7880 It is easier to understand how @code{progn} works after learning about
7881 @code{search-forward}, so we will look at @code{search-forward} and
7882 then at @code{progn}.
7884 @node search-forward, progn, zap-to-char body, zap-to-char
7885 @comment node-name, next, previous, up
7886 @subsection The @code{search-forward} Function
7887 @findex search-forward
7889 The @code{search-forward} function is used to locate the
7890 zapped-for-character in @code{zap-to-char}. If the search is
7891 successful, @code{search-forward} leaves point immediately after the
7892 last character in the target string. (In @code{zap-to-char}, the
7893 target string is just one character long. @code{zap-to-char} uses the
7894 function @code{char-to-string} to ensure that the computer treats that
7895 character as a string.) If the search is backwards,
7896 @code{search-forward} leaves point just before the first character in
7897 the target. Also, @code{search-forward} returns @code{t} for true.
7898 (Moving point is therefore a `side effect'.)
7901 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7904 (search-forward (char-to-string char) nil nil arg)
7907 The @code{search-forward} function takes four arguments:
7911 The first argument is the target, what is searched for. This must be a
7912 string, such as @samp{"z"}.
7914 As it happens, the argument passed to @code{zap-to-char} is a single
7915 character. Because of the way computers are built, the Lisp
7916 interpreter may treat a single character as being different from a
7917 string of characters. Inside the computer, a single character has a
7918 different electronic format than a string of one character. (A single
7919 character can often be recorded in the computer using exactly one
7920 byte; but a string may be longer, and the computer needs to be ready
7921 for this.) Since the @code{search-forward} function searches for a
7922 string, the character that the @code{zap-to-char} function receives as
7923 its argument must be converted inside the computer from one format to
7924 the other; otherwise the @code{search-forward} function will fail.
7925 The @code{char-to-string} function is used to make this conversion.
7928 The second argument bounds the search; it is specified as a position in
7929 the buffer. In this case, the search can go to the end of the buffer,
7930 so no bound is set and the second argument is @code{nil}.
7933 The third argument tells the function what it should do if the search
7934 fails---it can signal an error (and print a message) or it can return
7935 @code{nil}. A @code{nil} as the third argument causes the function to
7936 signal an error when the search fails.
7939 The fourth argument to @code{search-forward} is the repeat count---how
7940 many occurrences of the string to look for. This argument is optional
7941 and if the function is called without a repeat count, this argument is
7942 passed the value 1. If this argument is negative, the search goes
7947 In template form, a @code{search-forward} expression looks like this:
7951 (search-forward "@var{target-string}"
7952 @var{limit-of-search}
7953 @var{what-to-do-if-search-fails}
7958 We will look at @code{progn} next.
7960 @node progn, Summing up zap-to-char, search-forward, zap-to-char
7961 @comment node-name, next, previous, up
7962 @subsection The @code{progn} Special Form
7965 @code{progn} is a special form that causes each of its arguments to be
7966 evaluated in sequence and then returns the value of the last one. The
7967 preceding expressions are evaluated only for the side effects they
7968 perform. The values produced by them are discarded.
7971 The template for a @code{progn} expression is very simple:
7980 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7981 put point in exactly the right position; and return the location of
7982 point so that @code{kill-region} will know how far to kill to.
7984 The first argument to the @code{progn} is @code{search-forward}. When
7985 @code{search-forward} finds the string, the function leaves point
7986 immediately after the last character in the target string. (In this
7987 case the target string is just one character long.) If the search is
7988 backwards, @code{search-forward} leaves point just before the first
7989 character in the target. The movement of point is a side effect.
7991 The second and last argument to @code{progn} is the expression
7992 @code{(point)}. This expression returns the value of point, which in
7993 this case will be the location to which it has been moved by
7994 @code{search-forward}. (In the source, a line that tells the function
7995 to go to the previous character, if it is going forward, was commented
7996 out in 1999; I don't remember whether that feature or mis-feature was
7997 ever a part of the distributed source.) The value of @code{point} is
7998 returned by the @code{progn} expression and is passed to
7999 @code{kill-region} as @code{kill-region}'s second argument.
8001 @node Summing up zap-to-char, , progn, zap-to-char
8002 @comment node-name, next, previous, up
8003 @subsection Summing up @code{zap-to-char}
8005 Now that we have seen how @code{search-forward} and @code{progn} work,
8006 we can see how the @code{zap-to-char} function works as a whole.
8008 The first argument to @code{kill-region} is the position of the cursor
8009 when the @code{zap-to-char} command is given---the value of point at
8010 that time. Within the @code{progn}, the search function then moves
8011 point to just after the zapped-to-character and @code{point} returns the
8012 value of this location. The @code{kill-region} function puts together
8013 these two values of point, the first one as the beginning of the region
8014 and the second one as the end of the region, and removes the region.
8016 The @code{progn} special form is necessary because the
8017 @code{kill-region} command takes two arguments; and it would fail if
8018 @code{search-forward} and @code{point} expressions were written in
8019 sequence as two additional arguments. The @code{progn} expression is
8020 a single argument to @code{kill-region} and returns the one value that
8021 @code{kill-region} needs for its second argument.
8023 @node kill-region, copy-region-as-kill, zap-to-char, Cutting & Storing Text
8024 @comment node-name, next, previous, up
8025 @section @code{kill-region}
8028 The @code{zap-to-char} function uses the @code{kill-region} function.
8029 This function clips text from a region and copies that text to
8030 the kill ring, from which it may be retrieved.
8035 (defun kill-region (beg end &optional yank-handler)
8036 "Kill (\"cut\") text between point and mark.
8037 This deletes the text from the buffer and saves it in the kill ring.
8038 The command \\[yank] can retrieve it from there.
8039 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
8041 If you want to append the killed region to the last killed text,
8042 use \\[append-next-kill] before \\[kill-region].
8044 If the buffer is read-only, Emacs will beep and refrain from deleting
8045 the text, but put the text in the kill ring anyway. This means that
8046 you can use the killing commands to copy text from a read-only buffer.
8048 This is the primitive for programs to kill text (as opposed to deleting it).
8049 Supply two arguments, character positions indicating the stretch of text
8051 Any command that calls this function is a \"kill command\".
8052 If the previous command was also a kill command,
8053 the text killed this time appends to the text killed last time
8054 to make one entry in the kill ring.
8056 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
8057 specifies the yank-handler text property to be set on the killed
8058 text. See `insert-for-yank'."
8059 ;; Pass point first, then mark, because the order matters
8060 ;; when calling kill-append.
8061 (interactive (list (point) (mark)))
8062 (unless (and beg end)
8063 (error "The mark is not set now, so there is no region"))
8065 (let ((string (filter-buffer-substring beg end t)))
8066 (when string ;STRING is nil if BEG = END
8067 ;; Add that string to the kill ring, one way or another.
8068 (if (eq last-command 'kill-region)
8069 (kill-append string (< end beg) yank-handler)
8070 (kill-new string nil yank-handler)))
8071 (when (or string (eq last-command 'kill-region))
8072 (setq this-command 'kill-region))
8074 ((buffer-read-only text-read-only)
8075 ;; The code above failed because the buffer, or some of the characters
8076 ;; in the region, are read-only.
8077 ;; We should beep, in case the user just isn't aware of this.
8078 ;; However, there's no harm in putting
8079 ;; the region's text in the kill ring, anyway.
8080 (copy-region-as-kill beg end)
8081 ;; Set this-command now, so it will be set even if we get an error.
8082 (setq this-command 'kill-region)
8083 ;; This should barf, if appropriate, and give us the correct error.
8084 (if kill-read-only-ok
8085 (progn (message "Read only text copied to kill ring") nil)
8086 ;; Signal an error if the buffer is read-only.
8087 (barf-if-buffer-read-only)
8088 ;; If the buffer isn't read-only, the text is.
8089 (signal 'text-read-only (list (current-buffer)))))))
8092 The Emacs 22 version of that function uses @code{condition-case} and
8093 @code{copy-region-as-kill}, both of which we will explain.
8094 @code{condition-case} is an important special form.
8096 In essence, the @code{kill-region} function calls
8097 @code{condition-case}, which takes three arguments. In this function,
8098 the first argument does nothing. The second argument contains the
8099 code that does the work when all goes well. The third argument
8100 contains the code that is called in the event of an error.
8103 * Complete kill-region:: The function definition.
8104 * condition-case:: Dealing with a problem.
8108 @node Complete kill-region, condition-case, kill-region, kill-region
8110 @unnumberedsubsec The Complete @code{kill-region} Definition
8114 We will go through the @code{condition-case} code in a moment. First,
8115 let us look at the definition of @code{kill-region}, with comments
8121 (defun kill-region (beg end)
8122 "Kill (\"cut\") text between point and mark.
8123 This deletes the text from the buffer and saves it in the kill ring.
8124 The command \\[yank] can retrieve it from there. @dots{} "
8128 ;; @bullet{} Since order matters, pass point first.
8129 (interactive (list (point) (mark)))
8130 ;; @bullet{} And tell us if we cannot cut the text.
8131 ;; `unless' is an `if' without a then-part.
8132 (unless (and beg end)
8133 (error "The mark is not set now, so there is no region"))
8137 ;; @bullet{} `condition-case' takes three arguments.
8138 ;; If the first argument is nil, as it is here,
8139 ;; information about the error signal is not
8140 ;; stored for use by another function.
8145 ;; @bullet{} The second argument to `condition-case' tells the
8146 ;; Lisp interpreter what to do when all goes well.
8150 ;; It starts with a `let' function that extracts the string
8151 ;; and tests whether it exists. If so (that is what the
8152 ;; `when' checks), it calls an `if' function that determines
8153 ;; whether the previous command was another call to
8154 ;; `kill-region'; if it was, then the new text is appended to
8155 ;; the previous text; if not, then a different function,
8156 ;; `kill-new', is called.
8160 ;; The `kill-append' function concatenates the new string and
8161 ;; the old. The `kill-new' function inserts text into a new
8162 ;; item in the kill ring.
8166 ;; `when' is an `if' without an else-part. The second `when'
8167 ;; again checks whether the current string exists; in
8168 ;; addition, it checks whether the previous command was
8169 ;; another call to `kill-region'. If one or the other
8170 ;; condition is true, then it sets the current command to
8171 ;; be `kill-region'.
8174 (let ((string (filter-buffer-substring beg end t)))
8175 (when string ;STRING is nil if BEG = END
8176 ;; Add that string to the kill ring, one way or another.
8177 (if (eq last-command 'kill-region)
8180 ;; @minus{} `yank-handler' is an optional argument to
8181 ;; `kill-region' that tells the `kill-append' and
8182 ;; `kill-new' functions how deal with properties
8183 ;; added to the text, such as `bold' or `italics'.
8184 (kill-append string (< end beg) yank-handler)
8185 (kill-new string nil yank-handler)))
8186 (when (or string (eq last-command 'kill-region))
8187 (setq this-command 'kill-region))
8192 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8193 ;; what to do with an error.
8196 ;; The third argument has a conditions part and a body part.
8197 ;; If the conditions are met (in this case,
8198 ;; if text or buffer are read-only)
8199 ;; then the body is executed.
8202 ;; The first part of the third argument is the following:
8203 ((buffer-read-only text-read-only) ;; the if-part
8204 ;; @dots{} the then-part
8205 (copy-region-as-kill beg end)
8208 ;; Next, also as part of the then-part, set this-command, so
8209 ;; it will be set in an error
8210 (setq this-command 'kill-region)
8211 ;; Finally, in the then-part, send a message if you may copy
8212 ;; the text to the kill ring without signaling an error, but
8213 ;; don't if you may not.
8216 (if kill-read-only-ok
8217 (progn (message "Read only text copied to kill ring") nil)
8218 (barf-if-buffer-read-only)
8219 ;; If the buffer isn't read-only, the text is.
8220 (signal 'text-read-only (list (current-buffer)))))
8228 (defun kill-region (beg end)
8229 "Kill between point and mark.
8230 The text is deleted but saved in the kill ring."
8235 ;; 1. `condition-case' takes three arguments.
8236 ;; If the first argument is nil, as it is here,
8237 ;; information about the error signal is not
8238 ;; stored for use by another function.
8243 ;; 2. The second argument to `condition-case'
8244 ;; tells the Lisp interpreter what to do when all goes well.
8248 ;; The `delete-and-extract-region' function usually does the
8249 ;; work. If the beginning and ending of the region are both
8250 ;; the same, then the variable `string' will be empty, or nil
8251 (let ((string (delete-and-extract-region beg end)))
8255 ;; `when' is an `if' clause that cannot take an `else-part'.
8256 ;; Emacs normally sets the value of `last-command' to the
8257 ;; previous command.
8260 ;; `kill-append' concatenates the new string and the old.
8261 ;; `kill-new' inserts text into a new item in the kill ring.
8263 (if (eq last-command 'kill-region)
8264 ;; if true, prepend string
8265 (kill-append string (< end beg))
8267 (setq this-command 'kill-region))
8271 ;; 3. The third argument to `condition-case' tells the interpreter
8272 ;; what to do with an error.
8275 ;; The third argument has a conditions part and a body part.
8276 ;; If the conditions are met (in this case,
8277 ;; if text or buffer are read-only)
8278 ;; then the body is executed.
8281 ((buffer-read-only text-read-only) ;; this is the if-part
8283 (copy-region-as-kill beg end)
8286 (if kill-read-only-ok ;; usually this variable is nil
8287 (message "Read only text copied to kill ring")
8288 ;; or else, signal an error if the buffer is read-only;
8289 (barf-if-buffer-read-only)
8290 ;; and, in any case, signal that the text is read-only.
8291 (signal 'text-read-only (list (current-buffer)))))))
8296 @node condition-case, Lisp macro, Complete kill-region, kill-region
8297 @comment node-name, next, previous, up
8298 @subsection @code{condition-case}
8299 @findex condition-case
8301 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8302 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8303 expression, it provides you with help; in the jargon, this is called
8304 ``signaling an error''. Usually, the computer stops the program and
8305 shows you a message.
8307 However, some programs undertake complicated actions. They should not
8308 simply stop on an error. In the @code{kill-region} function, the most
8309 likely error is that you will try to kill text that is read-only and
8310 cannot be removed. So the @code{kill-region} function contains code
8311 to handle this circumstance. This code, which makes up the body of
8312 the @code{kill-region} function, is inside of a @code{condition-case}
8316 The template for @code{condition-case} looks like this:
8323 @var{error-handler}@dots{})
8327 The second argument, @var{bodyform}, is straightforward. The
8328 @code{condition-case} special form causes the Lisp interpreter to
8329 evaluate the code in @var{bodyform}. If no error occurs, the special
8330 form returns the code's value and produces the side-effects, if any.
8332 In short, the @var{bodyform} part of a @code{condition-case}
8333 expression determines what should happen when everything works
8336 However, if an error occurs, among its other actions, the function
8337 generating the error signal will define one or more error condition
8340 An error handler is the third argument to @code{condition case}.
8341 An error handler has two parts, a @var{condition-name} and a
8342 @var{body}. If the @var{condition-name} part of an error handler
8343 matches a condition name generated by an error, then the @var{body}
8344 part of the error handler is run.
8346 As you will expect, the @var{condition-name} part of an error handler
8347 may be either a single condition name or a list of condition names.
8349 Also, a complete @code{condition-case} expression may contain more
8350 than one error handler. When an error occurs, the first applicable
8353 Lastly, the first argument to the @code{condition-case} expression,
8354 the @var{var} argument, is sometimes bound to a variable that
8355 contains information about the error. However, if that argument is
8356 nil, as is the case in @code{kill-region}, that information is
8360 In brief, in the @code{kill-region} function, the code
8361 @code{condition-case} works like this:
8365 @var{If no errors}, @var{run only this code}
8366 @var{but}, @var{if errors}, @var{run this other code}.
8373 copy-region-as-kill is short, 12 lines, and uses
8374 filter-buffer-substring, which is longer, 39 lines
8375 and has delete-and-extract-region in it.
8376 delete-and-extract-region is written in C.
8378 see Initializing a Variable with @code{defvar}
8380 Initializing a Variable with @code{defvar} includes line 8350
8383 @node Lisp macro, , condition-case, kill-region
8384 @comment node-name, next, previous, up
8385 @subsection Lisp macro
8389 The part of the @code{condition-case} expression that is evaluated in
8390 the expectation that all goes well has a @code{when}. The code uses
8391 @code{when} to determine whether the @code{string} variable points to
8394 A @code{when} expression is simply a programmers' convenience. It is
8395 an @code{if} without the possibility of an else clause. In your mind,
8396 you can replace @code{when} with @code{if} and understand what goes
8397 on. That is what the Lisp interpreter does.
8399 Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro}
8400 enables you to define new control constructs and other language
8401 features. It tells the interpreter how to compute another Lisp
8402 expression which will in turn compute the value. In this case, the
8403 `other expression' is an @code{if} expression.
8405 The @code{kill-region} function definition also has an @code{unless}
8406 macro; it is the converse of @code{when}. The @code{unless} macro is
8407 an @code{if} without a then clause
8409 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8410 Emacs Lisp Reference Manual}. The C programming language also
8411 provides macros. These are different, but also useful.
8414 We will briefly look at C macros in
8415 @ref{Digression into C}.
8419 Regarding the @code{when} macro, in the @code{condition-case}
8420 expression, when the string has content, then another conditional
8421 expression is executed. This is an @code{if} with both a then-part
8426 (if (eq last-command 'kill-region)
8427 (kill-append string (< end beg) yank-handler)
8428 (kill-new string nil yank-handler))
8432 The then-part is evaluated if the previous command was another call to
8433 @code{kill-region}; if not, the else-part is evaluated.
8435 @code{yank-handler} is an optional argument to @code{kill-region} that
8436 tells the @code{kill-append} and @code{kill-new} functions how deal
8437 with properties added to the text, such as `bold' or `italics'.
8439 @code{last-command} is a variable that comes with Emacs that we have
8440 not seen before. Normally, whenever a function is executed, Emacs
8441 sets the value of @code{last-command} to the previous command.
8444 In this segment of the definition, the @code{if} expression checks
8445 whether the previous command was @code{kill-region}. If it was,
8448 (kill-append string (< end beg) yank-handler)
8452 concatenates a copy of the newly clipped text to the just previously
8453 clipped text in the kill ring.
8455 @node copy-region-as-kill, Digression into C, kill-region, Cutting & Storing Text
8456 @comment node-name, next, previous, up
8457 @section @code{copy-region-as-kill}
8458 @findex copy-region-as-kill
8461 The @code{copy-region-as-kill} function copies a region of text from a
8462 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8463 in the @code{kill-ring}.
8465 If you call @code{copy-region-as-kill} immediately after a
8466 @code{kill-region} command, Emacs appends the newly copied text to the
8467 previously copied text. This means that if you yank back the text, you
8468 get it all, from both this and the previous operation. On the other
8469 hand, if some other command precedes the @code{copy-region-as-kill},
8470 the function copies the text into a separate entry in the kill ring.
8473 * Complete copy-region-as-kill:: The complete function definition.
8474 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8477 @node Complete copy-region-as-kill, copy-region-as-kill body, copy-region-as-kill, copy-region-as-kill
8479 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8483 Here is the complete text of the version 22 @code{copy-region-as-kill}
8488 (defun copy-region-as-kill (beg end)
8489 "Save the region as if killed, but don't kill it.
8490 In Transient Mark mode, deactivate the mark.
8491 If `interprogram-cut-function' is non-nil, also save the text for a window
8492 system cut and paste."
8496 (if (eq last-command 'kill-region)
8497 (kill-append (filter-buffer-substring beg end) (< end beg))
8498 (kill-new (filter-buffer-substring beg end)))
8501 (if transient-mark-mode
8502 (setq deactivate-mark t))
8508 As usual, this function can be divided into its component parts:
8512 (defun copy-region-as-kill (@var{argument-list})
8513 "@var{documentation}@dots{}"
8519 The arguments are @code{beg} and @code{end} and the function is
8520 interactive with @code{"r"}, so the two arguments must refer to the
8521 beginning and end of the region. If you have been reading though this
8522 document from the beginning, understanding these parts of a function is
8523 almost becoming routine.
8525 The documentation is somewhat confusing unless you remember that the
8526 word `kill' has a meaning different from usual. The `Transient Mark'
8527 and @code{interprogram-cut-function} comments explain certain
8530 After you once set a mark, a buffer always contains a region. If you
8531 wish, you can use Transient Mark mode to highlight the region
8532 temporarily. (No one wants to highlight the region all the time, so
8533 Transient Mark mode highlights it only at appropriate times. Many
8534 people turn off Transient Mark mode, so the region is never
8537 Also, a windowing system allows you to copy, cut, and paste among
8538 different programs. In the X windowing system, for example, the
8539 @code{interprogram-cut-function} function is @code{x-select-text},
8540 which works with the windowing system's equivalent of the Emacs kill
8543 The body of the @code{copy-region-as-kill} function starts with an
8544 @code{if} clause. What this clause does is distinguish between two
8545 different situations: whether or not this command is executed
8546 immediately after a previous @code{kill-region} command. In the first
8547 case, the new region is appended to the previously copied text.
8548 Otherwise, it is inserted into the beginning of the kill ring as a
8549 separate piece of text from the previous piece.
8551 The last two lines of the function prevent the region from lighting up
8552 if Transient Mark mode is turned on.
8554 The body of @code{copy-region-as-kill} merits discussion in detail.
8556 @node copy-region-as-kill body, , Complete copy-region-as-kill, copy-region-as-kill
8557 @comment node-name, next, previous, up
8558 @subsection The Body of @code{copy-region-as-kill}
8560 The @code{copy-region-as-kill} function works in much the same way as
8561 the @code{kill-region} function. Both are written so that two or more
8562 kills in a row combine their text into a single entry. If you yank
8563 back the text from the kill ring, you get it all in one piece.
8564 Moreover, kills that kill forward from the current position of the
8565 cursor are added to the end of the previously copied text and commands
8566 that copy text backwards add it to the beginning of the previously
8567 copied text. This way, the words in the text stay in the proper
8570 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8571 use of the @code{last-command} variable that keeps track of the
8572 previous Emacs command.
8575 * last-command & this-command::
8576 * kill-append function::
8577 * kill-new function::
8580 @node last-command & this-command, kill-append function, copy-region-as-kill body, copy-region-as-kill body
8582 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8585 Normally, whenever a function is executed, Emacs sets the value of
8586 @code{this-command} to the function being executed (which in this case
8587 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8588 the value of @code{last-command} to the previous value of
8589 @code{this-command}.
8591 In the first part of the body of the @code{copy-region-as-kill}
8592 function, an @code{if} expression determines whether the value of
8593 @code{last-command} is @code{kill-region}. If so, the then-part of
8594 the @code{if} expression is evaluated; it uses the @code{kill-append}
8595 function to concatenate the text copied at this call to the function
8596 with the text already in the first element (the @sc{car}) of the kill
8597 ring. On the other hand, if the value of @code{last-command} is not
8598 @code{kill-region}, then the @code{copy-region-as-kill} function
8599 attaches a new element to the kill ring using the @code{kill-new}
8603 The @code{if} expression reads as follows; it uses @code{eq}:
8607 (if (eq last-command 'kill-region)
8609 (kill-append (filter-buffer-substring beg end) (< end beg))
8611 (kill-new (filter-buffer-substring beg end)))
8615 @findex filter-buffer-substring
8616 (The @code{filter-buffer-substring} function returns a filtered
8617 substring of the buffer, if any. Optionally---the arguments are not
8618 here, so neither is done---the function may delete the initial text or
8619 return the text without its properties; this function is a replacement
8620 for the older @code{buffer-substring} function, which came before text
8621 properties were implemented.)
8623 @findex eq @r{(example of use)}
8625 The @code{eq} function tests whether its first argument is the same Lisp
8626 object as its second argument. The @code{eq} function is similar to the
8627 @code{equal} function in that it is used to test for equality, but
8628 differs in that it determines whether two representations are actually
8629 the same object inside the computer, but with different names.
8630 @code{equal} determines whether the structure and contents of two
8631 expressions are the same.
8633 If the previous command was @code{kill-region}, then the Emacs Lisp
8634 interpreter calls the @code{kill-append} function
8636 @node kill-append function, kill-new function, last-command & this-command, copy-region-as-kill body
8637 @unnumberedsubsubsec The @code{kill-append} function
8641 The @code{kill-append} function looks like this:
8646 (defun kill-append (string before-p &optional yank-handler)
8647 "Append STRING to the end of the latest kill in the kill ring.
8648 If BEFORE-P is non-nil, prepend STRING to the kill.
8650 (let* ((cur (car kill-ring)))
8651 (kill-new (if before-p (concat string cur) (concat cur string))
8652 (or (= (length cur) 0)
8654 (get-text-property 0 'yank-handler cur)))
8661 (defun kill-append (string before-p)
8662 "Append STRING to the end of the latest kill in the kill ring.
8663 If BEFORE-P is non-nil, prepend STRING to the kill.
8664 If `interprogram-cut-function' is set, pass the resulting kill to
8666 (kill-new (if before-p
8667 (concat string (car kill-ring))
8668 (concat (car kill-ring) string))
8673 The @code{kill-append} function is fairly straightforward. It uses
8674 the @code{kill-new} function, which we will discuss in more detail in
8677 (Also, the function provides an optional argument called
8678 @code{yank-handler}; when invoked, this argument tells the function
8679 how to deal with properties added to the text, such as `bold' or
8682 @c !!! bug in GNU Emacs 22 version of kill-append ?
8683 It has a @code{let*} function to set the value of the first element of
8684 the kill ring to @code{cur}. (I do not know why the function does not
8685 use @code{let} instead; only one value is set in the expression.
8686 Perhaps this is a bug that produces no problems?)
8688 Consider the conditional that is one of the two arguments to
8689 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8690 the @sc{car} of the kill ring. Whether it prepends or appends the
8691 text depends on the results of an @code{if} expression:
8695 (if before-p ; @r{if-part}
8696 (concat string cur) ; @r{then-part}
8697 (concat cur string)) ; @r{else-part}
8702 If the region being killed is before the region that was killed in the
8703 last command, then it should be prepended before the material that was
8704 saved in the previous kill; and conversely, if the killed text follows
8705 what was just killed, it should be appended after the previous text.
8706 The @code{if} expression depends on the predicate @code{before-p} to
8707 decide whether the newly saved text should be put before or after the
8708 previously saved text.
8710 The symbol @code{before-p} is the name of one of the arguments to
8711 @code{kill-append}. When the @code{kill-append} function is
8712 evaluated, it is bound to the value returned by evaluating the actual
8713 argument. In this case, this is the expression @code{(< end beg)}.
8714 This expression does not directly determine whether the killed text in
8715 this command is located before or after the kill text of the last
8716 command; what it does is determine whether the value of the variable
8717 @code{end} is less than the value of the variable @code{beg}. If it
8718 is, it means that the user is most likely heading towards the
8719 beginning of the buffer. Also, the result of evaluating the predicate
8720 expression, @code{(< end beg)}, will be true and the text will be
8721 prepended before the previous text. On the other hand, if the value of
8722 the variable @code{end} is greater than the value of the variable
8723 @code{beg}, the text will be appended after the previous text.
8726 When the newly saved text will be prepended, then the string with the new
8727 text will be concatenated before the old text:
8735 But if the text will be appended, it will be concatenated
8739 (concat cur string))
8742 To understand how this works, we first need to review the
8743 @code{concat} function. The @code{concat} function links together or
8744 unites two strings of text. The result is a string. For example:
8748 (concat "abc" "def")
8754 (car '("first element" "second element")))
8755 @result{} "new first element"
8758 '("first element" "second element")) " modified")
8759 @result{} "first element modified"
8763 We can now make sense of @code{kill-append}: it modifies the contents
8764 of the kill ring. The kill ring is a list, each element of which is
8765 saved text. The @code{kill-append} function uses the @code{kill-new}
8766 function which in turn uses the @code{setcar} function.
8768 @node kill-new function, , kill-append function, copy-region-as-kill body
8769 @unnumberedsubsubsec The @code{kill-new} function
8772 @c in GNU Emacs 22, additional documentation to kill-new:
8774 Optional third arguments YANK-HANDLER controls how the STRING is later
8775 inserted into a buffer; see `insert-for-yank' for details.
8776 When a yank handler is specified, STRING must be non-empty (the yank
8777 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8779 When the yank handler has a non-nil PARAM element, the original STRING
8780 argument is not used by `insert-for-yank'. However, since Lisp code
8781 may access and use elements from the kill ring directly, the STRING
8782 argument should still be a \"useful\" string for such uses."
8785 The @code{kill-new} function looks like this:
8789 (defun kill-new (string &optional replace yank-handler)
8790 "Make STRING the latest kill in the kill ring.
8791 Set `kill-ring-yank-pointer' to point to it.
8793 If `interprogram-cut-function' is non-nil, apply it to STRING.
8794 Optional second argument REPLACE non-nil means that STRING will replace
8795 the front of the kill ring, rather than being added to the list.
8799 (if (> (length string) 0)
8801 (put-text-property 0 (length string)
8802 'yank-handler yank-handler string))
8804 (signal 'args-out-of-range
8805 (list string "yank-handler specified for empty string"))))
8808 (if (fboundp 'menu-bar-update-yank-menu)
8809 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8812 (if (and replace kill-ring)
8813 (setcar kill-ring string)
8814 (push string kill-ring)
8815 (if (> (length kill-ring) kill-ring-max)
8816 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8819 (setq kill-ring-yank-pointer kill-ring)
8820 (if interprogram-cut-function
8821 (funcall interprogram-cut-function string (not replace))))
8826 (defun kill-new (string &optional replace)
8827 "Make STRING the latest kill in the kill ring.
8828 Set the kill-ring-yank pointer to point to it.
8829 If `interprogram-cut-function' is non-nil, apply it to STRING.
8830 Optional second argument REPLACE non-nil means that STRING will replace
8831 the front of the kill ring, rather than being added to the list."
8832 (and (fboundp 'menu-bar-update-yank-menu)
8833 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8834 (if (and replace kill-ring)
8835 (setcar kill-ring string)
8836 (setq kill-ring (cons string kill-ring))
8837 (if (> (length kill-ring) kill-ring-max)
8838 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8839 (setq kill-ring-yank-pointer kill-ring)
8840 (if interprogram-cut-function
8841 (funcall interprogram-cut-function string (not replace))))
8844 (Notice that the function is not interactive.)
8846 As usual, we can look at this function in parts.
8848 The function definition has an optional @code{yank-handler} argument,
8849 which when invoked tells the function how to deal with properties
8850 added to the text, such as `bold' or `italics'. We will skip that.
8853 The first line of the documentation makes sense:
8856 Make STRING the latest kill in the kill ring.
8860 Let's skip over the rest of the documentation for the moment.
8863 Also, let's skip over the initial @code{if} expression and those lines
8864 of code involving @code{menu-bar-update-yank-menu}. We will explain
8868 The critical lines are these:
8872 (if (and replace kill-ring)
8874 (setcar kill-ring string)
8878 (push string kill-ring)
8881 (setq kill-ring (cons string kill-ring))
8882 (if (> (length kill-ring) kill-ring-max)
8883 ;; @r{avoid overly long kill ring}
8884 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8887 (setq kill-ring-yank-pointer kill-ring)
8888 (if interprogram-cut-function
8889 (funcall interprogram-cut-function string (not replace))))
8893 The conditional test is @w{@code{(and replace kill-ring)}}.
8894 This will be true when two conditions are met: the kill ring has
8895 something in it, and the @code{replace} variable is true.
8898 When the @code{kill-append} function sets @code{replace} to be true
8899 and when the kill ring has at least one item in it, the @code{setcar}
8900 expression is executed:
8903 (setcar kill-ring string)
8906 The @code{setcar} function actually changes the first element of the
8907 @code{kill-ring} list to the value of @code{string}. It replaces the
8911 On the other hand, if the kill ring is empty, or replace is false, the
8912 else-part of the condition is executed:
8915 (push string kill-ring)
8920 @code{push} puts its first argument onto the second. It is similar to
8924 (setq kill-ring (cons string kill-ring))
8932 (add-to-list kill-ring string)
8936 When it is false, the expression first constructs a new version of the
8937 kill ring by prepending @code{string} to the existing kill ring as a
8938 new element (that is what the @code{push} does). Then it executes a
8939 second @code{if} clause. This second @code{if} clause keeps the kill
8940 ring from growing too long.
8942 Let's look at these two expressions in order.
8944 The @code{push} line of the else-part sets the new value of the kill
8945 ring to what results from adding the string being killed to the old
8948 We can see how this works with an example.
8954 (setq example-list '("here is a clause" "another clause"))
8959 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8960 @code{example-list} and see what it returns:
8965 @result{} ("here is a clause" "another clause")
8971 Now, we can add a new element on to this list by evaluating the
8972 following expression:
8973 @findex push, @r{example}
8976 (push "a third clause" example-list)
8981 When we evaluate @code{example-list}, we find its value is:
8986 @result{} ("a third clause" "here is a clause" "another clause")
8991 Thus, the third clause is added to the list by @code{push}.
8994 Now for the second part of the @code{if} clause. This expression
8995 keeps the kill ring from growing too long. It looks like this:
8999 (if (> (length kill-ring) kill-ring-max)
9000 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
9004 The code checks whether the length of the kill ring is greater than
9005 the maximum permitted length. This is the value of
9006 @code{kill-ring-max} (which is 60, by default). If the length of the
9007 kill ring is too long, then this code sets the last element of the
9008 kill ring to @code{nil}. It does this by using two functions,
9009 @code{nthcdr} and @code{setcdr}.
9011 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
9012 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
9013 @sc{car} of a list. In this case, however, @code{setcdr} will not be
9014 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
9015 function is used to cause it to set the @sc{cdr} of the next to last
9016 element of the kill ring---this means that since the @sc{cdr} of the
9017 next to last element is the last element of the kill ring, it will set
9018 the last element of the kill ring.
9020 @findex nthcdr, @r{example}
9021 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
9022 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
9023 @dots{} It does this @var{N} times and returns the results.
9024 (@xref{nthcdr, , @code{nthcdr}}.)
9026 @findex setcdr, @r{example}
9027 Thus, if we had a four element list that was supposed to be three
9028 elements long, we could set the @sc{cdr} of the next to last element
9029 to @code{nil}, and thereby shorten the list. (If you set the last
9030 element to some other value than @code{nil}, which you could do, then
9031 you would not have shortened the list. @xref{setcdr, ,
9034 You can see shortening by evaluating the following three expressions
9035 in turn. First set the value of @code{trees} to @code{(maple oak pine
9036 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
9037 and then find the value of @code{trees}:
9041 (setq trees '(maple oak pine birch))
9042 @result{} (maple oak pine birch)
9046 (setcdr (nthcdr 2 trees) nil)
9050 @result{} (maple oak pine)
9055 (The value returned by the @code{setcdr} expression is @code{nil} since
9056 that is what the @sc{cdr} is set to.)
9058 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
9059 @sc{cdr} a number of times that is one less than the maximum permitted
9060 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
9061 element (which will be the rest of the elements in the kill ring) to
9062 @code{nil}. This prevents the kill ring from growing too long.
9065 The next to last expression in the @code{kill-new} function is
9068 (setq kill-ring-yank-pointer kill-ring)
9071 The @code{kill-ring-yank-pointer} is a global variable that is set to be
9072 the @code{kill-ring}.
9074 Even though the @code{kill-ring-yank-pointer} is called a
9075 @samp{pointer}, it is a variable just like the kill ring. However, the
9076 name has been chosen to help humans understand how the variable is used.
9079 Now, to return to an early expression in the body of the function:
9083 (if (fboundp 'menu-bar-update-yank-menu)
9084 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9089 It starts with an @code{if} expression
9091 In this case, the expression tests first to see whether
9092 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9093 calls it. The @code{fboundp} function returns true if the symbol it
9094 is testing has a function definition that `is not void'. If the
9095 symbol's function definition were void, we would receive an error
9096 message, as we did when we created errors intentionally (@pxref{Making
9097 Errors, , Generate an Error Message}).
9100 The then-part contains an expression whose first element is the
9101 function @code{and}.
9104 The @code{and} special form evaluates each of its arguments until one
9105 of the arguments returns a value of @code{nil}, in which case the
9106 @code{and} expression returns @code{nil}; however, if none of the
9107 arguments returns a value of @code{nil}, the value resulting from
9108 evaluating the last argument is returned. (Since such a value is not
9109 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9110 @code{and} expression returns a true value only if all its arguments
9111 are true. (@xref{Second Buffer Related Review}.)
9113 The expression determines whether the second argument to
9114 @code{menu-bar-update-yank-menu} is true or not.
9116 ;; If we're supposed to be extending an existing string, and that
9117 ;; string really is at the front of the menu, then update it in place.
9120 @code{menu-bar-update-yank-menu} is one of the functions that make it
9121 possible to use the `Select and Paste' menu in the Edit item of a menu
9122 bar; using a mouse, you can look at the various pieces of text you
9123 have saved and select one piece to paste.
9125 The last expression in the @code{kill-new} function adds the newly
9126 copied string to whatever facility exists for copying and pasting
9127 among different programs running in a windowing system. In the X
9128 Windowing system, for example, the @code{x-select-text} function takes
9129 the string and stores it in memory operated by X. You can paste the
9130 string in another program, such as an Xterm.
9133 The expression looks like this:
9137 (if interprogram-cut-function
9138 (funcall interprogram-cut-function string (not replace))))
9142 If an @code{interprogram-cut-function} exists, then Emacs executes
9143 @code{funcall}, which in turn calls its first argument as a function
9144 and passes the remaining arguments to it. (Incidentally, as far as I
9145 can see, this @code{if} expression could be replaced by an @code{and}
9146 expression similar to the one in the first part of the function.)
9148 We are not going to discuss windowing systems and other programs
9149 further, but merely note that this is a mechanism that enables GNU
9150 Emacs to work easily and well with other programs.
9152 This code for placing text in the kill ring, either concatenated with
9153 an existing element or as a new element, leads us to the code for
9154 bringing back text that has been cut out of the buffer---the yank
9155 commands. However, before discussing the yank commands, it is better
9156 to learn how lists are implemented in a computer. This will make
9157 clear such mysteries as the use of the term `pointer'. But before
9158 that, we will digress into C.
9161 @c is this true in Emacs 22? Does not seems to be
9163 (If the @w{@code{(< end beg))}}
9164 expression is true, @code{kill-append} prepends the string to the just
9165 previously clipped text. For a detailed discussion, see
9166 @ref{kill-append function, , The @code{kill-append} function}.)
9168 If you then yank back the text, i.e., `paste' it, you get both
9169 pieces of text at once. That way, if you delete two words in a row,
9170 and then yank them back, you get both words, in their proper order,
9171 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9174 On the other hand, if the previous command is not @code{kill-region},
9175 then the @code{kill-new} function is called, which adds the text to
9176 the kill ring as the latest item, and sets the
9177 @code{kill-ring-yank-pointer} variable to point to it.
9181 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9182 @c use the delete-and-extract-region function
9184 2006 Oct 26, the Digression into C is now OK but should come after
9185 copy-region-as-kill and filter-buffer-substring
9189 copy-region-as-kill is short, 12 lines, and uses
9190 filter-buffer-substring, which is longer, 39 lines
9191 and has delete-and-extract-region in it.
9192 delete-and-extract-region is written in C.
9194 see Initializing a Variable with @code{defvar}
9197 @node Digression into C, defvar, copy-region-as-kill, Cutting & Storing Text
9198 @comment node-name, next, previous, up
9199 @section Digression into C
9200 @findex delete-and-extract-region
9201 @cindex C, a digression into
9202 @cindex Digression into C
9204 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9205 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9206 function, which in turn uses the @code{delete-and-extract-region}
9207 function. It removes the contents of a region and you cannot get them
9210 Unlike the other code discussed here, the
9211 @code{delete-and-extract-region} function is not written in Emacs
9212 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9213 system. Since it is very simple, I will digress briefly from Lisp and
9216 @c GNU Emacs 22 in /usr/local/src/emacs/src/editfns.c
9217 @c the DEFUN for buffer-substring-no-properties
9220 Like many of the other Emacs primitives,
9221 @code{delete-and-extract-region} is written as an instance of a C
9222 macro, a macro being a template for code. The complete macro looks
9227 DEFUN ("buffer-substring-no-properties", Fbuffer_substring_no_properties,
9228 Sbuffer_substring_no_properties, 2, 2, 0,
9229 doc: /* Return the characters of part of the buffer,
9230 without the text properties.
9231 The two arguments START and END are character positions;
9232 they can be in either order. */)
9234 Lisp_Object start, end;
9238 validate_region (&start, &end);
9242 return make_buffer_string (b, e, 0);
9247 Without going into the details of the macro writing process, let me
9248 point out that this macro starts with the word @code{DEFUN}. The word
9249 @code{DEFUN} was chosen since the code serves the same purpose as
9250 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9251 @file{emacs/src/lisp.h}.)
9253 The word @code{DEFUN} is followed by seven parts inside of
9258 The first part is the name given to the function in Lisp,
9259 @code{delete-and-extract-region}.
9262 The second part is the name of the function in C,
9263 @code{Fdelete_and_extract_region}. By convention, it starts with
9264 @samp{F}. Since C does not use hyphens in names, underscores are used
9268 The third part is the name for the C constant structure that records
9269 information on this function for internal use. It is the name of the
9270 function in C but begins with an @samp{S} instead of an @samp{F}.
9273 The fourth and fifth parts specify the minimum and maximum number of
9274 arguments the function can have. This function demands exactly 2
9278 The sixth part is nearly like the argument that follows the
9279 @code{interactive} declaration in a function written in Lisp: a letter
9280 followed, perhaps, by a prompt. The only difference from the Lisp is
9281 when the macro is called with no arguments. Then you write a @code{0}
9282 (which is a `null string'), as in this macro.
9284 If you were to specify arguments, you would place them between
9285 quotation marks. The C macro for @code{goto-char} includes
9286 @code{"NGoto char: "} in this position to indicate that the function
9287 expects a raw prefix, in this case, a numerical location in a buffer,
9288 and provides a prompt.
9291 The seventh part is a documentation string, just like the one for a
9292 function written in Emacs Lisp, except that every newline must be
9293 written explicitly as @samp{\n} followed by a backslash and carriage
9297 Thus, the first two lines of documentation for @code{goto-char} are
9302 "Set point to POSITION, a number or marker.\n\
9303 Beginning of buffer is position (point-min), end is (point-max)."
9309 In a C macro, the formal parameters come next, with a statement of
9310 what kind of object they are, followed by what might be called the `body'
9311 of the macro. For @code{delete-and-extract-region} the `body'
9312 consists of the following four lines:
9316 validate_region (&start, &end);
9317 if (XINT (start) == XINT (end))
9318 return build_string ("");
9319 return del_range_1 (XINT (start), XINT (end), 1, 1);
9323 The @code{validate_region} function checks whether the values
9324 passed as the beginning and end of the region are the proper type and
9325 are within range. If the beginning and end positions are the same,
9326 then return and empty string.
9328 The @code{del_range_1} function actually deletes the text. It is a
9329 complex function we will not look into. It updates the buffer and
9330 does other things. However, it is worth looking at the two arguments
9331 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9332 @w{@code{XINT (end)}}.
9334 As far as the C language is concerned, @code{start} and @code{end} are
9335 two integers that mark the beginning and end of the region to be
9336 deleted@footnote{More precisely, and requiring more expert knowledge
9337 to understand, the two integers are of type `Lisp_Object', which can
9338 also be a C union instead of an integer type.}.
9340 In early versions of Emacs, these two numbers were thirty-two bits
9341 long, but the code is slowly being generalized to handle other
9342 lengths. Three of the available bits are used to specify the type of
9343 information; the remaining bits are used as `content'.
9345 @samp{XINT} is a C macro that extracts the relevant number from the
9346 longer collection of bits; the three other bits are discarded.
9349 The command in @code{delete-and-extract-region} looks like this:
9352 del_range_1 (XINT (start), XINT (end), 1, 1);
9356 It deletes the region between the beginning position, @code{start},
9357 and the ending position, @code{end}.
9359 From the point of view of the person writing Lisp, Emacs is all very
9360 simple; but hidden underneath is a great deal of complexity to make it
9363 @node defvar, cons & search-fwd Review, Digression into C, Cutting & Storing Text
9364 @comment node-name, next, previous, up
9365 @section Initializing a Variable with @code{defvar}
9367 @cindex Initializing a variable
9368 @cindex Variable initialization
9373 copy-region-as-kill is short, 12 lines, and uses
9374 filter-buffer-substring, which is longer, 39 lines
9375 and has delete-and-extract-region in it.
9376 delete-and-extract-region is written in C.
9378 see Initializing a Variable with @code{defvar}
9382 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9383 functions within it, @code{kill-append} and @code{kill-new}, copy a
9384 region in a buffer and save it in a variable called the
9385 @code{kill-ring}. This section describes how the @code{kill-ring}
9386 variable is created and initialized using the @code{defvar} special
9389 (Again we note that the term @code{kill-ring} is a misnomer. The text
9390 that is clipped out of the buffer can be brought back; it is not a ring
9391 of corpses, but a ring of resurrectable text.)
9393 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9394 given an initial value by using the @code{defvar} special form. The
9395 name comes from ``define variable''.
9397 The @code{defvar} special form is similar to @code{setq} in that it sets
9398 the value of a variable. It is unlike @code{setq} in two ways: first,
9399 it only sets the value of the variable if the variable does not already
9400 have a value. If the variable already has a value, @code{defvar} does
9401 not override the existing value. Second, @code{defvar} has a
9402 documentation string.
9404 (Another special form, @code{defcustom}, is designed for variables
9405 that people customize. It has more features than @code{defvar}.
9406 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9409 * See variable current value::
9410 * defvar and asterisk::
9413 @node See variable current value, defvar and asterisk, defvar, defvar
9415 @unnumberedsubsec Seeing the Current Value of a Variable
9418 You can see the current value of a variable, any variable, by using
9419 the @code{describe-variable} function, which is usually invoked by
9420 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9421 (followed by @key{RET}) when prompted, you will see what is in your
9422 current kill ring---this may be quite a lot! Conversely, if you have
9423 been doing nothing this Emacs session except read this document, you
9424 may have nothing in it. Also, you will see the documentation for
9430 List of killed text sequences.
9431 Since the kill ring is supposed to interact nicely with cut-and-paste
9432 facilities offered by window systems, use of this variable should
9435 interact nicely with `interprogram-cut-function' and
9436 `interprogram-paste-function'. The functions `kill-new',
9437 `kill-append', and `current-kill' are supposed to implement this
9438 interaction; you may want to use them instead of manipulating the kill
9444 The kill ring is defined by a @code{defvar} in the following way:
9448 (defvar kill-ring nil
9449 "List of killed text sequences.
9455 In this variable definition, the variable is given an initial value of
9456 @code{nil}, which makes sense, since if you have saved nothing, you want
9457 nothing back if you give a @code{yank} command. The documentation
9458 string is written just like the documentation string of a @code{defun}.
9459 As with the documentation string of the @code{defun}, the first line of
9460 the documentation should be a complete sentence, since some commands,
9461 like @code{apropos}, print only the first line of documentation.
9462 Succeeding lines should not be indented; otherwise they look odd when
9463 you use @kbd{C-h v} (@code{describe-variable}).
9465 @node defvar and asterisk, , See variable current value, defvar
9466 @subsection @code{defvar} and an asterisk
9467 @findex defvar @r{for a user customizable variable}
9468 @findex defvar @r{with an asterisk}
9470 In the past, Emacs used the @code{defvar} special form both for
9471 internal variables that you would not expect a user to change and for
9472 variables that you do expect a user to change. Although you can still
9473 use @code{defvar} for user customizable variables, please use
9474 @code{defcustom} instead, since that special form provides a path into
9475 the Customization commands. (@xref{defcustom, , Specifying Variables
9476 using @code{defcustom}}.)
9478 When you specified a variable using the @code{defvar} special form,
9479 you could distinguish a variable that a user might want to change from
9480 others by typing an asterisk, @samp{*}, in the first column of its
9481 documentation string. For example:
9485 (defvar shell-command-default-error-buffer nil
9486 "*Buffer name for `shell-command' @dots{} error output.
9491 @findex set-variable
9493 You could (and still can) use the @code{set-variable} command to
9494 change the value of @code{shell-command-default-error-buffer}
9495 temporarily. However, options set using @code{set-variable} are set
9496 only for the duration of your editing session. The new values are not
9497 saved between sessions. Each time Emacs starts, it reads the original
9498 value, unless you change the value within your @file{.emacs} file,
9499 either by setting it manually or by using @code{customize}.
9500 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9502 For me, the major use of the @code{set-variable} command is to suggest
9503 variables that I might want to set in my @file{.emacs} file. There
9504 are now more than 700 such variables --- far too many to remember
9505 readily. Fortunately, you can press @key{TAB} after calling the
9506 @code{M-x set-variable} command to see the list of variables.
9507 (@xref{Examining, , Examining and Setting Variables, emacs,
9508 The GNU Emacs Manual}.)
9511 @node cons & search-fwd Review, search Exercises, defvar, Cutting & Storing Text
9512 @comment node-name, next, previous, up
9515 Here is a brief summary of some recently introduced functions.
9520 @code{car} returns the first element of a list; @code{cdr} returns the
9521 second and subsequent elements of a list.
9528 (car '(1 2 3 4 5 6 7))
9530 (cdr '(1 2 3 4 5 6 7))
9531 @result{} (2 3 4 5 6 7)
9536 @code{cons} constructs a list by prepending its first argument to its
9550 @code{funcall} evaluates its first argument as a function. It passes
9551 its remaining arguments to its first argument.
9554 Return the result of taking @sc{cdr} `n' times on a list.
9562 The `rest of the rest', as it were.
9569 (nthcdr 3 '(1 2 3 4 5 6 7))
9576 @code{setcar} changes the first element of a list; @code{setcdr}
9577 changes the second and subsequent elements of a list.
9584 (setq triple '(1 2 3))
9591 (setcdr triple '("foo" "bar"))
9594 @result{} (37 "foo" "bar")
9599 Evaluate each argument in sequence and then return the value of the
9612 @item save-restriction
9613 Record whatever narrowing is in effect in the current buffer, if any,
9614 and restore that narrowing after evaluating the arguments.
9616 @item search-forward
9617 Search for a string, and if the string is found, move point. With a
9618 regular expression, use the similar @code{re-search-forward}.
9619 (@xref{Regexp Search, , Regular Expression Searches}, for an
9620 explanation of regular expression patterns and searches.)
9624 @code{search-forward} and @code{re-search-forward} take four
9629 The string or regular expression to search for.
9632 Optionally, the limit of the search.
9635 Optionally, what to do if the search fails, return @code{nil} or an
9639 Optionally, how many times to repeat the search; if negative, the
9640 search goes backwards.
9644 @itemx delete-and-extract-region
9645 @itemx copy-region-as-kill
9647 @code{kill-region} cuts the text between point and mark from the
9648 buffer and stores that text in the kill ring, so you can get it back
9651 @code{copy-region-as-kill} copies the text between point and mark into
9652 the kill ring, from which you can get it by yanking. The function
9653 does not cut or remove the text from the buffer.
9656 @code{delete-and-extract-region} removes the text between point and
9657 mark from the buffer and throws it away. You cannot get it back.
9658 (This is not an interactive command.)
9661 @node search Exercises, , cons & search-fwd Review, Cutting & Storing Text
9662 @section Searching Exercises
9666 Write an interactive function that searches for a string. If the
9667 search finds the string, leave point after it and display a message
9668 that says ``Found!''. (Do not use @code{search-forward} for the name
9669 of this function; if you do, you will overwrite the existing version of
9670 @code{search-forward} that comes with Emacs. Use a name such as
9671 @code{test-search} instead.)
9674 Write a function that prints the third element of the kill ring in the
9675 echo area, if any; if the kill ring does not contain a third element,
9676 print an appropriate message.
9679 @node List Implementation, Yanking, Cutting & Storing Text, Top
9680 @comment node-name, next, previous, up
9681 @chapter How Lists are Implemented
9682 @cindex Lists in a computer
9684 In Lisp, atoms are recorded in a straightforward fashion; if the
9685 implementation is not straightforward in practice, it is, nonetheless,
9686 straightforward in theory. The atom @samp{rose}, for example, is
9687 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9688 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9689 is equally simple, but it takes a moment to get used to the idea. A
9690 list is kept using a series of pairs of pointers. In the series, the
9691 first pointer in each pair points to an atom or to another list, and the
9692 second pointer in each pair points to the next pair, or to the symbol
9693 @code{nil}, which marks the end of the list.
9695 A pointer itself is quite simply the electronic address of what is
9696 pointed to. Hence, a list is kept as a series of electronic addresses.
9699 * Lists diagrammed::
9700 * Symbols as Chest:: Exploring a powerful metaphor.
9704 @node Lists diagrammed, Symbols as Chest, List Implementation, List Implementation
9706 @unnumberedsec Lists diagrammed
9709 For example, the list @code{(rose violet buttercup)} has three elements,
9710 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9711 electronic address of @samp{rose} is recorded in a segment of computer
9712 memory along with the address that gives the electronic address of where
9713 the atom @samp{violet} is located; and that address (the one that tells
9714 where @samp{violet} is located) is kept along with an address that tells
9715 where the address for the atom @samp{buttercup} is located.
9718 This sounds more complicated than it is and is easier seen in a diagram:
9720 @c clear print-postscript-figures
9721 @c !!! cons-cell-diagram #1
9725 ___ ___ ___ ___ ___ ___
9726 |___|___|--> |___|___|--> |___|___|--> nil
9729 --> rose --> violet --> buttercup
9733 @ifset print-postscript-figures
9736 @center @image{cons-1}
9737 %%%% old method of including an image
9738 % \input /usr/local/lib/tex/inputs/psfig.tex
9739 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9744 @ifclear print-postscript-figures
9748 ___ ___ ___ ___ ___ ___
9749 |___|___|--> |___|___|--> |___|___|--> nil
9752 --> rose --> violet --> buttercup
9759 In the diagram, each box represents a word of computer memory that
9760 holds a Lisp object, usually in the form of a memory address. The boxes,
9761 i.e.@: the addresses, are in pairs. Each arrow points to what the address
9762 is the address of, either an atom or another pair of addresses. The
9763 first box is the electronic address of @samp{rose} and the arrow points
9764 to @samp{rose}; the second box is the address of the next pair of boxes,
9765 the first part of which is the address of @samp{violet} and the second
9766 part of which is the address of the next pair. The very last box
9767 points to the symbol @code{nil}, which marks the end of the list.
9770 When a variable is set to a list with a function such as @code{setq},
9771 it stores the address of the first box in the variable. Thus,
9772 evaluation of the expression
9775 (setq bouquet '(rose violet buttercup))
9780 creates a situation like this:
9782 @c cons-cell-diagram #2
9788 | ___ ___ ___ ___ ___ ___
9789 --> |___|___|--> |___|___|--> |___|___|--> nil
9792 --> rose --> violet --> buttercup
9796 @ifset print-postscript-figures
9799 @center @image{cons-2}
9800 %%%% old method of including an image
9801 % \input /usr/local/lib/tex/inputs/psfig.tex
9802 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9807 @ifclear print-postscript-figures
9813 | ___ ___ ___ ___ ___ ___
9814 --> |___|___|--> |___|___|--> |___|___|--> nil
9817 --> rose --> violet --> buttercup
9824 In this example, the symbol @code{bouquet} holds the address of the first
9828 This same list can be illustrated in a different sort of box notation
9831 @c cons-cell-diagram #2a
9837 | -------------- --------------- ----------------
9838 | | car | cdr | | car | cdr | | car | cdr |
9839 -->| rose | o------->| violet | o------->| butter- | nil |
9840 | | | | | | | cup | |
9841 -------------- --------------- ----------------
9845 @ifset print-postscript-figures
9848 @center @image{cons-2a}
9849 %%%% old method of including an image
9850 % \input /usr/local/lib/tex/inputs/psfig.tex
9851 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9856 @ifclear print-postscript-figures
9862 | -------------- --------------- ----------------
9863 | | car | cdr | | car | cdr | | car | cdr |
9864 -->| rose | o------->| violet | o------->| butter- | nil |
9865 | | | | | | | cup | |
9866 -------------- --------------- ----------------
9872 (Symbols consist of more than pairs of addresses, but the structure of
9873 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9874 consists of a group of address-boxes, one of which is the address of
9875 the printed word @samp{bouquet}, a second of which is the address of a
9876 function definition attached to the symbol, if any, a third of which
9877 is the address of the first pair of address-boxes for the list
9878 @code{(rose violet buttercup)}, and so on. Here we are showing that
9879 the symbol's third address-box points to the first pair of
9880 address-boxes for the list.)
9882 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9883 changed; the symbol simply has an address further down the list. (In
9884 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9885 evaluation of the following expression
9888 (setq flowers (cdr bouquet))
9895 @c cons-cell-diagram #3
9902 | ___ ___ | ___ ___ ___ ___
9903 --> | | | --> | | | | | |
9904 |___|___|----> |___|___|--> |___|___|--> nil
9907 --> rose --> violet --> buttercup
9912 @ifset print-postscript-figures
9915 @center @image{cons-3}
9916 %%%% old method of including an image
9917 % \input /usr/local/lib/tex/inputs/psfig.tex
9918 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9923 @ifclear print-postscript-figures
9930 | ___ ___ | ___ ___ ___ ___
9931 --> | | | --> | | | | | |
9932 |___|___|----> |___|___|--> |___|___|--> nil
9935 --> rose --> violet --> buttercup
9943 The value of @code{flowers} is @code{(violet buttercup)}, which is
9944 to say, the symbol @code{flowers} holds the address of the pair of
9945 address-boxes, the first of which holds the address of @code{violet},
9946 and the second of which holds the address of @code{buttercup}.
9948 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9949 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9950 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9951 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9952 information about cons cells and dotted pairs.
9955 The function @code{cons} adds a new pair of addresses to the front of
9956 a series of addresses like that shown above. For example, evaluating
9960 (setq bouquet (cons 'lily bouquet))
9967 @c cons-cell-diagram #4
9974 | ___ ___ ___ ___ | ___ ___ ___ ___
9975 --> | | | | | | --> | | | | | |
9976 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9979 --> lily --> rose --> violet --> buttercup
9984 @ifset print-postscript-figures
9987 @center @image{cons-4}
9988 %%%% old method of including an image
9989 % \input /usr/local/lib/tex/inputs/psfig.tex
9990 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9995 @ifclear print-postscript-figures
10002 | ___ ___ ___ ___ | ___ ___ ___ ___
10003 --> | | | | | | --> | | | | | |
10004 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
10007 --> lily --> rose --> violet --> buttercup
10016 However, this does not change the value of the symbol
10017 @code{flowers}, as you can see by evaluating the following,
10020 (eq (cdr (cdr bouquet)) flowers)
10024 which returns @code{t} for true.
10026 Until it is reset, @code{flowers} still has the value
10027 @code{(violet buttercup)}; that is, it has the address of the cons
10028 cell whose first address is of @code{violet}. Also, this does not
10029 alter any of the pre-existing cons cells; they are all still there.
10031 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
10032 of the next cons cell in the series; to get the @sc{car} of a list,
10033 you get the address of the first element of the list; to @code{cons} a
10034 new element on a list, you add a new cons cell to the front of the list.
10035 That is all there is to it! The underlying structure of Lisp is
10036 brilliantly simple!
10038 And what does the last address in a series of cons cells refer to? It
10039 is the address of the empty list, of @code{nil}.
10041 In summary, when a Lisp variable is set to a value, it is provided with
10042 the address of the list to which the variable refers.
10044 @node Symbols as Chest, List Exercise, Lists diagrammed, List Implementation
10045 @section Symbols as a Chest of Drawers
10046 @cindex Symbols as a Chest of Drawers
10047 @cindex Chest of Drawers, metaphor for a symbol
10048 @cindex Drawers, Chest of, metaphor for a symbol
10050 In an earlier section, I suggested that you might imagine a symbol as
10051 being a chest of drawers. The function definition is put in one
10052 drawer, the value in another, and so on. What is put in the drawer
10053 holding the value can be changed without affecting the contents of the
10054 drawer holding the function definition, and vice-verse.
10056 Actually, what is put in each drawer is the address of the value or
10057 function definition. It is as if you found an old chest in the attic,
10058 and in one of its drawers you found a map giving you directions to
10059 where the buried treasure lies.
10061 (In addition to its name, symbol definition, and variable value, a
10062 symbol has a `drawer' for a @dfn{property list} which can be used to
10063 record other information. Property lists are not discussed here; see
10064 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
10065 Reference Manual}.)
10068 Here is a fanciful representation:
10070 @c chest-of-drawers diagram
10075 Chest of Drawers Contents of Drawers
10079 ---------------------
10080 | directions to | [map to]
10081 | symbol name | bouquet
10083 +---------------------+
10085 | symbol definition | [none]
10087 +---------------------+
10088 | directions to | [map to]
10089 | variable value | (rose violet buttercup)
10091 +---------------------+
10093 | property list | [not described here]
10095 +---------------------+
10101 @ifset print-postscript-figures
10104 @center @image{drawers}
10105 %%%% old method of including an image
10106 % \input /usr/local/lib/tex/inputs/psfig.tex
10107 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10112 @ifclear print-postscript-figures
10117 Chest of Drawers Contents of Drawers
10121 ---------------------
10122 | directions to | [map to]
10123 | symbol name | bouquet
10125 +---------------------+
10127 | symbol definition | [none]
10129 +---------------------+
10130 | directions to | [map to]
10131 | variable value | (rose violet buttercup)
10133 +---------------------+
10135 | property list | [not described here]
10137 +---------------------+
10145 @node List Exercise, , Symbols as Chest, List Implementation
10148 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10149 more flowers on to this list and set this new list to
10150 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10151 What does the @code{more-flowers} list now contain?
10153 @node Yanking, Loops & Recursion, List Implementation, Top
10154 @comment node-name, next, previous, up
10155 @chapter Yanking Text Back
10157 @cindex Text retrieval
10158 @cindex Retrieving text
10159 @cindex Pasting text
10161 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10162 you can bring it back with a `yank' command. The text that is cut out of
10163 the buffer is put in the kill ring and the yank commands insert the
10164 appropriate contents of the kill ring back into a buffer (not necessarily
10165 the original buffer).
10167 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10168 the kill ring into the current buffer. If the @kbd{C-y} command is
10169 followed immediately by @kbd{M-y}, the first element is replaced by
10170 the second element. Successive @kbd{M-y} commands replace the second
10171 element with the third, fourth, or fifth element, and so on. When the
10172 last element in the kill ring is reached, it is replaced by the first
10173 element and the cycle is repeated. (Thus the kill ring is called a
10174 `ring' rather than just a `list'. However, the actual data structure
10175 that holds the text is a list.
10176 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10177 list is handled as a ring.)
10180 * Kill Ring Overview::
10181 * kill-ring-yank-pointer:: The kill ring is a list.
10182 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
10185 @node Kill Ring Overview, kill-ring-yank-pointer, Yanking, Yanking
10186 @comment node-name, next, previous, up
10187 @section Kill Ring Overview
10188 @cindex Kill ring overview
10190 The kill ring is a list of textual strings. This is what it looks like:
10193 ("some text" "a different piece of text" "yet more text")
10196 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10197 string of characters saying @samp{some text} would be inserted in this
10198 buffer where my cursor is located.
10200 The @code{yank} command is also used for duplicating text by copying it.
10201 The copied text is not cut from the buffer, but a copy of it is put on the
10202 kill ring and is inserted by yanking it back.
10204 Three functions are used for bringing text back from the kill ring:
10205 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10206 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10207 which is used by the two other functions.
10209 These functions refer to the kill ring through a variable called the
10210 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10211 @code{yank} and @code{yank-pop} functions is:
10214 (insert (car kill-ring-yank-pointer))
10218 (Well, no more. In GNU Emacs 22, the function has been replaced by
10219 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10220 repetitively for each @code{yank-handler} segment. In turn,
10221 @code{insert-for-yank-1} strips text properties from the inserted text
10222 according to @code{yank-excluded-properties}. Otherwise, it is just
10223 like @code{insert}. We will stick with plain @code{insert} since it
10224 is easier to understand.)
10226 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10227 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10229 @node kill-ring-yank-pointer, yank nthcdr Exercises, Kill Ring Overview, Yanking
10230 @comment node-name, next, previous, up
10231 @section The @code{kill-ring-yank-pointer} Variable
10233 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10234 a variable. It points to something by being bound to the value of what
10235 it points to, like any other Lisp variable.
10238 Thus, if the value of the kill ring is:
10241 ("some text" "a different piece of text" "yet more text")
10246 and the @code{kill-ring-yank-pointer} points to the second clause, the
10247 value of @code{kill-ring-yank-pointer} is:
10250 ("a different piece of text" "yet more text")
10253 As explained in the previous chapter (@pxref{List Implementation}), the
10254 computer does not keep two different copies of the text being pointed to
10255 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10256 words ``a different piece of text'' and ``yet more text'' are not
10257 duplicated. Instead, the two Lisp variables point to the same pieces of
10258 text. Here is a diagram:
10260 @c cons-cell-diagram #5
10264 kill-ring kill-ring-yank-pointer
10266 | ___ ___ | ___ ___ ___ ___
10267 ---> | | | --> | | | | | |
10268 |___|___|----> |___|___|--> |___|___|--> nil
10271 | | --> "yet more text"
10273 | --> "a different piece of text"
10280 @ifset print-postscript-figures
10283 @center @image{cons-5}
10284 %%%% old method of including an image
10285 % \input /usr/local/lib/tex/inputs/psfig.tex
10286 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10291 @ifclear print-postscript-figures
10295 kill-ring kill-ring-yank-pointer
10297 | ___ ___ | ___ ___ ___ ___
10298 ---> | | | --> | | | | | |
10299 |___|___|----> |___|___|--> |___|___|--> nil
10302 | | --> "yet more text"
10304 | --> "a different piece of text
10313 Both the variable @code{kill-ring} and the variable
10314 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10315 usually described as if it were actually what it is composed of. The
10316 @code{kill-ring} is spoken of as if it were the list rather than that it
10317 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10318 spoken of as pointing to a list.
10320 These two ways of talking about the same thing sound confusing at first but
10321 make sense on reflection. The kill ring is generally thought of as the
10322 complete structure of data that holds the information of what has recently
10323 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10324 on the other hand, serves to indicate---that is, to `point to'---that part
10325 of the kill ring of which the first element (the @sc{car}) will be
10329 In GNU Emacs 22, the @code{kill-new} function calls
10331 @code{(setq kill-ring-yank-pointer kill-ring)}
10333 (defun rotate-yank-pointer (arg)
10334 "Rotate the yanking point in the kill ring.
10335 With argument, rotate that many kills forward (or backward, if negative)."
10337 (current-kill arg))
10339 (defun current-kill (n &optional do-not-move)
10340 "Rotate the yanking point by N places, and then return that kill.
10341 If N is zero, `interprogram-paste-function' is set, and calling it
10342 returns a string, then that string is added to the front of the
10343 kill ring and returned as the latest kill.
10344 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10345 yanking point; just return the Nth kill forward."
10346 (let ((interprogram-paste (and (= n 0)
10347 interprogram-paste-function
10348 (funcall interprogram-paste-function))))
10349 (if interprogram-paste
10351 ;; Disable the interprogram cut function when we add the new
10352 ;; text to the kill ring, so Emacs doesn't try to own the
10353 ;; selection, with identical text.
10354 (let ((interprogram-cut-function nil))
10355 (kill-new interprogram-paste))
10356 interprogram-paste)
10357 (or kill-ring (error "Kill ring is empty"))
10358 (let ((ARGth-kill-element
10359 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10360 (length kill-ring))
10363 (setq kill-ring-yank-pointer ARGth-kill-element))
10364 (car ARGth-kill-element)))))
10369 @node yank nthcdr Exercises, , kill-ring-yank-pointer, Yanking
10370 @section Exercises with @code{yank} and @code{nthcdr}
10374 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10375 your kill ring. Add several items to your kill ring; look at its
10376 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10377 around the kill ring. How many items were in your kill ring? Find
10378 the value of @code{kill-ring-max}. Was your kill ring full, or could
10379 you have kept more blocks of text within it?
10382 Using @code{nthcdr} and @code{car}, construct a series of expressions
10383 to return the first, second, third, and fourth elements of a list.
10386 @node Loops & Recursion, Regexp Search, Yanking, Top
10387 @comment node-name, next, previous, up
10388 @chapter Loops and Recursion
10389 @cindex Loops and recursion
10390 @cindex Recursion and loops
10391 @cindex Repetition (loops)
10393 Emacs Lisp has two primary ways to cause an expression, or a series of
10394 expressions, to be evaluated repeatedly: one uses a @code{while}
10395 loop, and the other uses @dfn{recursion}.
10397 Repetition can be very valuable. For example, to move forward four
10398 sentences, you need only write a program that will move forward one
10399 sentence and then repeat the process four times. Since a computer does
10400 not get bored or tired, such repetitive action does not have the
10401 deleterious effects that excessive or the wrong kinds of repetition can
10404 People mostly write Emacs Lisp functions using @code{while} loops and
10405 their kin; but you can use recursion, which provides a very powerful
10406 way to think about and then to solve problems@footnote{You can write
10407 recursive functions to be frugal or wasteful of mental or computer
10408 resources; as it happens, methods that people find easy---that are
10409 frugal of `mental resources'---sometimes use considerable computer
10410 resources. Emacs was designed to run on machines that we now consider
10411 limited and its default settings are conservative. You may want to
10412 increase the values of @code{max-specpdl-size} and
10413 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10414 15 and 30 times their default value.}.
10417 * while:: Causing a stretch of code to repeat.
10419 * Recursion:: Causing a function to call itself.
10420 * Looping exercise::
10423 @node while, dolist dotimes, Loops & Recursion, Loops & Recursion
10424 @comment node-name, next, previous, up
10425 @section @code{while}
10429 The @code{while} special form tests whether the value returned by
10430 evaluating its first argument is true or false. This is similar to what
10431 the Lisp interpreter does with an @code{if}; what the interpreter does
10432 next, however, is different.
10434 In a @code{while} expression, if the value returned by evaluating the
10435 first argument is false, the Lisp interpreter skips the rest of the
10436 expression (the @dfn{body} of the expression) and does not evaluate it.
10437 However, if the value is true, the Lisp interpreter evaluates the body
10438 of the expression and then again tests whether the first argument to
10439 @code{while} is true or false. If the value returned by evaluating the
10440 first argument is again true, the Lisp interpreter again evaluates the
10441 body of the expression.
10444 The template for a @code{while} expression looks like this:
10448 (while @var{true-or-false-test}
10454 * Looping with while:: Repeat so long as test returns true.
10455 * Loop Example:: A @code{while} loop that uses a list.
10456 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10457 * Incrementing Loop:: A loop with an incrementing counter.
10458 * Incrementing Loop Details::
10459 * Decrementing Loop:: A loop with a decrementing counter.
10462 @node Looping with while, Loop Example, while, while
10464 @unnumberedsubsec Looping with @code{while}
10467 So long as the true-or-false-test of the @code{while} expression
10468 returns a true value when it is evaluated, the body is repeatedly
10469 evaluated. This process is called a loop since the Lisp interpreter
10470 repeats the same thing again and again, like an airplane doing a loop.
10471 When the result of evaluating the true-or-false-test is false, the
10472 Lisp interpreter does not evaluate the rest of the @code{while}
10473 expression and `exits the loop'.
10475 Clearly, if the value returned by evaluating the first argument to
10476 @code{while} is always true, the body following will be evaluated
10477 again and again @dots{} and again @dots{} forever. Conversely, if the
10478 value returned is never true, the expressions in the body will never
10479 be evaluated. The craft of writing a @code{while} loop consists of
10480 choosing a mechanism such that the true-or-false-test returns true
10481 just the number of times that you want the subsequent expressions to
10482 be evaluated, and then have the test return false.
10484 The value returned by evaluating a @code{while} is the value of the
10485 true-or-false-test. An interesting consequence of this is that a
10486 @code{while} loop that evaluates without error will return @code{nil}
10487 or false regardless of whether it has looped 1 or 100 times or none at
10488 all. A @code{while} expression that evaluates successfully never
10489 returns a true value! What this means is that @code{while} is always
10490 evaluated for its side effects, which is to say, the consequences of
10491 evaluating the expressions within the body of the @code{while} loop.
10492 This makes sense. It is not the mere act of looping that is desired,
10493 but the consequences of what happens when the expressions in the loop
10494 are repeatedly evaluated.
10496 @node Loop Example, print-elements-of-list, Looping with while, while
10497 @comment node-name, next, previous, up
10498 @subsection A @code{while} Loop and a List
10500 A common way to control a @code{while} loop is to test whether a list
10501 has any elements. If it does, the loop is repeated; but if it does not,
10502 the repetition is ended. Since this is an important technique, we will
10503 create a short example to illustrate it.
10505 A simple way to test whether a list has elements is to evaluate the
10506 list: if it has no elements, it is an empty list and will return the
10507 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10508 the other hand, a list with elements will return those elements when it
10509 is evaluated. Since Emacs Lisp considers as true any value that is not
10510 @code{nil}, a list that returns elements will test true in a
10514 For example, you can set the variable @code{empty-list} to @code{nil} by
10515 evaluating the following @code{setq} expression:
10518 (setq empty-list ())
10522 After evaluating the @code{setq} expression, you can evaluate the
10523 variable @code{empty-list} in the usual way, by placing the cursor after
10524 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10531 On the other hand, if you set a variable to be a list with elements, the
10532 list will appear when you evaluate the variable, as you can see by
10533 evaluating the following two expressions:
10537 (setq animals '(gazelle giraffe lion tiger))
10543 Thus, to create a @code{while} loop that tests whether there are any
10544 items in the list @code{animals}, the first part of the loop will be
10555 When the @code{while} tests its first argument, the variable
10556 @code{animals} is evaluated. It returns a list. So long as the list
10557 has elements, the @code{while} considers the results of the test to be
10558 true; but when the list is empty, it considers the results of the test
10561 To prevent the @code{while} loop from running forever, some mechanism
10562 needs to be provided to empty the list eventually. An oft-used
10563 technique is to have one of the subsequent forms in the @code{while}
10564 expression set the value of the list to be the @sc{cdr} of the list.
10565 Each time the @code{cdr} function is evaluated, the list will be made
10566 shorter, until eventually only the empty list will be left. At this
10567 point, the test of the @code{while} loop will return false, and the
10568 arguments to the @code{while} will no longer be evaluated.
10570 For example, the list of animals bound to the variable @code{animals}
10571 can be set to be the @sc{cdr} of the original list with the
10572 following expression:
10575 (setq animals (cdr animals))
10579 If you have evaluated the previous expressions and then evaluate this
10580 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10581 area. If you evaluate the expression again, @code{(lion tiger)} will
10582 appear in the echo area. If you evaluate it again and yet again,
10583 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10585 A template for a @code{while} loop that uses the @code{cdr} function
10586 repeatedly to cause the true-or-false-test eventually to test false
10591 (while @var{test-whether-list-is-empty}
10593 @var{set-list-to-cdr-of-list})
10597 This test and use of @code{cdr} can be put together in a function that
10598 goes through a list and prints each element of the list on a line of its
10601 @node print-elements-of-list, Incrementing Loop, Loop Example, while
10602 @subsection An Example: @code{print-elements-of-list}
10603 @findex print-elements-of-list
10605 The @code{print-elements-of-list} function illustrates a @code{while}
10608 @cindex @file{*scratch*} buffer
10609 The function requires several lines for its output. If you are
10610 reading this in a recent instance of GNU Emacs,
10611 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10612 you can evaluate the following expression inside of Info, as usual.
10614 If you are using an earlier version of Emacs, you need to copy the
10615 necessary expressions to your @file{*scratch*} buffer and evaluate
10616 them there. This is because the echo area had only one line in the
10619 You can copy the expressions by marking the beginning of the region
10620 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10621 the end of the region and then copying the region using @kbd{M-w}
10622 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10623 then provides visual feedback). In the @file{*scratch*}
10624 buffer, you can yank the expressions back by typing @kbd{C-y}
10627 After you have copied the expressions to the @file{*scratch*} buffer,
10628 evaluate each expression in turn. Be sure to evaluate the last
10629 expression, @code{(print-elements-of-list animals)}, by typing
10630 @kbd{C-u C-x C-e}, that is, by giving an argument to
10631 @code{eval-last-sexp}. This will cause the result of the evaluation
10632 to be printed in the @file{*scratch*} buffer instead of being printed
10633 in the echo area. (Otherwise you will see something like this in your
10634 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10635 each @samp{^J} stands for a `newline'.)
10638 In a recent instance of GNU Emacs, you can evaluate these expressions
10639 directly in the Info buffer, and the echo area will grow to show the
10644 (setq animals '(gazelle giraffe lion tiger))
10646 (defun print-elements-of-list (list)
10647 "Print each element of LIST on a line of its own."
10650 (setq list (cdr list))))
10652 (print-elements-of-list animals)
10658 When you evaluate the three expressions in sequence, you will see
10674 Each element of the list is printed on a line of its own (that is what
10675 the function @code{print} does) and then the value returned by the
10676 function is printed. Since the last expression in the function is the
10677 @code{while} loop, and since @code{while} loops always return
10678 @code{nil}, a @code{nil} is printed after the last element of the list.
10680 @node Incrementing Loop, Incrementing Loop Details, print-elements-of-list, while
10681 @comment node-name, next, previous, up
10682 @subsection A Loop with an Incrementing Counter
10684 A loop is not useful unless it stops when it ought. Besides
10685 controlling a loop with a list, a common way of stopping a loop is to
10686 write the first argument as a test that returns false when the correct
10687 number of repetitions are complete. This means that the loop must
10688 have a counter---an expression that counts how many times the loop
10691 @node Incrementing Loop Details, Decrementing Loop, Incrementing Loop, while
10693 @unnumberedsubsec Details of an Incrementing Loop
10696 The test for a loop with an incrementing counter can be an expression
10697 such as @code{(< count desired-number)} which returns @code{t} for
10698 true if the value of @code{count} is less than the
10699 @code{desired-number} of repetitions and @code{nil} for false if the
10700 value of @code{count} is equal to or is greater than the
10701 @code{desired-number}. The expression that increments the count can
10702 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10703 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10704 argument. (The expression @w{@code{(1+ count)}} has the same result
10705 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10708 The template for a @code{while} loop controlled by an incrementing
10709 counter looks like this:
10713 @var{set-count-to-initial-value}
10714 (while (< count desired-number) ; @r{true-or-false-test}
10716 (setq count (1+ count))) ; @r{incrementer}
10721 Note that you need to set the initial value of @code{count}; usually it
10725 * Incrementing Example:: Counting pebbles in a triangle.
10726 * Inc Example parts:: The parts of the function definition.
10727 * Inc Example altogether:: Putting the function definition together.
10730 @node Incrementing Example, Inc Example parts, Incrementing Loop Details, Incrementing Loop Details
10731 @unnumberedsubsubsec Example with incrementing counter
10733 Suppose you are playing on the beach and decide to make a triangle of
10734 pebbles, putting one pebble in the first row, two in the second row,
10735 three in the third row and so on, like this:
10753 @bullet{} @bullet{}
10754 @bullet{} @bullet{} @bullet{}
10755 @bullet{} @bullet{} @bullet{} @bullet{}
10762 (About 2500 years ago, Pythagoras and others developed the beginnings of
10763 number theory by considering questions such as this.)
10765 Suppose you want to know how many pebbles you will need to make a
10766 triangle with 7 rows?
10768 Clearly, what you need to do is add up the numbers from 1 to 7. There
10769 are two ways to do this; start with the smallest number, one, and add up
10770 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10771 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10772 mechanisms illustrate common ways of writing @code{while} loops, we will
10773 create two examples, one counting up and the other counting down. In
10774 this first example, we will start with 1 and add 2, 3, 4 and so on.
10776 If you are just adding up a short list of numbers, the easiest way to do
10777 it is to add up all the numbers at once. However, if you do not know
10778 ahead of time how many numbers your list will have, or if you want to be
10779 prepared for a very long list, then you need to design your addition so
10780 that what you do is repeat a simple process many times instead of doing
10781 a more complex process once.
10783 For example, instead of adding up all the pebbles all at once, what you
10784 can do is add the number of pebbles in the first row, 1, to the number
10785 in the second row, 2, and then add the total of those two rows to the
10786 third row, 3. Then you can add the number in the fourth row, 4, to the
10787 total of the first three rows; and so on.
10789 The critical characteristic of the process is that each repetitive
10790 action is simple. In this case, at each step we add only two numbers,
10791 the number of pebbles in the row and the total already found. This
10792 process of adding two numbers is repeated again and again until the last
10793 row has been added to the total of all the preceding rows. In a more
10794 complex loop the repetitive action might not be so simple, but it will
10795 be simpler than doing everything all at once.
10797 @node Inc Example parts, Inc Example altogether, Incrementing Example, Incrementing Loop Details
10798 @unnumberedsubsubsec The parts of the function definition
10800 The preceding analysis gives us the bones of our function definition:
10801 first, we will need a variable that we can call @code{total} that will
10802 be the total number of pebbles. This will be the value returned by
10805 Second, we know that the function will require an argument: this
10806 argument will be the total number of rows in the triangle. It can be
10807 called @code{number-of-rows}.
10809 Finally, we need a variable to use as a counter. We could call this
10810 variable @code{counter}, but a better name is @code{row-number}. That
10811 is because what the counter does in this function is count rows, and a
10812 program should be written to be as understandable as possible.
10814 When the Lisp interpreter first starts evaluating the expressions in the
10815 function, the value of @code{total} should be set to zero, since we have
10816 not added anything to it. Then the function should add the number of
10817 pebbles in the first row to the total, and then add the number of
10818 pebbles in the second to the total, and then add the number of
10819 pebbles in the third row to the total, and so on, until there are no
10820 more rows left to add.
10822 Both @code{total} and @code{row-number} are used only inside the
10823 function, so they can be declared as local variables with @code{let}
10824 and given initial values. Clearly, the initial value for @code{total}
10825 should be 0. The initial value of @code{row-number} should be 1,
10826 since we start with the first row. This means that the @code{let}
10827 statement will look like this:
10837 After the internal variables are declared and bound to their initial
10838 values, we can begin the @code{while} loop. The expression that serves
10839 as the test should return a value of @code{t} for true so long as the
10840 @code{row-number} is less than or equal to the @code{number-of-rows}.
10841 (If the expression tests true only so long as the row number is less
10842 than the number of rows in the triangle, the last row will never be
10843 added to the total; hence the row number has to be either less than or
10844 equal to the number of rows.)
10847 @findex <= @r{(less than or equal)}
10848 Lisp provides the @code{<=} function that returns true if the value of
10849 its first argument is less than or equal to the value of its second
10850 argument and false otherwise. So the expression that the @code{while}
10851 will evaluate as its test should look like this:
10854 (<= row-number number-of-rows)
10857 The total number of pebbles can be found by repeatedly adding the number
10858 of pebbles in a row to the total already found. Since the number of
10859 pebbles in the row is equal to the row number, the total can be found by
10860 adding the row number to the total. (Clearly, in a more complex
10861 situation, the number of pebbles in the row might be related to the row
10862 number in a more complicated way; if this were the case, the row number
10863 would be replaced by the appropriate expression.)
10866 (setq total (+ total row-number))
10870 What this does is set the new value of @code{total} to be equal to the
10871 sum of adding the number of pebbles in the row to the previous total.
10873 After setting the value of @code{total}, the conditions need to be
10874 established for the next repetition of the loop, if there is one. This
10875 is done by incrementing the value of the @code{row-number} variable,
10876 which serves as a counter. After the @code{row-number} variable has
10877 been incremented, the true-or-false-test at the beginning of the
10878 @code{while} loop tests whether its value is still less than or equal to
10879 the value of the @code{number-of-rows} and if it is, adds the new value
10880 of the @code{row-number} variable to the @code{total} of the previous
10881 repetition of the loop.
10884 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10885 @code{row-number} variable can be incremented with this expression:
10888 (setq row-number (1+ row-number))
10891 @node Inc Example altogether, , Inc Example parts, Incrementing Loop Details
10892 @unnumberedsubsubsec Putting the function definition together
10894 We have created the parts for the function definition; now we need to
10898 First, the contents of the @code{while} expression:
10902 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10903 (setq total (+ total row-number))
10904 (setq row-number (1+ row-number))) ; @r{incrementer}
10908 Along with the @code{let} expression varlist, this very nearly
10909 completes the body of the function definition. However, it requires
10910 one final element, the need for which is somewhat subtle.
10912 The final touch is to place the variable @code{total} on a line by
10913 itself after the @code{while} expression. Otherwise, the value returned
10914 by the whole function is the value of the last expression that is
10915 evaluated in the body of the @code{let}, and this is the value
10916 returned by the @code{while}, which is always @code{nil}.
10918 This may not be evident at first sight. It almost looks as if the
10919 incrementing expression is the last expression of the whole function.
10920 But that expression is part of the body of the @code{while}; it is the
10921 last element of the list that starts with the symbol @code{while}.
10922 Moreover, the whole of the @code{while} loop is a list within the body
10926 In outline, the function will look like this:
10930 (defun @var{name-of-function} (@var{argument-list})
10931 "@var{documentation}@dots{}"
10932 (let (@var{varlist})
10933 (while (@var{true-or-false-test})
10934 @var{body-of-while}@dots{} )
10935 @dots{} )) ; @r{Need final expression here.}
10939 The result of evaluating the @code{let} is what is going to be returned
10940 by the @code{defun} since the @code{let} is not embedded within any
10941 containing list, except for the @code{defun} as a whole. However, if
10942 the @code{while} is the last element of the @code{let} expression, the
10943 function will always return @code{nil}. This is not what we want!
10944 Instead, what we want is the value of the variable @code{total}. This
10945 is returned by simply placing the symbol as the last element of the list
10946 starting with @code{let}. It gets evaluated after the preceding
10947 elements of the list are evaluated, which means it gets evaluated after
10948 it has been assigned the correct value for the total.
10950 It may be easier to see this by printing the list starting with
10951 @code{let} all on one line. This format makes it evident that the
10952 @var{varlist} and @code{while} expressions are the second and third
10953 elements of the list starting with @code{let}, and the @code{total} is
10958 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10963 Putting everything together, the @code{triangle} function definition
10968 (defun triangle (number-of-rows) ; @r{Version with}
10969 ; @r{ incrementing counter.}
10970 "Add up the number of pebbles in a triangle.
10971 The first row has one pebble, the second row two pebbles,
10972 the third row three pebbles, and so on.
10973 The argument is NUMBER-OF-ROWS."
10978 (while (<= row-number number-of-rows)
10979 (setq total (+ total row-number))
10980 (setq row-number (1+ row-number)))
10986 After you have installed @code{triangle} by evaluating the function, you
10987 can try it out. Here are two examples:
10998 The sum of the first four numbers is 10 and the sum of the first seven
11001 @node Decrementing Loop, , Incrementing Loop Details, while
11002 @comment node-name, next, previous, up
11003 @subsection Loop with a Decrementing Counter
11005 Another common way to write a @code{while} loop is to write the test
11006 so that it determines whether a counter is greater than zero. So long
11007 as the counter is greater than zero, the loop is repeated. But when
11008 the counter is equal to or less than zero, the loop is stopped. For
11009 this to work, the counter has to start out greater than zero and then
11010 be made smaller and smaller by a form that is evaluated
11013 The test will be an expression such as @code{(> counter 0)} which
11014 returns @code{t} for true if the value of @code{counter} is greater
11015 than zero, and @code{nil} for false if the value of @code{counter} is
11016 equal to or less than zero. The expression that makes the number
11017 smaller and smaller can be a simple @code{setq} such as @code{(setq
11018 counter (1- counter))}, where @code{1-} is a built-in function in
11019 Emacs Lisp that subtracts 1 from its argument.
11022 The template for a decrementing @code{while} loop looks like this:
11026 (while (> counter 0) ; @r{true-or-false-test}
11028 (setq counter (1- counter))) ; @r{decrementer}
11033 * Decrementing Example:: More pebbles on the beach.
11034 * Dec Example parts:: The parts of the function definition.
11035 * Dec Example altogether:: Putting the function definition together.
11038 @node Decrementing Example, Dec Example parts, Decrementing Loop, Decrementing Loop
11039 @unnumberedsubsubsec Example with decrementing counter
11041 To illustrate a loop with a decrementing counter, we will rewrite the
11042 @code{triangle} function so the counter decreases to zero.
11044 This is the reverse of the earlier version of the function. In this
11045 case, to find out how many pebbles are needed to make a triangle with
11046 3 rows, add the number of pebbles in the third row, 3, to the number
11047 in the preceding row, 2, and then add the total of those two rows to
11048 the row that precedes them, which is 1.
11050 Likewise, to find the number of pebbles in a triangle with 7 rows, add
11051 the number of pebbles in the seventh row, 7, to the number in the
11052 preceding row, which is 6, and then add the total of those two rows to
11053 the row that precedes them, which is 5, and so on. As in the previous
11054 example, each addition only involves adding two numbers, the total of
11055 the rows already added up and the number of pebbles in the row that is
11056 being added to the total. This process of adding two numbers is
11057 repeated again and again until there are no more pebbles to add.
11059 We know how many pebbles to start with: the number of pebbles in the
11060 last row is equal to the number of rows. If the triangle has seven
11061 rows, the number of pebbles in the last row is 7. Likewise, we know how
11062 many pebbles are in the preceding row: it is one less than the number in
11065 @node Dec Example parts, Dec Example altogether, Decrementing Example, Decrementing Loop
11066 @unnumberedsubsubsec The parts of the function definition
11068 We start with three variables: the total number of rows in the
11069 triangle; the number of pebbles in a row; and the total number of
11070 pebbles, which is what we want to calculate. These variables can be
11071 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
11072 @code{total}, respectively.
11074 Both @code{total} and @code{number-of-pebbles-in-row} are used only
11075 inside the function and are declared with @code{let}. The initial
11076 value of @code{total} should, of course, be zero. However, the
11077 initial value of @code{number-of-pebbles-in-row} should be equal to
11078 the number of rows in the triangle, since the addition will start with
11082 This means that the beginning of the @code{let} expression will look
11088 (number-of-pebbles-in-row number-of-rows))
11093 The total number of pebbles can be found by repeatedly adding the number
11094 of pebbles in a row to the total already found, that is, by repeatedly
11095 evaluating the following expression:
11098 (setq total (+ total number-of-pebbles-in-row))
11102 After the @code{number-of-pebbles-in-row} is added to the @code{total},
11103 the @code{number-of-pebbles-in-row} should be decremented by one, since
11104 the next time the loop repeats, the preceding row will be
11105 added to the total.
11107 The number of pebbles in a preceding row is one less than the number of
11108 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
11109 used to compute the number of pebbles in the preceding row. This can be
11110 done with the following expression:
11114 (setq number-of-pebbles-in-row
11115 (1- number-of-pebbles-in-row))
11119 Finally, we know that the @code{while} loop should stop making repeated
11120 additions when there are no pebbles in a row. So the test for
11121 the @code{while} loop is simply:
11124 (while (> number-of-pebbles-in-row 0)
11127 @node Dec Example altogether, , Dec Example parts, Decrementing Loop
11128 @unnumberedsubsubsec Putting the function definition together
11130 We can put these expressions together to create a function definition
11131 that works. However, on examination, we find that one of the local
11132 variables is unneeded!
11135 The function definition looks like this:
11139 ;;; @r{First subtractive version.}
11140 (defun triangle (number-of-rows)
11141 "Add up the number of pebbles in a triangle."
11143 (number-of-pebbles-in-row number-of-rows))
11144 (while (> number-of-pebbles-in-row 0)
11145 (setq total (+ total number-of-pebbles-in-row))
11146 (setq number-of-pebbles-in-row
11147 (1- number-of-pebbles-in-row)))
11152 As written, this function works.
11154 However, we do not need @code{number-of-pebbles-in-row}.
11156 @cindex Argument as local variable
11157 When the @code{triangle} function is evaluated, the symbol
11158 @code{number-of-rows} will be bound to a number, giving it an initial
11159 value. That number can be changed in the body of the function as if
11160 it were a local variable, without any fear that such a change will
11161 effect the value of the variable outside of the function. This is a
11162 very useful characteristic of Lisp; it means that the variable
11163 @code{number-of-rows} can be used anywhere in the function where
11164 @code{number-of-pebbles-in-row} is used.
11167 Here is a second version of the function written a bit more cleanly:
11171 (defun triangle (number) ; @r{Second version.}
11172 "Return sum of numbers 1 through NUMBER inclusive."
11174 (while (> number 0)
11175 (setq total (+ total number))
11176 (setq number (1- number)))
11181 In brief, a properly written @code{while} loop will consist of three parts:
11185 A test that will return false after the loop has repeated itself the
11186 correct number of times.
11189 An expression the evaluation of which will return the value desired
11190 after being repeatedly evaluated.
11193 An expression to change the value passed to the true-or-false-test so
11194 that the test returns false after the loop has repeated itself the right
11198 @node dolist dotimes, Recursion, while, Loops & Recursion
11199 @comment node-name, next, previous, up
11200 @section Save your time: @code{dolist} and @code{dotimes}
11202 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11203 provide for looping. Sometimes these are quicker to write than the
11204 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11205 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11207 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11208 list': @code{dolist} automatically shortens the list each time it
11209 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11210 each shorter version of the list to the first of its arguments.
11212 @code{dotimes} loops a specific number of times: you specify the number.
11219 @node dolist, dotimes, dolist dotimes, dolist dotimes
11220 @unnumberedsubsubsec The @code{dolist} Macro
11223 Suppose, for example, you want to reverse a list, so that
11224 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11227 In practice, you would use the @code{reverse} function, like this:
11231 (setq animals '(gazelle giraffe lion tiger))
11239 Here is how you could reverse the list using a @code{while} loop:
11243 (setq animals '(gazelle giraffe lion tiger))
11245 (defun reverse-list-with-while (list)
11246 "Using while, reverse the order of LIST."
11247 (let (value) ; make sure list starts empty
11249 (setq value (cons (car list) value))
11250 (setq list (cdr list)))
11253 (reverse-list-with-while animals)
11259 And here is how you could use the @code{dolist} macro:
11263 (setq animals '(gazelle giraffe lion tiger))
11265 (defun reverse-list-with-dolist (list)
11266 "Using dolist, reverse the order of LIST."
11267 (let (value) ; make sure list starts empty
11268 (dolist (element list value)
11269 (setq value (cons element value)))))
11271 (reverse-list-with-dolist animals)
11277 In Info, you can place your cursor after the closing parenthesis of
11278 each expression and type @kbd{C-x C-e}; in each case, you should see
11281 (tiger lion giraffe gazelle)
11287 For this example, the existing @code{reverse} function is obviously best.
11288 The @code{while} loop is just like our first example (@pxref{Loop
11289 Example, , A @code{while} Loop and a List}). The @code{while} first
11290 checks whether the list has elements; if so, it constructs a new list
11291 by adding the first element of the list to the existing list (which in
11292 the first iteration of the loop is @code{nil}). Since the second
11293 element is prepended in front of the first element, and the third
11294 element is prepended in front of the second element, the list is reversed.
11296 In the expression using a @code{while} loop,
11297 the @w{@code{(setq list (cdr list))}}
11298 expression shortens the list, so the @code{while} loop eventually
11299 stops. In addition, it provides the @code{cons} expression with a new
11300 first element by creating a new and shorter list at each repetition of
11303 The @code{dolist} expression does very much the same as the
11304 @code{while} expression, except that the @code{dolist} macro does some
11305 of the work you have to do when writing a @code{while} expression.
11307 Like a @code{while} loop, a @code{dolist} loops. What is different is
11308 that it automatically shortens the list each time it loops --- it
11309 `@sc{cdr}s down the list' on its own --- and it automatically binds
11310 the @sc{car} of each shorter version of the list to the first of its
11313 In the example, the @sc{car} of each shorter version of the list is
11314 referred to using the symbol @samp{element}, the list itself is called
11315 @samp{list}, and the value returned is called @samp{value}. The
11316 remainder of the @code{dolist} expression is the body.
11318 The @code{dolist} expression binds the @sc{car} of each shorter
11319 version of the list to @code{element} and then evaluates the body of
11320 the expression; and repeats the loop. The result is returned in
11323 @node dotimes, , dolist, dolist dotimes
11324 @unnumberedsubsubsec The @code{dotimes} Macro
11327 The @code{dotimes} macro is similar to @code{dolist}, except that it
11328 loops a specific number of times.
11330 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11331 and so forth each time around the loop, and the value of the third
11332 argument is returned. You need to provide the value of the second
11333 argument, which is how many times the macro loops.
11336 For example, the following binds the numbers from 0 up to, but not
11337 including, the number 3 to the first argument, @var{number}, and then
11338 constructs a list of the three numbers. (The first number is 0, the
11339 second number is 1, and the third number is 2; this makes a total of
11340 three numbers in all, starting with zero as the first number.)
11344 (let (value) ; otherwise a value is a void variable
11345 (dotimes (number 3 value)
11346 (setq value (cons number value))))
11353 @code{dotimes} returns @code{value}, so the way to use
11354 @code{dotimes} is to operate on some expression @var{number} number of
11355 times and then return the result, either as a list or an atom.
11358 Here is an example of a @code{defun} that uses @code{dotimes} to add
11359 up the number of pebbles in a triangle.
11363 (defun triangle-using-dotimes (number-of-rows)
11364 "Using dotimes, add up the number of pebbles in a triangle."
11365 (let ((total 0)) ; otherwise a total is a void variable
11366 (dotimes (number number-of-rows total)
11367 (setq total (+ total (1+ number))))))
11369 (triangle-using-dotimes 4)
11373 @node Recursion, Looping exercise, dolist dotimes, Loops & Recursion
11374 @comment node-name, next, previous, up
11378 A recursive function contains code that tells the Lisp interpreter to
11379 call a program that runs exactly like itself, but with slightly
11380 different arguments. The code runs exactly the same because it has
11381 the same name. However, even though the program has the same name, it
11382 is not the same entity. It is different. In the jargon, it is a
11383 different `instance'.
11385 Eventually, if the program is written correctly, the `slightly
11386 different arguments' will become sufficiently different from the first
11387 arguments that the final instance will stop.
11390 * Building Robots:: Same model, different serial number ...
11391 * Recursive Definition Parts:: Walk until you stop ...
11392 * Recursion with list:: Using a list as the test whether to recurse.
11393 * Recursive triangle function::
11394 * Recursion with cond::
11395 * Recursive Patterns:: Often used templates.
11396 * No Deferment:: Don't store up work ...
11397 * No deferment solution::
11400 @node Building Robots, Recursive Definition Parts, Recursion, Recursion
11401 @comment node-name, next, previous, up
11402 @subsection Building Robots: Extending the Metaphor
11403 @cindex Building robots
11404 @cindex Robots, building
11406 It is sometimes helpful to think of a running program as a robot that
11407 does a job. In doing its job, a recursive function calls on a second
11408 robot to help it. The second robot is identical to the first in every
11409 way, except that the second robot helps the first and has been
11410 passed different arguments than the first.
11412 In a recursive function, the second robot may call a third; and the
11413 third may call a fourth, and so on. Each of these is a different
11414 entity; but all are clones.
11416 Since each robot has slightly different instructions---the arguments
11417 will differ from one robot to the next---the last robot should know
11420 Let's expand on the metaphor in which a computer program is a robot.
11422 A function definition provides the blueprints for a robot. When you
11423 install a function definition, that is, when you evaluate a
11424 @code{defun} special form, you install the necessary equipment to
11425 build robots. It is as if you were in a factory, setting up an
11426 assembly line. Robots with the same name are built according to the
11427 same blueprints. So they have, as it were, the same `model number',
11428 but a different `serial number'.
11430 We often say that a recursive function `calls itself'. What we mean
11431 is that the instructions in a recursive function cause the Lisp
11432 interpreter to run a different function that has the same name and
11433 does the same job as the first, but with different arguments.
11435 It is important that the arguments differ from one instance to the
11436 next; otherwise, the process will never stop.
11438 @node Recursive Definition Parts, Recursion with list, Building Robots, Recursion
11439 @comment node-name, next, previous, up
11440 @subsection The Parts of a Recursive Definition
11441 @cindex Parts of a Recursive Definition
11442 @cindex Recursive Definition Parts
11444 A recursive function typically contains a conditional expression which
11449 A true-or-false-test that determines whether the function is called
11450 again, here called the @dfn{do-again-test}.
11453 The name of the function. When this name is called, a new instance of
11454 the function---a new robot, as it were---is created and told what to do.
11457 An expression that returns a different value each time the function is
11458 called, here called the @dfn{next-step-expression}. Consequently, the
11459 argument (or arguments) passed to the new instance of the function
11460 will be different from that passed to the previous instance. This
11461 causes the conditional expression, the @dfn{do-again-test}, to test
11462 false after the correct number of repetitions.
11465 Recursive functions can be much simpler than any other kind of
11466 function. Indeed, when people first start to use them, they often look
11467 so mysteriously simple as to be incomprehensible. Like riding a
11468 bicycle, reading a recursive function definition takes a certain knack
11469 which is hard at first but then seems simple.
11472 There are several different common recursive patterns. A very simple
11473 pattern looks like this:
11477 (defun @var{name-of-recursive-function} (@var{argument-list})
11478 "@var{documentation}@dots{}"
11479 (if @var{do-again-test}
11481 (@var{name-of-recursive-function}
11482 @var{next-step-expression})))
11486 Each time a recursive function is evaluated, a new instance of it is
11487 created and told what to do. The arguments tell the instance what to do.
11489 An argument is bound to the value of the next-step-expression. Each
11490 instance runs with a different value of the next-step-expression.
11492 The value in the next-step-expression is used in the do-again-test.
11494 The value returned by the next-step-expression is passed to the new
11495 instance of the function, which evaluates it (or some
11496 transmogrification of it) to determine whether to continue or stop.
11497 The next-step-expression is designed so that the do-again-test returns
11498 false when the function should no longer be repeated.
11500 The do-again-test is sometimes called the @dfn{stop condition},
11501 since it stops the repetitions when it tests false.
11503 @node Recursion with list, Recursive triangle function, Recursive Definition Parts, Recursion
11504 @comment node-name, next, previous, up
11505 @subsection Recursion with a List
11507 The example of a @code{while} loop that printed the elements of a list
11508 of numbers can be written recursively. Here is the code, including
11509 an expression to set the value of the variable @code{animals} to a list.
11511 If you are using GNU Emacs 20 or before, this example must be copied
11512 to the @file{*scratch*} buffer and each expression must be evaluated
11513 there. Use @kbd{C-u C-x C-e} to evaluate the
11514 @code{(print-elements-recursively animals)} expression so that the
11515 results are printed in the buffer; otherwise the Lisp interpreter will
11516 try to squeeze the results into the one line of the echo area.
11518 Also, place your cursor immediately after the last closing parenthesis
11519 of the @code{print-elements-recursively} function, before the comment.
11520 Otherwise, the Lisp interpreter will try to evaluate the comment.
11522 If you are using a more recent version of Emacs, you can evaluate this
11523 expression directly in Info.
11525 @findex print-elements-recursively
11528 (setq animals '(gazelle giraffe lion tiger))
11530 (defun print-elements-recursively (list)
11531 "Print each element of LIST on a line of its own.
11533 (when list ; @r{do-again-test}
11534 (print (car list)) ; @r{body}
11535 (print-elements-recursively ; @r{recursive call}
11536 (cdr list)))) ; @r{next-step-expression}
11538 (print-elements-recursively animals)
11542 The @code{print-elements-recursively} function first tests whether
11543 there is any content in the list; if there is, the function prints the
11544 first element of the list, the @sc{car} of the list. Then the
11545 function `invokes itself', but gives itself as its argument, not the
11546 whole list, but the second and subsequent elements of the list, the
11547 @sc{cdr} of the list.
11549 Put another way, if the list is not empty, the function invokes
11550 another instance of code that is similar to the initial code, but is a
11551 different thread of execution, with different arguments than the first
11554 Put in yet another way, if the list is not empty, the first robot
11555 assembles a second robot and tells it what to do; the second robot is
11556 a different individual from the first, but is the same model.
11558 When the second evaluation occurs, the @code{when} expression is
11559 evaluated and if true, prints the first element of the list it
11560 receives as its argument (which is the second element of the original
11561 list). Then the function `calls itself' with the @sc{cdr} of the list
11562 it is invoked with, which (the second time around) is the @sc{cdr} of
11563 the @sc{cdr} of the original list.
11565 Note that although we say that the function `calls itself', what we
11566 mean is that the Lisp interpreter assembles and instructs a new
11567 instance of the program. The new instance is a clone of the first,
11568 but is a separate individual.
11570 Each time the function `invokes itself', it invokes itself on a
11571 shorter version of the original list. It creates a new instance that
11572 works on a shorter list.
11574 Eventually, the function invokes itself on an empty list. It creates
11575 a new instance whose argument is @code{nil}. The conditional expression
11576 tests the value of @code{list}. Since the value of @code{list} is
11577 @code{nil}, the @code{when} expression tests false so the then-part is
11578 not evaluated. The function as a whole then returns @code{nil}.
11581 When you evaluate the expression @code{(print-elements-recursively
11582 animals)} in the @file{*scratch*} buffer, you see this result:
11598 @node Recursive triangle function, Recursion with cond, Recursion with list, Recursion
11599 @comment node-name, next, previous, up
11600 @subsection Recursion in Place of a Counter
11601 @findex triangle-recursively
11604 The @code{triangle} function described in a previous section can also
11605 be written recursively. It looks like this:
11609 (defun triangle-recursively (number)
11610 "Return the sum of the numbers 1 through NUMBER inclusive.
11612 (if (= number 1) ; @r{do-again-test}
11614 (+ number ; @r{else-part}
11615 (triangle-recursively ; @r{recursive call}
11616 (1- number))))) ; @r{next-step-expression}
11618 (triangle-recursively 7)
11623 You can install this function by evaluating it and then try it by
11624 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11625 cursor immediately after the last parenthesis of the function
11626 definition, before the comment.) The function evaluates to 28.
11628 To understand how this function works, let's consider what happens in the
11629 various cases when the function is passed 1, 2, 3, or 4 as the value of
11633 * Recursive Example arg of 1 or 2::
11634 * Recursive Example arg of 3 or 4::
11637 @node Recursive Example arg of 1 or 2, Recursive Example arg of 3 or 4, Recursive triangle function, Recursive triangle function
11639 @unnumberedsubsubsec An argument of 1 or 2
11642 First, what happens if the value of the argument is 1?
11644 The function has an @code{if} expression after the documentation
11645 string. It tests whether the value of @code{number} is equal to 1; if
11646 so, Emacs evaluates the then-part of the @code{if} expression, which
11647 returns the number 1 as the value of the function. (A triangle with
11648 one row has one pebble in it.)
11650 Suppose, however, that the value of the argument is 2. In this case,
11651 Emacs evaluates the else-part of the @code{if} expression.
11654 The else-part consists of an addition, the recursive call to
11655 @code{triangle-recursively} and a decrementing action; and it looks like
11659 (+ number (triangle-recursively (1- number)))
11662 When Emacs evaluates this expression, the innermost expression is
11663 evaluated first; then the other parts in sequence. Here are the steps
11667 @item Step 1 @w{ } Evaluate the innermost expression.
11669 The innermost expression is @code{(1- number)} so Emacs decrements the
11670 value of @code{number} from 2 to 1.
11672 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11674 The Lisp interpreter creates an individual instance of
11675 @code{triangle-recursively}. It does not matter that this function is
11676 contained within itself. Emacs passes the result Step 1 as the
11677 argument used by this instance of the @code{triangle-recursively}
11680 In this case, Emacs evaluates @code{triangle-recursively} with an
11681 argument of 1. This means that this evaluation of
11682 @code{triangle-recursively} returns 1.
11684 @item Step 3 @w{ } Evaluate the value of @code{number}.
11686 The variable @code{number} is the second element of the list that
11687 starts with @code{+}; its value is 2.
11689 @item Step 4 @w{ } Evaluate the @code{+} expression.
11691 The @code{+} expression receives two arguments, the first
11692 from the evaluation of @code{number} (Step 3) and the second from the
11693 evaluation of @code{triangle-recursively} (Step 2).
11695 The result of the addition is the sum of 2 plus 1, and the number 3 is
11696 returned, which is correct. A triangle with two rows has three
11700 @node Recursive Example arg of 3 or 4, , Recursive Example arg of 1 or 2, Recursive triangle function
11701 @unnumberedsubsubsec An argument of 3 or 4
11703 Suppose that @code{triangle-recursively} is called with an argument of
11707 @item Step 1 @w{ } Evaluate the do-again-test.
11709 The @code{if} expression is evaluated first. This is the do-again
11710 test and returns false, so the else-part of the @code{if} expression
11711 is evaluated. (Note that in this example, the do-again-test causes
11712 the function to call itself when it tests false, not when it tests
11715 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11717 The innermost expression of the else-part is evaluated, which decrements
11718 3 to 2. This is the next-step-expression.
11720 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11722 The number 2 is passed to the @code{triangle-recursively} function.
11724 We already know what happens when Emacs evaluates @code{triangle-recursively} with
11725 an argument of 2. After going through the sequence of actions described
11726 earlier, it returns a value of 3. So that is what will happen here.
11728 @item Step 4 @w{ } Evaluate the addition.
11730 3 will be passed as an argument to the addition and will be added to the
11731 number with which the function was called, which is 3.
11735 The value returned by the function as a whole will be 6.
11737 Now that we know what will happen when @code{triangle-recursively} is
11738 called with an argument of 3, it is evident what will happen if it is
11739 called with an argument of 4:
11743 In the recursive call, the evaluation of
11746 (triangle-recursively (1- 4))
11751 will return the value of evaluating
11754 (triangle-recursively 3)
11758 which is 6 and this value will be added to 4 by the addition in the
11763 The value returned by the function as a whole will be 10.
11765 Each time @code{triangle-recursively} is evaluated, it evaluates a
11766 version of itself---a different instance of itself---with a smaller
11767 argument, until the argument is small enough so that it does not
11770 Note that this particular design for a recursive function
11771 requires that operations be deferred.
11773 Before @code{(triangle-recursively 7)} can calculate its answer, it
11774 must call @code{(triangle-recursively 6)}; and before
11775 @code{(triangle-recursively 6)} can calculate its answer, it must call
11776 @code{(triangle-recursively 5)}; and so on. That is to say, the
11777 calculation that @code{(triangle-recursively 7)} makes must be
11778 deferred until @code{(triangle-recursively 6)} makes its calculation;
11779 and @code{(triangle-recursively 6)} must defer until
11780 @code{(triangle-recursively 5)} completes; and so on.
11782 If each of these instances of @code{triangle-recursively} are thought
11783 of as different robots, the first robot must wait for the second to
11784 complete its job, which must wait until the third completes, and so
11787 There is a way around this kind of waiting, which we will discuss in
11788 @ref{No Deferment, , Recursion without Deferments}.
11790 @node Recursion with cond, Recursive Patterns, Recursive triangle function, Recursion
11791 @comment node-name, next, previous, up
11792 @subsection Recursion Example Using @code{cond}
11795 The version of @code{triangle-recursively} described earlier is written
11796 with the @code{if} special form. It can also be written using another
11797 special form called @code{cond}. The name of the special form
11798 @code{cond} is an abbreviation of the word @samp{conditional}.
11800 Although the @code{cond} special form is not used as often in the
11801 Emacs Lisp sources as @code{if}, it is used often enough to justify
11805 The template for a @code{cond} expression looks like this:
11815 where the @var{body} is a series of lists.
11818 Written out more fully, the template looks like this:
11823 (@var{first-true-or-false-test} @var{first-consequent})
11824 (@var{second-true-or-false-test} @var{second-consequent})
11825 (@var{third-true-or-false-test} @var{third-consequent})
11830 When the Lisp interpreter evaluates the @code{cond} expression, it
11831 evaluates the first element (the @sc{car} or true-or-false-test) of
11832 the first expression in a series of expressions within the body of the
11835 If the true-or-false-test returns @code{nil} the rest of that
11836 expression, the consequent, is skipped and the true-or-false-test of the
11837 next expression is evaluated. When an expression is found whose
11838 true-or-false-test returns a value that is not @code{nil}, the
11839 consequent of that expression is evaluated. The consequent can be one
11840 or more expressions. If the consequent consists of more than one
11841 expression, the expressions are evaluated in sequence and the value of
11842 the last one is returned. If the expression does not have a consequent,
11843 the value of the true-or-false-test is returned.
11845 If none of the true-or-false-tests test true, the @code{cond} expression
11846 returns @code{nil}.
11849 Written using @code{cond}, the @code{triangle} function looks like this:
11853 (defun triangle-using-cond (number)
11854 (cond ((<= number 0) 0)
11857 (+ number (triangle-using-cond (1- number))))))
11862 In this example, the @code{cond} returns 0 if the number is less than or
11863 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11864 number (triangle-using-cond (1- number)))} if the number is greater than
11867 @node Recursive Patterns, No Deferment, Recursion with cond, Recursion
11868 @comment node-name, next, previous, up
11869 @subsection Recursive Patterns
11870 @cindex Recursive Patterns
11872 Here are three common recursive patterns. Each involves a list.
11873 Recursion does not need to involve lists, but Lisp is designed for lists
11874 and this provides a sense of its primal capabilities.
11882 @node Every, Accumulate, Recursive Patterns, Recursive Patterns
11883 @comment node-name, next, previous, up
11884 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11885 @cindex Every, type of recursive pattern
11886 @cindex Recursive pattern: every
11888 In the @code{every} recursive pattern, an action is performed on every
11892 The basic pattern is:
11896 If a list be empty, return @code{nil}.
11898 Else, act on the beginning of the list (the @sc{car} of the list)
11901 through a recursive call by the function on the rest (the
11902 @sc{cdr}) of the list,
11904 and, optionally, combine the acted-on element, using @code{cons},
11905 with the results of acting on the rest.
11914 (defun square-each (numbers-list)
11915 "Square each of a NUMBERS LIST, recursively."
11916 (if (not numbers-list) ; do-again-test
11919 (* (car numbers-list) (car numbers-list))
11920 (square-each (cdr numbers-list))))) ; next-step-expression
11924 (square-each '(1 2 3))
11931 If @code{numbers-list} is empty, do nothing. But if it has content,
11932 construct a list combining the square of the first number in the list
11933 with the result of the recursive call.
11935 (The example follows the pattern exactly: @code{nil} is returned if
11936 the numbers' list is empty. In practice, you would write the
11937 conditional so it carries out the action when the numbers' list is not
11940 The @code{print-elements-recursively} function (@pxref{Recursion with
11941 list, , Recursion with a List}) is another example of an @code{every}
11942 pattern, except in this case, rather than bring the results together
11943 using @code{cons}, we print each element of output.
11946 The @code{print-elements-recursively} function looks like this:
11950 (setq animals '(gazelle giraffe lion tiger))
11954 (defun print-elements-recursively (list)
11955 "Print each element of LIST on a line of its own.
11957 (when list ; @r{do-again-test}
11958 (print (car list)) ; @r{body}
11959 (print-elements-recursively ; @r{recursive call}
11960 (cdr list)))) ; @r{next-step-expression}
11962 (print-elements-recursively animals)
11967 The pattern for @code{print-elements-recursively} is:
11971 When the list is empty, do nothing.
11973 But when the list has at least one element,
11976 act on the beginning of the list (the @sc{car} of the list),
11978 and make a recursive call on the rest (the @sc{cdr}) of the list.
11982 @node Accumulate, Keep, Every, Recursive Patterns
11983 @comment node-name, next, previous, up
11984 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11985 @cindex Accumulate, type of recursive pattern
11986 @cindex Recursive pattern: accumulate
11988 Another recursive pattern is called the @code{accumulate} pattern. In
11989 the @code{accumulate} recursive pattern, an action is performed on
11990 every element of a list and the result of that action is accumulated
11991 with the results of performing the action on the other elements.
11993 This is very like the `every' pattern using @code{cons}, except that
11994 @code{cons} is not used, but some other combiner.
12001 If a list be empty, return zero or some other constant.
12003 Else, act on the beginning of the list (the @sc{car} of the list),
12006 and combine that acted-on element, using @code{+} or
12007 some other combining function, with
12009 a recursive call by the function on the rest (the @sc{cdr}) of the list.
12014 Here is an example:
12018 (defun add-elements (numbers-list)
12019 "Add the elements of NUMBERS-LIST together."
12020 (if (not numbers-list)
12022 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
12026 (add-elements '(1 2 3 4))
12031 @xref{Files List, , Making a List of Files}, for an example of the
12032 accumulate pattern.
12034 @node Keep, , Accumulate, Recursive Patterns
12035 @comment node-name, next, previous, up
12036 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
12037 @cindex Keep, type of recursive pattern
12038 @cindex Recursive pattern: keep
12040 A third recursive pattern is called the @code{keep} pattern.
12041 In the @code{keep} recursive pattern, each element of a list is tested;
12042 the element is acted on and the results are kept only if the element
12045 Again, this is very like the `every' pattern, except the element is
12046 skipped unless it meets a criterion.
12049 The pattern has three parts:
12053 If a list be empty, return @code{nil}.
12055 Else, if the beginning of the list (the @sc{car} of the list) passes
12059 act on that element and combine it, using @code{cons} with
12061 a recursive call by the function on the rest (the @sc{cdr}) of the list.
12064 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
12068 skip on that element,
12070 and, recursively call the function on the rest (the @sc{cdr}) of the list.
12075 Here is an example that uses @code{cond}:
12079 (defun keep-three-letter-words (word-list)
12080 "Keep three letter words in WORD-LIST."
12082 ;; First do-again-test: stop-condition
12083 ((not word-list) nil)
12085 ;; Second do-again-test: when to act
12086 ((eq 3 (length (symbol-name (car word-list))))
12087 ;; combine acted-on element with recursive call on shorter list
12088 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
12090 ;; Third do-again-test: when to skip element;
12091 ;; recursively call shorter list with next-step expression
12092 (t (keep-three-letter-words (cdr word-list)))))
12096 (keep-three-letter-words '(one two three four five six))
12097 @result{} (one two six)
12101 It goes without saying that you need not use @code{nil} as the test for
12102 when to stop; and you can, of course, combine these patterns.
12104 @node No Deferment, No deferment solution, Recursive Patterns, Recursion
12105 @subsection Recursion without Deferments
12106 @cindex Deferment in recursion
12107 @cindex Recursion without Deferments
12109 Let's consider again what happens with the @code{triangle-recursively}
12110 function. We will find that the intermediate calculations are
12111 deferred until all can be done.
12114 Here is the function definition:
12118 (defun triangle-recursively (number)
12119 "Return the sum of the numbers 1 through NUMBER inclusive.
12121 (if (= number 1) ; @r{do-again-test}
12123 (+ number ; @r{else-part}
12124 (triangle-recursively ; @r{recursive call}
12125 (1- number))))) ; @r{next-step-expression}
12129 What happens when we call this function with a argument of 7?
12131 The first instance of the @code{triangle-recursively} function adds
12132 the number 7 to the value returned by a second instance of
12133 @code{triangle-recursively}, an instance that has been passed an
12134 argument of 6. That is to say, the first calculation is:
12137 (+ 7 (triangle-recursively 6))
12141 The first instance of @code{triangle-recursively}---you may want to
12142 think of it as a little robot---cannot complete its job. It must hand
12143 off the calculation for @code{(triangle-recursively 6)} to a second
12144 instance of the program, to a second robot. This second individual is
12145 completely different from the first one; it is, in the jargon, a
12146 `different instantiation'. Or, put another way, it is a different
12147 robot. It is the same model as the first; it calculates triangle
12148 numbers recursively; but it has a different serial number.
12150 And what does @code{(triangle-recursively 6)} return? It returns the
12151 number 6 added to the value returned by evaluating
12152 @code{triangle-recursively} with an argument of 5. Using the robot
12153 metaphor, it asks yet another robot to help it.
12159 (+ 7 6 (triangle-recursively 5))
12163 And what happens next?
12166 (+ 7 6 5 (triangle-recursively 4))
12169 Each time @code{triangle-recursively} is called, except for the last
12170 time, it creates another instance of the program---another robot---and
12171 asks it to make a calculation.
12174 Eventually, the full addition is set up and performed:
12180 This design for the function defers the calculation of the first step
12181 until the second can be done, and defers that until the third can be
12182 done, and so on. Each deferment means the computer must remember what
12183 is being waited on. This is not a problem when there are only a few
12184 steps, as in this example. But it can be a problem when there are
12187 @node No deferment solution, , No Deferment, Recursion
12188 @subsection No Deferment Solution
12189 @cindex No deferment solution
12190 @cindex Defermentless solution
12191 @cindex Solution without deferment
12193 The solution to the problem of deferred operations is to write in a
12194 manner that does not defer operations@footnote{The phrase @dfn{tail
12195 recursive} is used to describe such a process, one that uses
12196 `constant space'.}. This requires
12197 writing to a different pattern, often one that involves writing two
12198 function definitions, an `initialization' function and a `helper'
12201 The `initialization' function sets up the job; the `helper' function
12205 Here are the two function definitions for adding up numbers. They are
12206 so simple, I find them hard to understand.
12210 (defun triangle-initialization (number)
12211 "Return the sum of the numbers 1 through NUMBER inclusive.
12212 This is the `initialization' component of a two function
12213 duo that uses recursion."
12214 (triangle-recursive-helper 0 0 number))
12220 (defun triangle-recursive-helper (sum counter number)
12221 "Return SUM, using COUNTER, through NUMBER inclusive.
12222 This is the `helper' component of a two function duo
12223 that uses recursion."
12224 (if (> counter number)
12226 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12227 (1+ counter) ; @r{counter}
12228 number))) ; @r{number}
12233 Install both function definitions by evaluating them, then call
12234 @code{triangle-initialization} with 2 rows:
12238 (triangle-initialization 2)
12243 The `initialization' function calls the first instance of the `helper'
12244 function with three arguments: zero, zero, and a number which is the
12245 number of rows in the triangle.
12247 The first two arguments passed to the `helper' function are
12248 initialization values. These values are changed when
12249 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12250 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12251 process that is iterative in a procedure that is recursive. The
12252 process is called iterative because the computer need only record the
12253 three values, @code{sum}, @code{counter}, and @code{number}; the
12254 procedure is recursive because the function `calls itself'. On the
12255 other hand, both the process and the procedure used by
12256 @code{triangle-recursively} are called recursive. The word
12257 `recursive' has different meanings in the two contexts.}
12259 Let's see what happens when we have a triangle that has one row. (This
12260 triangle will have one pebble in it!)
12263 @code{triangle-initialization} will call its helper with
12264 the arguments @w{@code{0 0 1}}. That function will run the conditional
12265 test whether @code{(> counter number)}:
12273 and find that the result is false, so it will invoke
12274 the else-part of the @code{if} clause:
12278 (triangle-recursive-helper
12279 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12280 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12281 number) ; @r{number stays the same}
12287 which will first compute:
12291 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12292 (1+ 0) ; @r{counter}
12296 (triangle-recursive-helper 0 1 1)
12300 Again, @code{(> counter number)} will be false, so again, the Lisp
12301 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12302 new instance with new arguments.
12305 This new instance will be;
12309 (triangle-recursive-helper
12310 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12311 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12312 number) ; @r{number stays the same}
12316 (triangle-recursive-helper 1 2 1)
12320 In this case, the @code{(> counter number)} test will be true! So the
12321 instance will return the value of the sum, which will be 1, as
12324 Now, let's pass @code{triangle-initialization} an argument
12325 of 2, to find out how many pebbles there are in a triangle with two rows.
12327 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12330 In stages, the instances called will be:
12334 @r{sum counter number}
12335 (triangle-recursive-helper 0 1 2)
12337 (triangle-recursive-helper 1 2 2)
12339 (triangle-recursive-helper 3 3 2)
12343 When the last instance is called, the @code{(> counter number)} test
12344 will be true, so the instance will return the value of @code{sum},
12347 This kind of pattern helps when you are writing functions that can use
12348 many resources in a computer.
12351 @node Looping exercise, , Recursion, Loops & Recursion
12352 @section Looping Exercise
12356 Write a function similar to @code{triangle} in which each row has a
12357 value which is the square of the row number. Use a @code{while} loop.
12360 Write a function similar to @code{triangle} that multiplies instead of
12364 Rewrite these two functions recursively. Rewrite these functions
12367 @c comma in printed title causes problem in Info cross reference
12369 Write a function for Texinfo mode that creates an index entry at the
12370 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12371 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12372 written in Texinfo.)
12374 Many of the functions you will need are described in two of the
12375 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12376 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12377 @code{forward-paragraph} to put the index entry at the beginning of
12378 the paragraph, you will have to use @w{@kbd{C-h f}}
12379 (@code{describe-function}) to find out how to make the command go
12382 For more information, see
12384 @ref{Indicating, , Indicating Definitions, texinfo}.
12387 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12388 a Texinfo manual in the current directory. Or, if you are on the
12390 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12393 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12394 Documentation Format}.
12398 @node Regexp Search, Counting Words, Loops & Recursion, Top
12399 @comment node-name, next, previous, up
12400 @chapter Regular Expression Searches
12401 @cindex Searches, illustrating
12402 @cindex Regular expression searches
12403 @cindex Patterns, searching for
12404 @cindex Motion by sentence and paragraph
12405 @cindex Sentences, movement by
12406 @cindex Paragraphs, movement by
12408 Regular expression searches are used extensively in GNU Emacs. The
12409 two functions, @code{forward-sentence} and @code{forward-paragraph},
12410 illustrate these searches well. They use regular expressions to find
12411 where to move point. The phrase `regular expression' is often written
12414 Regular expression searches are described in @ref{Regexp Search, ,
12415 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12416 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12417 Manual}. In writing this chapter, I am presuming that you have at
12418 least a mild acquaintance with them. The major point to remember is
12419 that regular expressions permit you to search for patterns as well as
12420 for literal strings of characters. For example, the code in
12421 @code{forward-sentence} searches for the pattern of possible
12422 characters that could mark the end of a sentence, and moves point to
12425 Before looking at the code for the @code{forward-sentence} function, it
12426 is worth considering what the pattern that marks the end of a sentence
12427 must be. The pattern is discussed in the next section; following that
12428 is a description of the regular expression search function,
12429 @code{re-search-forward}. The @code{forward-sentence} function
12430 is described in the section following. Finally, the
12431 @code{forward-paragraph} function is described in the last section of
12432 this chapter. @code{forward-paragraph} is a complex function that
12433 introduces several new features.
12436 * sentence-end:: The regular expression for @code{sentence-end}.
12437 * re-search-forward:: Very similar to @code{search-forward}.
12438 * forward-sentence:: A straightforward example of regexp search.
12439 * forward-paragraph:: A somewhat complex example.
12440 * etags:: How to create your own @file{TAGS} table.
12442 * re-search Exercises::
12445 @node sentence-end, re-search-forward, Regexp Search, Regexp Search
12446 @comment node-name, next, previous, up
12447 @section The Regular Expression for @code{sentence-end}
12448 @findex sentence-end
12450 The symbol @code{sentence-end} is bound to the pattern that marks the
12451 end of a sentence. What should this regular expression be?
12453 Clearly, a sentence may be ended by a period, a question mark, or an
12454 exclamation mark. Indeed, in English, only clauses that end with one
12455 of those three characters should be considered the end of a sentence.
12456 This means that the pattern should include the character set:
12462 However, we do not want @code{forward-sentence} merely to jump to a
12463 period, a question mark, or an exclamation mark, because such a character
12464 might be used in the middle of a sentence. A period, for example, is
12465 used after abbreviations. So other information is needed.
12467 According to convention, you type two spaces after every sentence, but
12468 only one space after a period, a question mark, or an exclamation mark in
12469 the body of a sentence. So a period, a question mark, or an exclamation
12470 mark followed by two spaces is a good indicator of an end of sentence.
12471 However, in a file, the two spaces may instead be a tab or the end of a
12472 line. This means that the regular expression should include these three
12473 items as alternatives.
12476 This group of alternatives will look like this:
12487 Here, @samp{$} indicates the end of the line, and I have pointed out
12488 where the tab and two spaces are inserted in the expression. Both are
12489 inserted by putting the actual characters into the expression.
12491 Two backslashes, @samp{\\}, are required before the parentheses and
12492 vertical bars: the first backslash quotes the following backslash in
12493 Emacs; and the second indicates that the following character, the
12494 parenthesis or the vertical bar, is special.
12497 Also, a sentence may be followed by one or more carriage returns, like
12508 Like tabs and spaces, a carriage return is inserted into a regular
12509 expression by inserting it literally. The asterisk indicates that the
12510 @key{RET} is repeated zero or more times.
12512 But a sentence end does not consist only of a period, a question mark or
12513 an exclamation mark followed by appropriate space: a closing quotation
12514 mark or a closing brace of some kind may precede the space. Indeed more
12515 than one such mark or brace may precede the space. These require a
12516 expression that looks like this:
12522 In this expression, the first @samp{]} is the first character in the
12523 expression; the second character is @samp{"}, which is preceded by a
12524 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12525 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12527 All this suggests what the regular expression pattern for matching the
12528 end of a sentence should be; and, indeed, if we evaluate
12529 @code{sentence-end} we find that it returns the following value:
12534 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12540 (Well, not in GNU Emacs 22; that is because of an effort to make the
12541 process simpler and to handle more glyphs and languages. When the
12542 value of @code{sentence-end} is @code{nil}, then use the value defined
12543 by the function @code{sentence-end}. (Here is a use of the difference
12544 between a value and a function in Emacs Lisp.) The function returns a
12545 value constructed from the variables @code{sentence-end-base},
12546 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12547 and @code{sentence-end-without-space}. The critical variable is
12548 @code{sentence-end-base}; its global value is similar to the one
12549 described above but it also contains two additional quotation marks.
12550 These have differing degrees of curliness. The
12551 @code{sentence-end-without-period} variable, when true, tells Emacs
12552 that a sentence may end without a period, such as text in Thai.)
12556 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12557 literally in the pattern.)
12559 This regular expression can be deciphered as follows:
12563 The first part of the pattern is the three characters, a period, a question
12564 mark and an exclamation mark, within square brackets. The pattern must
12565 begin with one or other of these characters.
12568 The second part of the pattern is the group of closing braces and
12569 quotation marks, which can appear zero or more times. These may follow
12570 the period, question mark or exclamation mark. In a regular expression,
12571 the backslash, @samp{\}, followed by the double quotation mark,
12572 @samp{"}, indicates the class of string-quote characters. Usually, the
12573 double quotation mark is the only character in this class. The
12574 asterisk, @samp{*}, indicates that the items in the previous group (the
12575 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12578 @item \\($\\| \\| \\)
12579 The third part of the pattern is one or other of: either the end of a
12580 line, or two blank spaces, or a tab. The double back-slashes are used
12581 to prevent Emacs from reading the parentheses and vertical bars as part
12582 of the search pattern; the parentheses are used to mark the group and
12583 the vertical bars are used to indicated that the patterns to either side
12584 of them are alternatives. The dollar sign is used to indicate the end
12585 of a line and both the two spaces and the tab are each inserted as is to
12586 indicate what they are.
12589 Finally, the last part of the pattern indicates that the end of the line
12590 or the whitespace following the period, question mark or exclamation
12591 mark may, but need not, be followed by one or more carriage returns. In
12592 the pattern, the carriage return is inserted as an actual carriage
12593 return between square brackets but here it is shown as @key{RET}.
12597 @node re-search-forward, forward-sentence, sentence-end, Regexp Search
12598 @comment node-name, next, previous, up
12599 @section The @code{re-search-forward} Function
12600 @findex re-search-forward
12602 The @code{re-search-forward} function is very like the
12603 @code{search-forward} function. (@xref{search-forward, , The
12604 @code{search-forward} Function}.)
12606 @code{re-search-forward} searches for a regular expression. If the
12607 search is successful, it leaves point immediately after the last
12608 character in the target. If the search is backwards, it leaves point
12609 just before the first character in the target. You may tell
12610 @code{re-search-forward} to return @code{t} for true. (Moving point
12611 is therefore a `side effect'.)
12613 Like @code{search-forward}, the @code{re-search-forward} function takes
12618 The first argument is the regular expression that the function searches
12619 for. The regular expression will be a string between quotation marks.
12622 The optional second argument limits how far the function will search; it is a
12623 bound, which is specified as a position in the buffer.
12626 The optional third argument specifies how the function responds to
12627 failure: @code{nil} as the third argument causes the function to
12628 signal an error (and print a message) when the search fails; any other
12629 value causes it to return @code{nil} if the search fails and @code{t}
12630 if the search succeeds.
12633 The optional fourth argument is the repeat count. A negative repeat
12634 count causes @code{re-search-forward} to search backwards.
12638 The template for @code{re-search-forward} looks like this:
12642 (re-search-forward "@var{regular-expression}"
12643 @var{limit-of-search}
12644 @var{what-to-do-if-search-fails}
12645 @var{repeat-count})
12649 The second, third, and fourth arguments are optional. However, if you
12650 want to pass a value to either or both of the last two arguments, you
12651 must also pass a value to all the preceding arguments. Otherwise, the
12652 Lisp interpreter will mistake which argument you are passing the value
12656 In the @code{forward-sentence} function, the regular expression will be
12657 the value of the variable @code{sentence-end}. In simple form, that is:
12661 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12667 The limit of the search will be the end of the paragraph (since a
12668 sentence cannot go beyond a paragraph). If the search fails, the
12669 function will return @code{nil}; and the repeat count will be provided
12670 by the argument to the @code{forward-sentence} function.
12672 @node forward-sentence, forward-paragraph, re-search-forward, Regexp Search
12673 @comment node-name, next, previous, up
12674 @section @code{forward-sentence}
12675 @findex forward-sentence
12677 The command to move the cursor forward a sentence is a straightforward
12678 illustration of how to use regular expression searches in Emacs Lisp.
12679 Indeed, the function looks longer and more complicated than it is; this
12680 is because the function is designed to go backwards as well as forwards;
12681 and, optionally, over more than one sentence. The function is usually
12682 bound to the key command @kbd{M-e}.
12685 * Complete forward-sentence::
12686 * fwd-sentence while loops:: Two @code{while} loops.
12687 * fwd-sentence re-search:: A regular expression search.
12690 @node Complete forward-sentence, fwd-sentence while loops, forward-sentence, forward-sentence
12692 @unnumberedsubsec Complete @code{forward-sentence} function definition
12696 Here is the code for @code{forward-sentence}:
12701 (defun forward-sentence (&optional arg)
12702 "Move forward to next `sentence-end'. With argument, repeat.
12703 With negative argument, move backward repeatedly to `sentence-beginning'.
12705 The variable `sentence-end' is a regular expression that matches ends of
12706 sentences. Also, every paragraph boundary terminates sentences as well."
12710 (or arg (setq arg 1))
12711 (let ((opoint (point))
12712 (sentence-end (sentence-end)))
12714 (let ((pos (point))
12715 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12716 (if (and (re-search-backward sentence-end par-beg t)
12717 (or (< (match-end 0) pos)
12718 (re-search-backward sentence-end par-beg t)))
12719 (goto-char (match-end 0))
12720 (goto-char par-beg)))
12721 (setq arg (1+ arg)))
12725 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12726 (if (re-search-forward sentence-end par-end t)
12727 (skip-chars-backward " \t\n")
12728 (goto-char par-end)))
12729 (setq arg (1- arg)))
12730 (constrain-to-field nil opoint t)))
12738 (defun forward-sentence (&optional arg)
12739 "Move forward to next sentence-end. With argument, repeat.
12740 With negative argument, move backward repeatedly to sentence-beginning.
12741 Sentence ends are identified by the value of sentence-end
12742 treated as a regular expression. Also, every paragraph boundary
12743 terminates sentences as well."
12747 (or arg (setq arg 1))
12750 (save-excursion (start-of-paragraph-text) (point))))
12751 (if (re-search-backward
12752 (concat sentence-end "[^ \t\n]") par-beg t)
12753 (goto-char (1- (match-end 0)))
12754 (goto-char par-beg)))
12755 (setq arg (1+ arg)))
12758 (save-excursion (end-of-paragraph-text) (point))))
12759 (if (re-search-forward sentence-end par-end t)
12760 (skip-chars-backward " \t\n")
12761 (goto-char par-end)))
12762 (setq arg (1- arg))))
12767 The function looks long at first sight and it is best to look at its
12768 skeleton first, and then its muscle. The way to see the skeleton is to
12769 look at the expressions that start in the left-most columns:
12773 (defun forward-sentence (&optional arg)
12774 "@var{documentation}@dots{}"
12776 (or arg (setq arg 1))
12777 (let ((opoint (point)) (sentence-end (sentence-end)))
12779 (let ((pos (point))
12780 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12781 @var{rest-of-body-of-while-loop-when-going-backwards}
12783 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12784 @var{rest-of-body-of-while-loop-when-going-forwards}
12785 @var{handle-forms-and-equivalent}
12789 This looks much simpler! The function definition consists of
12790 documentation, an @code{interactive} expression, an @code{or}
12791 expression, a @code{let} expression, and @code{while} loops.
12793 Let's look at each of these parts in turn.
12795 We note that the documentation is thorough and understandable.
12797 The function has an @code{interactive "p"} declaration. This means
12798 that the processed prefix argument, if any, is passed to the
12799 function as its argument. (This will be a number.) If the function
12800 is not passed an argument (it is optional) then the argument
12801 @code{arg} will be bound to 1.
12803 When @code{forward-sentence} is called non-interactively without an
12804 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12805 handles this. What it does is either leave the value of @code{arg} as
12806 it is, but only if @code{arg} is bound to a value; or it sets the
12807 value of @code{arg} to 1, in the case when @code{arg} is bound to
12810 Next is a @code{let}. That specifies the values of two local
12811 variables, @code{point} and @code{sentence-end}. The local value of
12812 point, from before the search, is used in the
12813 @code{constrain-to-field} function which handles forms and
12814 equivalents. The @code{sentence-end} variable is set by the
12815 @code{sentence-end} function.
12817 @node fwd-sentence while loops, fwd-sentence re-search, Complete forward-sentence, forward-sentence
12818 @unnumberedsubsec The @code{while} loops
12820 Two @code{while} loops follow. The first @code{while} has a
12821 true-or-false-test that tests true if the prefix argument for
12822 @code{forward-sentence} is a negative number. This is for going
12823 backwards. The body of this loop is similar to the body of the second
12824 @code{while} clause, but it is not exactly the same. We will skip
12825 this @code{while} loop and concentrate on the second @code{while}
12829 The second @code{while} loop is for moving point forward. Its skeleton
12834 (while (> arg 0) ; @r{true-or-false-test}
12836 (if (@var{true-or-false-test})
12839 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12843 The @code{while} loop is of the decrementing kind.
12844 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12845 has a true-or-false-test that tests true so long as the counter (in
12846 this case, the variable @code{arg}) is greater than zero; and it has a
12847 decrementer that subtracts 1 from the value of the counter every time
12850 If no prefix argument is given to @code{forward-sentence}, which is
12851 the most common way the command is used, this @code{while} loop will
12852 run once, since the value of @code{arg} will be 1.
12854 The body of the @code{while} loop consists of a @code{let} expression,
12855 which creates and binds a local variable, and has, as its body, an
12856 @code{if} expression.
12859 The body of the @code{while} loop looks like this:
12864 (save-excursion (end-of-paragraph-text) (point))))
12865 (if (re-search-forward sentence-end par-end t)
12866 (skip-chars-backward " \t\n")
12867 (goto-char par-end)))
12871 The @code{let} expression creates and binds the local variable
12872 @code{par-end}. As we shall see, this local variable is designed to
12873 provide a bound or limit to the regular expression search. If the
12874 search fails to find a proper sentence ending in the paragraph, it will
12875 stop on reaching the end of the paragraph.
12877 But first, let us examine how @code{par-end} is bound to the value of
12878 the end of the paragraph. What happens is that the @code{let} sets the
12879 value of @code{par-end} to the value returned when the Lisp interpreter
12880 evaluates the expression
12884 (save-excursion (end-of-paragraph-text) (point))
12889 In this expression, @code{(end-of-paragraph-text)} moves point to the
12890 end of the paragraph, @code{(point)} returns the value of point, and then
12891 @code{save-excursion} restores point to its original position. Thus,
12892 the @code{let} binds @code{par-end} to the value returned by the
12893 @code{save-excursion} expression, which is the position of the end of
12894 the paragraph. (The @code{end-of-paragraph-text} function uses
12895 @code{forward-paragraph}, which we will discuss shortly.)
12898 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12899 expression that looks like this:
12903 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12904 (skip-chars-backward " \t\n") ; @r{then-part}
12905 (goto-char par-end))) ; @r{else-part}
12909 The @code{if} tests whether its first argument is true and if so,
12910 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12911 evaluates the else-part. The true-or-false-test of the @code{if}
12912 expression is the regular expression search.
12914 It may seem odd to have what looks like the `real work' of
12915 the @code{forward-sentence} function buried here, but this is a common
12916 way this kind of operation is carried out in Lisp.
12918 @node fwd-sentence re-search, , fwd-sentence while loops, forward-sentence
12919 @unnumberedsubsec The regular expression search
12921 The @code{re-search-forward} function searches for the end of the
12922 sentence, that is, for the pattern defined by the @code{sentence-end}
12923 regular expression. If the pattern is found---if the end of the sentence is
12924 found---then the @code{re-search-forward} function does two things:
12928 The @code{re-search-forward} function carries out a side effect, which
12929 is to move point to the end of the occurrence found.
12932 The @code{re-search-forward} function returns a value of true. This is
12933 the value received by the @code{if}, and means that the search was
12938 The side effect, the movement of point, is completed before the
12939 @code{if} function is handed the value returned by the successful
12940 conclusion of the search.
12942 When the @code{if} function receives the value of true from a successful
12943 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12944 which is the expression @code{(skip-chars-backward " \t\n")}. This
12945 expression moves backwards over any blank spaces, tabs or carriage
12946 returns until a printed character is found and then leaves point after
12947 the character. Since point has already been moved to the end of the
12948 pattern that marks the end of the sentence, this action leaves point
12949 right after the closing printed character of the sentence, which is
12952 On the other hand, if the @code{re-search-forward} function fails to
12953 find a pattern marking the end of the sentence, the function returns
12954 false. The false then causes the @code{if} to evaluate its third
12955 argument, which is @code{(goto-char par-end)}: it moves point to the
12956 end of the paragraph.
12958 (And if the text is in a form or equivalent, and point may not move
12959 fully, then the @code{constrain-to-field} function comes into play.)
12961 Regular expression searches are exceptionally useful and the pattern
12962 illustrated by @code{re-search-forward}, in which the search is the
12963 test of an @code{if} expression, is handy. You will see or write code
12964 incorporating this pattern often.
12966 @node forward-paragraph, etags, forward-sentence, Regexp Search
12967 @comment node-name, next, previous, up
12968 @section @code{forward-paragraph}: a Goldmine of Functions
12969 @findex forward-paragraph
12973 (defun forward-paragraph (&optional arg)
12974 "Move forward to end of paragraph.
12975 With argument ARG, do it ARG times;
12976 a negative argument ARG = -N means move backward N paragraphs.
12978 A line which `paragraph-start' matches either separates paragraphs
12979 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12980 A paragraph end is the beginning of a line which is not part of the paragraph
12981 to which the end of the previous line belongs, or the end of the buffer.
12982 Returns the count of paragraphs left to move."
12984 (or arg (setq arg 1))
12985 (let* ((opoint (point))
12986 (fill-prefix-regexp
12987 (and fill-prefix (not (equal fill-prefix ""))
12988 (not paragraph-ignore-fill-prefix)
12989 (regexp-quote fill-prefix)))
12990 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12991 ;; These regexps shouldn't be anchored, because we look for them
12992 ;; starting at the left-margin. This allows paragraph commands to
12993 ;; work normally with indented text.
12994 ;; This hack will not find problem cases like "whatever\\|^something".
12995 (parstart (if (and (not (equal "" paragraph-start))
12996 (equal ?^ (aref paragraph-start 0)))
12997 (substring paragraph-start 1)
12999 (parsep (if (and (not (equal "" paragraph-separate))
13000 (equal ?^ (aref paragraph-separate 0)))
13001 (substring paragraph-separate 1)
13002 paragraph-separate))
13004 (if fill-prefix-regexp
13005 (concat parsep "\\|"
13006 fill-prefix-regexp "[ \t]*$")
13008 ;; This is used for searching.
13009 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
13011 (while (and (< arg 0) (not (bobp)))
13012 (if (and (not (looking-at parsep))
13013 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
13014 (looking-at parsep))
13015 (setq arg (1+ arg))
13016 (setq start (point))
13017 ;; Move back over paragraph-separating lines.
13018 (forward-char -1) (beginning-of-line)
13019 (while (and (not (bobp))
13020 (progn (move-to-left-margin)
13021 (looking-at parsep)))
13025 (setq arg (1+ arg))
13026 ;; Go to end of the previous (non-separating) line.
13028 ;; Search back for line that starts or separates paragraphs.
13029 (if (if fill-prefix-regexp
13030 ;; There is a fill prefix; it overrides parstart.
13031 (let (multiple-lines)
13032 (while (and (progn (beginning-of-line) (not (bobp)))
13033 (progn (move-to-left-margin)
13034 (not (looking-at parsep)))
13035 (looking-at fill-prefix-regexp))
13036 (unless (= (point) start)
13037 (setq multiple-lines t))
13039 (move-to-left-margin)
13040 ;; This deleted code caused a long hanging-indent line
13041 ;; not to be filled together with the following lines.
13042 ;; ;; Don't move back over a line before the paragraph
13043 ;; ;; which doesn't start with fill-prefix
13044 ;; ;; unless that is the only line we've moved over.
13045 ;; (and (not (looking-at fill-prefix-regexp))
13047 ;; (forward-line 1))
13049 (while (and (re-search-backward sp-parstart nil 1)
13050 (setq found-start t)
13051 ;; Found a candidate, but need to check if it is a
13053 (progn (setq start (point))
13054 (move-to-left-margin)
13055 (not (looking-at parsep)))
13056 (not (and (looking-at parstart)
13057 (or (not use-hard-newlines)
13060 (1- start) 'hard)))))
13061 (setq found-start nil)
13066 ;; Move forward over paragraph separators.
13067 ;; We know this cannot reach the place we started
13068 ;; because we know we moved back over a non-separator.
13069 (while (and (not (eobp))
13070 (progn (move-to-left-margin)
13071 (looking-at parsep)))
13073 ;; If line before paragraph is just margin, back up to there.
13075 (if (> (current-column) (current-left-margin))
13077 (skip-chars-backward " \t")
13079 (forward-line 1))))
13080 ;; No starter or separator line => use buffer beg.
13081 (goto-char (point-min))))))
13083 (while (and (> arg 0) (not (eobp)))
13084 ;; Move forward over separator lines...
13085 (while (and (not (eobp))
13086 (progn (move-to-left-margin) (not (eobp)))
13087 (looking-at parsep))
13089 (unless (eobp) (setq arg (1- arg)))
13090 ;; ... and one more line.
13092 (if fill-prefix-regexp
13093 ;; There is a fill prefix; it overrides parstart.
13094 (while (and (not (eobp))
13095 (progn (move-to-left-margin) (not (eobp)))
13096 (not (looking-at parsep))
13097 (looking-at fill-prefix-regexp))
13099 (while (and (re-search-forward sp-parstart nil 1)
13100 (progn (setq start (match-beginning 0))
13103 (progn (move-to-left-margin)
13104 (not (looking-at parsep)))
13105 (or (not (looking-at parstart))
13106 (and use-hard-newlines
13107 (not (get-text-property (1- start) 'hard)))))
13109 (if (< (point) (point-max))
13110 (goto-char start))))
13111 (constrain-to-field nil opoint t)
13112 ;; Return the number of steps that could not be done.
13116 The @code{forward-paragraph} function moves point forward to the end
13117 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
13118 number of functions that are important in themselves, including
13119 @code{let*}, @code{match-beginning}, and @code{looking-at}.
13121 The function definition for @code{forward-paragraph} is considerably
13122 longer than the function definition for @code{forward-sentence}
13123 because it works with a paragraph, each line of which may begin with a
13126 A fill prefix consists of a string of characters that are repeated at
13127 the beginning of each line. For example, in Lisp code, it is a
13128 convention to start each line of a paragraph-long comment with
13129 @samp{;;; }. In Text mode, four blank spaces make up another common
13130 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
13131 emacs, The GNU Emacs Manual}, for more information about fill
13134 The existence of a fill prefix means that in addition to being able to
13135 find the end of a paragraph whose lines begin on the left-most
13136 column, the @code{forward-paragraph} function must be able to find the
13137 end of a paragraph when all or many of the lines in the buffer begin
13138 with the fill prefix.
13140 Moreover, it is sometimes practical to ignore a fill prefix that
13141 exists, especially when blank lines separate paragraphs.
13142 This is an added complication.
13145 * forward-paragraph in brief:: Key parts of the function definition.
13146 * fwd-para let:: The @code{let*} expression.
13147 * fwd-para while:: The forward motion @code{while} loop.
13150 @node forward-paragraph in brief, fwd-para let, forward-paragraph, forward-paragraph
13152 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13155 Rather than print all of the @code{forward-paragraph} function, we
13156 will only print parts of it. Read without preparation, the function
13160 In outline, the function looks like this:
13164 (defun forward-paragraph (&optional arg)
13165 "@var{documentation}@dots{}"
13167 (or arg (setq arg 1))
13170 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13172 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13177 The first parts of the function are routine: the function's argument
13178 list consists of one optional argument. Documentation follows.
13180 The lower case @samp{p} in the @code{interactive} declaration means
13181 that the processed prefix argument, if any, is passed to the function.
13182 This will be a number, and is the repeat count of how many paragraphs
13183 point will move. The @code{or} expression in the next line handles
13184 the common case when no argument is passed to the function, which occurs
13185 if the function is called from other code rather than interactively.
13186 This case was described earlier. (@xref{forward-sentence, The
13187 @code{forward-sentence} function}.) Now we reach the end of the
13188 familiar part of this function.
13190 @node fwd-para let, fwd-para while, forward-paragraph in brief, forward-paragraph
13191 @unnumberedsubsec The @code{let*} expression
13193 The next line of the @code{forward-paragraph} function begins a
13194 @code{let*} expression. This is a different than @code{let}. The
13195 symbol is @code{let*} not @code{let}.
13197 The @code{let*} special form is like @code{let} except that Emacs sets
13198 each variable in sequence, one after another, and variables in the
13199 latter part of the varlist can make use of the values to which Emacs
13200 set variables in the earlier part of the varlist.
13203 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13206 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13208 In the @code{let*} expression in this function, Emacs binds a total of
13209 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13210 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13211 @code{found-start}.
13213 The variable @code{parsep} appears twice, first, to remove instances
13214 of @samp{^}, and second, to handle fill prefixes.
13216 The variable @code{opoint} is just the value of @code{point}. As you
13217 can guess, it is used in a @code{constrain-to-field} expression, just
13218 as in @code{forward-sentence}.
13220 The variable @code{fill-prefix-regexp} is set to the value returned by
13221 evaluating the following list:
13226 (not (equal fill-prefix ""))
13227 (not paragraph-ignore-fill-prefix)
13228 (regexp-quote fill-prefix))
13233 This is an expression whose first element is the @code{and} special form.
13235 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13236 function}), the @code{and} special form evaluates each of its
13237 arguments until one of the arguments returns a value of @code{nil}, in
13238 which case the @code{and} expression returns @code{nil}; however, if
13239 none of the arguments returns a value of @code{nil}, the value
13240 resulting from evaluating the last argument is returned. (Since such
13241 a value is not @code{nil}, it is considered true in Lisp.) In other
13242 words, an @code{and} expression returns a true value only if all its
13243 arguments are true.
13246 In this case, the variable @code{fill-prefix-regexp} is bound to a
13247 non-@code{nil} value only if the following four expressions produce a
13248 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13249 @code{fill-prefix-regexp} is bound to @code{nil}.
13253 When this variable is evaluated, the value of the fill prefix, if any,
13254 is returned. If there is no fill prefix, this variable returns
13257 @item (not (equal fill-prefix "")
13258 This expression checks whether an existing fill prefix is an empty
13259 string, that is, a string with no characters in it. An empty string is
13260 not a useful fill prefix.
13262 @item (not paragraph-ignore-fill-prefix)
13263 This expression returns @code{nil} if the variable
13264 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13265 true value such as @code{t}.
13267 @item (regexp-quote fill-prefix)
13268 This is the last argument to the @code{and} special form. If all the
13269 arguments to the @code{and} are true, the value resulting from
13270 evaluating this expression will be returned by the @code{and} expression
13271 and bound to the variable @code{fill-prefix-regexp},
13274 @findex regexp-quote
13276 The result of evaluating this @code{and} expression successfully is that
13277 @code{fill-prefix-regexp} will be bound to the value of
13278 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13279 What @code{regexp-quote} does is read a string and return a regular
13280 expression that will exactly match the string and match nothing else.
13281 This means that @code{fill-prefix-regexp} will be set to a value that
13282 will exactly match the fill prefix if the fill prefix exists.
13283 Otherwise, the variable will be set to @code{nil}.
13285 The next two local variables in the @code{let*} expression are
13286 designed to remove instances of @samp{^} from @code{parstart} and
13287 @code{parsep}, the local variables which indicate the paragraph start
13288 and the paragraph separator. The next expression sets @code{parsep}
13289 again. That is to handle fill prefixes.
13291 This is the setting that requires the definition call @code{let*}
13292 rather than @code{let}. The true-or-false-test for the @code{if}
13293 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13294 @code{nil} or some other value.
13296 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13297 the else-part of the @code{if} expression and binds @code{parsep} to
13298 its local value. (@code{parsep} is a regular expression that matches
13299 what separates paragraphs.)
13301 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13302 the then-part of the @code{if} expression and binds @code{parsep} to a
13303 regular expression that includes the @code{fill-prefix-regexp} as part
13306 Specifically, @code{parsep} is set to the original value of the
13307 paragraph separate regular expression concatenated with an alternative
13308 expression that consists of the @code{fill-prefix-regexp} followed by
13309 optional whitespace to the end of the line. The whitespace is defined
13310 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13311 regexp as an alternative to @code{parsep}.
13313 According to a comment in the code, the next local variable,
13314 @code{sp-parstart}, is used for searching, and then the final two,
13315 @code{start} and @code{found-start}, are set to @code{nil}.
13317 Now we get into the body of the @code{let*}. The first part of the body
13318 of the @code{let*} deals with the case when the function is given a
13319 negative argument and is therefore moving backwards. We will skip this
13322 @node fwd-para while, , fwd-para let, forward-paragraph
13323 @unnumberedsubsec The forward motion @code{while} loop
13325 The second part of the body of the @code{let*} deals with forward
13326 motion. It is a @code{while} loop that repeats itself so long as the
13327 value of @code{arg} is greater than zero. In the most common use of
13328 the function, the value of the argument is 1, so the body of the
13329 @code{while} loop is evaluated exactly once, and the cursor moves
13330 forward one paragraph.
13333 (while (and (> arg 0) (not (eobp)))
13335 ;; Move forward over separator lines...
13336 (while (and (not (eobp))
13337 (progn (move-to-left-margin) (not (eobp)))
13338 (looking-at parsep))
13340 (unless (eobp) (setq arg (1- arg)))
13341 ;; ... and one more line.
13344 (if fill-prefix-regexp
13345 ;; There is a fill prefix; it overrides parstart.
13346 (while (and (not (eobp))
13347 (progn (move-to-left-margin) (not (eobp)))
13348 (not (looking-at parsep))
13349 (looking-at fill-prefix-regexp))
13352 (while (and (re-search-forward sp-parstart nil 1)
13353 (progn (setq start (match-beginning 0))
13356 (progn (move-to-left-margin)
13357 (not (looking-at parsep)))
13358 (or (not (looking-at parstart))
13359 (and use-hard-newlines
13360 (not (get-text-property (1- start) 'hard)))))
13363 (if (< (point) (point-max))
13364 (goto-char start))))
13367 This part handles three situations: when point is between paragraphs,
13368 when there is a fill prefix and when there is no fill prefix.
13371 The @code{while} loop looks like this:
13375 ;; @r{going forwards and not at the end of the buffer}
13376 (while (and (> arg 0) (not (eobp)))
13378 ;; @r{between paragraphs}
13379 ;; Move forward over separator lines...
13380 (while (and (not (eobp))
13381 (progn (move-to-left-margin) (not (eobp)))
13382 (looking-at parsep))
13384 ;; @r{This decrements the loop}
13385 (unless (eobp) (setq arg (1- arg)))
13386 ;; ... and one more line.
13391 (if fill-prefix-regexp
13392 ;; There is a fill prefix; it overrides parstart;
13393 ;; we go forward line by line
13394 (while (and (not (eobp))
13395 (progn (move-to-left-margin) (not (eobp)))
13396 (not (looking-at parsep))
13397 (looking-at fill-prefix-regexp))
13402 ;; There is no fill prefix;
13403 ;; we go forward character by character
13404 (while (and (re-search-forward sp-parstart nil 1)
13405 (progn (setq start (match-beginning 0))
13408 (progn (move-to-left-margin)
13409 (not (looking-at parsep)))
13410 (or (not (looking-at parstart))
13411 (and use-hard-newlines
13412 (not (get-text-property (1- start) 'hard)))))
13417 ;; and if there is no fill prefix and if we are not at the end,
13418 ;; go to whatever was found in the regular expression search
13420 (if (< (point) (point-max))
13421 (goto-char start))))
13426 We can see that this is a decrementing counter @code{while} loop,
13427 using the expression @code{(setq arg (1- arg))} as the decrementer.
13428 That expression is not far from the @code{while}, but is hidden in
13429 another Lisp macro, an @code{unless} macro. Unless we are at the end
13430 of the buffer --- that is what the @code{eobp} function determines; it
13431 is an abbreviation of @samp{End Of Buffer P} --- we decrease the value
13432 of @code{arg} by one.
13434 (If we are at the end of the buffer, we cannot go forward any more and
13435 the next loop of the @code{while} expression will test false since the
13436 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13437 function means exactly as you expect; it is another name for
13438 @code{null}, a function that returns true when its argument is false.)
13440 Interestingly, the loop count is not decremented until we leave the
13441 space between paragraphs, unless we come to the end of buffer or stop
13442 seeing the local value of the paragraph separator.
13444 That second @code{while} also has a @code{(move-to-left-margin)}
13445 expression. The function is self-explanatory. It is inside a
13446 @code{progn} expression and not the last element of its body, so it is
13447 only invoked for its side effect, which is to move point to the left
13448 margin of the current line.
13451 The @code{looking-at} function is also self-explanatory; it returns
13452 true if the text after point matches the regular expression given as
13455 The rest of the body of the loop looks difficult at first, but makes
13456 sense as you come to understand it.
13459 First consider what happens if there is a fill prefix:
13463 (if fill-prefix-regexp
13464 ;; There is a fill prefix; it overrides parstart;
13465 ;; we go forward line by line
13466 (while (and (not (eobp))
13467 (progn (move-to-left-margin) (not (eobp)))
13468 (not (looking-at parsep))
13469 (looking-at fill-prefix-regexp))
13475 This expression moves point forward line by line so long
13476 as four conditions are true:
13480 Point is not at the end of the buffer.
13483 We can move to the left margin of the text and are
13484 not at the end of the buffer.
13487 The text following point does not separate paragraphs.
13490 The pattern following point is the fill prefix regular expression.
13493 The last condition may be puzzling, until you remember that point was
13494 moved to the beginning of the line early in the @code{forward-paragraph}
13495 function. This means that if the text has a fill prefix, the
13496 @code{looking-at} function will see it.
13499 Consider what happens when there is no fill prefix.
13503 (while (and (re-search-forward sp-parstart nil 1)
13504 (progn (setq start (match-beginning 0))
13507 (progn (move-to-left-margin)
13508 (not (looking-at parsep)))
13509 (or (not (looking-at parstart))
13510 (and use-hard-newlines
13511 (not (get-text-property (1- start) 'hard)))))
13517 This @code{while} loop has us searching forward for
13518 @code{sp-parstart}, which is the combination of possible whitespace
13519 with a the local value of the start of a paragraph or of a paragraph
13520 separator. (The latter two are within an expression starting
13521 @code{\(?:} so that they are not referenced by the
13522 @code{match-beginning} function.)
13525 The two expressions,
13529 (setq start (match-beginning 0))
13535 mean go to the start of the text matched by the regular expression
13538 The @code{(match-beginning 0)} expression is new. It returns a number
13539 specifying the location of the start of the text that was matched by
13542 The @code{match-beginning} function is used here because of a
13543 characteristic of a forward search: a successful forward search,
13544 regardless of whether it is a plain search or a regular expression
13545 search, moves point to the end of the text that is found. In this
13546 case, a successful search moves point to the end of the pattern for
13547 @code{sp-parstart}.
13549 However, we want to put point at the end of the current paragraph, not
13550 somewhere else. Indeed, since the search possibly includes the
13551 paragraph separator, point may end up at the beginning of the next one
13552 unless we use an expression that includes @code{match-beginning}.
13554 @findex match-beginning
13555 When given an argument of 0, @code{match-beginning} returns the
13556 position that is the start of the text matched by the most recent
13557 search. In this case, the most recent search looks for
13558 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13559 the beginning position of that pattern, rather than the end position
13562 (Incidentally, when passed a positive number as an argument, the
13563 @code{match-beginning} function returns the location of point at that
13564 parenthesized expression in the last search unless that parenthesized
13565 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13566 appears here since the argument is 0.)
13569 The last expression when there is no fill prefix is
13573 (if (< (point) (point-max))
13574 (goto-char start))))
13579 This says that if there is no fill prefix and if we are not at the
13580 end, point should move to the beginning of whatever was found by the
13581 regular expression search for @code{sp-parstart}.
13583 The full definition for the @code{forward-paragraph} function not only
13584 includes code for going forwards, but also code for going backwards.
13586 If you are reading this inside of GNU Emacs and you want to see the
13587 whole function, you can type @kbd{C-h f} (@code{describe-function})
13588 and the name of the function. This gives you the function
13589 documentation and the name of the library containing the function's
13590 source. Place point over the name of the library and press the RET
13591 key; you will be taken directly to the source. (Be sure to install
13592 your sources! Without them, you are like a person who tries to drive
13593 a car with his eyes shut!)
13595 @node etags, Regexp Review, forward-paragraph, Regexp Search
13596 @section Create Your Own @file{TAGS} File
13598 @cindex @file{TAGS} file, create own
13600 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13601 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13602 name of the function when prompted for it. This is a good habit to
13603 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13604 to the source for a function, variable, or node. The function depends
13605 on tags tables to tell it where to go.
13607 If the @code{find-tag} function first asks you for the name of a
13608 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13609 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13610 @file{TAGS} file depends on how your copy of Emacs was installed. I
13611 just told you the location that provides both my C and my Emacs Lisp
13614 You can also create your own @file{TAGS} file for directories that
13617 You often need to build and install tags tables yourself. They are
13618 not built automatically. A tags table is called a @file{TAGS} file;
13619 the name is in upper case letters.
13621 You can create a @file{TAGS} file by calling the @code{etags} program
13622 that comes as a part of the Emacs distribution. Usually, @code{etags}
13623 is compiled and installed when Emacs is built. (@code{etags} is not
13624 an Emacs Lisp function or a part of Emacs; it is a C program.)
13627 To create a @file{TAGS} file, first switch to the directory in which
13628 you want to create the file. In Emacs you can do this with the
13629 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13630 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13631 compile command, with @w{@code{etags *.el}} as the command to execute
13634 M-x compile RET etags *.el RET
13638 to create a @file{TAGS} file for Emacs Lisp.
13640 For example, if you have a large number of files in your
13641 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13642 of which I load 12---you can create a @file{TAGS} file for the Emacs
13643 Lisp files in that directory.
13646 The @code{etags} program takes all the usual shell `wildcards'. For
13647 example, if you have two directories for which you want a single
13648 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13649 @file{../elisp/} is the second directory:
13652 M-x compile RET etags *.el ../elisp/*.el RET
13659 M-x compile RET etags --help RET
13663 to see a list of the options accepted by @code{etags} as well as a
13664 list of supported languages.
13666 The @code{etags} program handles more than 20 languages, including
13667 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13668 LaTeX, Pascal, Perl, PostScript, Python, TeX, Texinfo, makefiles, and
13669 most assemblers. The program has no switches for specifying the
13670 language; it recognizes the language in an input file according to its
13671 file name and contents.
13673 @file{etags} is very helpful when you are writing code yourself and
13674 want to refer back to functions you have already written. Just run
13675 @code{etags} again at intervals as you write new functions, so they
13676 become part of the @file{TAGS} file.
13678 If you think an appropriate @file{TAGS} file already exists for what
13679 you want, but do not know where it is, you can use the @code{locate}
13680 program to attempt to find it.
13682 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13683 for you the full path names of all your @file{TAGS} files. On my
13684 system, this command lists 34 @file{TAGS} files. On the other hand, a
13685 `plain vanilla' system I recently installed did not contain any
13688 If the tags table you want has been created, you can use the @code{M-x
13689 visit-tags-table} command to specify it. Otherwise, you will need to
13690 create the tag table yourself and then use @code{M-x
13693 @subsubheading Building Tags in the Emacs sources
13694 @cindex Building Tags in the Emacs sources
13695 @cindex Tags in the Emacs sources
13698 The GNU Emacs sources come with a @file{Makefile} that contains a
13699 sophisticated @code{etags} command that creates, collects, and merges
13700 tags tables from all over the Emacs sources and puts the information
13701 into one @file{TAGS} file in the @file{src/} directory. (The
13702 @file{src/} directory is below the top level of your Emacs directory.)
13705 To build this @file{TAGS} file, go to the top level of your Emacs
13706 source directory and run the compile command @code{make tags}:
13709 M-x compile RET make tags RET
13713 (The @code{make tags} command works well with the GNU Emacs sources,
13714 as well as with some other source packages.)
13716 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13719 @node Regexp Review, re-search Exercises, etags, Regexp Search
13720 @comment node-name, next, previous, up
13723 Here is a brief summary of some recently introduced functions.
13727 Repeatedly evaluate the body of the expression so long as the first
13728 element of the body tests true. Then return @code{nil}. (The
13729 expression is evaluated only for its side effects.)
13738 (insert (format "foo is %d.\n" foo))
13739 (setq foo (1- foo))))
13741 @result{} foo is 2.
13748 (The @code{insert} function inserts its arguments at point; the
13749 @code{format} function returns a string formatted from its arguments
13750 the way @code{message} formats its arguments; @code{\n} produces a new
13753 @item re-search-forward
13754 Search for a pattern, and if the pattern is found, move point to rest
13758 Takes four arguments, like @code{search-forward}:
13762 A regular expression that specifies the pattern to search for.
13763 (Remember to put quotation marks around this argument!)
13766 Optionally, the limit of the search.
13769 Optionally, what to do if the search fails, return @code{nil} or an
13773 Optionally, how many times to repeat the search; if negative, the
13774 search goes backwards.
13778 Bind some variables locally to particular values,
13779 and then evaluate the remaining arguments, returning the value of the
13780 last one. While binding the local variables, use the local values of
13781 variables bound earlier, if any.
13790 (message "`bar' is %d." bar))
13791 @result{} `bar' is 21.
13795 @item match-beginning
13796 Return the position of the start of the text found by the last regular
13800 Return @code{t} for true if the text after point matches the argument,
13801 which should be a regular expression.
13804 Return @code{t} for true if point is at the end of the accessible part
13805 of a buffer. The end of the accessible part is the end of the buffer
13806 if the buffer is not narrowed; it is the end of the narrowed part if
13807 the buffer is narrowed.
13811 @node re-search Exercises, , Regexp Review, Regexp Search
13812 @section Exercises with @code{re-search-forward}
13816 Write a function to search for a regular expression that matches two
13817 or more blank lines in sequence.
13820 Write a function to search for duplicated words, such as `the the'.
13821 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13822 Manual}, for information on how to write a regexp (a regular
13823 expression) to match a string that is composed of two identical
13824 halves. You can devise several regexps; some are better than others.
13825 The function I use is described in an appendix, along with several
13826 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13829 @node Counting Words, Words in a defun, Regexp Search, Top
13830 @chapter Counting: Repetition and Regexps
13831 @cindex Repetition for word counting
13832 @cindex Regular expressions for word counting
13834 Repetition and regular expression searches are powerful tools that you
13835 often use when you write code in Emacs Lisp. This chapter illustrates
13836 the use of regular expression searches through the construction of
13837 word count commands using @code{while} loops and recursion.
13840 * Why Count Words::
13841 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
13842 * recursive-count-words:: Start with case of no words in region.
13843 * Counting Exercise::
13846 @node Why Count Words, @value{COUNT-WORDS}, Counting Words, Counting Words
13848 @unnumberedsec Counting words
13851 The standard Emacs distribution contains functions for counting the
13852 number of lines and words within a region.
13854 Certain types of writing ask you to count words. Thus, if you write
13855 an essay, you may be limited to 800 words; if you write a novel, you
13856 may discipline yourself to write 1000 words a day. It seems odd, but
13857 for a long time, Emacs lacked a word count command. Perhaps people used
13858 Emacs mostly for code or types of documentation that did not require
13859 word counts; or perhaps they restricted themselves to the operating
13860 system word count command, @code{wc}. Alternatively, people may have
13861 followed the publishers' convention and computed a word count by
13862 dividing the number of characters in a document by five.
13864 There are many ways to implement a command to count words. Here are
13865 some examples, which you may wish to compare with the standard Emacs
13866 command, @code{count-words-region}.
13868 @node @value{COUNT-WORDS}, recursive-count-words, Why Count Words, Counting Words
13869 @comment node-name, next, previous, up
13870 @section The @code{@value{COUNT-WORDS}} Function
13871 @findex @value{COUNT-WORDS}
13873 A word count command could count words in a line, paragraph, region,
13874 or buffer. What should the command cover? You could design the
13875 command to count the number of words in a complete buffer. However,
13876 the Emacs tradition encourages flexibility---you may want to count
13877 words in just a section, rather than all of a buffer. So it makes
13878 more sense to design the command to count the number of words in a
13879 region. Once you have a command to count words in a region, you can,
13880 if you wish, count words in a whole buffer by marking it with
13881 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13883 Clearly, counting words is a repetitive act: starting from the
13884 beginning of the region, you count the first word, then the second
13885 word, then the third word, and so on, until you reach the end of the
13886 region. This means that word counting is ideally suited to recursion
13887 or to a @code{while} loop.
13890 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
13891 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
13894 @node Design @value{COUNT-WORDS}, Whitespace Bug, @value{COUNT-WORDS}, @value{COUNT-WORDS}
13896 @unnumberedsubsec Designing @code{@value{COUNT-WORDS}}
13899 First, we will implement the word count command with a @code{while}
13900 loop, then with recursion. The command will, of course, be
13904 The template for an interactive function definition is, as always:
13908 (defun @var{name-of-function} (@var{argument-list})
13909 "@var{documentation}@dots{}"
13910 (@var{interactive-expression}@dots{})
13915 What we need to do is fill in the slots.
13917 The name of the function should be self-explanatory and similar to the
13918 existing @code{count-lines-region} name. This makes the name easier
13919 to remember. @code{count-words-region} is the obvious choice. Since
13920 that name is now used for the standard Emacs command to count words, we
13921 will name our implementation @code{@value{COUNT-WORDS}}.
13923 The function counts words within a region. This means that the
13924 argument list must contain symbols that are bound to the two
13925 positions, the beginning and end of the region. These two positions
13926 can be called @samp{beginning} and @samp{end} respectively. The first
13927 line of the documentation should be a single sentence, since that is
13928 all that is printed as documentation by a command such as
13929 @code{apropos}. The interactive expression will be of the form
13930 @samp{(interactive "r")}, since that will cause Emacs to pass the
13931 beginning and end of the region to the function's argument list. All
13934 The body of the function needs to be written to do three tasks:
13935 first, to set up conditions under which the @code{while} loop can
13936 count words, second, to run the @code{while} loop, and third, to send
13937 a message to the user.
13939 When a user calls @code{@value{COUNT-WORDS}}, point may be at the
13940 beginning or the end of the region. However, the counting process
13941 must start at the beginning of the region. This means we will want
13942 to put point there if it is not already there. Executing
13943 @code{(goto-char beginning)} ensures this. Of course, we will want to
13944 return point to its expected position when the function finishes its
13945 work. For this reason, the body must be enclosed in a
13946 @code{save-excursion} expression.
13948 The central part of the body of the function consists of a
13949 @code{while} loop in which one expression jumps point forward word by
13950 word, and another expression counts those jumps. The true-or-false-test
13951 of the @code{while} loop should test true so long as point should jump
13952 forward, and false when point is at the end of the region.
13954 We could use @code{(forward-word 1)} as the expression for moving point
13955 forward word by word, but it is easier to see what Emacs identifies as a
13956 `word' if we use a regular expression search.
13958 A regular expression search that finds the pattern for which it is
13959 searching leaves point after the last character matched. This means
13960 that a succession of successful word searches will move point forward
13963 As a practical matter, we want the regular expression search to jump
13964 over whitespace and punctuation between words as well as over the
13965 words themselves. A regexp that refuses to jump over interword
13966 whitespace would never jump more than one word! This means that
13967 the regexp should include the whitespace and punctuation that follows
13968 a word, if any, as well as the word itself. (A word may end a buffer
13969 and not have any following whitespace or punctuation, so that part of
13970 the regexp must be optional.)
13972 Thus, what we want for the regexp is a pattern defining one or more
13973 word constituent characters followed, optionally, by one or more
13974 characters that are not word constituents. The regular expression for
13982 The buffer's syntax table determines which characters are and are not
13983 word constituents. For more information about syntax,
13984 @pxref{Syntax Tables, , Syntax Tables, elisp, The GNU Emacs Lisp
13988 The search expression looks like this:
13991 (re-search-forward "\\w+\\W*")
13995 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13996 single backslash has special meaning to the Emacs Lisp interpreter.
13997 It indicates that the following character is interpreted differently
13998 than usual. For example, the two characters, @samp{\n}, stand for
13999 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
14000 backslashes in a row stand for an ordinary, `unspecial' backslash, so
14001 Emacs Lisp interpreter ends of seeing a single backslash followed by a
14002 letter. So it discovers the letter is special.)
14004 We need a counter to count how many words there are; this variable
14005 must first be set to 0 and then incremented each time Emacs goes
14006 around the @code{while} loop. The incrementing expression is simply:
14009 (setq count (1+ count))
14012 Finally, we want to tell the user how many words there are in the
14013 region. The @code{message} function is intended for presenting this
14014 kind of information to the user. The message has to be phrased so
14015 that it reads properly regardless of how many words there are in the
14016 region: we don't want to say that ``there are 1 words in the region''.
14017 The conflict between singular and plural is ungrammatical. We can
14018 solve this problem by using a conditional expression that evaluates
14019 different messages depending on the number of words in the region.
14020 There are three possibilities: no words in the region, one word in the
14021 region, and more than one word. This means that the @code{cond}
14022 special form is appropriate.
14025 All this leads to the following function definition:
14029 ;;; @r{First version; has bugs!}
14030 (defun @value{COUNT-WORDS} (beginning end)
14031 "Print number of words in the region.
14032 Words are defined as at least one word-constituent
14033 character followed by at least one character that
14034 is not a word-constituent. The buffer's syntax
14035 table determines which characters these are."
14037 (message "Counting words in region ... ")
14041 ;;; @r{1. Set up appropriate conditions.}
14043 (goto-char beginning)
14048 ;;; @r{2. Run the} while @r{loop.}
14049 (while (< (point) end)
14050 (re-search-forward "\\w+\\W*")
14051 (setq count (1+ count)))
14055 ;;; @r{3. Send a message to the user.}
14056 (cond ((zerop count)
14058 "The region does NOT have any words."))
14061 "The region has 1 word."))
14064 "The region has %d words." count))))))
14069 As written, the function works, but not in all circumstances.
14071 @node Whitespace Bug, , Design @value{COUNT-WORDS}, @value{COUNT-WORDS}
14072 @comment node-name, next, previous, up
14073 @subsection The Whitespace Bug in @code{@value{COUNT-WORDS}}
14075 The @code{@value{COUNT-WORDS}} command described in the preceding
14076 section has two bugs, or rather, one bug with two manifestations.
14077 First, if you mark a region containing only whitespace in the middle
14078 of some text, the @code{@value{COUNT-WORDS}} command tells you that the
14079 region contains one word! Second, if you mark a region containing
14080 only whitespace at the end of the buffer or the accessible portion of
14081 a narrowed buffer, the command displays an error message that looks
14085 Search failed: "\\w+\\W*"
14088 If you are reading this in Info in GNU Emacs, you can test for these
14091 First, evaluate the function in the usual manner to install it.
14093 Here is a copy of the definition. Place your cursor after the closing
14094 parenthesis and type @kbd{C-x C-e} to install it.
14098 ;; @r{First version; has bugs!}
14099 (defun @value{COUNT-WORDS} (beginning end)
14100 "Print number of words in the region.
14101 Words are defined as at least one word-constituent character followed
14102 by at least one character that is not a word-constituent. The buffer's
14103 syntax table determines which characters these are."
14107 (message "Counting words in region ... ")
14111 ;;; @r{1. Set up appropriate conditions.}
14113 (goto-char beginning)
14118 ;;; @r{2. Run the} while @r{loop.}
14119 (while (< (point) end)
14120 (re-search-forward "\\w+\\W*")
14121 (setq count (1+ count)))
14125 ;;; @r{3. Send a message to the user.}
14126 (cond ((zerop count)
14127 (message "The region does NOT have any words."))
14128 ((= 1 count) (message "The region has 1 word."))
14129 (t (message "The region has %d words." count))))))
14135 If you wish, you can also install this keybinding by evaluating it:
14138 (global-set-key "\C-c=" '@value{COUNT-WORDS})
14141 To conduct the first test, set mark and point to the beginning and end
14142 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14143 @value{COUNT-WORDS}} if you have not bound @kbd{C-c =}):
14150 Emacs will tell you, correctly, that the region has three words.
14152 Repeat the test, but place mark at the beginning of the line and place
14153 point just @emph{before} the word @samp{one}. Again type the command
14154 @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}). Emacs should tell you
14155 that the region has no words, since it is composed only of the
14156 whitespace at the beginning of the line. But instead Emacs tells you
14157 that the region has one word!
14159 For the third test, copy the sample line to the end of the
14160 @file{*scratch*} buffer and then type several spaces at the end of the
14161 line. Place mark right after the word @samp{three} and point at the
14162 end of line. (The end of the line will be the end of the buffer.)
14163 Type @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}) as you did before.
14164 Again, Emacs should tell you that the region has no words, since it is
14165 composed only of the whitespace at the end of the line. Instead,
14166 Emacs displays an error message saying @samp{Search failed}.
14168 The two bugs stem from the same problem.
14170 Consider the first manifestation of the bug, in which the command
14171 tells you that the whitespace at the beginning of the line contains
14172 one word. What happens is this: The @code{M-x @value{COUNT-WORDS}}
14173 command moves point to the beginning of the region. The @code{while}
14174 tests whether the value of point is smaller than the value of
14175 @code{end}, which it is. Consequently, the regular expression search
14176 looks for and finds the first word. It leaves point after the word.
14177 @code{count} is set to one. The @code{while} loop repeats; but this
14178 time the value of point is larger than the value of @code{end}, the
14179 loop is exited; and the function displays a message saying the number
14180 of words in the region is one. In brief, the regular expression
14181 search looks for and finds the word even though it is outside
14184 In the second manifestation of the bug, the region is whitespace at
14185 the end of the buffer. Emacs says @samp{Search failed}. What happens
14186 is that the true-or-false-test in the @code{while} loop tests true, so
14187 the search expression is executed. But since there are no more words
14188 in the buffer, the search fails.
14190 In both manifestations of the bug, the search extends or attempts to
14191 extend outside of the region.
14193 The solution is to limit the search to the region---this is a fairly
14194 simple action, but as you may have come to expect, it is not quite as
14195 simple as you might think.
14197 As we have seen, the @code{re-search-forward} function takes a search
14198 pattern as its first argument. But in addition to this first,
14199 mandatory argument, it accepts three optional arguments. The optional
14200 second argument bounds the search. The optional third argument, if
14201 @code{t}, causes the function to return @code{nil} rather than signal
14202 an error if the search fails. The optional fourth argument is a
14203 repeat count. (In Emacs, you can see a function's documentation by
14204 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14206 In the @code{@value{COUNT-WORDS}} definition, the value of the end of
14207 the region is held by the variable @code{end} which is passed as an
14208 argument to the function. Thus, we can add @code{end} as an argument
14209 to the regular expression search expression:
14212 (re-search-forward "\\w+\\W*" end)
14215 However, if you make only this change to the @code{@value{COUNT-WORDS}}
14216 definition and then test the new version of the definition on a
14217 stretch of whitespace, you will receive an error message saying
14218 @samp{Search failed}.
14220 What happens is this: the search is limited to the region, and fails
14221 as you expect because there are no word-constituent characters in the
14222 region. Since it fails, we receive an error message. But we do not
14223 want to receive an error message in this case; we want to receive the
14224 message that "The region does NOT have any words."
14226 The solution to this problem is to provide @code{re-search-forward}
14227 with a third argument of @code{t}, which causes the function to return
14228 @code{nil} rather than signal an error if the search fails.
14230 However, if you make this change and try it, you will see the message
14231 ``Counting words in region ... '' and @dots{} you will keep on seeing
14232 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14234 Here is what happens: the search is limited to the region, as before,
14235 and it fails because there are no word-constituent characters in the
14236 region, as expected. Consequently, the @code{re-search-forward}
14237 expression returns @code{nil}. It does nothing else. In particular,
14238 it does not move point, which it does as a side effect if it finds the
14239 search target. After the @code{re-search-forward} expression returns
14240 @code{nil}, the next expression in the @code{while} loop is evaluated.
14241 This expression increments the count. Then the loop repeats. The
14242 true-or-false-test tests true because the value of point is still less
14243 than the value of end, since the @code{re-search-forward} expression
14244 did not move point. @dots{} and the cycle repeats @dots{}
14246 The @code{@value{COUNT-WORDS}} definition requires yet another
14247 modification, to cause the true-or-false-test of the @code{while} loop
14248 to test false if the search fails. Put another way, there are two
14249 conditions that must be satisfied in the true-or-false-test before the
14250 word count variable is incremented: point must still be within the
14251 region and the search expression must have found a word to count.
14253 Since both the first condition and the second condition must be true
14254 together, the two expressions, the region test and the search
14255 expression, can be joined with an @code{and} special form and embedded in
14256 the @code{while} loop as the true-or-false-test, like this:
14259 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14262 @c colon in printed section title causes problem in Info cross reference
14263 @c also trouble with an overfull hbox
14266 (For information about @code{and}, see
14267 @ref{kill-new function, , The @code{kill-new} function}.)
14271 (@xref{kill-new function, , The @code{kill-new} function}, for
14272 information about @code{and}.)
14275 The @code{re-search-forward} expression returns @code{t} if the search
14276 succeeds and as a side effect moves point. Consequently, as words are
14277 found, point is moved through the region. When the search expression
14278 fails to find another word, or when point reaches the end of the
14279 region, the true-or-false-test tests false, the @code{while} loop
14280 exits, and the @code{@value{COUNT-WORDS}} function displays one or
14281 other of its messages.
14283 After incorporating these final changes, the @code{@value{COUNT-WORDS}}
14284 works without bugs (or at least, without bugs that I have found!).
14285 Here is what it looks like:
14289 ;;; @r{Final version:} @code{while}
14290 (defun @value{COUNT-WORDS} (beginning end)
14291 "Print number of words in the region."
14293 (message "Counting words in region ... ")
14297 ;;; @r{1. Set up appropriate conditions.}
14300 (goto-char beginning)
14304 ;;; @r{2. Run the} while @r{loop.}
14305 (while (and (< (point) end)
14306 (re-search-forward "\\w+\\W*" end t))
14307 (setq count (1+ count)))
14311 ;;; @r{3. Send a message to the user.}
14312 (cond ((zerop count)
14314 "The region does NOT have any words."))
14317 "The region has 1 word."))
14320 "The region has %d words." count))))))
14324 @node recursive-count-words, Counting Exercise, @value{COUNT-WORDS}, Counting Words
14325 @comment node-name, next, previous, up
14326 @section Count Words Recursively
14327 @cindex Count words recursively
14328 @cindex Recursively counting words
14329 @cindex Words, counted recursively
14331 You can write the function for counting words recursively as well as
14332 with a @code{while} loop. Let's see how this is done.
14334 First, we need to recognize that the @code{@value{COUNT-WORDS}}
14335 function has three jobs: it sets up the appropriate conditions for
14336 counting to occur; it counts the words in the region; and it sends a
14337 message to the user telling how many words there are.
14339 If we write a single recursive function to do everything, we will
14340 receive a message for every recursive call. If the region contains 13
14341 words, we will receive thirteen messages, one right after the other.
14342 We don't want this! Instead, we must write two functions to do the
14343 job, one of which (the recursive function) will be used inside of the
14344 other. One function will set up the conditions and display the
14345 message; the other will return the word count.
14347 Let us start with the function that causes the message to be displayed.
14348 We can continue to call this @code{@value{COUNT-WORDS}}.
14350 This is the function that the user will call. It will be interactive.
14351 Indeed, it will be similar to our previous versions of this
14352 function, except that it will call @code{recursive-count-words} to
14353 determine how many words are in the region.
14356 We can readily construct a template for this function, based on our
14361 ;; @r{Recursive version; uses regular expression search}
14362 (defun @value{COUNT-WORDS} (beginning end)
14363 "@var{documentation}@dots{}"
14364 (@var{interactive-expression}@dots{})
14368 ;;; @r{1. Set up appropriate conditions.}
14369 (@var{explanatory message})
14370 (@var{set-up functions}@dots{}
14374 ;;; @r{2. Count the words.}
14375 @var{recursive call}
14379 ;;; @r{3. Send a message to the user.}
14380 @var{message providing word count}))
14384 The definition looks straightforward, except that somehow the count
14385 returned by the recursive call must be passed to the message
14386 displaying the word count. A little thought suggests that this can be
14387 done by making use of a @code{let} expression: we can bind a variable
14388 in the varlist of a @code{let} expression to the number of words in
14389 the region, as returned by the recursive call; and then the
14390 @code{cond} expression, using binding, can display the value to the
14393 Often, one thinks of the binding within a @code{let} expression as
14394 somehow secondary to the `primary' work of a function. But in this
14395 case, what you might consider the `primary' job of the function,
14396 counting words, is done within the @code{let} expression.
14399 Using @code{let}, the function definition looks like this:
14403 (defun @value{COUNT-WORDS} (beginning end)
14404 "Print number of words in the region."
14409 ;;; @r{1. Set up appropriate conditions.}
14410 (message "Counting words in region ... ")
14412 (goto-char beginning)
14416 ;;; @r{2. Count the words.}
14417 (let ((count (recursive-count-words end)))
14421 ;;; @r{3. Send a message to the user.}
14422 (cond ((zerop count)
14424 "The region does NOT have any words."))
14427 "The region has 1 word."))
14430 "The region has %d words." count))))))
14434 Next, we need to write the recursive counting function.
14436 A recursive function has at least three parts: the `do-again-test', the
14437 `next-step-expression', and the recursive call.
14439 The do-again-test determines whether the function will or will not be
14440 called again. Since we are counting words in a region and can use a
14441 function that moves point forward for every word, the do-again-test
14442 can check whether point is still within the region. The do-again-test
14443 should find the value of point and determine whether point is before,
14444 at, or after the value of the end of the region. We can use the
14445 @code{point} function to locate point. Clearly, we must pass the
14446 value of the end of the region to the recursive counting function as an
14449 In addition, the do-again-test should also test whether the search finds a
14450 word. If it does not, the function should not call itself again.
14452 The next-step-expression changes a value so that when the recursive
14453 function is supposed to stop calling itself, it stops. More
14454 precisely, the next-step-expression changes a value so that at the
14455 right time, the do-again-test stops the recursive function from
14456 calling itself again. In this case, the next-step-expression can be
14457 the expression that moves point forward, word by word.
14459 The third part of a recursive function is the recursive call.
14461 Somewhere, also, we also need a part that does the `work' of the
14462 function, a part that does the counting. A vital part!
14465 But already, we have an outline of the recursive counting function:
14469 (defun recursive-count-words (region-end)
14470 "@var{documentation}@dots{}"
14471 @var{do-again-test}
14472 @var{next-step-expression}
14473 @var{recursive call})
14477 Now we need to fill in the slots. Let's start with the simplest cases
14478 first: if point is at or beyond the end of the region, there cannot
14479 be any words in the region, so the function should return zero.
14480 Likewise, if the search fails, there are no words to count, so the
14481 function should return zero.
14483 On the other hand, if point is within the region and the search
14484 succeeds, the function should call itself again.
14487 Thus, the do-again-test should look like this:
14491 (and (< (point) region-end)
14492 (re-search-forward "\\w+\\W*" region-end t))
14496 Note that the search expression is part of the do-again-test---the
14497 function returns @code{t} if its search succeeds and @code{nil} if it
14498 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14499 @code{@value{COUNT-WORDS}}}, for an explanation of how
14500 @code{re-search-forward} works.)
14502 The do-again-test is the true-or-false test of an @code{if} clause.
14503 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14504 clause should call the function again; but if it fails, the else-part
14505 should return zero since either point is outside the region or the
14506 search failed because there were no words to find.
14508 But before considering the recursive call, we need to consider the
14509 next-step-expression. What is it? Interestingly, it is the search
14510 part of the do-again-test.
14512 In addition to returning @code{t} or @code{nil} for the
14513 do-again-test, @code{re-search-forward} moves point forward as a side
14514 effect of a successful search. This is the action that changes the
14515 value of point so that the recursive function stops calling itself
14516 when point completes its movement through the region. Consequently,
14517 the @code{re-search-forward} expression is the next-step-expression.
14520 In outline, then, the body of the @code{recursive-count-words}
14521 function looks like this:
14525 (if @var{do-again-test-and-next-step-combined}
14527 @var{recursive-call-returning-count}
14533 How to incorporate the mechanism that counts?
14535 If you are not used to writing recursive functions, a question like
14536 this can be troublesome. But it can and should be approached
14539 We know that the counting mechanism should be associated in some way
14540 with the recursive call. Indeed, since the next-step-expression moves
14541 point forward by one word, and since a recursive call is made for
14542 each word, the counting mechanism must be an expression that adds one
14543 to the value returned by a call to @code{recursive-count-words}.
14546 Consider several cases:
14550 If there are two words in the region, the function should return
14551 a value resulting from adding one to the value returned when it counts
14552 the first word, plus the number returned when it counts the remaining
14553 words in the region, which in this case is one.
14556 If there is one word in the region, the function should return
14557 a value resulting from adding one to the value returned when it counts
14558 that word, plus the number returned when it counts the remaining
14559 words in the region, which in this case is zero.
14562 If there are no words in the region, the function should return zero.
14565 From the sketch we can see that the else-part of the @code{if} returns
14566 zero for the case of no words. This means that the then-part of the
14567 @code{if} must return a value resulting from adding one to the value
14568 returned from a count of the remaining words.
14571 The expression will look like this, where @code{1+} is a function that
14572 adds one to its argument.
14575 (1+ (recursive-count-words region-end))
14579 The whole @code{recursive-count-words} function will then look like
14584 (defun recursive-count-words (region-end)
14585 "@var{documentation}@dots{}"
14587 ;;; @r{1. do-again-test}
14588 (if (and (< (point) region-end)
14589 (re-search-forward "\\w+\\W*" region-end t))
14593 ;;; @r{2. then-part: the recursive call}
14594 (1+ (recursive-count-words region-end))
14596 ;;; @r{3. else-part}
14602 Let's examine how this works:
14604 If there are no words in the region, the else part of the @code{if}
14605 expression is evaluated and consequently the function returns zero.
14607 If there is one word in the region, the value of point is less than
14608 the value of @code{region-end} and the search succeeds. In this case,
14609 the true-or-false-test of the @code{if} expression tests true, and the
14610 then-part of the @code{if} expression is evaluated. The counting
14611 expression is evaluated. This expression returns a value (which will
14612 be the value returned by the whole function) that is the sum of one
14613 added to the value returned by a recursive call.
14615 Meanwhile, the next-step-expression has caused point to jump over the
14616 first (and in this case only) word in the region. This means that
14617 when @code{(recursive-count-words region-end)} is evaluated a second
14618 time, as a result of the recursive call, the value of point will be
14619 equal to or greater than the value of region end. So this time,
14620 @code{recursive-count-words} will return zero. The zero will be added
14621 to one, and the original evaluation of @code{recursive-count-words}
14622 will return one plus zero, which is one, which is the correct amount.
14624 Clearly, if there are two words in the region, the first call to
14625 @code{recursive-count-words} returns one added to the value returned
14626 by calling @code{recursive-count-words} on a region containing the
14627 remaining word---that is, it adds one to one, producing two, which is
14628 the correct amount.
14630 Similarly, if there are three words in the region, the first call to
14631 @code{recursive-count-words} returns one added to the value returned
14632 by calling @code{recursive-count-words} on a region containing the
14633 remaining two words---and so on and so on.
14637 With full documentation the two functions look like this:
14641 The recursive function:
14643 @findex recursive-count-words
14646 (defun recursive-count-words (region-end)
14647 "Number of words between point and REGION-END."
14651 ;;; @r{1. do-again-test}
14652 (if (and (< (point) region-end)
14653 (re-search-forward "\\w+\\W*" region-end t))
14657 ;;; @r{2. then-part: the recursive call}
14658 (1+ (recursive-count-words region-end))
14660 ;;; @r{3. else-part}
14671 ;;; @r{Recursive version}
14672 (defun @value{COUNT-WORDS} (beginning end)
14673 "Print number of words in the region.
14677 Words are defined as at least one word-constituent
14678 character followed by at least one character that is
14679 not a word-constituent. The buffer's syntax table
14680 determines which characters these are."
14684 (message "Counting words in region ... ")
14686 (goto-char beginning)
14687 (let ((count (recursive-count-words end)))
14690 (cond ((zerop count)
14692 "The region does NOT have any words."))
14696 (message "The region has 1 word."))
14699 "The region has %d words." count))))))
14703 @node Counting Exercise, , recursive-count-words, Counting Words
14704 @section Exercise: Counting Punctuation
14706 Using a @code{while} loop, write a function to count the number of
14707 punctuation marks in a region---period, comma, semicolon, colon,
14708 exclamation mark, and question mark. Do the same using recursion.
14710 @node Words in a defun, Readying a Graph, Counting Words, Top
14711 @chapter Counting Words in a @code{defun}
14712 @cindex Counting words in a @code{defun}
14713 @cindex Word counting in a @code{defun}
14715 Our next project is to count the number of words in a function
14716 definition. Clearly, this can be done using some variant of
14717 @code{@value{COUNT-WORDS}}. @xref{Counting Words, , Counting Words:
14718 Repetition and Regexps}. If we are just going to count the words in
14719 one definition, it is easy enough to mark the definition with the
14720 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14721 @code{@value{COUNT-WORDS}}.
14723 However, I am more ambitious: I want to count the words and symbols in
14724 every definition in the Emacs sources and then print a graph that
14725 shows how many functions there are of each length: how many contain 40
14726 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14727 and so on. I have often been curious how long a typical function is,
14728 and this will tell.
14731 * Divide and Conquer::
14732 * Words and Symbols:: What to count?
14733 * Syntax:: What constitutes a word or symbol?
14734 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
14735 * Several defuns:: Counting several defuns in a file.
14736 * Find a File:: Do you want to look at a file?
14737 * lengths-list-file:: A list of the lengths of many definitions.
14738 * Several files:: Counting in definitions in different files.
14739 * Several files recursively:: Recursively counting in different files.
14740 * Prepare the data:: Prepare the data for display in a graph.
14743 @node Divide and Conquer, Words and Symbols, Words in a defun, Words in a defun
14745 @unnumberedsec Divide and Conquer
14748 Described in one phrase, the histogram project is daunting; but
14749 divided into numerous small steps, each of which we can take one at a
14750 time, the project becomes less fearsome. Let us consider what the
14755 First, write a function to count the words in one definition. This
14756 includes the problem of handling symbols as well as words.
14759 Second, write a function to list the numbers of words in each function
14760 in a file. This function can use the @code{count-words-in-defun}
14764 Third, write a function to list the numbers of words in each function
14765 in each of several files. This entails automatically finding the
14766 various files, switching to them, and counting the words in the
14767 definitions within them.
14770 Fourth, write a function to convert the list of numbers that we
14771 created in step three to a form that will be suitable for printing as
14775 Fifth, write a function to print the results as a graph.
14778 This is quite a project! But if we take each step slowly, it will not
14781 @node Words and Symbols, Syntax, Divide and Conquer, Words in a defun
14782 @section What to Count?
14783 @cindex Words and symbols in defun
14785 When we first start thinking about how to count the words in a
14786 function definition, the first question is (or ought to be) what are
14787 we going to count? When we speak of `words' with respect to a Lisp
14788 function definition, we are actually speaking, in large part, of
14789 `symbols'. For example, the following @code{multiply-by-seven}
14790 function contains the five symbols @code{defun},
14791 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14792 addition, in the documentation string, it contains the four words
14793 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14794 symbol @samp{number} is repeated, so the definition contains a total
14795 of ten words and symbols.
14799 (defun multiply-by-seven (number)
14800 "Multiply NUMBER by seven."
14806 However, if we mark the @code{multiply-by-seven} definition with
14807 @kbd{C-M-h} (@code{mark-defun}), and then call
14808 @code{@value{COUNT-WORDS}} on it, we will find that
14809 @code{@value{COUNT-WORDS}} claims the definition has eleven words, not
14810 ten! Something is wrong!
14812 The problem is twofold: @code{@value{COUNT-WORDS}} does not count the
14813 @samp{*} as a word, and it counts the single symbol,
14814 @code{multiply-by-seven}, as containing three words. The hyphens are
14815 treated as if they were interword spaces rather than intraword
14816 connectors: @samp{multiply-by-seven} is counted as if it were written
14817 @samp{multiply by seven}.
14819 The cause of this confusion is the regular expression search within
14820 the @code{@value{COUNT-WORDS}} definition that moves point forward word
14821 by word. In the canonical version of @code{@value{COUNT-WORDS}}, the
14829 This regular expression is a pattern defining one or more word
14830 constituent characters possibly followed by one or more characters
14831 that are not word constituents. What is meant by `word constituent
14832 characters' brings us to the issue of syntax, which is worth a section
14835 @node Syntax, count-words-in-defun, Words and Symbols, Words in a defun
14836 @section What Constitutes a Word or Symbol?
14837 @cindex Syntax categories and tables
14839 Emacs treats different characters as belonging to different
14840 @dfn{syntax categories}. For example, the regular expression,
14841 @samp{\\w+}, is a pattern specifying one or more @emph{word
14842 constituent} characters. Word constituent characters are members of
14843 one syntax category. Other syntax categories include the class of
14844 punctuation characters, such as the period and the comma, and the
14845 class of whitespace characters, such as the blank space and the tab
14846 character. (For more information, @pxref{Syntax Tables, , Syntax
14847 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14849 Syntax tables specify which characters belong to which categories.
14850 Usually, a hyphen is not specified as a `word constituent character'.
14851 Instead, it is specified as being in the `class of characters that are
14852 part of symbol names but not words.' This means that the
14853 @code{@value{COUNT-WORDS}} function treats it in the same way it treats
14854 an interword white space, which is why @code{@value{COUNT-WORDS}}
14855 counts @samp{multiply-by-seven} as three words.
14857 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14858 one symbol: modify the syntax table or modify the regular expression.
14860 We could redefine a hyphen as a word constituent character by
14861 modifying the syntax table that Emacs keeps for each mode. This
14862 action would serve our purpose, except that a hyphen is merely the
14863 most common character within symbols that is not typically a word
14864 constituent character; there are others, too.
14866 Alternatively, we can redefine the regexp used in the
14867 @code{@value{COUNT-WORDS}} definition so as to include symbols. This
14868 procedure has the merit of clarity, but the task is a little tricky.
14871 The first part is simple enough: the pattern must match ``at least one
14872 character that is a word or symbol constituent''. Thus:
14875 "\\(\\w\\|\\s_\\)+"
14879 The @samp{\\(} is the first part of the grouping construct that
14880 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14881 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14882 character and the @samp{\\s_} matches any character that is part of a
14883 symbol name but not a word-constituent character. The @samp{+}
14884 following the group indicates that the word or symbol constituent
14885 characters must be matched at least once.
14887 However, the second part of the regexp is more difficult to design.
14888 What we want is to follow the first part with ``optionally one or more
14889 characters that are not constituents of a word or symbol''. At first,
14890 I thought I could define this with the following:
14893 "\\(\\W\\|\\S_\\)*"
14897 The upper case @samp{W} and @samp{S} match characters that are
14898 @emph{not} word or symbol constituents. Unfortunately, this
14899 expression matches any character that is either not a word constituent
14900 or not a symbol constituent. This matches any character!
14902 I then noticed that every word or symbol in my test region was
14903 followed by white space (blank space, tab, or newline). So I tried
14904 placing a pattern to match one or more blank spaces after the pattern
14905 for one or more word or symbol constituents. This failed, too. Words
14906 and symbols are often separated by whitespace, but in actual code
14907 parentheses may follow symbols and punctuation may follow words. So
14908 finally, I designed a pattern in which the word or symbol constituents
14909 are followed optionally by characters that are not white space and
14910 then followed optionally by white space.
14913 Here is the full regular expression:
14916 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14919 @node count-words-in-defun, Several defuns, Syntax, Words in a defun
14920 @section The @code{count-words-in-defun} Function
14921 @cindex Counting words in a @code{defun}
14923 We have seen that there are several ways to write a
14924 @code{count-words-region} function. To write a
14925 @code{count-words-in-defun}, we need merely adapt one of these
14928 The version that uses a @code{while} loop is easy to understand, so I
14929 am going to adapt that. Because @code{count-words-in-defun} will be
14930 part of a more complex program, it need not be interactive and it need
14931 not display a message but just return the count. These considerations
14932 simplify the definition a little.
14934 On the other hand, @code{count-words-in-defun} will be used within a
14935 buffer that contains function definitions. Consequently, it is
14936 reasonable to ask that the function determine whether it is called
14937 when point is within a function definition, and if it is, to return
14938 the count for that definition. This adds complexity to the
14939 definition, but saves us from needing to pass arguments to the
14943 These considerations lead us to prepare the following template:
14947 (defun count-words-in-defun ()
14948 "@var{documentation}@dots{}"
14949 (@var{set up}@dots{}
14950 (@var{while loop}@dots{})
14951 @var{return count})
14956 As usual, our job is to fill in the slots.
14960 We are presuming that this function will be called within a buffer
14961 containing function definitions. Point will either be within a
14962 function definition or not. For @code{count-words-in-defun} to work,
14963 point must move to the beginning of the definition, a counter must
14964 start at zero, and the counting loop must stop when point reaches the
14965 end of the definition.
14967 The @code{beginning-of-defun} function searches backwards for an
14968 opening delimiter such as a @samp{(} at the beginning of a line, and
14969 moves point to that position, or else to the limit of the search. In
14970 practice, this means that @code{beginning-of-defun} moves point to the
14971 beginning of an enclosing or preceding function definition, or else to
14972 the beginning of the buffer. We can use @code{beginning-of-defun} to
14973 place point where we wish to start.
14975 The @code{while} loop requires a counter to keep track of the words or
14976 symbols being counted. A @code{let} expression can be used to create
14977 a local variable for this purpose, and bind it to an initial value of zero.
14979 The @code{end-of-defun} function works like @code{beginning-of-defun}
14980 except that it moves point to the end of the definition.
14981 @code{end-of-defun} can be used as part of an expression that
14982 determines the position of the end of the definition.
14984 The set up for @code{count-words-in-defun} takes shape rapidly: first
14985 we move point to the beginning of the definition, then we create a
14986 local variable to hold the count, and finally, we record the position
14987 of the end of the definition so the @code{while} loop will know when to stop
14991 The code looks like this:
14995 (beginning-of-defun)
14997 (end (save-excursion (end-of-defun) (point))))
15002 The code is simple. The only slight complication is likely to concern
15003 @code{end}: it is bound to the position of the end of the definition
15004 by a @code{save-excursion} expression that returns the value of point
15005 after @code{end-of-defun} temporarily moves it to the end of the
15008 The second part of the @code{count-words-in-defun}, after the set up,
15009 is the @code{while} loop.
15011 The loop must contain an expression that jumps point forward word by
15012 word and symbol by symbol, and another expression that counts the
15013 jumps. The true-or-false-test for the @code{while} loop should test
15014 true so long as point should jump forward, and false when point is at
15015 the end of the definition. We have already redefined the regular
15016 expression for this, so the loop is straightforward:
15020 (while (and (< (point) end)
15022 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t))
15023 (setq count (1+ count)))
15027 The third part of the function definition returns the count of words
15028 and symbols. This part is the last expression within the body of the
15029 @code{let} expression, and can be, very simply, the local variable
15030 @code{count}, which when evaluated returns the count.
15033 Put together, the @code{count-words-in-defun} definition looks like this:
15035 @findex count-words-in-defun
15038 (defun count-words-in-defun ()
15039 "Return the number of words and symbols in a defun."
15040 (beginning-of-defun)
15042 (end (save-excursion (end-of-defun) (point))))
15046 (and (< (point) end)
15048 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
15050 (setq count (1+ count)))
15055 How to test this? The function is not interactive, but it is easy to
15056 put a wrapper around the function to make it interactive; we can use
15057 almost the same code as for the recursive version of
15058 @code{@value{COUNT-WORDS}}:
15062 ;;; @r{Interactive version.}
15063 (defun count-words-defun ()
15064 "Number of words and symbols in a function definition."
15067 "Counting words and symbols in function definition ... ")
15070 (let ((count (count-words-in-defun)))
15074 "The definition does NOT have any words or symbols."))
15079 "The definition has 1 word or symbol."))
15082 "The definition has %d words or symbols." count)))))
15088 Let's re-use @kbd{C-c =} as a convenient keybinding:
15091 (global-set-key "\C-c=" 'count-words-defun)
15094 Now we can try out @code{count-words-defun}: install both
15095 @code{count-words-in-defun} and @code{count-words-defun}, and set the
15096 keybinding, and then place the cursor within the following definition:
15100 (defun multiply-by-seven (number)
15101 "Multiply NUMBER by seven."
15108 Success! The definition has 10 words and symbols.
15110 The next problem is to count the numbers of words and symbols in
15111 several definitions within a single file.
15113 @node Several defuns, Find a File, count-words-in-defun, Words in a defun
15114 @section Count Several @code{defuns} Within a File
15116 A file such as @file{simple.el} may have a hundred or more function
15117 definitions within it. Our long term goal is to collect statistics on
15118 many files, but as a first step, our immediate goal is to collect
15119 statistics on one file.
15121 The information will be a series of numbers, each number being the
15122 length of a function definition. We can store the numbers in a list.
15124 We know that we will want to incorporate the information regarding one
15125 file with information about many other files; this means that the
15126 function for counting definition lengths within one file need only
15127 return the list of lengths. It need not and should not display any
15130 The word count commands contain one expression to jump point forward
15131 word by word and another expression to count the jumps. The function
15132 to return the lengths of definitions can be designed to work the same
15133 way, with one expression to jump point forward definition by
15134 definition and another expression to construct the lengths' list.
15136 This statement of the problem makes it elementary to write the
15137 function definition. Clearly, we will start the count at the
15138 beginning of the file, so the first command will be @code{(goto-char
15139 (point-min))}. Next, we start the @code{while} loop; and the
15140 true-or-false test of the loop can be a regular expression search for
15141 the next function definition---so long as the search succeeds, point
15142 is moved forward and then the body of the loop is evaluated. The body
15143 needs an expression that constructs the lengths' list. @code{cons},
15144 the list construction command, can be used to create the list. That
15145 is almost all there is to it.
15148 Here is what this fragment of code looks like:
15152 (goto-char (point-min))
15153 (while (re-search-forward "^(defun" nil t)
15155 (cons (count-words-in-defun) lengths-list)))
15159 What we have left out is the mechanism for finding the file that
15160 contains the function definitions.
15162 In previous examples, we either used this, the Info file, or we
15163 switched back and forth to some other buffer, such as the
15164 @file{*scratch*} buffer.
15166 Finding a file is a new process that we have not yet discussed.
15168 @node Find a File, lengths-list-file, Several defuns, Words in a defun
15169 @comment node-name, next, previous, up
15170 @section Find a File
15171 @cindex Find a File
15173 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15174 command. This command is almost, but not quite right for the lengths
15178 Let's look at the source for @code{find-file}:
15182 (defun find-file (filename)
15183 "Edit file FILENAME.
15184 Switch to a buffer visiting file FILENAME,
15185 creating one if none already exists."
15186 (interactive "FFind file: ")
15187 (switch-to-buffer (find-file-noselect filename)))
15192 (The most recent version of the @code{find-file} function definition
15193 permits you to specify optional wildcards to visit multiple files; that
15194 makes the definition more complex and we will not discuss it here,
15195 since it is not relevant. You can see its source using either
15196 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15200 (defun find-file (filename &optional wildcards)
15201 "Edit file FILENAME.
15202 Switch to a buffer visiting file FILENAME,
15203 creating one if none already exists.
15204 Interactively, the default if you just type RET is the current directory,
15205 but the visited file name is available through the minibuffer history:
15206 type M-n to pull it into the minibuffer.
15208 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15209 expand wildcards (if any) and visit multiple files. You can
15210 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15212 To visit a file without any kind of conversion and without
15213 automatically choosing a major mode, use \\[find-file-literally]."
15214 (interactive (find-file-read-args "Find file: " nil))
15215 (let ((value (find-file-noselect filename nil nil wildcards)))
15217 (mapcar 'switch-to-buffer (nreverse value))
15218 (switch-to-buffer value))))
15221 The definition I am showing possesses short but complete documentation
15222 and an interactive specification that prompts you for a file name when
15223 you use the command interactively. The body of the definition
15224 contains two functions, @code{find-file-noselect} and
15225 @code{switch-to-buffer}.
15227 According to its documentation as shown by @kbd{C-h f} (the
15228 @code{describe-function} command), the @code{find-file-noselect}
15229 function reads the named file into a buffer and returns the buffer.
15230 (Its most recent version includes an optional wildcards argument,
15231 too, as well as another to read a file literally and an other you
15232 suppress warning messages. These optional arguments are irrelevant.)
15234 However, the @code{find-file-noselect} function does not select the
15235 buffer in which it puts the file. Emacs does not switch its attention
15236 (or yours if you are using @code{find-file-noselect}) to the selected
15237 buffer. That is what @code{switch-to-buffer} does: it switches the
15238 buffer to which Emacs attention is directed; and it switches the
15239 buffer displayed in the window to the new buffer. We have discussed
15240 buffer switching elsewhere. (@xref{Switching Buffers}.)
15242 In this histogram project, we do not need to display each file on the
15243 screen as the program determines the length of each definition within
15244 it. Instead of employing @code{switch-to-buffer}, we can work with
15245 @code{set-buffer}, which redirects the attention of the computer
15246 program to a different buffer but does not redisplay it on the screen.
15247 So instead of calling on @code{find-file} to do the job, we must write
15248 our own expression.
15250 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15252 @node lengths-list-file, Several files, Find a File, Words in a defun
15253 @section @code{lengths-list-file} in Detail
15255 The core of the @code{lengths-list-file} function is a @code{while}
15256 loop containing a function to move point forward `defun by defun' and
15257 a function to count the number of words and symbols in each defun.
15258 This core must be surrounded by functions that do various other tasks,
15259 including finding the file, and ensuring that point starts out at the
15260 beginning of the file. The function definition looks like this:
15261 @findex lengths-list-file
15265 (defun lengths-list-file (filename)
15266 "Return list of definitions' lengths within FILE.
15267 The returned list is a list of numbers.
15268 Each number is the number of words or
15269 symbols in one function definition."
15272 (message "Working on `%s' ... " filename)
15274 (let ((buffer (find-file-noselect filename))
15276 (set-buffer buffer)
15277 (setq buffer-read-only t)
15279 (goto-char (point-min))
15280 (while (re-search-forward "^(defun" nil t)
15282 (cons (count-words-in-defun) lengths-list)))
15283 (kill-buffer buffer)
15289 The function is passed one argument, the name of the file on which it
15290 will work. It has four lines of documentation, but no interactive
15291 specification. Since people worry that a computer is broken if they
15292 don't see anything going on, the first line of the body is a
15295 The next line contains a @code{save-excursion} that returns Emacs's
15296 attention to the current buffer when the function completes. This is
15297 useful in case you embed this function in another function that
15298 presumes point is restored to the original buffer.
15300 In the varlist of the @code{let} expression, Emacs finds the file and
15301 binds the local variable @code{buffer} to the buffer containing the
15302 file. At the same time, Emacs creates @code{lengths-list} as a local
15305 Next, Emacs switches its attention to the buffer.
15307 In the following line, Emacs makes the buffer read-only. Ideally,
15308 this line is not necessary. None of the functions for counting words
15309 and symbols in a function definition should change the buffer.
15310 Besides, the buffer is not going to be saved, even if it were changed.
15311 This line is entirely the consequence of great, perhaps excessive,
15312 caution. The reason for the caution is that this function and those
15313 it calls work on the sources for Emacs and it is inconvenient if they
15314 are inadvertently modified. It goes without saying that I did not
15315 realize a need for this line until an experiment went awry and started
15316 to modify my Emacs source files @dots{}
15318 Next comes a call to widen the buffer if it is narrowed. This
15319 function is usually not needed---Emacs creates a fresh buffer if none
15320 already exists; but if a buffer visiting the file already exists Emacs
15321 returns that one. In this case, the buffer may be narrowed and must
15322 be widened. If we wanted to be fully `user-friendly', we would
15323 arrange to save the restriction and the location of point, but we
15326 The @code{(goto-char (point-min))} expression moves point to the
15327 beginning of the buffer.
15329 Then comes a @code{while} loop in which the `work' of the function is
15330 carried out. In the loop, Emacs determines the length of each
15331 definition and constructs a lengths' list containing the information.
15333 Emacs kills the buffer after working through it. This is to save
15334 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15335 source files of interest; GNU Emacs 22 contains over a thousand source
15336 files. Another function will apply @code{lengths-list-file} to each
15339 Finally, the last expression within the @code{let} expression is the
15340 @code{lengths-list} variable; its value is returned as the value of
15341 the whole function.
15343 You can try this function by installing it in the usual fashion. Then
15344 place your cursor after the following expression and type @kbd{C-x
15345 C-e} (@code{eval-last-sexp}).
15347 @c !!! 22.1.1 lisp sources location here
15350 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15354 (You may need to change the pathname of the file; the one here is for
15355 GNU Emacs version 22.1.1. To change the expression, copy it to
15356 the @file{*scratch*} buffer and edit it.
15360 (Also, to see the full length of the list, rather than a truncated
15361 version, you may have to evaluate the following:
15364 (custom-set-variables '(eval-expression-print-length nil))
15368 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15369 Then evaluate the @code{lengths-list-file} expression.)
15372 The lengths' list for @file{debug.el} takes less than a second to
15373 produce and looks like this in GNU Emacs 22:
15376 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15380 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15381 took seven seconds to produce and looked like this:
15384 (75 41 80 62 20 45 44 68 45 12 34 235)
15387 (The newer version of @file{debug.el} contains more defuns than the
15388 earlier one; and my new machine is much faster than the old one.)
15390 Note that the length of the last definition in the file is first in
15393 @node Several files, Several files recursively, lengths-list-file, Words in a defun
15394 @section Count Words in @code{defuns} in Different Files
15396 In the previous section, we created a function that returns a list of
15397 the lengths of each definition in a file. Now, we want to define a
15398 function to return a master list of the lengths of the definitions in
15401 Working on each of a list of files is a repetitious act, so we can use
15402 either a @code{while} loop or recursion.
15405 * lengths-list-many-files:: Return a list of the lengths of defuns.
15406 * append:: Attach one list to another.
15409 @node lengths-list-many-files, append, Several files, Several files
15411 @unnumberedsubsec Determine the lengths of @code{defuns}
15414 The design using a @code{while} loop is routine. The argument passed
15415 the function is a list of files. As we saw earlier (@pxref{Loop
15416 Example}), you can write a @code{while} loop so that the body of the
15417 loop is evaluated if such a list contains elements, but to exit the
15418 loop if the list is empty. For this design to work, the body of the
15419 loop must contain an expression that shortens the list each time the
15420 body is evaluated, so that eventually the list is empty. The usual
15421 technique is to set the value of the list to the value of the @sc{cdr}
15422 of the list each time the body is evaluated.
15425 The template looks like this:
15429 (while @var{test-whether-list-is-empty}
15431 @var{set-list-to-cdr-of-list})
15435 Also, we remember that a @code{while} loop returns @code{nil} (the
15436 result of evaluating the true-or-false-test), not the result of any
15437 evaluation within its body. (The evaluations within the body of the
15438 loop are done for their side effects.) However, the expression that
15439 sets the lengths' list is part of the body---and that is the value
15440 that we want returned by the function as a whole. To do this, we
15441 enclose the @code{while} loop within a @code{let} expression, and
15442 arrange that the last element of the @code{let} expression contains
15443 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15444 Example with an Incrementing Counter}.)
15446 @findex lengths-list-many-files
15448 These considerations lead us directly to the function itself:
15452 ;;; @r{Use @code{while} loop.}
15453 (defun lengths-list-many-files (list-of-files)
15454 "Return list of lengths of defuns in LIST-OF-FILES."
15457 (let (lengths-list)
15459 ;;; @r{true-or-false-test}
15460 (while list-of-files
15465 ;;; @r{Generate a lengths' list.}
15467 (expand-file-name (car list-of-files)))))
15471 ;;; @r{Make files' list shorter.}
15472 (setq list-of-files (cdr list-of-files)))
15474 ;;; @r{Return final value of lengths' list.}
15479 @code{expand-file-name} is a built-in function that converts a file
15480 name to the absolute, long, path name form. The function employs the
15481 name of the directory in which the function is called.
15483 @c !!! 22.1.1 lisp sources location here
15485 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15486 Emacs is visiting the
15487 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15497 @c !!! 22.1.1 lisp sources location here
15499 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15502 The only other new element of this function definition is the as yet
15503 unstudied function @code{append}, which merits a short section for
15506 @node append, , lengths-list-many-files, Several files
15507 @subsection The @code{append} Function
15510 The @code{append} function attaches one list to another. Thus,
15513 (append '(1 2 3 4) '(5 6 7 8))
15524 This is exactly how we want to attach two lengths' lists produced by
15525 @code{lengths-list-file} to each other. The results contrast with
15529 (cons '(1 2 3 4) '(5 6 7 8))
15534 which constructs a new list in which the first argument to @code{cons}
15535 becomes the first element of the new list:
15538 ((1 2 3 4) 5 6 7 8)
15541 @node Several files recursively, Prepare the data, Several files, Words in a defun
15542 @section Recursively Count Words in Different Files
15544 Besides a @code{while} loop, you can work on each of a list of files
15545 with recursion. A recursive version of @code{lengths-list-many-files}
15546 is short and simple.
15548 The recursive function has the usual parts: the `do-again-test', the
15549 `next-step-expression', and the recursive call. The `do-again-test'
15550 determines whether the function should call itself again, which it
15551 will do if the @code{list-of-files} contains any remaining elements;
15552 the `next-step-expression' resets the @code{list-of-files} to the
15553 @sc{cdr} of itself, so eventually the list will be empty; and the
15554 recursive call calls itself on the shorter list. The complete
15555 function is shorter than this description!
15556 @findex recursive-lengths-list-many-files
15560 (defun recursive-lengths-list-many-files (list-of-files)
15561 "Return list of lengths of each defun in LIST-OF-FILES."
15562 (if list-of-files ; @r{do-again-test}
15565 (expand-file-name (car list-of-files)))
15566 (recursive-lengths-list-many-files
15567 (cdr list-of-files)))))
15572 In a sentence, the function returns the lengths' list for the first of
15573 the @code{list-of-files} appended to the result of calling itself on
15574 the rest of the @code{list-of-files}.
15576 Here is a test of @code{recursive-lengths-list-many-files}, along with
15577 the results of running @code{lengths-list-file} on each of the files
15580 Install @code{recursive-lengths-list-many-files} and
15581 @code{lengths-list-file}, if necessary, and then evaluate the
15582 following expressions. You may need to change the files' pathnames;
15583 those here work when this Info file and the Emacs sources are located
15584 in their customary places. To change the expressions, copy them to
15585 the @file{*scratch*} buffer, edit them, and then evaluate them.
15587 The results are shown after the @samp{@result{}}. (These results are
15588 for files from Emacs version 22.1.1; files from other versions of
15589 Emacs may produce different results.)
15591 @c !!! 22.1.1 lisp sources location here
15594 (cd "/usr/local/share/emacs/22.1.1/")
15596 (lengths-list-file "./lisp/macros.el")
15597 @result{} (283 263 480 90)
15601 (lengths-list-file "./lisp/mail/mailalias.el")
15602 @result{} (38 32 29 95 178 180 321 218 324)
15606 (lengths-list-file "./lisp/makesum.el")
15611 (recursive-lengths-list-many-files
15612 '("./lisp/macros.el"
15613 "./lisp/mail/mailalias.el"
15614 "./lisp/makesum.el"))
15615 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15619 The @code{recursive-lengths-list-many-files} function produces the
15622 The next step is to prepare the data in the list for display in a graph.
15624 @node Prepare the data, , Several files recursively, Words in a defun
15625 @section Prepare the Data for Display in a Graph
15627 The @code{recursive-lengths-list-many-files} function returns a list
15628 of numbers. Each number records the length of a function definition.
15629 What we need to do now is transform this data into a list of numbers
15630 suitable for generating a graph. The new list will tell how many
15631 functions definitions contain less than 10 words and
15632 symbols, how many contain between 10 and 19 words and symbols, how
15633 many contain between 20 and 29 words and symbols, and so on.
15635 In brief, we need to go through the lengths' list produced by the
15636 @code{recursive-lengths-list-many-files} function and count the number
15637 of defuns within each range of lengths, and produce a list of those
15641 * Data for Display in Detail::
15642 * Sorting:: Sorting lists.
15643 * Files List:: Making a list of files.
15644 * Counting function definitions::
15647 @node Data for Display in Detail, Sorting, Prepare the data, Prepare the data
15649 @unnumberedsubsec The Data for Display in Detail
15652 Based on what we have done before, we can readily foresee that it
15653 should not be too hard to write a function that `@sc{cdr}s' down the
15654 lengths' list, looks at each element, determines which length range it
15655 is in, and increments a counter for that range.
15657 However, before beginning to write such a function, we should consider
15658 the advantages of sorting the lengths' list first, so the numbers are
15659 ordered from smallest to largest. First, sorting will make it easier
15660 to count the numbers in each range, since two adjacent numbers will
15661 either be in the same length range or in adjacent ranges. Second, by
15662 inspecting a sorted list, we can discover the highest and lowest
15663 number, and thereby determine the largest and smallest length range
15666 @node Sorting, Files List, Data for Display in Detail, Prepare the data
15667 @subsection Sorting Lists
15670 Emacs contains a function to sort lists, called (as you might guess)
15671 @code{sort}. The @code{sort} function takes two arguments, the list
15672 to be sorted, and a predicate that determines whether the first of
15673 two list elements is ``less'' than the second.
15675 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15676 Type Object as an Argument}), a predicate is a function that
15677 determines whether some property is true or false. The @code{sort}
15678 function will reorder a list according to whatever property the
15679 predicate uses; this means that @code{sort} can be used to sort
15680 non-numeric lists by non-numeric criteria---it can, for example,
15681 alphabetize a list.
15684 The @code{<} function is used when sorting a numeric list. For example,
15687 (sort '(4 8 21 17 33 7 21 7) '<)
15695 (4 7 7 8 17 21 21 33)
15699 (Note that in this example, both the arguments are quoted so that the
15700 symbols are not evaluated before being passed to @code{sort} as
15703 Sorting the list returned by the
15704 @code{recursive-lengths-list-many-files} function is straightforward;
15705 it uses the @code{<} function:
15709 In GNU Emacs 22, eval
15711 (cd "/usr/local/share/emacs/22.0.50/")
15713 (recursive-lengths-list-many-files
15714 '("./lisp/macros.el"
15715 "./lisp/mail/mailalias.el"
15716 "./lisp/makesum.el"))
15724 (recursive-lengths-list-many-files
15725 '("./lisp/macros.el"
15726 "./lisp/mailalias.el"
15727 "./lisp/makesum.el"))
15737 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15741 (Note that in this example, the first argument to @code{sort} is not
15742 quoted, since the expression must be evaluated so as to produce the
15743 list that is passed to @code{sort}.)
15745 @node Files List, Counting function definitions, Sorting, Prepare the data
15746 @subsection Making a List of Files
15748 The @code{recursive-lengths-list-many-files} function requires a list
15749 of files as its argument. For our test examples, we constructed such
15750 a list by hand; but the Emacs Lisp source directory is too large for
15751 us to do for that. Instead, we will write a function to do the job
15752 for us. In this function, we will use both a @code{while} loop and a
15755 @findex directory-files
15756 We did not have to write a function like this for older versions of
15757 GNU Emacs, since they placed all the @samp{.el} files in one
15758 directory. Instead, we were able to use the @code{directory-files}
15759 function, which lists the names of files that match a specified
15760 pattern within a single directory.
15762 However, recent versions of Emacs place Emacs Lisp files in
15763 sub-directories of the top level @file{lisp} directory. This
15764 re-arrangement eases navigation. For example, all the mail related
15765 files are in a @file{lisp} sub-directory called @file{mail}. But at
15766 the same time, this arrangement forces us to create a file listing
15767 function that descends into the sub-directories.
15769 @findex files-in-below-directory
15770 We can create this function, called @code{files-in-below-directory},
15771 using familiar functions such as @code{car}, @code{nthcdr}, and
15772 @code{substring} in conjunction with an existing function called
15773 @code{directory-files-and-attributes}. This latter function not only
15774 lists all the filenames in a directory, including the names
15775 of sub-directories, but also their attributes.
15777 To restate our goal: to create a function that will enable us
15778 to feed filenames to @code{recursive-lengths-list-many-files}
15779 as a list that looks like this (but with more elements):
15783 ("./lisp/macros.el"
15784 "./lisp/mail/rmail.el"
15785 "./lisp/makesum.el")
15789 The @code{directory-files-and-attributes} function returns a list of
15790 lists. Each of the lists within the main list consists of 13
15791 elements. The first element is a string that contains the name of the
15792 file -- which, in GNU/Linux, may be a `directory file', that is to
15793 say, a file with the special attributes of a directory. The second
15794 element of the list is @code{t} for a directory, a string
15795 for symbolic link (the string is the name linked to), or @code{nil}.
15797 For example, the first @samp{.el} file in the @file{lisp/} directory
15798 is @file{abbrev.el}. Its name is
15799 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15800 directory or a symbolic link.
15803 This is how @code{directory-files-and-attributes} lists that file and
15829 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15830 directory. The beginning of its listing looks like this:
15841 (To learn about the different attributes, look at the documentation of
15842 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15843 function does not list the filename, so its first element is
15844 @code{directory-files-and-attributes}'s second element.)
15846 We will want our new function, @code{files-in-below-directory}, to
15847 list the @samp{.el} files in the directory it is told to check, and in
15848 any directories below that directory.
15850 This gives us a hint on how to construct
15851 @code{files-in-below-directory}: within a directory, the function
15852 should add @samp{.el} filenames to a list; and if, within a directory,
15853 the function comes upon a sub-directory, it should go into that
15854 sub-directory and repeat its actions.
15856 However, we should note that every directory contains a name that
15857 refers to itself, called @file{.}, (``dot'') and a name that refers to
15858 its parent directory, called @file{..} (``double dot''). (In
15859 @file{/}, the root directory, @file{..} refers to itself, since
15860 @file{/} has no parent.) Clearly, we do not want our
15861 @code{files-in-below-directory} function to enter those directories,
15862 since they always lead us, directly or indirectly, to the current
15865 Consequently, our @code{files-in-below-directory} function must do
15870 Check to see whether it is looking at a filename that ends in
15871 @samp{.el}; and if so, add its name to a list.
15874 Check to see whether it is looking at a filename that is the name of a
15875 directory; and if so,
15879 Check to see whether it is looking at @file{.} or @file{..}; and if
15883 Or else, go into that directory and repeat the process.
15887 Let's write a function definition to do these tasks. We will use a
15888 @code{while} loop to move from one filename to another within a
15889 directory, checking what needs to be done; and we will use a recursive
15890 call to repeat the actions on each sub-directory. The recursive
15891 pattern is `accumulate'
15892 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15893 using @code{append} as the combiner.
15896 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15897 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15899 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15900 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15903 @c /usr/local/share/emacs/22.1.1/lisp/
15906 Here is the function:
15910 (defun files-in-below-directory (directory)
15911 "List the .el files in DIRECTORY and in its sub-directories."
15912 ;; Although the function will be used non-interactively,
15913 ;; it will be easier to test if we make it interactive.
15914 ;; The directory will have a name such as
15915 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15916 (interactive "DDirectory name: ")
15919 (let (el-files-list
15920 (current-directory-list
15921 (directory-files-and-attributes directory t)))
15922 ;; while we are in the current directory
15923 (while current-directory-list
15927 ;; check to see whether filename ends in `.el'
15928 ;; and if so, append its name to a list.
15929 ((equal ".el" (substring (car (car current-directory-list)) -3))
15930 (setq el-files-list
15931 (cons (car (car current-directory-list)) el-files-list)))
15934 ;; check whether filename is that of a directory
15935 ((eq t (car (cdr (car current-directory-list))))
15936 ;; decide whether to skip or recurse
15939 (substring (car (car current-directory-list)) -1))
15940 ;; then do nothing since filename is that of
15941 ;; current directory or parent, "." or ".."
15945 ;; else descend into the directory and repeat the process
15946 (setq el-files-list
15948 (files-in-below-directory
15949 (car (car current-directory-list)))
15951 ;; move to the next filename in the list; this also
15952 ;; shortens the list so the while loop eventually comes to an end
15953 (setq current-directory-list (cdr current-directory-list)))
15954 ;; return the filenames
15959 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15960 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15962 The @code{files-in-below-directory} @code{directory-files} function
15963 takes one argument, the name of a directory.
15966 Thus, on my system,
15968 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15970 @c !!! 22.1.1 lisp sources location here
15974 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15979 tells me that in and below my Lisp sources directory are 1031
15982 @code{files-in-below-directory} returns a list in reverse alphabetical
15983 order. An expression to sort the list in alphabetical order looks
15989 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15996 "Test how long it takes to find lengths of all sorted elisp defuns."
15997 (insert "\n" (current-time-string) "\n")
16000 (recursive-lengths-list-many-files
16001 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
16003 (insert (format "%s" (current-time-string))))
16006 @node Counting function definitions, , Files List, Prepare the data
16007 @subsection Counting function definitions
16009 Our immediate goal is to generate a list that tells us how many
16010 function definitions contain fewer than 10 words and symbols, how many
16011 contain between 10 and 19 words and symbols, how many contain between
16012 20 and 29 words and symbols, and so on.
16014 With a sorted list of numbers, this is easy: count how many elements
16015 of the list are smaller than 10, then, after moving past the numbers
16016 just counted, count how many are smaller than 20, then, after moving
16017 past the numbers just counted, count how many are smaller than 30, and
16018 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
16019 larger than the top of that range. We can call the list of such
16020 numbers the @code{top-of-ranges} list.
16023 If we wished, we could generate this list automatically, but it is
16024 simpler to write a list manually. Here it is:
16025 @vindex top-of-ranges
16029 (defvar top-of-ranges
16032 110 120 130 140 150
16033 160 170 180 190 200
16034 210 220 230 240 250
16035 260 270 280 290 300)
16036 "List specifying ranges for `defuns-per-range'.")
16040 To change the ranges, we edit this list.
16042 Next, we need to write the function that creates the list of the
16043 number of definitions within each range. Clearly, this function must
16044 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
16047 The @code{defuns-per-range} function must do two things again and
16048 again: it must count the number of definitions within a range
16049 specified by the current top-of-range value; and it must shift to the
16050 next higher value in the @code{top-of-ranges} list after counting the
16051 number of definitions in the current range. Since each of these
16052 actions is repetitive, we can use @code{while} loops for the job.
16053 One loop counts the number of definitions in the range defined by the
16054 current top-of-range value, and the other loop selects each of the
16055 top-of-range values in turn.
16057 Several entries of the @code{sorted-lengths} list are counted for each
16058 range; this means that the loop for the @code{sorted-lengths} list
16059 will be inside the loop for the @code{top-of-ranges} list, like a
16060 small gear inside a big gear.
16062 The inner loop counts the number of definitions within the range. It
16063 is a simple counting loop of the type we have seen before.
16064 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
16065 The true-or-false test of the loop tests whether the value from the
16066 @code{sorted-lengths} list is smaller than the current value of the
16067 top of the range. If it is, the function increments the counter and
16068 tests the next value from the @code{sorted-lengths} list.
16071 The inner loop looks like this:
16075 (while @var{length-element-smaller-than-top-of-range}
16076 (setq number-within-range (1+ number-within-range))
16077 (setq sorted-lengths (cdr sorted-lengths)))
16081 The outer loop must start with the lowest value of the
16082 @code{top-of-ranges} list, and then be set to each of the succeeding
16083 higher values in turn. This can be done with a loop like this:
16087 (while top-of-ranges
16088 @var{body-of-loop}@dots{}
16089 (setq top-of-ranges (cdr top-of-ranges)))
16094 Put together, the two loops look like this:
16098 (while top-of-ranges
16100 ;; @r{Count the number of elements within the current range.}
16101 (while @var{length-element-smaller-than-top-of-range}
16102 (setq number-within-range (1+ number-within-range))
16103 (setq sorted-lengths (cdr sorted-lengths)))
16105 ;; @r{Move to next range.}
16106 (setq top-of-ranges (cdr top-of-ranges)))
16110 In addition, in each circuit of the outer loop, Emacs should record
16111 the number of definitions within that range (the value of
16112 @code{number-within-range}) in a list. We can use @code{cons} for
16113 this purpose. (@xref{cons, , @code{cons}}.)
16115 The @code{cons} function works fine, except that the list it
16116 constructs will contain the number of definitions for the highest
16117 range at its beginning and the number of definitions for the lowest
16118 range at its end. This is because @code{cons} attaches new elements
16119 of the list to the beginning of the list, and since the two loops are
16120 working their way through the lengths' list from the lower end first,
16121 the @code{defuns-per-range-list} will end up largest number first.
16122 But we will want to print our graph with smallest values first and the
16123 larger later. The solution is to reverse the order of the
16124 @code{defuns-per-range-list}. We can do this using the
16125 @code{nreverse} function, which reverses the order of a list.
16132 (nreverse '(1 2 3 4))
16143 Note that the @code{nreverse} function is ``destructive''---that is,
16144 it changes the list to which it is applied; this contrasts with the
16145 @code{car} and @code{cdr} functions, which are non-destructive. In
16146 this case, we do not want the original @code{defuns-per-range-list},
16147 so it does not matter that it is destroyed. (The @code{reverse}
16148 function provides a reversed copy of a list, leaving the original list
16153 Put all together, the @code{defuns-per-range} looks like this:
16157 (defun defuns-per-range (sorted-lengths top-of-ranges)
16158 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16159 (let ((top-of-range (car top-of-ranges))
16160 (number-within-range 0)
16161 defuns-per-range-list)
16166 (while top-of-ranges
16172 ;; @r{Need number for numeric test.}
16173 (car sorted-lengths)
16174 (< (car sorted-lengths) top-of-range))
16178 ;; @r{Count number of definitions within current range.}
16179 (setq number-within-range (1+ number-within-range))
16180 (setq sorted-lengths (cdr sorted-lengths)))
16182 ;; @r{Exit inner loop but remain within outer loop.}
16186 (setq defuns-per-range-list
16187 (cons number-within-range defuns-per-range-list))
16188 (setq number-within-range 0) ; @r{Reset count to zero.}
16192 ;; @r{Move to next range.}
16193 (setq top-of-ranges (cdr top-of-ranges))
16194 ;; @r{Specify next top of range value.}
16195 (setq top-of-range (car top-of-ranges)))
16199 ;; @r{Exit outer loop and count the number of defuns larger than}
16200 ;; @r{ the largest top-of-range value.}
16201 (setq defuns-per-range-list
16203 (length sorted-lengths)
16204 defuns-per-range-list))
16208 ;; @r{Return a list of the number of definitions within each range,}
16209 ;; @r{ smallest to largest.}
16210 (nreverse defuns-per-range-list)))
16216 The function is straightforward except for one subtle feature. The
16217 true-or-false test of the inner loop looks like this:
16221 (and (car sorted-lengths)
16222 (< (car sorted-lengths) top-of-range))
16228 instead of like this:
16231 (< (car sorted-lengths) top-of-range)
16234 The purpose of the test is to determine whether the first item in the
16235 @code{sorted-lengths} list is less than the value of the top of the
16238 The simple version of the test works fine unless the
16239 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16240 @code{(car sorted-lengths)} expression function returns
16241 @code{nil}. The @code{<} function cannot compare a number to
16242 @code{nil}, which is an empty list, so Emacs signals an error and
16243 stops the function from attempting to continue to execute.
16245 The @code{sorted-lengths} list always becomes @code{nil} when the
16246 counter reaches the end of the list. This means that any attempt to
16247 use the @code{defuns-per-range} function with the simple version of
16248 the test will fail.
16250 We solve the problem by using the @code{(car sorted-lengths)}
16251 expression in conjunction with the @code{and} expression. The
16252 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16253 value so long as the list has at least one number within it, but
16254 returns @code{nil} if the list is empty. The @code{and} expression
16255 first evaluates the @code{(car sorted-lengths)} expression, and
16256 if it is @code{nil}, returns false @emph{without} evaluating the
16257 @code{<} expression. But if the @code{(car sorted-lengths)}
16258 expression returns a non-@code{nil} value, the @code{and} expression
16259 evaluates the @code{<} expression, and returns that value as the value
16260 of the @code{and} expression.
16262 @c colon in printed section title causes problem in Info cross reference
16263 This way, we avoid an error.
16266 (For information about @code{and}, see
16267 @ref{kill-new function, , The @code{kill-new} function}.)
16271 (@xref{kill-new function, , The @code{kill-new} function}, for
16272 information about @code{and}.)
16275 Here is a short test of the @code{defuns-per-range} function. First,
16276 evaluate the expression that binds (a shortened)
16277 @code{top-of-ranges} list to the list of values, then evaluate the
16278 expression for binding the @code{sorted-lengths} list, and then
16279 evaluate the @code{defuns-per-range} function.
16283 ;; @r{(Shorter list than we will use later.)}
16284 (setq top-of-ranges
16285 '(110 120 130 140 150
16286 160 170 180 190 200))
16288 (setq sorted-lengths
16289 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16291 (defuns-per-range sorted-lengths top-of-ranges)
16297 The list returned looks like this:
16300 (2 2 2 0 0 1 0 2 0 0 4)
16304 Indeed, there are two elements of the @code{sorted-lengths} list
16305 smaller than 110, two elements between 110 and 119, two elements
16306 between 120 and 129, and so on. There are four elements with a value
16309 @c The next step is to turn this numbers' list into a graph.
16310 @node Readying a Graph, Emacs Initialization, Words in a defun, Top
16311 @chapter Readying a Graph
16312 @cindex Readying a graph
16313 @cindex Graph prototype
16314 @cindex Prototype graph
16315 @cindex Body of graph
16317 Our goal is to construct a graph showing the numbers of function
16318 definitions of various lengths in the Emacs lisp sources.
16320 As a practical matter, if you were creating a graph, you would
16321 probably use a program such as @code{gnuplot} to do the job.
16322 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16323 however, we create one from scratch, and in the process we will
16324 re-acquaint ourselves with some of what we learned before and learn
16327 In this chapter, we will first write a simple graph printing function.
16328 This first definition will be a @dfn{prototype}, a rapidly written
16329 function that enables us to reconnoiter this unknown graph-making
16330 territory. We will discover dragons, or find that they are myth.
16331 After scouting the terrain, we will feel more confident and enhance
16332 the function to label the axes automatically.
16335 * Columns of a graph::
16336 * graph-body-print:: How to print the body of a graph.
16337 * recursive-graph-body-print::
16339 * Line Graph Exercise::
16342 @node Columns of a graph, graph-body-print, Readying a Graph, Readying a Graph
16344 @unnumberedsec Printing the Columns of a Graph
16347 Since Emacs is designed to be flexible and work with all kinds of
16348 terminals, including character-only terminals, the graph will need to
16349 be made from one of the `typewriter' symbols. An asterisk will do; as
16350 we enhance the graph-printing function, we can make the choice of
16351 symbol a user option.
16353 We can call this function @code{graph-body-print}; it will take a
16354 @code{numbers-list} as its only argument. At this stage, we will not
16355 label the graph, but only print its body.
16357 The @code{graph-body-print} function inserts a vertical column of
16358 asterisks for each element in the @code{numbers-list}. The height of
16359 each line is determined by the value of that element of the
16360 @code{numbers-list}.
16362 Inserting columns is a repetitive act; that means that this function can
16363 be written either with a @code{while} loop or recursively.
16365 Our first challenge is to discover how to print a column of asterisks.
16366 Usually, in Emacs, we print characters onto a screen horizontally,
16367 line by line, by typing. We have two routes we can follow: write our
16368 own column-insertion function or discover whether one exists in Emacs.
16370 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16371 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16372 command, except that the latter finds only those functions that are
16373 commands. The @kbd{M-x apropos} command lists all symbols that match
16374 a regular expression, including functions that are not interactive.
16377 What we want to look for is some command that prints or inserts
16378 columns. Very likely, the name of the function will contain either
16379 the word `print' or the word `insert' or the word `column'.
16380 Therefore, we can simply type @kbd{M-x apropos RET
16381 print\|insert\|column RET} and look at the result. On my system, this
16382 command once too takes quite some time, and then produced a list of 79
16383 functions and variables. Now it does not take much time at all and
16384 produces a list of 211 functions and variables. Scanning down the
16385 list, the only function that looks as if it might do the job is
16386 @code{insert-rectangle}.
16389 Indeed, this is the function we want; its documentation says:
16394 Insert text of RECTANGLE with upper left corner at point.
16395 RECTANGLE's first line is inserted at point,
16396 its second line is inserted at a point vertically under point, etc.
16397 RECTANGLE should be a list of strings.
16398 After this command, the mark is at the upper left corner
16399 and point is at the lower right corner.
16403 We can run a quick test, to make sure it does what we expect of it.
16405 Here is the result of placing the cursor after the
16406 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16407 (@code{eval-last-sexp}). The function inserts the strings
16408 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16409 point. Also the function returns @code{nil}.
16413 (insert-rectangle '("first" "second" "third"))first
16420 Of course, we won't be inserting the text of the
16421 @code{insert-rectangle} expression itself into the buffer in which we
16422 are making the graph, but will call the function from our program. We
16423 shall, however, have to make sure that point is in the buffer at the
16424 place where the @code{insert-rectangle} function will insert its
16427 If you are reading this in Info, you can see how this works by
16428 switching to another buffer, such as the @file{*scratch*} buffer,
16429 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16430 @code{insert-rectangle} expression into the minibuffer at the prompt,
16431 and then typing @key{RET}. This causes Emacs to evaluate the
16432 expression in the minibuffer, but to use as the value of point the
16433 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16434 keybinding for @code{eval-expression}. Also, @code{nil} does not
16435 appear in the @file{*scratch*} buffer since the expression is
16436 evaluated in the minibuffer.)
16438 We find when we do this that point ends up at the end of the last
16439 inserted line---that is to say, this function moves point as a
16440 side-effect. If we were to repeat the command, with point at this
16441 position, the next insertion would be below and to the right of the
16442 previous insertion. We don't want this! If we are going to make a
16443 bar graph, the columns need to be beside each other.
16445 So we discover that each cycle of the column-inserting @code{while}
16446 loop must reposition point to the place we want it, and that place
16447 will be at the top, not the bottom, of the column. Moreover, we
16448 remember that when we print a graph, we do not expect all the columns
16449 to be the same height. This means that the top of each column may be
16450 at a different height from the previous one. We cannot simply
16451 reposition point to the same line each time, but moved over to the
16452 right---or perhaps we can@dots{}
16454 We are planning to make the columns of the bar graph out of asterisks.
16455 The number of asterisks in the column is the number specified by the
16456 current element of the @code{numbers-list}. We need to construct a
16457 list of asterisks of the right length for each call to
16458 @code{insert-rectangle}. If this list consists solely of the requisite
16459 number of asterisks, then we will have position point the right number
16460 of lines above the base for the graph to print correctly. This could
16463 Alternatively, if we can figure out some way to pass
16464 @code{insert-rectangle} a list of the same length each time, then we
16465 can place point on the same line each time, but move it over one
16466 column to the right for each new column. If we do this, however, some
16467 of the entries in the list passed to @code{insert-rectangle} must be
16468 blanks rather than asterisks. For example, if the maximum height of
16469 the graph is 5, but the height of the column is 3, then
16470 @code{insert-rectangle} requires an argument that looks like this:
16473 (" " " " "*" "*" "*")
16476 This last proposal is not so difficult, so long as we can determine
16477 the column height. There are two ways for us to specify the column
16478 height: we can arbitrarily state what it will be, which would work
16479 fine for graphs of that height; or we can search through the list of
16480 numbers and use the maximum height of the list as the maximum height
16481 of the graph. If the latter operation were difficult, then the former
16482 procedure would be easiest, but there is a function built into Emacs
16483 that determines the maximum of its arguments. We can use that
16484 function. The function is called @code{max} and it returns the
16485 largest of all its arguments, which must be numbers. Thus, for
16493 returns 7. (A corresponding function called @code{min} returns the
16494 smallest of all its arguments.)
16498 However, we cannot simply call @code{max} on the @code{numbers-list};
16499 the @code{max} function expects numbers as its argument, not a list of
16500 numbers. Thus, the following expression,
16503 (max '(3 4 6 5 7 3))
16508 produces the following error message;
16511 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16515 We need a function that passes a list of arguments to a function.
16516 This function is @code{apply}. This function `applies' its first
16517 argument (a function) to its remaining arguments, the last of which
16524 (apply 'max 3 4 7 3 '(4 8 5))
16530 (Incidentally, I don't know how you would learn of this function
16531 without a book such as this. It is possible to discover other
16532 functions, like @code{search-forward} or @code{insert-rectangle}, by
16533 guessing at a part of their names and then using @code{apropos}. Even
16534 though its base in metaphor is clear---`apply' its first argument to
16535 the rest---I doubt a novice would come up with that particular word
16536 when using @code{apropos} or other aid. Of course, I could be wrong;
16537 after all, the function was first named by someone who had to invent
16540 The second and subsequent arguments to @code{apply} are optional, so
16541 we can use @code{apply} to call a function and pass the elements of a
16542 list to it, like this, which also returns 8:
16545 (apply 'max '(4 8 5))
16548 This latter way is how we will use @code{apply}. The
16549 @code{recursive-lengths-list-many-files} function returns a numbers'
16550 list to which we can apply @code{max} (we could also apply @code{max} to
16551 the sorted numbers' list; it does not matter whether the list is
16555 Hence, the operation for finding the maximum height of the graph is this:
16558 (setq max-graph-height (apply 'max numbers-list))
16561 Now we can return to the question of how to create a list of strings
16562 for a column of the graph. Told the maximum height of the graph
16563 and the number of asterisks that should appear in the column, the
16564 function should return a list of strings for the
16565 @code{insert-rectangle} command to insert.
16567 Each column is made up of asterisks or blanks. Since the function is
16568 passed the value of the height of the column and the number of
16569 asterisks in the column, the number of blanks can be found by
16570 subtracting the number of asterisks from the height of the column.
16571 Given the number of blanks and the number of asterisks, two
16572 @code{while} loops can be used to construct the list:
16576 ;;; @r{First version.}
16577 (defun column-of-graph (max-graph-height actual-height)
16578 "Return list of strings that is one column of a graph."
16579 (let ((insert-list nil)
16580 (number-of-top-blanks
16581 (- max-graph-height actual-height)))
16585 ;; @r{Fill in asterisks.}
16586 (while (> actual-height 0)
16587 (setq insert-list (cons "*" insert-list))
16588 (setq actual-height (1- actual-height)))
16592 ;; @r{Fill in blanks.}
16593 (while (> number-of-top-blanks 0)
16594 (setq insert-list (cons " " insert-list))
16595 (setq number-of-top-blanks
16596 (1- number-of-top-blanks)))
16600 ;; @r{Return whole list.}
16605 If you install this function and then evaluate the following
16606 expression you will see that it returns the list as desired:
16609 (column-of-graph 5 3)
16617 (" " " " "*" "*" "*")
16620 As written, @code{column-of-graph} contains a major flaw: the symbols
16621 used for the blank and for the marked entries in the column are
16622 `hard-coded' as a space and asterisk. This is fine for a prototype,
16623 but you, or another user, may wish to use other symbols. For example,
16624 in testing the graph function, you many want to use a period in place
16625 of the space, to make sure the point is being repositioned properly
16626 each time the @code{insert-rectangle} function is called; or you might
16627 want to substitute a @samp{+} sign or other symbol for the asterisk.
16628 You might even want to make a graph-column that is more than one
16629 display column wide. The program should be more flexible. The way to
16630 do that is to replace the blank and the asterisk with two variables
16631 that we can call @code{graph-blank} and @code{graph-symbol} and define
16632 those variables separately.
16634 Also, the documentation is not well written. These considerations
16635 lead us to the second version of the function:
16639 (defvar graph-symbol "*"
16640 "String used as symbol in graph, usually an asterisk.")
16644 (defvar graph-blank " "
16645 "String used as blank in graph, usually a blank space.
16646 graph-blank must be the same number of columns wide
16652 (For an explanation of @code{defvar}, see
16653 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16657 ;;; @r{Second version.}
16658 (defun column-of-graph (max-graph-height actual-height)
16659 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16663 The graph-symbols are contiguous entries at the end
16665 The list will be inserted as one column of a graph.
16666 The strings are either graph-blank or graph-symbol."
16670 (let ((insert-list nil)
16671 (number-of-top-blanks
16672 (- max-graph-height actual-height)))
16676 ;; @r{Fill in @code{graph-symbols}.}
16677 (while (> actual-height 0)
16678 (setq insert-list (cons graph-symbol insert-list))
16679 (setq actual-height (1- actual-height)))
16683 ;; @r{Fill in @code{graph-blanks}.}
16684 (while (> number-of-top-blanks 0)
16685 (setq insert-list (cons graph-blank insert-list))
16686 (setq number-of-top-blanks
16687 (1- number-of-top-blanks)))
16689 ;; @r{Return whole list.}
16694 If we wished, we could rewrite @code{column-of-graph} a third time to
16695 provide optionally for a line graph as well as for a bar graph. This
16696 would not be hard to do. One way to think of a line graph is that it
16697 is no more than a bar graph in which the part of each bar that is
16698 below the top is blank. To construct a column for a line graph, the
16699 function first constructs a list of blanks that is one shorter than
16700 the value, then it uses @code{cons} to attach a graph symbol to the
16701 list; then it uses @code{cons} again to attach the `top blanks' to
16704 It is easy to see how to write such a function, but since we don't
16705 need it, we will not do it. But the job could be done, and if it were
16706 done, it would be done with @code{column-of-graph}. Even more
16707 important, it is worth noting that few changes would have to be made
16708 anywhere else. The enhancement, if we ever wish to make it, is
16711 Now, finally, we come to our first actual graph printing function.
16712 This prints the body of a graph, not the labels for the vertical and
16713 horizontal axes, so we can call this @code{graph-body-print}.
16715 @node graph-body-print, recursive-graph-body-print, Columns of a graph, Readying a Graph
16716 @section The @code{graph-body-print} Function
16717 @findex graph-body-print
16719 After our preparation in the preceding section, the
16720 @code{graph-body-print} function is straightforward. The function
16721 will print column after column of asterisks and blanks, using the
16722 elements of a numbers' list to specify the number of asterisks in each
16723 column. This is a repetitive act, which means we can use a
16724 decrementing @code{while} loop or recursive function for the job. In
16725 this section, we will write the definition using a @code{while} loop.
16727 The @code{column-of-graph} function requires the height of the graph
16728 as an argument, so we should determine and record that as a local variable.
16730 This leads us to the following template for the @code{while} loop
16731 version of this function:
16735 (defun graph-body-print (numbers-list)
16736 "@var{documentation}@dots{}"
16737 (let ((height @dots{}
16742 (while numbers-list
16743 @var{insert-columns-and-reposition-point}
16744 (setq numbers-list (cdr numbers-list)))))
16749 We need to fill in the slots of the template.
16751 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16752 determine the height of the graph.
16754 The @code{while} loop will cycle through the @code{numbers-list} one
16755 element at a time. As it is shortened by the @code{(setq numbers-list
16756 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16757 list is the value of the argument for @code{column-of-graph}.
16759 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16760 function inserts the list returned by @code{column-of-graph}. Since
16761 the @code{insert-rectangle} function moves point to the lower right of
16762 the inserted rectangle, we need to save the location of point at the
16763 time the rectangle is inserted, move back to that position after the
16764 rectangle is inserted, and then move horizontally to the next place
16765 from which @code{insert-rectangle} is called.
16767 If the inserted columns are one character wide, as they will be if
16768 single blanks and asterisks are used, the repositioning command is
16769 simply @code{(forward-char 1)}; however, the width of a column may be
16770 greater than one. This means that the repositioning command should be
16771 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16772 itself is the length of a @code{graph-blank} and can be found using
16773 the expression @code{(length graph-blank)}. The best place to bind
16774 the @code{symbol-width} variable to the value of the width of graph
16775 column is in the varlist of the @code{let} expression.
16778 These considerations lead to the following function definition:
16782 (defun graph-body-print (numbers-list)
16783 "Print a bar graph of the NUMBERS-LIST.
16784 The numbers-list consists of the Y-axis values."
16786 (let ((height (apply 'max numbers-list))
16787 (symbol-width (length graph-blank))
16792 (while numbers-list
16793 (setq from-position (point))
16795 (column-of-graph height (car numbers-list)))
16796 (goto-char from-position)
16797 (forward-char symbol-width)
16800 ;; @r{Draw graph column by column.}
16802 (setq numbers-list (cdr numbers-list)))
16805 ;; @r{Place point for X axis labels.}
16806 (forward-line height)
16813 The one unexpected expression in this function is the
16814 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16815 expression makes the graph printing operation more interesting to
16816 watch than it would be otherwise. The expression causes Emacs to
16817 `sit' or do nothing for a zero length of time and then redraw the
16818 screen. Placed here, it causes Emacs to redraw the screen column by
16819 column. Without it, Emacs would not redraw the screen until the
16822 We can test @code{graph-body-print} with a short list of numbers.
16826 Install @code{graph-symbol}, @code{graph-blank},
16827 @code{column-of-graph}, which are in
16829 @ref{Readying a Graph, , Readying a Graph},
16832 @ref{Columns of a graph},
16834 and @code{graph-body-print}.
16838 Copy the following expression:
16841 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16845 Switch to the @file{*scratch*} buffer and place the cursor where you
16846 want the graph to start.
16849 Type @kbd{M-:} (@code{eval-expression}).
16852 Yank the @code{graph-body-print} expression into the minibuffer
16853 with @kbd{C-y} (@code{yank)}.
16856 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16860 Emacs will print a graph like this:
16874 @node recursive-graph-body-print, Printed Axes, graph-body-print, Readying a Graph
16875 @section The @code{recursive-graph-body-print} Function
16876 @findex recursive-graph-body-print
16878 The @code{graph-body-print} function may also be written recursively.
16879 The recursive solution is divided into two parts: an outside `wrapper'
16880 that uses a @code{let} expression to determine the values of several
16881 variables that need only be found once, such as the maximum height of
16882 the graph, and an inside function that is called recursively to print
16886 The `wrapper' is uncomplicated:
16890 (defun recursive-graph-body-print (numbers-list)
16891 "Print a bar graph of the NUMBERS-LIST.
16892 The numbers-list consists of the Y-axis values."
16893 (let ((height (apply 'max numbers-list))
16894 (symbol-width (length graph-blank))
16896 (recursive-graph-body-print-internal
16903 The recursive function is a little more difficult. It has four parts:
16904 the `do-again-test', the printing code, the recursive call, and the
16905 `next-step-expression'. The `do-again-test' is a @code{when}
16906 expression that determines whether the @code{numbers-list} contains
16907 any remaining elements; if it does, the function prints one column of
16908 the graph using the printing code and calls itself again. The
16909 function calls itself again according to the value produced by the
16910 `next-step-expression' which causes the call to act on a shorter
16911 version of the @code{numbers-list}.
16915 (defun recursive-graph-body-print-internal
16916 (numbers-list height symbol-width)
16917 "Print a bar graph.
16918 Used within recursive-graph-body-print function."
16923 (setq from-position (point))
16925 (column-of-graph height (car numbers-list)))
16928 (goto-char from-position)
16929 (forward-char symbol-width)
16930 (sit-for 0) ; @r{Draw graph column by column.}
16931 (recursive-graph-body-print-internal
16932 (cdr numbers-list) height symbol-width)))
16937 After installation, this expression can be tested; here is a sample:
16940 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16944 Here is what @code{recursive-graph-body-print} produces:
16958 Either of these two functions, @code{graph-body-print} or
16959 @code{recursive-graph-body-print}, create the body of a graph.
16961 @node Printed Axes, Line Graph Exercise, recursive-graph-body-print, Readying a Graph
16962 @section Need for Printed Axes
16964 A graph needs printed axes, so you can orient yourself. For a do-once
16965 project, it may be reasonable to draw the axes by hand using Emacs's
16966 Picture mode; but a graph drawing function may be used more than once.
16968 For this reason, I have written enhancements to the basic
16969 @code{print-graph-body} function that automatically print labels for
16970 the horizontal and vertical axes. Since the label printing functions
16971 do not contain much new material, I have placed their description in
16972 an appendix. @xref{Full Graph, , A Graph with Labeled Axes}.
16974 @node Line Graph Exercise, , Printed Axes, Readying a Graph
16977 Write a line graph version of the graph printing functions.
16979 @node Emacs Initialization, Debugging, Readying a Graph, Top
16980 @chapter Your @file{.emacs} File
16981 @cindex @file{.emacs} file
16982 @cindex Customizing your @file{.emacs} file
16983 @cindex Initialization file
16985 ``You don't have to like Emacs to like it'' -- this seemingly
16986 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16987 the box' Emacs is a generic tool. Most people who use it, customize
16988 it to suit themselves.
16990 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16991 expressions in Emacs Lisp you can change or extend Emacs.
16994 * Default Configuration::
16995 * Site-wide Init:: You can write site-wide init files.
16996 * defcustom:: Emacs will write code for you.
16997 * Beginning a .emacs File:: How to write a @code{.emacs file}.
16998 * Text and Auto-fill:: Automatically wrap lines.
16999 * Mail Aliases:: Use abbreviations for email addresses.
17000 * Indent Tabs Mode:: Don't use tabs with @TeX{}
17001 * Keybindings:: Create some personal keybindings.
17002 * Keymaps:: More about key binding.
17003 * Loading Files:: Load (i.e., evaluate) files automatically.
17004 * Autoload:: Make functions available.
17005 * Simple Extension:: Define a function; bind it to a key.
17006 * X11 Colors:: Colors in X.
17008 * Mode Line:: How to customize your mode line.
17011 @node Default Configuration, Site-wide Init, Emacs Initialization, Emacs Initialization
17013 @unnumberedsec Emacs's Default Configuration
17016 There are those who appreciate Emacs's default configuration. After
17017 all, Emacs starts you in C mode when you edit a C file, starts you in
17018 Fortran mode when you edit a Fortran file, and starts you in
17019 Fundamental mode when you edit an unadorned file. This all makes
17020 sense, if you do not know who is going to use Emacs. Who knows what a
17021 person hopes to do with an unadorned file? Fundamental mode is the
17022 right default for such a file, just as C mode is the right default for
17023 editing C code. (Enough programming languages have syntaxes
17024 that enable them to share or nearly share features, so C mode is
17025 now provided by CC mode, the `C Collection'.)
17027 But when you do know who is going to use Emacs---you,
17028 yourself---then it makes sense to customize Emacs.
17030 For example, I seldom want Fundamental mode when I edit an
17031 otherwise undistinguished file; I want Text mode. This is why I
17032 customize Emacs: so it suits me.
17034 You can customize and extend Emacs by writing or adapting a
17035 @file{~/.emacs} file. This is your personal initialization file; its
17036 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
17037 may also add @file{.el} to @file{~/.emacs} and call it a
17038 @file{~/.emacs.el} file. In the past, you were forbidden to type the
17039 extra keystrokes that the name @file{~/.emacs.el} requires, but now
17040 you may. The new format is consistent with the Emacs Lisp file
17041 naming conventions; the old format saves typing.}
17043 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
17044 code yourself; or you can use Emacs's @code{customize} feature to write
17045 the code for you. You can combine your own expressions and
17046 auto-written Customize expressions in your @file{.emacs} file.
17048 (I myself prefer to write my own expressions, except for those,
17049 particularly fonts, that I find easier to manipulate using the
17050 @code{customize} command. I combine the two methods.)
17052 Most of this chapter is about writing expressions yourself. It
17053 describes a simple @file{.emacs} file; for more information, see
17054 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
17055 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
17058 @node Site-wide Init, defcustom, Default Configuration, Emacs Initialization
17059 @section Site-wide Initialization Files
17061 @cindex @file{default.el} init file
17062 @cindex @file{site-init.el} init file
17063 @cindex @file{site-load.el} init file
17064 In addition to your personal initialization file, Emacs automatically
17065 loads various site-wide initialization files, if they exist. These
17066 have the same form as your @file{.emacs} file, but are loaded by
17069 Two site-wide initialization files, @file{site-load.el} and
17070 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
17071 `dumped' version of Emacs is created, as is most common. (Dumped
17072 copies of Emacs load more quickly. However, once a file is loaded and
17073 dumped, a change to it does not lead to a change in Emacs unless you
17074 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
17075 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
17076 @file{INSTALL} file.)
17078 Three other site-wide initialization files are loaded automatically
17079 each time you start Emacs, if they exist. These are
17080 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
17081 file, and @file{default.el}, and the terminal type file, which are both
17082 loaded @emph{after} your @file{.emacs} file.
17084 Settings and definitions in your @file{.emacs} file will overwrite
17085 conflicting settings and definitions in a @file{site-start.el} file,
17086 if it exists; but the settings and definitions in a @file{default.el}
17087 or terminal type file will overwrite those in your @file{.emacs} file.
17088 (You can prevent interference from a terminal type file by setting
17089 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
17090 Simple Extension}.)
17092 @c Rewritten to avoid overfull hbox.
17093 The @file{INSTALL} file that comes in the distribution contains
17094 descriptions of the @file{site-init.el} and @file{site-load.el} files.
17096 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
17097 control loading. These files are in the @file{lisp} directory of the
17098 Emacs distribution and are worth perusing.
17100 The @file{loaddefs.el} file contains a good many suggestions as to
17101 what to put into your own @file{.emacs} file, or into a site-wide
17102 initialization file.
17104 @node defcustom, Beginning a .emacs File, Site-wide Init, Emacs Initialization
17105 @section Specifying Variables using @code{defcustom}
17108 You can specify variables using @code{defcustom} so that you and
17109 others can then use Emacs's @code{customize} feature to set their
17110 values. (You cannot use @code{customize} to write function
17111 definitions; but you can write @code{defuns} in your @file{.emacs}
17112 file. Indeed, you can write any Lisp expression in your @file{.emacs}
17115 The @code{customize} feature depends on the @code{defcustom} special
17116 form. Although you can use @code{defvar} or @code{setq} for variables
17117 that users set, the @code{defcustom} special form is designed for the
17120 You can use your knowledge of @code{defvar} for writing the
17121 first three arguments for @code{defcustom}. The first argument to
17122 @code{defcustom} is the name of the variable. The second argument is
17123 the variable's initial value, if any; and this value is set only if
17124 the value has not already been set. The third argument is the
17127 The fourth and subsequent arguments to @code{defcustom} specify types
17128 and options; these are not featured in @code{defvar}. (These
17129 arguments are optional.)
17131 Each of these arguments consists of a keyword followed by a value.
17132 Each keyword starts with the colon character @samp{:}.
17135 For example, the customizable user option variable
17136 @code{text-mode-hook} looks like this:
17140 (defcustom text-mode-hook nil
17141 "Normal hook run when entering Text mode and many related modes."
17143 :options '(turn-on-auto-fill flyspell-mode)
17149 The name of the variable is @code{text-mode-hook}; it has no default
17150 value; and its documentation string tells you what it does.
17152 The @code{:type} keyword tells Emacs the kind of data to which
17153 @code{text-mode-hook} should be set and how to display the value in a
17154 Customization buffer.
17156 The @code{:options} keyword specifies a suggested list of values for
17157 the variable. Usually, @code{:options} applies to a hook.
17158 The list is only a suggestion; it is not exclusive; a person who sets
17159 the variable may set it to other values; the list shown following the
17160 @code{:options} keyword is intended to offer convenient choices to a
17163 Finally, the @code{:group} keyword tells the Emacs Customization
17164 command in which group the variable is located. This tells where to
17167 The @code{defcustom} function recognizes more than a dozen keywords.
17168 For more information, see @ref{Customization, , Writing Customization
17169 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17171 Consider @code{text-mode-hook} as an example.
17173 There are two ways to customize this variable. You can use the
17174 customization command or write the appropriate expressions yourself.
17177 Using the customization command, you can type:
17184 and find that the group for editing files of data is called `data'.
17185 Enter that group. Text Mode Hook is the first member. You can click
17186 on its various options, such as @code{turn-on-auto-fill}, to set the
17187 values. After you click on the button to
17190 Save for Future Sessions
17194 Emacs will write an expression into your @file{.emacs} file.
17195 It will look like this:
17199 (custom-set-variables
17200 ;; custom-set-variables was added by Custom.
17201 ;; If you edit it by hand, you could mess it up, so be careful.
17202 ;; Your init file should contain only one such instance.
17203 ;; If there is more than one, they won't work right.
17204 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17209 (The @code{text-mode-hook-identify} function tells
17210 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17211 It comes on automatically.)
17213 The @code{custom-set-variables} function works somewhat differently
17214 than a @code{setq}. While I have never learned the differences, I
17215 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17216 file by hand: I make the changes in what appears to me to be a
17217 reasonable manner and have not had any problems. Others prefer to use
17218 the Customization command and let Emacs do the work for them.
17220 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17221 This function sets the various font faces. Over time, I have set a
17222 considerable number of faces. Some of the time, I re-set them using
17223 @code{customize}; other times, I simply edit the
17224 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17226 The second way to customize your @code{text-mode-hook} is to set it
17227 yourself in your @file{.emacs} file using code that has nothing to do
17228 with the @code{custom-set-@dots{}} functions.
17231 When you do this, and later use @code{customize}, you will see a
17235 CHANGED outside Customize; operating on it here may be unreliable.
17239 This message is only a warning. If you click on the button to
17242 Save for Future Sessions
17246 Emacs will write a @code{custom-set-@dots{}} expression near the end
17247 of your @file{.emacs} file that will be evaluated after your
17248 hand-written expression. It will, therefore, overrule your
17249 hand-written expression. No harm will be done. When you do this,
17250 however, be careful to remember which expression is active; if you
17251 forget, you may confuse yourself.
17253 So long as you remember where the values are set, you will have no
17254 trouble. In any event, the values are always set in your
17255 initialization file, which is usually called @file{.emacs}.
17257 I myself use @code{customize} for hardly anything. Mostly, I write
17258 expressions myself.
17262 Incidentally, to be more complete concerning defines: @code{defsubst}
17263 defines an inline function. The syntax is just like that of
17264 @code{defun}. @code{defconst} defines a symbol as a constant. The
17265 intent is that neither programs nor users should ever change a value
17266 set by @code{defconst}. (You can change it; the value set is a
17267 variable; but please do not.)
17269 @node Beginning a .emacs File, Text and Auto-fill, defcustom, Emacs Initialization
17270 @section Beginning a @file{.emacs} File
17271 @cindex @file{.emacs} file, beginning of
17273 When you start Emacs, it loads your @file{.emacs} file unless you tell
17274 it not to by specifying @samp{-q} on the command line. (The
17275 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17277 A @file{.emacs} file contains Lisp expressions. Often, these are no
17278 more than expressions to set values; sometimes they are function
17281 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17282 Manual}, for a short description of initialization files.
17284 This chapter goes over some of the same ground, but is a walk among
17285 extracts from a complete, long-used @file{.emacs} file---my own.
17287 The first part of the file consists of comments: reminders to myself.
17288 By now, of course, I remember these things, but when I started, I did
17294 ;;;; Bob's .emacs file
17295 ; Robert J. Chassell
17296 ; 26 September 1985
17301 Look at that date! I started this file a long time ago. I have been
17302 adding to it ever since.
17306 ; Each section in this file is introduced by a
17307 ; line beginning with four semicolons; and each
17308 ; entry is introduced by a line beginning with
17309 ; three semicolons.
17314 This describes the usual conventions for comments in Emacs Lisp.
17315 Everything on a line that follows a semicolon is a comment. Two,
17316 three, and four semicolons are used as subsection and section markers.
17317 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17318 more about comments.)
17323 ; Control-h is the help key;
17324 ; after typing control-h, type a letter to
17325 ; indicate the subject about which you want help.
17326 ; For an explanation of the help facility,
17327 ; type control-h two times in a row.
17332 Just remember: type @kbd{C-h} two times for help.
17336 ; To find out about any mode, type control-h m
17337 ; while in that mode. For example, to find out
17338 ; about mail mode, enter mail mode and then type
17344 `Mode help', as I call this, is very helpful. Usually, it tells you
17345 all you need to know.
17347 Of course, you don't need to include comments like these in your
17348 @file{.emacs} file. I included them in mine because I kept forgetting
17349 about Mode help or the conventions for comments---but I was able to
17350 remember to look here to remind myself.
17352 @node Text and Auto-fill, Mail Aliases, Beginning a .emacs File, Emacs Initialization
17353 @section Text and Auto Fill Mode
17355 Now we come to the part that `turns on' Text mode and
17360 ;;; Text mode and Auto Fill mode
17361 ;; The next two lines put Emacs into Text mode
17362 ;; and Auto Fill mode, and are for writers who
17363 ;; want to start writing prose rather than code.
17364 (setq-default major-mode 'text-mode)
17365 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17369 Here is the first part of this @file{.emacs} file that does something
17370 besides remind a forgetful human!
17372 The first of the two lines in parentheses tells Emacs to turn on Text
17373 mode when you find a file, @emph{unless} that file should go into some
17374 other mode, such as C mode.
17376 @cindex Per-buffer, local variables list
17377 @cindex Local variables list, per-buffer,
17378 @cindex Automatic mode selection
17379 @cindex Mode selection, automatic
17380 When Emacs reads a file, it looks at the extension to the file name,
17381 if any. (The extension is the part that comes after a @samp{.}.) If
17382 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17383 on C mode. Also, Emacs looks at first nonblank line of the file; if
17384 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17385 possesses a list of extensions and specifications that it uses
17386 automatically. In addition, Emacs looks near the last page for a
17387 per-buffer, ``local variables list'', if any.
17390 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17393 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17397 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17398 Files'' in @cite{The GNU Emacs Manual}.
17401 Now, back to the @file{.emacs} file.
17404 Here is the line again; how does it work?
17406 @cindex Text Mode turned on
17408 (setq major-mode 'text-mode)
17412 This line is a short, but complete Emacs Lisp expression.
17414 We are already familiar with @code{setq}. It sets the following variable,
17415 @code{major-mode}, to the subsequent value, which is @code{text-mode}.
17416 The single quote mark before @code{text-mode} tells Emacs to deal directly
17417 with the @code{text-mode} symbol, not with whatever it might stand for.
17418 @xref{set & setq, , Setting the Value of a Variable},
17419 for a reminder of how @code{setq} works.
17420 The main point is that there is no difference between the procedure you
17421 use to set a value in your @file{.emacs} file and the procedure you use
17422 anywhere else in Emacs.
17425 Here is the next line:
17427 @cindex Auto Fill mode turned on
17430 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17434 In this line, the @code{add-hook} command adds
17435 @code{turn-on-auto-fill} to the variable.
17437 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17438 it!, turns on Auto Fill mode.
17440 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17441 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17442 turns on Auto Fill mode.
17444 In brief, the first line causes Emacs to enter Text mode when you edit a
17445 file, unless the file name extension, a first non-blank line, or local
17446 variables to tell Emacs otherwise.
17448 Text mode among other actions, sets the syntax table to work
17449 conveniently for writers. In Text mode, Emacs considers an apostrophe
17450 as part of a word like a letter; but Emacs does not consider a period
17451 or a space as part of a word. Thus, @kbd{M-f} moves you over
17452 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17453 the @samp{t} of @samp{it's}.
17455 The second line causes Emacs to turn on Auto Fill mode when it turns
17456 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17457 that is too wide and brings the excessively wide part of the line down
17458 to the next line. Emacs breaks lines between words, not within them.
17460 When Auto Fill mode is turned off, lines continue to the right as you
17461 type them. Depending on how you set the value of
17462 @code{truncate-lines}, the words you type either disappear off the
17463 right side of the screen, or else are shown, in a rather ugly and
17464 unreadable manner, as a continuation line on the screen.
17467 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17468 fill commands to insert two spaces after a colon:
17471 (setq colon-double-space t)
17474 @node Mail Aliases, Indent Tabs Mode, Text and Auto-fill, Emacs Initialization
17475 @section Mail Aliases
17477 Here is a @code{setq} that `turns on' mail aliases, along with more
17483 ; To enter mail mode, type `C-x m'
17484 ; To enter RMAIL (for reading mail),
17486 (setq mail-aliases t)
17490 @cindex Mail aliases
17492 This @code{setq} command sets the value of the variable
17493 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17494 says, in effect, ``Yes, use mail aliases.''
17496 Mail aliases are convenient short names for long email addresses or
17497 for lists of email addresses. The file where you keep your `aliases'
17498 is @file{~/.mailrc}. You write an alias like this:
17501 alias geo george@@foobar.wiz.edu
17505 When you write a message to George, address it to @samp{geo}; the
17506 mailer will automatically expand @samp{geo} to the full address.
17508 @node Indent Tabs Mode, Keybindings, Mail Aliases, Emacs Initialization
17509 @section Indent Tabs Mode
17510 @cindex Tabs, preventing
17511 @findex indent-tabs-mode
17513 By default, Emacs inserts tabs in place of multiple spaces when it
17514 formats a region. (For example, you might indent many lines of text
17515 all at once with the @code{indent-region} command.) Tabs look fine on
17516 a terminal or with ordinary printing, but they produce badly indented
17517 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17520 The following turns off Indent Tabs mode:
17524 ;;; Prevent Extraneous Tabs
17525 (setq-default indent-tabs-mode nil)
17529 Note that this line uses @code{setq-default} rather than the
17530 @code{setq} command that we have seen before. The @code{setq-default}
17531 command sets values only in buffers that do not have their own local
17532 values for the variable.
17535 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17537 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17541 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17542 Files'' in @cite{The GNU Emacs Manual}.
17546 @node Keybindings, Keymaps, Indent Tabs Mode, Emacs Initialization
17547 @section Some Keybindings
17549 Now for some personal keybindings:
17553 ;;; Compare windows
17554 (global-set-key "\C-cw" 'compare-windows)
17558 @findex compare-windows
17559 @code{compare-windows} is a nifty command that compares the text in
17560 your current window with text in the next window. It makes the
17561 comparison by starting at point in each window, moving over text in
17562 each window as far as they match. I use this command all the time.
17564 This also shows how to set a key globally, for all modes.
17566 @cindex Setting a key globally
17567 @cindex Global set key
17568 @cindex Key setting globally
17569 @findex global-set-key
17570 The command is @code{global-set-key}. It is followed by the
17571 keybinding. In a @file{.emacs} file, the keybinding is written as
17572 shown: @code{\C-c} stands for `control-c', which means `press the
17573 control key and the @key{c} key at the same time'. The @code{w} means
17574 `press the @key{w} key'. The keybinding is surrounded by double
17575 quotation marks. In documentation, you would write this as
17576 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17577 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17578 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17579 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17582 The command invoked by the keys is @code{compare-windows}. Note that
17583 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17584 would first try to evaluate the symbol to determine its value.
17586 These three things, the double quotation marks, the backslash before
17587 the @samp{C}, and the single quote mark are necessary parts of
17588 keybinding that I tend to forget. Fortunately, I have come to
17589 remember that I should look at my existing @file{.emacs} file, and
17590 adapt what is there.
17592 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17593 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17594 set of keys, @kbd{C-c} followed by a single character, is strictly
17595 reserved for individuals' own use. (I call these `own' keys, since
17596 these are for my own use.) You should always be able to create such a
17597 keybinding for your own use without stomping on someone else's
17598 keybinding. If you ever write an extension to Emacs, please avoid
17599 taking any of these keys for public use. Create a key like @kbd{C-c
17600 C-w} instead. Otherwise, we will run out of `own' keys.
17603 Here is another keybinding, with a comment:
17607 ;;; Keybinding for `occur'
17608 ; I use occur a lot, so let's bind it to a key:
17609 (global-set-key "\C-co" 'occur)
17614 The @code{occur} command shows all the lines in the current buffer
17615 that contain a match for a regular expression. Matching lines are
17616 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17617 to jump to occurrences.
17619 @findex global-unset-key
17620 @cindex Unbinding key
17621 @cindex Key unbinding
17623 Here is how to unbind a key, so it does not
17629 (global-unset-key "\C-xf")
17633 There is a reason for this unbinding: I found I inadvertently typed
17634 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17635 file, as I intended, I accidentally set the width for filled text,
17636 almost always to a width I did not want. Since I hardly ever reset my
17637 default width, I simply unbound the key.
17639 @findex list-buffers, @r{rebound}
17640 @findex buffer-menu, @r{bound to key}
17642 The following rebinds an existing key:
17646 ;;; Rebind `C-x C-b' for `buffer-menu'
17647 (global-set-key "\C-x\C-b" 'buffer-menu)
17651 By default, @kbd{C-x C-b} runs the
17652 @code{list-buffers} command. This command lists
17653 your buffers in @emph{another} window. Since I
17654 almost always want to do something in that
17655 window, I prefer the @code{buffer-menu}
17656 command, which not only lists the buffers,
17657 but moves point into that window.
17659 @node Keymaps, Loading Files, Keybindings, Emacs Initialization
17662 @cindex Rebinding keys
17664 Emacs uses @dfn{keymaps} to record which keys call which commands.
17665 When you use @code{global-set-key} to set the keybinding for a single
17666 command in all parts of Emacs, you are specifying the keybinding in
17667 @code{current-global-map}.
17669 Specific modes, such as C mode or Text mode, have their own keymaps;
17670 the mode-specific keymaps override the global map that is shared by
17673 The @code{global-set-key} function binds, or rebinds, the global
17674 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17675 function @code{buffer-menu}:
17678 (global-set-key "\C-x\C-b" 'buffer-menu)
17681 Mode-specific keymaps are bound using the @code{define-key} function,
17682 which takes a specific keymap as an argument, as well as the key and
17683 the command. For example, my @file{.emacs} file contains the
17684 following expression to bind the @code{texinfo-insert-@@group} command
17685 to @kbd{C-c C-c g}:
17689 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17694 The @code{texinfo-insert-@@group} function itself is a little extension
17695 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17696 use this command all the time and prefer to type the three strokes
17697 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17698 (@samp{@@group} and its matching @samp{@@end group} are commands that
17699 keep all enclosed text together on one page; many multi-line examples
17700 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17703 Here is the @code{texinfo-insert-@@group} function definition:
17707 (defun texinfo-insert-@@group ()
17708 "Insert the string @@group in a Texinfo buffer."
17710 (beginning-of-line)
17711 (insert "@@group\n"))
17715 (Of course, I could have used Abbrev mode to save typing, rather than
17716 write a function to insert a word; but I prefer key strokes consistent
17717 with other Texinfo mode key bindings.)
17719 You will see numerous @code{define-key} expressions in
17720 @file{loaddefs.el} as well as in the various mode libraries, such as
17721 @file{cc-mode.el} and @file{lisp-mode.el}.
17723 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17724 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17725 Reference Manual}, for more information about keymaps.
17727 @node Loading Files, Autoload, Keymaps, Emacs Initialization
17728 @section Loading Files
17729 @cindex Loading files
17732 Many people in the GNU Emacs community have written extensions to
17733 Emacs. As time goes by, these extensions are often included in new
17734 releases. For example, the Calendar and Diary packages are now part
17735 of the standard GNU Emacs, as is Calc.
17737 You can use a @code{load} command to evaluate a complete file and
17738 thereby install all the functions and variables in the file into Emacs.
17741 @c (auto-compression-mode t)
17744 (load "~/emacs/slowsplit")
17747 This evaluates, i.e.@: loads, the @file{slowsplit.el} file or if it
17748 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17749 @file{emacs} sub-directory of your home directory. The file contains
17750 the function @code{split-window-quietly}, which John Robinson wrote in
17753 The @code{split-window-quietly} function splits a window with the
17754 minimum of redisplay. I installed it in 1989 because it worked well
17755 with the slow 1200 baud terminals I was then using. Nowadays, I only
17756 occasionally come across such a slow connection, but I continue to use
17757 the function because I like the way it leaves the bottom half of a
17758 buffer in the lower of the new windows and the top half in the upper
17762 To replace the key binding for the default
17763 @code{split-window-vertically}, you must also unset that key and bind
17764 the keys to @code{split-window-quietly}, like this:
17768 (global-unset-key "\C-x2")
17769 (global-set-key "\C-x2" 'split-window-quietly)
17774 If you load many extensions, as I do, then instead of specifying the
17775 exact location of the extension file, as shown above, you can specify
17776 that directory as part of Emacs's @code{load-path}. Then, when Emacs
17777 loads a file, it will search that directory as well as its default
17778 list of directories. (The default list is specified in @file{paths.h}
17779 when Emacs is built.)
17782 The following command adds your @file{~/emacs} directory to the
17783 existing load path:
17787 ;;; Emacs Load Path
17788 (setq load-path (cons "~/emacs" load-path))
17792 Incidentally, @code{load-library} is an interactive interface to the
17793 @code{load} function. The complete function looks like this:
17795 @findex load-library
17798 (defun load-library (library)
17799 "Load the library named LIBRARY.
17800 This is an interface to the function `load'."
17802 (list (completing-read "Load library: "
17803 (apply-partially 'locate-file-completion-table
17805 (get-load-suffixes)))))
17810 The name of the function, @code{load-library}, comes from the use of
17811 `library' as a conventional synonym for `file'. The source for the
17812 @code{load-library} command is in the @file{files.el} library.
17814 Another interactive command that does a slightly different job is
17815 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17816 Emacs, emacs, The GNU Emacs Manual}, for information on the
17817 distinction between @code{load-library} and this command.
17819 @node Autoload, Simple Extension, Loading Files, Emacs Initialization
17820 @section Autoloading
17823 Instead of installing a function by loading the file that contains it,
17824 or by evaluating the function definition, you can make the function
17825 available but not actually install it until it is first called. This
17826 is called @dfn{autoloading}.
17828 When you execute an autoloaded function, Emacs automatically evaluates
17829 the file that contains the definition, and then calls the function.
17831 Emacs starts quicker with autoloaded functions, since their libraries
17832 are not loaded right away; but you need to wait a moment when you
17833 first use such a function, while its containing file is evaluated.
17835 Rarely used functions are frequently autoloaded. The
17836 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17837 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17838 come to use a `rare' function frequently. When you do, you should
17839 load that function's file with a @code{load} expression in your
17840 @file{.emacs} file.
17842 In my @file{.emacs} file, I load 14 libraries that contain functions
17843 that would otherwise be autoloaded. (Actually, it would have been
17844 better to include these files in my `dumped' Emacs, but I forgot.
17845 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17846 Reference Manual}, and the @file{INSTALL} file for more about
17849 You may also want to include autoloaded expressions in your @file{.emacs}
17850 file. @code{autoload} is a built-in function that takes up to five
17851 arguments, the final three of which are optional. The first argument
17852 is the name of the function to be autoloaded; the second is the name
17853 of the file to be loaded. The third argument is documentation for the
17854 function, and the fourth tells whether the function can be called
17855 interactively. The fifth argument tells what type of
17856 object---@code{autoload} can handle a keymap or macro as well as a
17857 function (the default is a function).
17860 Here is a typical example:
17864 (autoload 'html-helper-mode
17865 "html-helper-mode" "Edit HTML documents" t)
17870 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17871 which is a standard part of the distribution.)
17874 This expression autoloads the @code{html-helper-mode} function. It
17875 takes it from the @file{html-helper-mode.el} file (or from the byte
17876 compiled version @file{html-helper-mode.elc}, if that exists.) The
17877 file must be located in a directory specified by @code{load-path}.
17878 The documentation says that this is a mode to help you edit documents
17879 written in the HyperText Markup Language. You can call this mode
17880 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17881 duplicate the function's regular documentation in the autoload
17882 expression because the regular function is not yet loaded, so its
17883 documentation is not available.)
17885 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17886 Manual}, for more information.
17888 @node Simple Extension, X11 Colors, Autoload, Emacs Initialization
17889 @section A Simple Extension: @code{line-to-top-of-window}
17890 @findex line-to-top-of-window
17891 @cindex Simple extension in @file{.emacs} file
17893 Here is a simple extension to Emacs that moves the line point is on to
17894 the top of the window. I use this all the time, to make text easier
17897 You can put the following code into a separate file and then load it
17898 from your @file{.emacs} file, or you can include it within your
17899 @file{.emacs} file.
17902 Here is the definition:
17906 ;;; Line to top of window;
17907 ;;; replace three keystroke sequence C-u 0 C-l
17908 (defun line-to-top-of-window ()
17909 "Move the line point is on to top of window."
17916 Now for the keybinding.
17918 Nowadays, function keys as well as mouse button events and
17919 non-@sc{ascii} characters are written within square brackets, without
17920 quotation marks. (In Emacs version 18 and before, you had to write
17921 different function key bindings for each different make of terminal.)
17923 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17927 (global-set-key [f6] 'line-to-top-of-window)
17930 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17931 Your Init File, emacs, The GNU Emacs Manual}.
17933 @cindex Conditional 'twixt two versions of Emacs
17934 @cindex Version of Emacs, choosing
17935 @cindex Emacs version, choosing
17936 If you run two versions of GNU Emacs, such as versions 22 and 23, and
17937 use one @file{.emacs} file, you can select which code to evaluate with
17938 the following conditional:
17943 ((= 22 emacs-major-version)
17944 ;; evaluate version 22 code
17946 ((= 23 emacs-major-version)
17947 ;; evaluate version 23 code
17952 For example, in contrast to version 20, more recent versions blink
17953 their cursors by default. I hate such blinking, as well as other
17954 features, so I placed the following in my @file{.emacs}
17955 file@footnote{When I start instances of Emacs that do not load my
17956 @file{.emacs} file or any site file, I also turn off blinking:
17959 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17961 @exdent Or nowadays, using an even more sophisticated set of options,
17969 (when (>= emacs-major-version 21)
17970 (blink-cursor-mode 0)
17971 ;; Insert newline when you press `C-n' (next-line)
17972 ;; at the end of the buffer
17973 (setq next-line-add-newlines t)
17976 ;; Turn on image viewing
17977 (auto-image-file-mode t)
17980 ;; Turn on menu bar (this bar has text)
17981 ;; (Use numeric argument to turn on)
17985 ;; Turn off tool bar (this bar has icons)
17986 ;; (Use numeric argument to turn on)
17987 (tool-bar-mode nil)
17990 ;; Turn off tooltip mode for tool bar
17991 ;; (This mode causes icon explanations to pop up)
17992 ;; (Use numeric argument to turn on)
17994 ;; If tooltips turned on, make tips appear promptly
17995 (setq tooltip-delay 0.1) ; default is 0.7 second
18000 @node X11 Colors, Miscellaneous, Simple Extension, Emacs Initialization
18001 @section X11 Colors
18003 You can specify colors when you use Emacs with the MIT X Windowing
18006 I dislike the default colors and specify my own.
18009 Here are the expressions in my @file{.emacs}
18010 file that set values:
18014 ;; Set cursor color
18015 (set-cursor-color "white")
18018 (set-mouse-color "white")
18020 ;; Set foreground and background
18021 (set-foreground-color "white")
18022 (set-background-color "darkblue")
18026 ;;; Set highlighting colors for isearch and drag
18027 (set-face-foreground 'highlight "white")
18028 (set-face-background 'highlight "blue")
18032 (set-face-foreground 'region "cyan")
18033 (set-face-background 'region "blue")
18037 (set-face-foreground 'secondary-selection "skyblue")
18038 (set-face-background 'secondary-selection "darkblue")
18042 ;; Set calendar highlighting colors
18043 (setq calendar-load-hook
18045 (set-face-foreground 'diary-face "skyblue")
18046 (set-face-background 'holiday-face "slate blue")
18047 (set-face-foreground 'holiday-face "white")))
18051 The various shades of blue soothe my eye and prevent me from seeing
18052 the screen flicker.
18054 Alternatively, I could have set my specifications in various X
18055 initialization files. For example, I could set the foreground,
18056 background, cursor, and pointer (i.e., mouse) colors in my
18057 @file{~/.Xresources} file like this:
18061 Emacs*foreground: white
18062 Emacs*background: darkblue
18063 Emacs*cursorColor: white
18064 Emacs*pointerColor: white
18068 In any event, since it is not part of Emacs, I set the root color of
18069 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
18070 run more modern window managers, such as Enlightenment, Gnome, or KDE;
18071 in those cases, I often specify an image rather than a plain color.}:
18074 xsetroot -solid Navy -fg white &
18078 @node Miscellaneous, Mode Line, X11 Colors, Emacs Initialization
18079 @section Miscellaneous Settings for a @file{.emacs} File
18082 Here are a few miscellaneous settings:
18087 Set the shape and color of the mouse cursor:
18091 ; Cursor shapes are defined in
18092 ; `/usr/include/X11/cursorfont.h';
18093 ; for example, the `target' cursor is number 128;
18094 ; the `top_left_arrow' cursor is number 132.
18098 (let ((mpointer (x-get-resource "*mpointer"
18099 "*emacs*mpointer")))
18100 ;; If you have not set your mouse pointer
18101 ;; then set it, otherwise leave as is:
18102 (if (eq mpointer nil)
18103 (setq mpointer "132")) ; top_left_arrow
18106 (setq x-pointer-shape (string-to-int mpointer))
18107 (set-mouse-color "white"))
18112 Or you can set the values of a variety of features in an alist, like
18118 default-frame-alist
18119 '((cursor-color . "white")
18120 (mouse-color . "white")
18121 (foreground-color . "white")
18122 (background-color . "DodgerBlue4")
18123 ;; (cursor-type . bar)
18124 (cursor-type . box)
18127 (tool-bar-lines . 0)
18128 (menu-bar-lines . 1)
18132 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
18138 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
18139 into @kbd{@key{CTRL}-h}.@*
18140 (Some older keyboards needed this, although I have not seen the
18145 ;; Translate `C-h' to <DEL>.
18146 ; (keyboard-translate ?\C-h ?\C-?)
18148 ;; Translate <DEL> to `C-h'.
18149 (keyboard-translate ?\C-? ?\C-h)
18153 @item Turn off a blinking cursor!
18157 (if (fboundp 'blink-cursor-mode)
18158 (blink-cursor-mode -1))
18163 or start GNU Emacs with the command @code{emacs -nbc}.
18166 @item When using `grep'@*
18167 @samp{-i}@w{ } Ignore case distinctions@*
18168 @samp{-n}@w{ } Prefix each line of output with line number@*
18169 @samp{-H}@w{ } Print the filename for each match.@*
18170 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18173 (setq grep-command "grep -i -nH -e ")
18177 @c Evidently, no longer needed in GNU Emacs 22
18179 item Automatically uncompress compressed files when visiting them
18182 (load "uncompress")
18187 @item Find an existing buffer, even if it has a different name@*
18188 This avoids problems with symbolic links.
18191 (setq find-file-existing-other-name t)
18194 @item Set your language environment and default input method
18198 (set-language-environment "latin-1")
18199 ;; Remember you can enable or disable multilingual text input
18200 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18201 (setq default-input-method "latin-1-prefix")
18205 If you want to write with Chinese `GB' characters, set this instead:
18209 (set-language-environment "Chinese-GB")
18210 (setq default-input-method "chinese-tonepy")
18215 @subsubheading Fixing Unpleasant Key Bindings
18216 @cindex Key bindings, fixing
18217 @cindex Bindings, key, fixing unpleasant
18219 Some systems bind keys unpleasantly. Sometimes, for example, the
18220 @key{CTRL} key appears in an awkward spot rather than at the far left
18223 Usually, when people fix these sorts of keybindings, they do not
18224 change their @file{~/.emacs} file. Instead, they bind the proper keys
18225 on their consoles with the @code{loadkeys} or @code{install-keymap}
18226 commands in their boot script and then include @code{xmodmap} commands
18227 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18235 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18237 install-keymap emacs2
18243 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18244 Lock} key is at the far left of the home row:
18248 # Bind the key labeled `Caps Lock' to `Control'
18249 # (Such a broken user interface suggests that keyboard manufacturers
18250 # think that computers are typewriters from 1885.)
18252 xmodmap -e "clear Lock"
18253 xmodmap -e "add Control = Caps_Lock"
18259 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18260 key to a @key{META} key:
18264 # Some ill designed keyboards have a key labeled ALT and no Meta
18265 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18270 @node Mode Line, , Miscellaneous, Emacs Initialization
18271 @section A Modified Mode Line
18272 @vindex mode-line-format
18273 @cindex Mode line format
18275 Finally, a feature I really like: a modified mode line.
18277 When I work over a network, I forget which machine I am using. Also,
18278 I tend to I lose track of where I am, and which line point is on.
18280 So I reset my mode line to look like this:
18283 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18286 I am visiting a file called @file{foo.texi}, on my machine
18287 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18288 Texinfo mode, and am at the top of the buffer.
18291 My @file{.emacs} file has a section that looks like this:
18295 ;; Set a Mode Line that tells me which machine, which directory,
18296 ;; and which line I am on, plus the other customary information.
18297 (setq-default mode-line-format
18301 "mouse-1: select window, mouse-2: delete others ..."))
18302 mode-line-mule-info
18304 mode-line-frame-identification
18308 mode-line-buffer-identification
18311 (system-name) 0 (string-match "\\..+" (system-name))))
18316 "mouse-1: select window, mouse-2: delete others ..."))
18317 (line-number-mode " Line %l ")
18323 "mouse-1: select window, mouse-2: delete others ..."))
18324 (:eval (mode-line-mode-name))
18327 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18336 Here, I redefine the default mode line. Most of the parts are from
18337 the original; but I make a few changes. I set the @emph{default} mode
18338 line format so as to permit various modes, such as Info, to override
18341 Many elements in the list are self-explanatory:
18342 @code{mode-line-modified} is a variable that tells whether the buffer
18343 has been modified, @code{mode-name} tells the name of the mode, and so
18344 on. However, the format looks complicated because of two features we
18345 have not discussed.
18347 @cindex Properties, in mode line example
18348 The first string in the mode line is a dash or hyphen, @samp{-}. In
18349 the old days, it would have been specified simply as @code{"-"}. But
18350 nowadays, Emacs can add properties to a string, such as highlighting
18351 or, as in this case, a help feature. If you place your mouse cursor
18352 over the hyphen, some help information appears (By default, you must
18353 wait seven-tenths of a second before the information appears. You can
18354 change that timing by changing the value of @code{tooltip-delay}.)
18357 The new string format has a special syntax:
18360 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18364 The @code{#(} begins a list. The first element of the list is the
18365 string itself, just one @samp{-}. The second and third
18366 elements specify the range over which the fourth element applies. A
18367 range starts @emph{after} a character, so a zero means the range
18368 starts just before the first character; a 1 means that the range ends
18369 just after the first character. The third element is the property for
18370 the range. It consists of a property list, a
18371 property name, in this case, @samp{help-echo}, followed by a value, in this
18372 case, a string. The second, third, and fourth elements of this new
18373 string format can be repeated.
18375 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18376 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18377 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18379 @code{mode-line-buffer-identification}
18380 displays the current buffer name. It is a list
18381 beginning @code{(#("%12b" 0 4 @dots{}}.
18382 The @code{#(} begins the list.
18384 The @samp{"%12b"} displays the current buffer name, using the
18385 @code{buffer-name} function with which we are familiar; the `12'
18386 specifies the maximum number of characters that will be displayed.
18387 When a name has fewer characters, whitespace is added to fill out to
18388 this number. (Buffer names can and often should be longer than 12
18389 characters; this length works well in a typical 80 column wide
18392 @code{:eval} says to evaluate the following form and use the result as
18393 a string to display. In this case, the expression displays the first
18394 component of the full system name. The end of the first component is
18395 a @samp{.} (`period'), so I use the @code{string-match} function to
18396 tell me the length of the first component. The substring from the
18397 zeroth character to that length is the name of the machine.
18400 This is the expression:
18405 (system-name) 0 (string-match "\\..+" (system-name))))
18409 @samp{%[} and @samp{%]} cause a pair of square brackets
18410 to appear for each recursive editing level. @samp{%n} says `Narrow'
18411 when narrowing is in effect. @samp{%P} tells you the percentage of
18412 the buffer that is above the bottom of the window, or `Top', `Bottom',
18413 or `All'. (A lower case @samp{p} tell you the percentage above the
18414 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18417 Remember, ``You don't have to like Emacs to like it'' --- your own
18418 Emacs can have different colors, different commands, and different
18419 keys than a default Emacs.
18421 On the other hand, if you want to bring up a plain `out of the box'
18422 Emacs, with no customization, type:
18429 This will start an Emacs that does @emph{not} load your
18430 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18433 @node Debugging, Conclusion, Emacs Initialization, Top
18437 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18438 first is built into the internals of Emacs and is always with you;
18439 the second requires that you instrument a function before you can use it.
18441 Both debuggers are described extensively in @ref{Debugging, ,
18442 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18443 In this chapter, I will walk through a short example of each.
18446 * debug:: How to use the built-in debugger.
18447 * debug-on-entry:: Start debugging when you call a function.
18448 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18449 * edebug:: How to use Edebug, a source level debugger.
18450 * Debugging Exercises::
18453 @node debug, debug-on-entry, Debugging, Debugging
18454 @section @code{debug}
18457 Suppose you have written a function definition that is intended to
18458 return the sum of the numbers 1 through a given number. (This is the
18459 @code{triangle} function discussed earlier. @xref{Decrementing
18460 Example, , Example with Decrementing Counter}, for a discussion.)
18461 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18463 However, your function definition has a bug. You have mistyped
18464 @samp{1=} for @samp{1-}. Here is the broken definition:
18466 @findex triangle-bugged
18469 (defun triangle-bugged (number)
18470 "Return sum of numbers 1 through NUMBER inclusive."
18472 (while (> number 0)
18473 (setq total (+ total number))
18474 (setq number (1= number))) ; @r{Error here.}
18479 If you are reading this in Info, you can evaluate this definition in
18480 the normal fashion. You will see @code{triangle-bugged} appear in the
18484 Now evaluate the @code{triangle-bugged} function with an
18488 (triangle-bugged 4)
18492 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18498 ---------- Buffer: *Backtrace* ----------
18499 Debugger entered--Lisp error: (void-function 1=)
18501 (setq number (1= number))
18502 (while (> number 0) (setq total (+ total number))
18503 (setq number (1= number)))
18504 (let ((total 0)) (while (> number 0) (setq total ...)
18505 (setq number ...)) total)
18509 eval((triangle-bugged 4))
18510 eval-last-sexp-1(nil)
18511 eval-last-sexp(nil)
18512 call-interactively(eval-last-sexp)
18513 ---------- Buffer: *Backtrace* ----------
18518 (I have reformatted this example slightly; the debugger does not fold
18519 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18520 the @file{*Backtrace*} buffer.)
18522 In practice, for a bug as simple as this, the `Lisp error' line will
18523 tell you what you need to know to correct the definition. The
18524 function @code{1=} is `void'.
18528 In GNU Emacs 20 and before, you will see:
18531 Symbol's function definition is void:@: 1=
18535 which has the same meaning as the @file{*Backtrace*} buffer line in
18539 However, suppose you are not quite certain what is going on?
18540 You can read the complete backtrace.
18542 In this case, you need to run a recent GNU Emacs, which automatically
18543 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18544 else, you need to start the debugger manually as described below.
18546 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18547 what Emacs did that led to the error. Emacs made an interactive call
18548 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18549 of the @code{triangle-bugged} expression. Each line above tells you
18550 what the Lisp interpreter evaluated next.
18553 The third line from the top of the buffer is
18556 (setq number (1= number))
18560 Emacs tried to evaluate this expression; in order to do so, it tried
18561 to evaluate the inner expression shown on the second line from the
18570 This is where the error occurred; as the top line says:
18573 Debugger entered--Lisp error: (void-function 1=)
18577 You can correct the mistake, re-evaluate the function definition, and
18578 then run your test again.
18580 @node debug-on-entry, debug-on-quit, debug, Debugging
18581 @section @code{debug-on-entry}
18582 @findex debug-on-entry
18584 A recent GNU Emacs starts the debugger automatically when your
18585 function has an error.
18588 GNU Emacs version 20 and before did not; it simply
18589 presented you with an error message. You had to start the debugger
18593 Incidentally, you can start the debugger manually for all versions of
18594 Emacs; the advantage is that the debugger runs even if you do not have
18595 a bug in your code. Sometimes your code will be free of bugs!
18597 You can enter the debugger when you call the function by calling
18598 @code{debug-on-entry}.
18605 M-x debug-on-entry RET triangle-bugged RET
18610 Now, evaluate the following:
18613 (triangle-bugged 5)
18617 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18618 you that it is beginning to evaluate the @code{triangle-bugged}
18623 ---------- Buffer: *Backtrace* ----------
18624 Debugger entered--entering a function:
18625 * triangle-bugged(5)
18626 eval((triangle-bugged 5))
18629 eval-last-sexp-1(nil)
18630 eval-last-sexp(nil)
18631 call-interactively(eval-last-sexp)
18632 ---------- Buffer: *Backtrace* ----------
18636 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18637 the first expression in @code{triangle-bugged}; the buffer will look
18642 ---------- Buffer: *Backtrace* ----------
18643 Debugger entered--beginning evaluation of function call form:
18644 * (let ((total 0)) (while (> number 0) (setq total ...)
18645 (setq number ...)) total)
18646 * triangle-bugged(5)
18647 eval((triangle-bugged 5))
18650 eval-last-sexp-1(nil)
18651 eval-last-sexp(nil)
18652 call-interactively(eval-last-sexp)
18653 ---------- Buffer: *Backtrace* ----------
18658 Now, type @kbd{d} again, eight times, slowly. Each time you type
18659 @kbd{d}, Emacs will evaluate another expression in the function
18663 Eventually, the buffer will look like this:
18667 ---------- Buffer: *Backtrace* ----------
18668 Debugger entered--beginning evaluation of function call form:
18669 * (setq number (1= number))
18670 * (while (> number 0) (setq total (+ total number))
18671 (setq number (1= number)))
18674 * (let ((total 0)) (while (> number 0) (setq total ...)
18675 (setq number ...)) total)
18676 * triangle-bugged(5)
18677 eval((triangle-bugged 5))
18680 eval-last-sexp-1(nil)
18681 eval-last-sexp(nil)
18682 call-interactively(eval-last-sexp)
18683 ---------- Buffer: *Backtrace* ----------
18689 Finally, after you type @kbd{d} two more times, Emacs will reach the
18690 error, and the top two lines of the @file{*Backtrace*} buffer will look
18695 ---------- Buffer: *Backtrace* ----------
18696 Debugger entered--Lisp error: (void-function 1=)
18699 ---------- Buffer: *Backtrace* ----------
18703 By typing @kbd{d}, you were able to step through the function.
18705 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18706 quits the trace, but does not cancel @code{debug-on-entry}.
18708 @findex cancel-debug-on-entry
18709 To cancel the effect of @code{debug-on-entry}, call
18710 @code{cancel-debug-on-entry} and the name of the function, like this:
18713 M-x cancel-debug-on-entry RET triangle-bugged RET
18717 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18719 @node debug-on-quit, edebug, debug-on-entry, Debugging
18720 @section @code{debug-on-quit} and @code{(debug)}
18722 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18723 there are two other ways to start @code{debug}.
18725 @findex debug-on-quit
18726 You can start @code{debug} whenever you type @kbd{C-g}
18727 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18728 @code{t}. This is useful for debugging infinite loops.
18731 @cindex @code{(debug)} in code
18732 Or, you can insert a line that says @code{(debug)} into your code
18733 where you want the debugger to start, like this:
18737 (defun triangle-bugged (number)
18738 "Return sum of numbers 1 through NUMBER inclusive."
18740 (while (> number 0)
18741 (setq total (+ total number))
18742 (debug) ; @r{Start debugger.}
18743 (setq number (1= number))) ; @r{Error here.}
18748 The @code{debug} function is described in detail in @ref{Debugger, ,
18749 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18751 @node edebug, Debugging Exercises, debug-on-quit, Debugging
18752 @section The @code{edebug} Source Level Debugger
18753 @cindex Source level debugger
18756 Edebug is a source level debugger. Edebug normally displays the
18757 source of the code you are debugging, with an arrow at the left that
18758 shows which line you are currently executing.
18760 You can walk through the execution of a function, line by line, or run
18761 quickly until reaching a @dfn{breakpoint} where execution stops.
18763 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18764 Lisp Reference Manual}.
18767 Here is a bugged function definition for @code{triangle-recursively}.
18768 @xref{Recursive triangle function, , Recursion in place of a counter},
18769 for a review of it.
18773 (defun triangle-recursively-bugged (number)
18774 "Return sum of numbers 1 through NUMBER inclusive.
18779 (triangle-recursively-bugged
18780 (1= number))))) ; @r{Error here.}
18785 Normally, you would install this definition by positioning your cursor
18786 after the function's closing parenthesis and typing @kbd{C-x C-e}
18787 (@code{eval-last-sexp}) or else by positioning your cursor within the
18788 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18789 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18793 However, to prepare this function definition for Edebug, you must
18794 first @dfn{instrument} the code using a different command. You can do
18795 this by positioning your cursor within or just after the definition
18799 M-x edebug-defun RET
18803 This will cause Emacs to load Edebug automatically if it is not
18804 already loaded, and properly instrument the function.
18806 After instrumenting the function, place your cursor after the
18807 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18810 (triangle-recursively-bugged 3)
18814 You will be jumped back to the source for
18815 @code{triangle-recursively-bugged} and the cursor positioned at the
18816 beginning of the @code{if} line of the function. Also, you will see
18817 an arrowhead at the left hand side of that line. The arrowhead marks
18818 the line where the function is executing. (In the following examples,
18819 we show the arrowhead with @samp{=>}; in a windowing system, you may
18820 see the arrowhead as a solid triangle in the window `fringe'.)
18823 =>@point{}(if (= number 1)
18828 In the example, the location of point is displayed with a star,
18829 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18832 In the example, the location of point is displayed as @samp{@point{}}
18833 (in a printed book, it is displayed with a five pointed star).
18836 If you now press @key{SPC}, point will move to the next expression to
18837 be executed; the line will look like this:
18840 =>(if @point{}(= number 1)
18844 As you continue to press @key{SPC}, point will move from expression to
18845 expression. At the same time, whenever an expression returns a value,
18846 that value will be displayed in the echo area. For example, after you
18847 move point past @code{number}, you will see the following:
18850 Result: 3 (#o3, #x3, ?\C-c)
18854 This means the value of @code{number} is 3, which is octal three,
18855 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18856 alphabet, in case you need to know this information).
18858 You can continue moving through the code until you reach the line with
18859 the error. Before evaluation, that line looks like this:
18862 => @point{}(1= number))))) ; @r{Error here.}
18867 When you press @key{SPC} once again, you will produce an error message
18871 Symbol's function definition is void:@: 1=
18877 Press @kbd{q} to quit Edebug.
18879 To remove instrumentation from a function definition, simply
18880 re-evaluate it with a command that does not instrument it.
18881 For example, you could place your cursor after the definition's
18882 closing parenthesis and type @kbd{C-x C-e}.
18884 Edebug does a great deal more than walk with you through a function.
18885 You can set it so it races through on its own, stopping only at an
18886 error or at specified stopping points; you can cause it to display the
18887 changing values of various expressions; you can find out how many
18888 times a function is called, and more.
18890 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18891 Lisp Reference Manual}.
18894 @node Debugging Exercises, , edebug, Debugging
18895 @section Debugging Exercises
18899 Install the @code{@value{COUNT-WORDS}} function and then cause it to
18900 enter the built-in debugger when you call it. Run the command on a
18901 region containing two words. You will need to press @kbd{d} a
18902 remarkable number of times. On your system, is a `hook' called after
18903 the command finishes? (For information on hooks, see @ref{Command
18904 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18908 Copy @code{@value{COUNT-WORDS}} into the @file{*scratch*} buffer,
18909 instrument the function for Edebug, and walk through its execution.
18910 The function does not need to have a bug, although you can introduce
18911 one if you wish. If the function lacks a bug, the walk-through
18912 completes without problems.
18915 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18916 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.@:
18917 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18918 for commands made outside of the Edebug debugging buffer.)
18921 In the Edebug debugging buffer, use the @kbd{p}
18922 (@code{edebug-bounce-point}) command to see where in the region the
18923 @code{@value{COUNT-WORDS}} is working.
18926 Move point to some spot further down the function and then type the
18927 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18930 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18931 walk through the function on its own; use an upper case @kbd{T} for
18932 @code{edebug-Trace-fast-mode}.
18935 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18939 @node Conclusion, the-the, Debugging, Top
18940 @chapter Conclusion
18942 We have now reached the end of this Introduction. You have now
18943 learned enough about programming in Emacs Lisp to set values, to write
18944 simple @file{.emacs} files for yourself and your friends, and write
18945 simple customizations and extensions to Emacs.
18947 This is a place to stop. Or, if you wish, you can now go onward, and
18950 You have learned some of the basic nuts and bolts of programming. But
18951 only some. There are a great many more brackets and hinges that are
18952 easy to use that we have not touched.
18954 A path you can follow right now lies among the sources to GNU Emacs
18957 @cite{The GNU Emacs Lisp Reference Manual}.
18960 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18961 Emacs Lisp Reference Manual}.
18964 The Emacs Lisp sources are an adventure. When you read the sources and
18965 come across a function or expression that is unfamiliar, you need to
18966 figure out or find out what it does.
18968 Go to the Reference Manual. It is a thorough, complete, and fairly
18969 easy-to-read description of Emacs Lisp. It is written not only for
18970 experts, but for people who know what you know. (The @cite{Reference
18971 Manual} comes with the standard GNU Emacs distribution. Like this
18972 introduction, it comes as a Texinfo source file, so you can read it
18973 on-line and as a typeset, printed book.)
18975 Go to the other on-line help that is part of GNU Emacs: the on-line
18976 documentation for all functions and variables, and @code{find-tag},
18977 the program that takes you to sources.
18979 Here is an example of how I explore the sources. Because of its name,
18980 @file{simple.el} is the file I looked at first, a long time ago. As
18981 it happens some of the functions in @file{simple.el} are complicated,
18982 or at least look complicated at first sight. The @code{open-line}
18983 function, for example, looks complicated.
18985 You may want to walk through this function slowly, as we did with the
18986 @code{forward-sentence} function. (@xref{forward-sentence, The
18987 @code{forward-sentence} function}.) Or you may want to skip that
18988 function and look at another, such as @code{split-line}. You don't
18989 need to read all the functions. According to
18990 @code{count-words-in-defun}, the @code{split-line} function contains
18991 102 words and symbols.
18993 Even though it is short, @code{split-line} contains expressions
18994 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18995 @code{current-column} and @code{insert-and-inherit}.
18997 Consider the @code{skip-chars-forward} function. (It is part of the
18998 function definition for @code{back-to-indentation}, which is shown in
18999 @ref{Review, , Review}.)
19001 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
19002 typing @kbd{C-h f} (@code{describe-function}) and the name of the
19003 function. This gives you the function documentation.
19005 You may be able to guess what is done by a well named function such as
19006 @code{indent-to}; or you can look it up, too. Incidentally, the
19007 @code{describe-function} function itself is in @file{help.el}; it is
19008 one of those long, but decipherable functions. You can look up
19009 @code{describe-function} using the @kbd{C-h f} command!
19011 In this instance, since the code is Lisp, the @file{*Help*} buffer
19012 contains the name of the library containing the function's source.
19013 You can put point over the name of the library and press the RET key,
19014 which in this situation is bound to @code{help-follow}, and be taken
19015 directly to the source, in the same way as @kbd{M-.}
19018 The definition for @code{describe-function} illustrates how to
19019 customize the @code{interactive} expression without using the standard
19020 character codes; and it shows how to create a temporary buffer.
19022 (The @code{indent-to} function is written in C rather than Emacs Lisp;
19023 it is a `built-in' function. @code{help-follow} takes you to its
19024 source as does @code{find-tag}, when properly set up.)
19026 You can look at a function's source using @code{find-tag}, which is
19027 bound to @kbd{M-.} Finally, you can find out what the Reference
19028 Manual has to say by visiting the manual in Info, and typing @kbd{i}
19029 (@code{Info-index}) and the name of the function, or by looking up the
19030 function in the index to a printed copy of the manual.
19032 Similarly, you can find out what is meant by
19033 @code{insert-and-inherit}.
19035 Other interesting source files include @file{paragraphs.el},
19036 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
19037 file includes short, easily understood functions as well as longer
19038 ones. The @file{loaddefs.el} file contains the many standard
19039 autoloads and many keymaps. I have never looked at it all; only at
19040 parts. @file{loadup.el} is the file that loads the standard parts of
19041 Emacs; it tells you a great deal about how Emacs is built.
19042 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
19043 Reference Manual}, for more about building.)
19045 As I said, you have learned some nuts and bolts; however, and very
19046 importantly, we have hardly touched major aspects of programming; I
19047 have said nothing about how to sort information, except to use the
19048 predefined @code{sort} function; I have said nothing about how to store
19049 information, except to use variables and lists; I have said nothing
19050 about how to write programs that write programs. These are topics for
19051 another, and different kind of book, a different kind of learning.
19053 What you have done is learn enough for much practical work with GNU
19054 Emacs. What you have done is get started. This is the end of a
19057 @c ================ Appendix ================
19059 @node the-the, Kill Ring, Conclusion, Top
19060 @appendix The @code{the-the} Function
19062 @cindex Duplicated words function
19063 @cindex Words, duplicated
19065 Sometimes when you you write text, you duplicate words---as with ``you
19066 you'' near the beginning of this sentence. I find that most
19067 frequently, I duplicate ``the''; hence, I call the function for
19068 detecting duplicated words, @code{the-the}.
19071 As a first step, you could use the following regular expression to
19072 search for duplicates:
19075 \\(\\w+[ \t\n]+\\)\\1
19079 This regexp matches one or more word-constituent characters followed
19080 by one or more spaces, tabs, or newlines. However, it does not detect
19081 duplicated words on different lines, since the ending of the first
19082 word, the end of the line, is different from the ending of the second
19083 word, a space. (For more information about regular expressions, see
19084 @ref{Regexp Search, , Regular Expression Searches}, as well as
19085 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
19086 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
19087 The GNU Emacs Lisp Reference Manual}.)
19089 You might try searching just for duplicated word-constituent
19090 characters but that does not work since the pattern detects doubles
19091 such as the two occurrences of `th' in `with the'.
19093 Another possible regexp searches for word-constituent characters
19094 followed by non-word-constituent characters, reduplicated. Here,
19095 @w{@samp{\\w+}} matches one or more word-constituent characters and
19096 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
19099 \\(\\(\\w+\\)\\W*\\)\\1
19105 Here is the pattern that I use. It is not perfect, but good enough.
19106 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
19107 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
19108 any characters that are @emph{not} an @@-sign, space, newline, or tab.
19111 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
19114 One can write more complicated expressions, but I found that this
19115 expression is good enough, so I use it.
19117 Here is the @code{the-the} function, as I include it in my
19118 @file{.emacs} file, along with a handy global key binding:
19123 "Search forward for for a duplicated word."
19125 (message "Searching for for duplicated words ...")
19129 ;; This regexp is not perfect
19130 ;; but is fairly good over all:
19131 (if (re-search-forward
19132 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
19133 (message "Found duplicated word.")
19134 (message "End of buffer")))
19138 ;; Bind `the-the' to C-c \
19139 (global-set-key "\C-c\\" 'the-the)
19148 one two two three four five
19153 You can substitute the other regular expressions shown above in the
19154 function definition and try each of them on this list.
19156 @node Kill Ring, Full Graph, the-the, Top
19157 @appendix Handling the Kill Ring
19158 @cindex Kill ring handling
19159 @cindex Handling the kill ring
19160 @cindex Ring, making a list like a
19162 The kill ring is a list that is transformed into a ring by the
19163 workings of the @code{current-kill} function. The @code{yank} and
19164 @code{yank-pop} commands use the @code{current-kill} function.
19166 This appendix describes the @code{current-kill} function as well as
19167 both the @code{yank} and the @code{yank-pop} commands, but first,
19168 consider the workings of the kill ring.
19171 * What the Kill Ring Does::
19173 * yank:: Paste a copy of a clipped element.
19174 * yank-pop:: Insert element pointed to.
19178 @node What the Kill Ring Does, current-kill, Kill Ring, Kill Ring
19180 @unnumberedsec What the Kill Ring Does
19184 The kill ring has a default maximum length of sixty items; this number
19185 is too large for an explanation. Instead, set it to four. Please
19186 evaluate the following:
19190 (setq old-kill-ring-max kill-ring-max)
19191 (setq kill-ring-max 4)
19196 Then, please copy each line of the following indented example into the
19197 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19201 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19202 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19203 merely copy it to the kill ring. However, your machine may beep at
19204 you. Alternatively, for silence, you may copy the region of each line
19205 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19206 each line for this command to succeed, but it does not matter at which
19207 end you put point or mark.)
19211 Please invoke the calls in order, so that five elements attempt to
19212 fill the kill ring:
19217 second piece of text
19219 fourth line of text
19226 Then find the value of @code{kill-ring} by evaluating
19238 ("fifth bit of text" "fourth line of text"
19239 "third line" "second piece of text")
19244 The first element, @samp{first some text}, was dropped.
19247 To return to the old value for the length of the kill ring, evaluate:
19250 (setq kill-ring-max old-kill-ring-max)
19253 @node current-kill, yank, What the Kill Ring Does, Kill Ring
19254 @comment node-name, next, previous, up
19255 @appendixsec The @code{current-kill} Function
19256 @findex current-kill
19258 The @code{current-kill} function changes the element in the kill ring
19259 to which @code{kill-ring-yank-pointer} points. (Also, the
19260 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19261 to the latest element of the kill ring. The @code{kill-new}
19262 function is used directly or indirectly by @code{kill-append},
19263 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19264 and @code{kill-region}.)
19267 * Code for current-kill::
19268 * Understanding current-kill::
19271 @node Code for current-kill, Understanding current-kill, current-kill, current-kill
19273 @unnumberedsubsec The code for @code{current-kill}
19278 The @code{current-kill} function is used by @code{yank} and by
19279 @code{yank-pop}. Here is the code for @code{current-kill}:
19283 (defun current-kill (n &optional do-not-move)
19284 "Rotate the yanking point by N places, and then return that kill.
19285 If N is zero, `interprogram-paste-function' is set, and calling it
19286 returns a string, then that string is added to the front of the
19287 kill ring and returned as the latest kill.
19290 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19291 yanking point; just return the Nth kill forward."
19292 (let ((interprogram-paste (and (= n 0)
19293 interprogram-paste-function
19294 (funcall interprogram-paste-function))))
19297 (if interprogram-paste
19299 ;; Disable the interprogram cut function when we add the new
19300 ;; text to the kill ring, so Emacs doesn't try to own the
19301 ;; selection, with identical text.
19302 (let ((interprogram-cut-function nil))
19303 (kill-new interprogram-paste))
19304 interprogram-paste)
19307 (or kill-ring (error "Kill ring is empty"))
19308 (let ((ARGth-kill-element
19309 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19310 (length kill-ring))
19313 (setq kill-ring-yank-pointer ARGth-kill-element))
19314 (car ARGth-kill-element)))))
19318 Remember also that the @code{kill-new} function sets
19319 @code{kill-ring-yank-pointer} to the latest element of the kill
19320 ring, which means that all the functions that call it set the value
19321 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19322 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19325 Here is the line in @code{kill-new}, which is explained in
19326 @ref{kill-new function, , The @code{kill-new} function}.
19329 (setq kill-ring-yank-pointer kill-ring)
19332 @node Understanding current-kill, , Code for current-kill, current-kill
19334 @unnumberedsubsec @code{current-kill} in Outline
19337 The @code{current-kill} function looks complex, but as usual, it can
19338 be understood by taking it apart piece by piece. First look at it in
19343 (defun current-kill (n &optional do-not-move)
19344 "Rotate the yanking point by N places, and then return that kill."
19350 This function takes two arguments, one of which is optional. It has a
19351 documentation string. It is @emph{not} interactive.
19354 * Body of current-kill::
19355 * Digression concerning error:: How to mislead humans, but not computers.
19356 * Determining the Element::
19359 @node Body of current-kill, Digression concerning error, Understanding current-kill, Understanding current-kill
19361 @unnumberedsubsubsec The Body of @code{current-kill}
19364 The body of the function definition is a @code{let} expression, which
19365 itself has a body as well as a @var{varlist}.
19367 The @code{let} expression declares a variable that will be only usable
19368 within the bounds of this function. This variable is called
19369 @code{interprogram-paste} and is for copying to another program. It
19370 is not for copying within this instance of GNU Emacs. Most window
19371 systems provide a facility for interprogram pasting. Sadly, that
19372 facility usually provides only for the last element. Most windowing
19373 systems have not adopted a ring of many possibilities, even though
19374 Emacs has provided it for decades.
19376 The @code{if} expression has two parts, one if there exists
19377 @code{interprogram-paste} and one if not.
19380 Let us consider the `if not' or else-part of the @code{current-kill}
19381 function. (The then-part uses the @code{kill-new} function, which
19382 we have already described. @xref{kill-new function, , The
19383 @code{kill-new} function}.)
19387 (or kill-ring (error "Kill ring is empty"))
19388 (let ((ARGth-kill-element
19389 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19390 (length kill-ring))
19393 (setq kill-ring-yank-pointer ARGth-kill-element))
19394 (car ARGth-kill-element))
19399 The code first checks whether the kill ring has content; otherwise it
19403 Note that the @code{or} expression is very similar to testing length
19410 (if (zerop (length kill-ring)) ; @r{if-part}
19411 (error "Kill ring is empty")) ; @r{then-part}
19417 If there is not anything in the kill ring, its length must be zero and
19418 an error message sent to the user: @samp{Kill ring is empty}. The
19419 @code{current-kill} function uses an @code{or} expression which is
19420 simpler. But an @code{if} expression reminds us what goes on.
19422 This @code{if} expression uses the function @code{zerop} which returns
19423 true if the value it is testing is zero. When @code{zerop} tests
19424 true, the then-part of the @code{if} is evaluated. The then-part is a
19425 list starting with the function @code{error}, which is a function that
19426 is similar to the @code{message} function
19427 (@pxref{message, , The @code{message} Function}) in that
19428 it prints a one-line message in the echo area. However, in addition
19429 to printing a message, @code{error} also stops evaluation of the
19430 function within which it is embedded. This means that the rest of the
19431 function will not be evaluated if the length of the kill ring is zero.
19433 Then the @code{current-kill} function selects the element to return.
19434 The selection depends on the number of places that @code{current-kill}
19435 rotates and on where @code{kill-ring-yank-pointer} points.
19437 Next, either the optional @code{do-not-move} argument is true or the
19438 current value of @code{kill-ring-yank-pointer} is set to point to the
19439 list. Finally, another expression returns the first element of the
19440 list even if the @code{do-not-move} argument is true.
19442 @node Digression concerning error, Determining the Element, Body of current-kill, Understanding current-kill
19444 @unnumberedsubsubsec Digression about the word `error'
19447 In my opinion, it is slightly misleading, at least to humans, to use
19448 the term `error' as the name of the @code{error} function. A better
19449 term would be `cancel'. Strictly speaking, of course, you cannot
19450 point to, much less rotate a pointer to a list that has no length, so
19451 from the point of view of the computer, the word `error' is correct.
19452 But a human expects to attempt this sort of thing, if only to find out
19453 whether the kill ring is full or empty. This is an act of
19456 From the human point of view, the act of exploration and discovery is
19457 not necessarily an error, and therefore should not be labeled as one,
19458 even in the bowels of a computer. As it is, the code in Emacs implies
19459 that a human who is acting virtuously, by exploring his or her
19460 environment, is making an error. This is bad. Even though the computer
19461 takes the same steps as it does when there is an `error', a term such as
19462 `cancel' would have a clearer connotation.
19464 @node Determining the Element, , Digression concerning error, Understanding current-kill
19466 @unnumberedsubsubsec Determining the Element
19469 Among other actions, the else-part of the @code{if} expression sets
19470 the value of @code{kill-ring-yank-pointer} to
19471 @code{ARGth-kill-element} when the kill ring has something in it and
19472 the value of @code{do-not-move} is @code{nil}.
19475 The code looks like this:
19479 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19480 (length kill-ring))
19485 This needs some examination. Unless it is not supposed to move the
19486 pointer, the @code{current-kill} function changes where
19487 @code{kill-ring-yank-pointer} points.
19489 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19490 expression does. Also, clearly, @code{ARGth-kill-element} is being
19491 set to be equal to some @sc{cdr} of the kill ring, using the
19492 @code{nthcdr} function that is described in an earlier section.
19493 (@xref{copy-region-as-kill}.) How does it do this?
19495 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19496 works by repeatedly taking the @sc{cdr} of a list---it takes the
19497 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19500 The two following expressions produce the same result:
19504 (setq kill-ring-yank-pointer (cdr kill-ring))
19506 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19510 However, the @code{nthcdr} expression is more complicated. It uses
19511 the @code{mod} function to determine which @sc{cdr} to select.
19513 (You will remember to look at inner functions first; indeed, we will
19514 have to go inside the @code{mod}.)
19516 The @code{mod} function returns the value of its first argument modulo
19517 the second; that is to say, it returns the remainder after dividing
19518 the first argument by the second. The value returned has the same
19519 sign as the second argument.
19527 @result{} 0 ;; @r{because there is no remainder}
19534 In this case, the first argument is often smaller than the second.
19546 We can guess what the @code{-} function does. It is like @code{+} but
19547 subtracts instead of adds; the @code{-} function subtracts its second
19548 argument from its first. Also, we already know what the @code{length}
19549 function does (@pxref{length}). It returns the length of a list.
19551 And @code{n} is the name of the required argument to the
19552 @code{current-kill} function.
19555 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19556 expression returns the whole list, as you can see by evaluating the
19561 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19562 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19563 (nthcdr (mod (- 0 4) 4)
19564 '("fourth line of text"
19566 "second piece of text"
19567 "first some text"))
19572 When the first argument to the @code{current-kill} function is one,
19573 the @code{nthcdr} expression returns the list without its first
19578 (nthcdr (mod (- 1 4) 4)
19579 '("fourth line of text"
19581 "second piece of text"
19582 "first some text"))
19586 @cindex @samp{global variable} defined
19587 @cindex @samp{variable, global}, defined
19588 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19589 are @dfn{global variables}. That means that any expression in Emacs
19590 Lisp can access them. They are not like the local variables set by
19591 @code{let} or like the symbols in an argument list.
19592 Local variables can only be accessed
19593 within the @code{let} that defines them or the function that specifies
19594 them in an argument list (and within expressions called by them).
19597 @c texi2dvi fails when the name of the section is within ifnottex ...
19598 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19599 @ref{defun, , The @code{defun} Special Form}.)
19602 @node yank, yank-pop, current-kill, Kill Ring
19603 @comment node-name, next, previous, up
19604 @appendixsec @code{yank}
19607 After learning about @code{current-kill}, the code for the
19608 @code{yank} function is almost easy.
19610 The @code{yank} function does not use the
19611 @code{kill-ring-yank-pointer} variable directly. It calls
19612 @code{insert-for-yank} which calls @code{current-kill} which sets the
19613 @code{kill-ring-yank-pointer} variable.
19616 The code looks like this:
19621 (defun yank (&optional arg)
19622 "Reinsert (\"paste\") the last stretch of killed text.
19623 More precisely, reinsert the stretch of killed text most recently
19624 killed OR yanked. Put point at end, and set mark at beginning.
19625 With just \\[universal-argument] as argument, same but put point at
19626 beginning (and mark at end). With argument N, reinsert the Nth most
19627 recently killed stretch of killed text.
19629 When this command inserts killed text into the buffer, it honors
19630 `yank-excluded-properties' and `yank-handler' as described in the
19631 doc string for `insert-for-yank-1', which see.
19633 See also the command \\[yank-pop]."
19637 (setq yank-window-start (window-start))
19638 ;; If we don't get all the way thru, make last-command indicate that
19639 ;; for the following command.
19640 (setq this-command t)
19641 (push-mark (point))
19644 (insert-for-yank (current-kill (cond
19649 ;; This is like exchange-point-and-mark,
19650 ;; but doesn't activate the mark.
19651 ;; It is cleaner to avoid activation, even though the command
19652 ;; loop would deactivate the mark because we inserted text.
19653 (goto-char (prog1 (mark t)
19654 (set-marker (mark-marker) (point) (current-buffer)))))
19657 ;; If we do get all the way thru, make this-command indicate that.
19658 (if (eq this-command t)
19659 (setq this-command 'yank))
19664 The key expression is @code{insert-for-yank}, which inserts the string
19665 returned by @code{current-kill}, but removes some text properties from
19668 However, before getting to that expression, the function sets the value
19669 of @code{yank-window-start} to the position returned by the
19670 @code{(window-start)} expression, the position at which the display
19671 currently starts. The @code{yank} function also sets
19672 @code{this-command} and pushes the mark.
19674 After it yanks the appropriate element, if the optional argument is a
19675 @sc{cons} rather than a number or nothing, it puts point at beginning
19676 of the yanked text and mark at its end.
19678 (The @code{prog1} function is like @code{progn} but returns the value
19679 of its first argument rather than the value of its last argument. Its
19680 first argument is forced to return the buffer's mark as an integer.
19681 You can see the documentation for these functions by placing point
19682 over them in this buffer and then typing @kbd{C-h f}
19683 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19686 The last part of the function tells what to do when it succeeds.
19688 @node yank-pop, ring file, yank, Kill Ring
19689 @comment node-name, next, previous, up
19690 @appendixsec @code{yank-pop}
19693 After understanding @code{yank} and @code{current-kill}, you know how
19694 to approach the @code{yank-pop} function. Leaving out the
19695 documentation to save space, it looks like this:
19700 (defun yank-pop (&optional arg)
19703 (if (not (eq last-command 'yank))
19704 (error "Previous command was not a yank"))
19707 (setq this-command 'yank)
19708 (unless arg (setq arg 1))
19709 (let ((inhibit-read-only t)
19710 (before (< (point) (mark t))))
19714 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19715 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19716 (setq yank-undo-function nil)
19719 (set-marker (mark-marker) (point) (current-buffer))
19720 (insert-for-yank (current-kill arg))
19721 ;; Set the window start back where it was in the yank command,
19723 (set-window-start (selected-window) yank-window-start t)
19727 ;; This is like exchange-point-and-mark,
19728 ;; but doesn't activate the mark.
19729 ;; It is cleaner to avoid activation, even though the command
19730 ;; loop would deactivate the mark because we inserted text.
19731 (goto-char (prog1 (mark t)
19732 (set-marker (mark-marker)
19734 (current-buffer))))))
19739 The function is interactive with a small @samp{p} so the prefix
19740 argument is processed and passed to the function. The command can
19741 only be used after a previous yank; otherwise an error message is
19742 sent. This check uses the variable @code{last-command} which is set
19743 by @code{yank} and is discussed elsewhere.
19744 (@xref{copy-region-as-kill}.)
19746 The @code{let} clause sets the variable @code{before} to true or false
19747 depending whether point is before or after mark and then the region
19748 between point and mark is deleted. This is the region that was just
19749 inserted by the previous yank and it is this text that will be
19752 @code{funcall} calls its first argument as a function, passing
19753 remaining arguments to it. The first argument is whatever the
19754 @code{or} expression returns. The two remaining arguments are the
19755 positions of point and mark set by the preceding @code{yank} command.
19757 There is more, but that is the hardest part.
19759 @node ring file, , yank-pop, Kill Ring
19760 @comment node-name, next, previous, up
19761 @appendixsec The @file{ring.el} File
19762 @cindex @file{ring.el} file
19764 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19765 provides many of the features we just discussed. But functions such
19766 as @code{kill-ring-yank-pointer} do not use this library, possibly
19767 because they were written earlier.
19769 @node Full Graph, Free Software and Free Manuals, Kill Ring, Top
19770 @appendix A Graph with Labeled Axes
19772 Printed axes help you understand a graph. They convey scale. In an
19773 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19774 wrote the code to print the body of a graph. Here we write the code
19775 for printing and labeling vertical and horizontal axes, along with the
19779 * Labeled Example::
19780 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19781 * print-Y-axis:: Print a label for the vertical axis.
19782 * print-X-axis:: Print a horizontal label.
19783 * Print Whole Graph:: The function to print a complete graph.
19786 @node Labeled Example, print-graph Varlist, Full Graph, Full Graph
19788 @unnumberedsec Labeled Example Graph
19791 Since insertions fill a buffer to the right and below point, the new
19792 graph printing function should first print the Y or vertical axis,
19793 then the body of the graph, and finally the X or horizontal axis.
19794 This sequence lays out for us the contents of the function:
19804 Print body of graph.
19811 Here is an example of how a finished graph should look:
19824 1 - ****************
19831 In this graph, both the vertical and the horizontal axes are labeled
19832 with numbers. However, in some graphs, the horizontal axis is time
19833 and would be better labeled with months, like this:
19847 Indeed, with a little thought, we can easily come up with a variety of
19848 vertical and horizontal labeling schemes. Our task could become
19849 complicated. But complications breed confusion. Rather than permit
19850 this, it is better choose a simple labeling scheme for our first
19851 effort, and to modify or replace it later.
19854 These considerations suggest the following outline for the
19855 @code{print-graph} function:
19859 (defun print-graph (numbers-list)
19860 "@var{documentation}@dots{}"
19861 (let ((height @dots{}
19865 (print-Y-axis height @dots{} )
19866 (graph-body-print numbers-list)
19867 (print-X-axis @dots{} )))
19871 We can work on each part of the @code{print-graph} function definition
19874 @node print-graph Varlist, print-Y-axis, Labeled Example, Full Graph
19875 @comment node-name, next, previous, up
19876 @appendixsec The @code{print-graph} Varlist
19877 @cindex @code{print-graph} varlist
19879 In writing the @code{print-graph} function, the first task is to write
19880 the varlist in the @code{let} expression. (We will leave aside for the
19881 moment any thoughts about making the function interactive or about the
19882 contents of its documentation string.)
19884 The varlist should set several values. Clearly, the top of the label
19885 for the vertical axis must be at least the height of the graph, which
19886 means that we must obtain this information here. Note that the
19887 @code{print-graph-body} function also requires this information. There
19888 is no reason to calculate the height of the graph in two different
19889 places, so we should change @code{print-graph-body} from the way we
19890 defined it earlier to take advantage of the calculation.
19892 Similarly, both the function for printing the X axis labels and the
19893 @code{print-graph-body} function need to learn the value of the width of
19894 each symbol. We can perform the calculation here and change the
19895 definition for @code{print-graph-body} from the way we defined it in the
19898 The length of the label for the horizontal axis must be at least as long
19899 as the graph. However, this information is used only in the function
19900 that prints the horizontal axis, so it does not need to be calculated here.
19902 These thoughts lead us directly to the following form for the varlist
19903 in the @code{let} for @code{print-graph}:
19907 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19908 (symbol-width (length graph-blank)))
19913 As we shall see, this expression is not quite right.
19916 @node print-Y-axis, print-X-axis, print-graph Varlist, Full Graph
19917 @comment node-name, next, previous, up
19918 @appendixsec The @code{print-Y-axis} Function
19919 @cindex Axis, print vertical
19920 @cindex Y axis printing
19921 @cindex Vertical axis printing
19922 @cindex Print vertical axis
19924 The job of the @code{print-Y-axis} function is to print a label for
19925 the vertical axis that looks like this:
19943 The function should be passed the height of the graph, and then should
19944 construct and insert the appropriate numbers and marks.
19947 * print-Y-axis in Detail::
19948 * Height of label:: What height for the Y axis?
19949 * Compute a Remainder:: How to compute the remainder of a division.
19950 * Y Axis Element:: Construct a line for the Y axis.
19951 * Y-axis-column:: Generate a list of Y axis labels.
19952 * print-Y-axis Penultimate:: A not quite final version.
19955 @node print-Y-axis in Detail, Height of label, print-Y-axis, print-Y-axis
19957 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19960 It is easy enough to see in the figure what the Y axis label should
19961 look like; but to say in words, and then to write a function
19962 definition to do the job is another matter. It is not quite true to
19963 say that we want a number and a tic every five lines: there are only
19964 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19965 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19966 and 9). It is better to say that we want a number and a tic mark on
19967 the base line (number 1) and then that we want a number and a tic on
19968 the fifth line from the bottom and on every line that is a multiple of
19971 @node Height of label, Compute a Remainder, print-Y-axis in Detail, print-Y-axis
19973 @unnumberedsubsec What height should the label be?
19976 The next issue is what height the label should be? Suppose the maximum
19977 height of tallest column of the graph is seven. Should the highest
19978 label on the Y axis be @samp{5 -}, and should the graph stick up above
19979 the label? Or should the highest label be @samp{7 -}, and mark the peak
19980 of the graph? Or should the highest label be @code{10 -}, which is a
19981 multiple of five, and be higher than the topmost value of the graph?
19983 The latter form is preferred. Most graphs are drawn within rectangles
19984 whose sides are an integral number of steps long---5, 10, 15, and so
19985 on for a step distance of five. But as soon as we decide to use a
19986 step height for the vertical axis, we discover that the simple
19987 expression in the varlist for computing the height is wrong. The
19988 expression is @code{(apply 'max numbers-list)}. This returns the
19989 precise height, not the maximum height plus whatever is necessary to
19990 round up to the nearest multiple of five. A more complex expression
19993 As usual in cases like this, a complex problem becomes simpler if it is
19994 divided into several smaller problems.
19996 First, consider the case when the highest value of the graph is an
19997 integral multiple of five---when it is 5, 10, 15, or some higher
19998 multiple of five. We can use this value as the Y axis height.
20000 A fairly simply way to determine whether a number is a multiple of
20001 five is to divide it by five and see if the division results in a
20002 remainder. If there is no remainder, the number is a multiple of
20003 five. Thus, seven divided by five has a remainder of two, and seven
20004 is not an integral multiple of five. Put in slightly different
20005 language, more reminiscent of the classroom, five goes into seven
20006 once, with a remainder of two. However, five goes into ten twice,
20007 with no remainder: ten is an integral multiple of five.
20009 @node Compute a Remainder, Y Axis Element, Height of label, print-Y-axis
20010 @appendixsubsec Side Trip: Compute a Remainder
20012 @findex % @r{(remainder function)}
20013 @cindex Remainder function, @code{%}
20014 In Lisp, the function for computing a remainder is @code{%}. The
20015 function returns the remainder of its first argument divided by its
20016 second argument. As it happens, @code{%} is a function in Emacs Lisp
20017 that you cannot discover using @code{apropos}: you find nothing if you
20018 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
20019 learn of the existence of @code{%} is to read about it in a book such
20020 as this or in the Emacs Lisp sources.
20022 You can try the @code{%} function by evaluating the following two
20034 The first expression returns 2 and the second expression returns 0.
20036 To test whether the returned value is zero or some other number, we
20037 can use the @code{zerop} function. This function returns @code{t} if
20038 its argument, which must be a number, is zero.
20050 Thus, the following expression will return @code{t} if the height
20051 of the graph is evenly divisible by five:
20054 (zerop (% height 5))
20058 (The value of @code{height}, of course, can be found from @code{(apply
20059 'max numbers-list)}.)
20061 On the other hand, if the value of @code{height} is not a multiple of
20062 five, we want to reset the value to the next higher multiple of five.
20063 This is straightforward arithmetic using functions with which we are
20064 already familiar. First, we divide the value of @code{height} by five
20065 to determine how many times five goes into the number. Thus, five
20066 goes into twelve twice. If we add one to this quotient and multiply by
20067 five, we will obtain the value of the next multiple of five that is
20068 larger than the height. Five goes into twelve twice. Add one to two,
20069 and multiply by five; the result is fifteen, which is the next multiple
20070 of five that is higher than twelve. The Lisp expression for this is:
20073 (* (1+ (/ height 5)) 5)
20077 For example, if you evaluate the following, the result is 15:
20080 (* (1+ (/ 12 5)) 5)
20083 All through this discussion, we have been using `five' as the value
20084 for spacing labels on the Y axis; but we may want to use some other
20085 value. For generality, we should replace `five' with a variable to
20086 which we can assign a value. The best name I can think of for this
20087 variable is @code{Y-axis-label-spacing}.
20090 Using this term, and an @code{if} expression, we produce the
20095 (if (zerop (% height Y-axis-label-spacing))
20098 (* (1+ (/ height Y-axis-label-spacing))
20099 Y-axis-label-spacing))
20104 This expression returns the value of @code{height} itself if the height
20105 is an even multiple of the value of the @code{Y-axis-label-spacing} or
20106 else it computes and returns a value of @code{height} that is equal to
20107 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
20109 We can now include this expression in the @code{let} expression of the
20110 @code{print-graph} function (after first setting the value of
20111 @code{Y-axis-label-spacing}):
20112 @vindex Y-axis-label-spacing
20116 (defvar Y-axis-label-spacing 5
20117 "Number of lines from one Y axis label to next.")
20122 (let* ((height (apply 'max numbers-list))
20123 (height-of-top-line
20124 (if (zerop (% height Y-axis-label-spacing))
20129 (* (1+ (/ height Y-axis-label-spacing))
20130 Y-axis-label-spacing)))
20131 (symbol-width (length graph-blank))))
20137 (Note use of the @code{let*} function: the initial value of height is
20138 computed once by the @code{(apply 'max numbers-list)} expression and
20139 then the resulting value of @code{height} is used to compute its
20140 final value. @xref{fwd-para let, , The @code{let*} expression}, for
20141 more about @code{let*}.)
20143 @node Y Axis Element, Y-axis-column, Compute a Remainder, print-Y-axis
20144 @appendixsubsec Construct a Y Axis Element
20146 When we print the vertical axis, we want to insert strings such as
20147 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20148 Moreover, we want the numbers and dashes to line up, so shorter
20149 numbers must be padded with leading spaces. If some of the strings
20150 use two digit numbers, the strings with single digit numbers must
20151 include a leading blank space before the number.
20153 @findex number-to-string
20154 To figure out the length of the number, the @code{length} function is
20155 used. But the @code{length} function works only with a string, not with
20156 a number. So the number has to be converted from being a number to
20157 being a string. This is done with the @code{number-to-string} function.
20162 (length (number-to-string 35))
20165 (length (number-to-string 100))
20171 (@code{number-to-string} is also called @code{int-to-string}; you will
20172 see this alternative name in various sources.)
20174 In addition, in each label, each number is followed by a string such
20175 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20176 This variable is defined with @code{defvar}:
20181 (defvar Y-axis-tic " - "
20182 "String that follows number in a Y axis label.")
20186 The length of the Y label is the sum of the length of the Y axis tic
20187 mark and the length of the number of the top of the graph.
20190 (length (concat (number-to-string height) Y-axis-tic)))
20193 This value will be calculated by the @code{print-graph} function in
20194 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20195 did not think to include this in the varlist when we first proposed it.)
20197 To make a complete vertical axis label, a tic mark is concatenated
20198 with a number; and the two together may be preceded by one or more
20199 spaces depending on how long the number is. The label consists of
20200 three parts: the (optional) leading spaces, the number, and the tic
20201 mark. The function is passed the value of the number for the specific
20202 row, and the value of the width of the top line, which is calculated
20203 (just once) by @code{print-graph}.
20207 (defun Y-axis-element (number full-Y-label-width)
20208 "Construct a NUMBERed label element.
20209 A numbered element looks like this ` 5 - ',
20210 and is padded as needed so all line up with
20211 the element for the largest number."
20214 (let* ((leading-spaces
20215 (- full-Y-label-width
20217 (concat (number-to-string number)
20222 (make-string leading-spaces ? )
20223 (number-to-string number)
20228 The @code{Y-axis-element} function concatenates together the leading
20229 spaces, if any; the number, as a string; and the tic mark.
20231 To figure out how many leading spaces the label will need, the
20232 function subtracts the actual length of the label---the length of the
20233 number plus the length of the tic mark---from the desired label width.
20235 @findex make-string
20236 Blank spaces are inserted using the @code{make-string} function. This
20237 function takes two arguments: the first tells it how long the string
20238 will be and the second is a symbol for the character to insert, in a
20239 special format. The format is a question mark followed by a blank
20240 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20241 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20242 syntax for characters. (Of course, you might want to replace the
20243 blank space by some other character @dots{} You know what to do.)
20245 The @code{number-to-string} function is used in the concatenation
20246 expression, to convert the number to a string that is concatenated
20247 with the leading spaces and the tic mark.
20249 @node Y-axis-column, print-Y-axis Penultimate, Y Axis Element, print-Y-axis
20250 @appendixsubsec Create a Y Axis Column
20252 The preceding functions provide all the tools needed to construct a
20253 function that generates a list of numbered and blank strings to insert
20254 as the label for the vertical axis:
20256 @findex Y-axis-column
20259 (defun Y-axis-column (height width-of-label)
20260 "Construct list of Y axis labels and blank strings.
20261 For HEIGHT of line above base and WIDTH-OF-LABEL."
20265 (while (> height 1)
20266 (if (zerop (% height Y-axis-label-spacing))
20267 ;; @r{Insert label.}
20270 (Y-axis-element height width-of-label)
20274 ;; @r{Else, insert blanks.}
20277 (make-string width-of-label ? )
20279 (setq height (1- height)))
20280 ;; @r{Insert base line.}
20282 (cons (Y-axis-element 1 width-of-label) Y-axis))
20283 (nreverse Y-axis)))
20287 In this function, we start with the value of @code{height} and
20288 repetitively subtract one from its value. After each subtraction, we
20289 test to see whether the value is an integral multiple of the
20290 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20291 using the @code{Y-axis-element} function; if not, we construct a
20292 blank label using the @code{make-string} function. The base line
20293 consists of the number one followed by a tic mark.
20296 @node print-Y-axis Penultimate, , Y-axis-column, print-Y-axis
20297 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20299 The list constructed by the @code{Y-axis-column} function is passed to
20300 the @code{print-Y-axis} function, which inserts the list as a column.
20302 @findex print-Y-axis
20305 (defun print-Y-axis (height full-Y-label-width)
20306 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20307 Height must be the maximum height of the graph.
20308 Full width is the width of the highest label element."
20309 ;; Value of height and full-Y-label-width
20310 ;; are passed by `print-graph'.
20313 (let ((start (point)))
20315 (Y-axis-column height full-Y-label-width))
20316 ;; @r{Place point ready for inserting graph.}
20318 ;; @r{Move point forward by value of} full-Y-label-width
20319 (forward-char full-Y-label-width)))
20323 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20324 insert the Y axis labels created by the @code{Y-axis-column} function.
20325 In addition, it places point at the correct position for printing the body of
20328 You can test @code{print-Y-axis}:
20336 Y-axis-label-spacing
20345 Copy the following expression:
20348 (print-Y-axis 12 5)
20352 Switch to the @file{*scratch*} buffer and place the cursor where you
20353 want the axis labels to start.
20356 Type @kbd{M-:} (@code{eval-expression}).
20359 Yank the @code{graph-body-print} expression into the minibuffer
20360 with @kbd{C-y} (@code{yank)}.
20363 Press @key{RET} to evaluate the expression.
20366 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20367 }}}. (The @code{print-graph} function will pass the value of
20368 @code{height-of-top-line}, which in this case will end up as 15,
20369 thereby getting rid of what might appear as a bug.)
20372 @node print-X-axis, Print Whole Graph, print-Y-axis, Full Graph
20373 @appendixsec The @code{print-X-axis} Function
20374 @cindex Axis, print horizontal
20375 @cindex X axis printing
20376 @cindex Print horizontal axis
20377 @cindex Horizontal axis printing
20379 X axis labels are much like Y axis labels, except that the ticks are on a
20380 line above the numbers. Labels should look like this:
20389 The first tic is under the first column of the graph and is preceded by
20390 several blank spaces. These spaces provide room in rows above for the Y
20391 axis labels. The second, third, fourth, and subsequent ticks are all
20392 spaced equally, according to the value of @code{X-axis-label-spacing}.
20394 The second row of the X axis consists of numbers, preceded by several
20395 blank spaces and also separated according to the value of the variable
20396 @code{X-axis-label-spacing}.
20398 The value of the variable @code{X-axis-label-spacing} should itself be
20399 measured in units of @code{symbol-width}, since you may want to change
20400 the width of the symbols that you are using to print the body of the
20401 graph without changing the ways the graph is labeled.
20404 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
20405 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20408 @node Similarities differences, X Axis Tic Marks, print-X-axis, print-X-axis
20410 @unnumberedsubsec Similarities and differences
20413 The @code{print-X-axis} function is constructed in more or less the
20414 same fashion as the @code{print-Y-axis} function except that it has
20415 two lines: the line of tic marks and the numbers. We will write a
20416 separate function to print each line and then combine them within the
20417 @code{print-X-axis} function.
20419 This is a three step process:
20423 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20426 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20429 Write a function to print both lines, the @code{print-X-axis} function,
20430 using @code{print-X-axis-tic-line} and
20431 @code{print-X-axis-numbered-line}.
20434 @node X Axis Tic Marks, , Similarities differences, print-X-axis
20435 @appendixsubsec X Axis Tic Marks
20437 The first function should print the X axis tic marks. We must specify
20438 the tic marks themselves and their spacing:
20442 (defvar X-axis-label-spacing
20443 (if (boundp 'graph-blank)
20444 (* 5 (length graph-blank)) 5)
20445 "Number of units from one X axis label to next.")
20450 (Note that the value of @code{graph-blank} is set by another
20451 @code{defvar}. The @code{boundp} predicate checks whether it has
20452 already been set; @code{boundp} returns @code{nil} if it has not. If
20453 @code{graph-blank} were unbound and we did not use this conditional
20454 construction, in a recent GNU Emacs, we would enter the debugger and
20455 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20456 @w{(void-variable graph-blank)}}.)
20459 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20463 (defvar X-axis-tic-symbol "|"
20464 "String to insert to point to a column in X axis.")
20469 The goal is to make a line that looks like this:
20475 The first tic is indented so that it is under the first column, which is
20476 indented to provide space for the Y axis labels.
20478 A tic element consists of the blank spaces that stretch from one tic to
20479 the next plus a tic symbol. The number of blanks is determined by the
20480 width of the tic symbol and the @code{X-axis-label-spacing}.
20483 The code looks like this:
20487 ;;; X-axis-tic-element
20491 ;; @r{Make a string of blanks.}
20492 (- (* symbol-width X-axis-label-spacing)
20493 (length X-axis-tic-symbol))
20495 ;; @r{Concatenate blanks with tic symbol.}
20501 Next, we determine how many blanks are needed to indent the first tic
20502 mark to the first column of the graph. This uses the value of
20503 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20506 The code to make @code{X-axis-leading-spaces}
20511 ;; X-axis-leading-spaces
20513 (make-string full-Y-label-width ? )
20518 We also need to determine the length of the horizontal axis, which is
20519 the length of the numbers list, and the number of ticks in the horizontal
20526 (length numbers-list)
20532 (* symbol-width X-axis-label-spacing)
20536 ;; number-of-X-ticks
20537 (if (zerop (% (X-length tic-width)))
20538 (/ (X-length tic-width))
20539 (1+ (/ (X-length tic-width))))
20544 All this leads us directly to the function for printing the X axis tic line:
20546 @findex print-X-axis-tic-line
20549 (defun print-X-axis-tic-line
20550 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20551 "Print ticks for X axis."
20552 (insert X-axis-leading-spaces)
20553 (insert X-axis-tic-symbol) ; @r{Under first column.}
20556 ;; @r{Insert second tic in the right spot.}
20559 (- (* symbol-width X-axis-label-spacing)
20560 ;; @r{Insert white space up to second tic symbol.}
20561 (* 2 (length X-axis-tic-symbol)))
20563 X-axis-tic-symbol))
20566 ;; @r{Insert remaining ticks.}
20567 (while (> number-of-X-tics 1)
20568 (insert X-axis-tic-element)
20569 (setq number-of-X-tics (1- number-of-X-tics))))
20573 The line of numbers is equally straightforward:
20576 First, we create a numbered element with blank spaces before each number:
20578 @findex X-axis-element
20581 (defun X-axis-element (number)
20582 "Construct a numbered X axis element."
20583 (let ((leading-spaces
20584 (- (* symbol-width X-axis-label-spacing)
20585 (length (number-to-string number)))))
20586 (concat (make-string leading-spaces ? )
20587 (number-to-string number))))
20591 Next, we create the function to print the numbered line, starting with
20592 the number ``1'' under the first column:
20594 @findex print-X-axis-numbered-line
20597 (defun print-X-axis-numbered-line
20598 (number-of-X-tics X-axis-leading-spaces)
20599 "Print line of X-axis numbers"
20600 (let ((number X-axis-label-spacing))
20601 (insert X-axis-leading-spaces)
20607 ;; @r{Insert white space up to next number.}
20608 (- (* symbol-width X-axis-label-spacing) 2)
20610 (number-to-string number)))
20613 ;; @r{Insert remaining numbers.}
20614 (setq number (+ number X-axis-label-spacing))
20615 (while (> number-of-X-tics 1)
20616 (insert (X-axis-element number))
20617 (setq number (+ number X-axis-label-spacing))
20618 (setq number-of-X-tics (1- number-of-X-tics)))))
20622 Finally, we need to write the @code{print-X-axis} that uses
20623 @code{print-X-axis-tic-line} and
20624 @code{print-X-axis-numbered-line}.
20626 The function must determine the local values of the variables used by both
20627 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20628 then it must call them. Also, it must print the carriage return that
20629 separates the two lines.
20631 The function consists of a varlist that specifies five local variables,
20632 and calls to each of the two line printing functions:
20634 @findex print-X-axis
20637 (defun print-X-axis (numbers-list)
20638 "Print X axis labels to length of NUMBERS-LIST."
20639 (let* ((leading-spaces
20640 (make-string full-Y-label-width ? ))
20643 ;; symbol-width @r{is provided by} graph-body-print
20644 (tic-width (* symbol-width X-axis-label-spacing))
20645 (X-length (length numbers-list))
20653 ;; @r{Make a string of blanks.}
20654 (- (* symbol-width X-axis-label-spacing)
20655 (length X-axis-tic-symbol))
20659 ;; @r{Concatenate blanks with tic symbol.}
20660 X-axis-tic-symbol))
20664 (if (zerop (% X-length tic-width))
20665 (/ X-length tic-width)
20666 (1+ (/ X-length tic-width)))))
20669 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20671 (print-X-axis-numbered-line tic-number leading-spaces)))
20676 You can test @code{print-X-axis}:
20680 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20681 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20682 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20685 Copy the following expression:
20690 (let ((full-Y-label-width 5)
20693 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20698 Switch to the @file{*scratch*} buffer and place the cursor where you
20699 want the axis labels to start.
20702 Type @kbd{M-:} (@code{eval-expression}).
20705 Yank the test expression into the minibuffer
20706 with @kbd{C-y} (@code{yank)}.
20709 Press @key{RET} to evaluate the expression.
20713 Emacs will print the horizontal axis like this:
20723 @node Print Whole Graph, , print-X-axis, Full Graph
20724 @appendixsec Printing the Whole Graph
20725 @cindex Printing the whole graph
20726 @cindex Whole graph printing
20727 @cindex Graph, printing all
20729 Now we are nearly ready to print the whole graph.
20731 The function to print the graph with the proper labels follows the
20732 outline we created earlier (@pxref{Full Graph, , A Graph with Labeled
20733 Axes}), but with additions.
20736 Here is the outline:
20740 (defun print-graph (numbers-list)
20741 "@var{documentation}@dots{}"
20742 (let ((height @dots{}
20746 (print-Y-axis height @dots{} )
20747 (graph-body-print numbers-list)
20748 (print-X-axis @dots{} )))
20753 * The final version:: A few changes.
20754 * Test print-graph:: Run a short test.
20755 * Graphing words in defuns:: Executing the final code.
20756 * lambda:: How to write an anonymous function.
20757 * mapcar:: Apply a function to elements of a list.
20758 * Another Bug:: Yet another bug @dots{} most insidious.
20759 * Final printed graph:: The graph itself!
20762 @node The final version, Test print-graph, Print Whole Graph, Print Whole Graph
20764 @unnumberedsubsec Changes for the Final Version
20767 The final version is different from what we planned in two ways:
20768 first, it contains additional values calculated once in the varlist;
20769 second, it carries an option to specify the labels' increment per row.
20770 This latter feature turns out to be essential; otherwise, a graph may
20771 have more rows than fit on a display or on a sheet of paper.
20774 This new feature requires a change to the @code{Y-axis-column}
20775 function, to add @code{vertical-step} to it. The function looks like
20778 @findex Y-axis-column @r{Final version.}
20781 ;;; @r{Final version.}
20782 (defun Y-axis-column
20783 (height width-of-label &optional vertical-step)
20784 "Construct list of labels for Y axis.
20785 HEIGHT is maximum height of graph.
20786 WIDTH-OF-LABEL is maximum width of label.
20787 VERTICAL-STEP, an option, is a positive integer
20788 that specifies how much a Y axis label increments
20789 for each line. For example, a step of 5 means
20790 that each line is five units of the graph."
20794 (number-per-line (or vertical-step 1)))
20795 (while (> height 1)
20796 (if (zerop (% height Y-axis-label-spacing))
20799 ;; @r{Insert label.}
20803 (* height number-per-line)
20808 ;; @r{Else, insert blanks.}
20811 (make-string width-of-label ? )
20813 (setq height (1- height)))
20816 ;; @r{Insert base line.}
20817 (setq Y-axis (cons (Y-axis-element
20818 (or vertical-step 1)
20821 (nreverse Y-axis)))
20825 The values for the maximum height of graph and the width of a symbol
20826 are computed by @code{print-graph} in its @code{let} expression; so
20827 @code{graph-body-print} must be changed to accept them.
20829 @findex graph-body-print @r{Final version.}
20832 ;;; @r{Final version.}
20833 (defun graph-body-print (numbers-list height symbol-width)
20834 "Print a bar graph of the NUMBERS-LIST.
20835 The numbers-list consists of the Y-axis values.
20836 HEIGHT is maximum height of graph.
20837 SYMBOL-WIDTH is number of each column."
20840 (let (from-position)
20841 (while numbers-list
20842 (setq from-position (point))
20844 (column-of-graph height (car numbers-list)))
20845 (goto-char from-position)
20846 (forward-char symbol-width)
20849 ;; @r{Draw graph column by column.}
20851 (setq numbers-list (cdr numbers-list)))
20852 ;; @r{Place point for X axis labels.}
20853 (forward-line height)
20859 Finally, the code for the @code{print-graph} function:
20861 @findex print-graph @r{Final version.}
20864 ;;; @r{Final version.}
20866 (numbers-list &optional vertical-step)
20867 "Print labeled bar graph of the NUMBERS-LIST.
20868 The numbers-list consists of the Y-axis values.
20872 Optionally, VERTICAL-STEP, a positive integer,
20873 specifies how much a Y axis label increments for
20874 each line. For example, a step of 5 means that
20875 each row is five units."
20878 (let* ((symbol-width (length graph-blank))
20879 ;; @code{height} @r{is both the largest number}
20880 ;; @r{and the number with the most digits.}
20881 (height (apply 'max numbers-list))
20884 (height-of-top-line
20885 (if (zerop (% height Y-axis-label-spacing))
20888 (* (1+ (/ height Y-axis-label-spacing))
20889 Y-axis-label-spacing)))
20892 (vertical-step (or vertical-step 1))
20893 (full-Y-label-width
20899 (* height-of-top-line vertical-step))
20905 height-of-top-line full-Y-label-width vertical-step)
20909 numbers-list height-of-top-line symbol-width)
20910 (print-X-axis numbers-list)))
20914 @node Test print-graph, Graphing words in defuns, The final version, Print Whole Graph
20915 @appendixsubsec Testing @code{print-graph}
20918 We can test the @code{print-graph} function with a short list of numbers:
20922 Install the final versions of @code{Y-axis-column},
20923 @code{graph-body-print}, and @code{print-graph} (in addition to the
20927 Copy the following expression:
20930 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20934 Switch to the @file{*scratch*} buffer and place the cursor where you
20935 want the axis labels to start.
20938 Type @kbd{M-:} (@code{eval-expression}).
20941 Yank the test expression into the minibuffer
20942 with @kbd{C-y} (@code{yank)}.
20945 Press @key{RET} to evaluate the expression.
20949 Emacs will print a graph that looks like this:
20970 On the other hand, if you pass @code{print-graph} a
20971 @code{vertical-step} value of 2, by evaluating this expression:
20974 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20979 The graph looks like this:
21000 (A question: is the `2' on the bottom of the vertical axis a bug or a
21001 feature? If you think it is a bug, and should be a `1' instead, (or
21002 even a `0'), you can modify the sources.)
21004 @node Graphing words in defuns, lambda, Test print-graph, Print Whole Graph
21005 @appendixsubsec Graphing Numbers of Words and Symbols
21007 Now for the graph for which all this code was written: a graph that
21008 shows how many function definitions contain fewer than 10 words and
21009 symbols, how many contain between 10 and 19 words and symbols, how
21010 many contain between 20 and 29 words and symbols, and so on.
21012 This is a multi-step process. First make sure you have loaded all the
21016 It is a good idea to reset the value of @code{top-of-ranges} in case
21017 you have set it to some different value. You can evaluate the
21022 (setq top-of-ranges
21025 110 120 130 140 150
21026 160 170 180 190 200
21027 210 220 230 240 250
21028 260 270 280 290 300)
21033 Next create a list of the number of words and symbols in each range.
21037 Evaluate the following:
21041 (setq list-for-graph
21044 (recursive-lengths-list-many-files
21045 (directory-files "/usr/local/emacs/lisp"
21053 On my old machine, this took about an hour. It looked though 303 Lisp
21054 files in my copy of Emacs version 19.23. After all that computing,
21055 the @code{list-for-graph} had this value:
21059 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
21060 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
21065 This means that my copy of Emacs had 537 function definitions with
21066 fewer than 10 words or symbols in them, 1,027 function definitions
21067 with 10 to 19 words or symbols in them, 955 function definitions with
21068 20 to 29 words or symbols in them, and so on.
21070 Clearly, just by looking at this list we can see that most function
21071 definitions contain ten to thirty words and symbols.
21073 Now for printing. We do @emph{not} want to print a graph that is
21074 1,030 lines high @dots{} Instead, we should print a graph that is
21075 fewer than twenty-five lines high. A graph that height can be
21076 displayed on almost any monitor, and easily printed on a sheet of paper.
21078 This means that each value in @code{list-for-graph} must be reduced to
21079 one-fiftieth its present value.
21081 Here is a short function to do just that, using two functions we have
21082 not yet seen, @code{mapcar} and @code{lambda}.
21086 (defun one-fiftieth (full-range)
21087 "Return list, each number one-fiftieth of previous."
21088 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21092 @node lambda, mapcar, Graphing words in defuns, Print Whole Graph
21093 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
21094 @cindex Anonymous function
21097 @code{lambda} is the symbol for an anonymous function, a function
21098 without a name. Every time you use an anonymous function, you need to
21099 include its whole body.
21106 (lambda (arg) (/ arg 50))
21110 is a function definition that says `return the value resulting from
21111 dividing whatever is passed to me as @code{arg} by 50'.
21114 Earlier, for example, we had a function @code{multiply-by-seven}; it
21115 multiplied its argument by 7. This function is similar, except it
21116 divides its argument by 50; and, it has no name. The anonymous
21117 equivalent of @code{multiply-by-seven} is:
21120 (lambda (number) (* 7 number))
21124 (@xref{defun, , The @code{defun} Special Form}.)
21128 If we want to multiply 3 by 7, we can write:
21130 @c !!! Clear print-postscript-figures if the computer formatting this
21131 @c document is too small and cannot handle all the diagrams and figures.
21132 @c clear print-postscript-figures
21133 @c set print-postscript-figures
21134 @c lambda example diagram #1
21138 (multiply-by-seven 3)
21139 \_______________/ ^
21145 @ifset print-postscript-figures
21148 @center @image{lambda-1}
21149 %%%% old method of including an image
21150 % \input /usr/local/lib/tex/inputs/psfig.tex
21151 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21156 @ifclear print-postscript-figures
21160 (multiply-by-seven 3)
21161 \_______________/ ^
21170 This expression returns 21.
21174 Similarly, we can write:
21176 @c lambda example diagram #2
21180 ((lambda (number) (* 7 number)) 3)
21181 \____________________________/ ^
21183 anonymous function argument
21187 @ifset print-postscript-figures
21190 @center @image{lambda-2}
21191 %%%% old method of including an image
21192 % \input /usr/local/lib/tex/inputs/psfig.tex
21193 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21198 @ifclear print-postscript-figures
21202 ((lambda (number) (* 7 number)) 3)
21203 \____________________________/ ^
21205 anonymous function argument
21213 If we want to divide 100 by 50, we can write:
21215 @c lambda example diagram #3
21219 ((lambda (arg) (/ arg 50)) 100)
21220 \______________________/ \_/
21222 anonymous function argument
21226 @ifset print-postscript-figures
21229 @center @image{lambda-3}
21230 %%%% old method of including an image
21231 % \input /usr/local/lib/tex/inputs/psfig.tex
21232 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21237 @ifclear print-postscript-figures
21241 ((lambda (arg) (/ arg 50)) 100)
21242 \______________________/ \_/
21244 anonymous function argument
21251 This expression returns 2. The 100 is passed to the function, which
21252 divides that number by 50.
21254 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21255 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21256 expressions derive from the Lambda Calculus.
21258 @node mapcar, Another Bug, lambda, Print Whole Graph
21259 @appendixsubsec The @code{mapcar} Function
21262 @code{mapcar} is a function that calls its first argument with each
21263 element of its second argument, in turn. The second argument must be
21266 The @samp{map} part of the name comes from the mathematical phrase,
21267 `mapping over a domain', meaning to apply a function to each of the
21268 elements in a domain. The mathematical phrase is based on the
21269 metaphor of a surveyor walking, one step at a time, over an area he is
21270 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21279 (mapcar '1+ '(2 4 6))
21285 The function @code{1+} which adds one to its argument, is executed on
21286 @emph{each} element of the list, and a new list is returned.
21288 Contrast this with @code{apply}, which applies its first argument to
21290 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21294 In the definition of @code{one-fiftieth}, the first argument is the
21295 anonymous function:
21298 (lambda (arg) (/ arg 50))
21302 and the second argument is @code{full-range}, which will be bound to
21303 @code{list-for-graph}.
21306 The whole expression looks like this:
21309 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21312 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21313 Lisp Reference Manual}, for more about @code{mapcar}.
21315 Using the @code{one-fiftieth} function, we can generate a list in
21316 which each element is one-fiftieth the size of the corresponding
21317 element in @code{list-for-graph}.
21321 (setq fiftieth-list-for-graph
21322 (one-fiftieth list-for-graph))
21327 The resulting list looks like this:
21331 (10 20 19 15 11 9 6 5 4 3 3 2 2
21332 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21337 This, we are almost ready to print! (We also notice the loss of
21338 information: many of the higher ranges are 0, meaning that fewer than
21339 50 defuns had that many words or symbols---but not necessarily meaning
21340 that none had that many words or symbols.)
21342 @node Another Bug, Final printed graph, mapcar, Print Whole Graph
21343 @appendixsubsec Another Bug @dots{} Most Insidious
21344 @cindex Bug, most insidious type
21345 @cindex Insidious type of bug
21347 I said `almost ready to print'! Of course, there is a bug in the
21348 @code{print-graph} function @dots{} It has a @code{vertical-step}
21349 option, but not a @code{horizontal-step} option. The
21350 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21351 @code{print-graph} function will print only by ones.
21353 This is a classic example of what some consider the most insidious
21354 type of bug, the bug of omission. This is not the kind of bug you can
21355 find by studying the code, for it is not in the code; it is an omitted
21356 feature. Your best actions are to try your program early and often;
21357 and try to arrange, as much as you can, to write code that is easy to
21358 understand and easy to change. Try to be aware, whenever you can,
21359 that whatever you have written, @emph{will} be rewritten, if not soon,
21360 eventually. A hard maxim to follow.
21362 It is the @code{print-X-axis-numbered-line} function that needs the
21363 work; and then the @code{print-X-axis} and the @code{print-graph}
21364 functions need to be adapted. Not much needs to be done; there is one
21365 nicety: the numbers ought to line up under the tic marks. This takes
21369 Here is the corrected @code{print-X-axis-numbered-line}:
21373 (defun print-X-axis-numbered-line
21374 (number-of-X-tics X-axis-leading-spaces
21375 &optional horizontal-step)
21376 "Print line of X-axis numbers"
21377 (let ((number X-axis-label-spacing)
21378 (horizontal-step (or horizontal-step 1)))
21381 (insert X-axis-leading-spaces)
21382 ;; @r{Delete extra leading spaces.}
21385 (length (number-to-string horizontal-step)))))
21390 ;; @r{Insert white space.}
21392 X-axis-label-spacing)
21395 (number-to-string horizontal-step)))
21399 (* number horizontal-step))))
21402 ;; @r{Insert remaining numbers.}
21403 (setq number (+ number X-axis-label-spacing))
21404 (while (> number-of-X-tics 1)
21405 (insert (X-axis-element
21406 (* number horizontal-step)))
21407 (setq number (+ number X-axis-label-spacing))
21408 (setq number-of-X-tics (1- number-of-X-tics)))))
21413 If you are reading this in Info, you can see the new versions of
21414 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21415 reading this in a printed book, you can see the changed lines here
21416 (the full text is too much to print).
21421 (defun print-X-axis (numbers-list horizontal-step)
21423 (print-X-axis-numbered-line
21424 tic-number leading-spaces horizontal-step))
21432 &optional vertical-step horizontal-step)
21434 (print-X-axis numbers-list horizontal-step))
21442 (defun print-X-axis (numbers-list horizontal-step)
21443 "Print X axis labels to length of NUMBERS-LIST.
21444 Optionally, HORIZONTAL-STEP, a positive integer,
21445 specifies how much an X axis label increments for
21449 ;; Value of symbol-width and full-Y-label-width
21450 ;; are passed by `print-graph'.
21451 (let* ((leading-spaces
21452 (make-string full-Y-label-width ? ))
21453 ;; symbol-width @r{is provided by} graph-body-print
21454 (tic-width (* symbol-width X-axis-label-spacing))
21455 (X-length (length numbers-list))
21461 ;; @r{Make a string of blanks.}
21462 (- (* symbol-width X-axis-label-spacing)
21463 (length X-axis-tic-symbol))
21467 ;; @r{Concatenate blanks with tic symbol.}
21468 X-axis-tic-symbol))
21470 (if (zerop (% X-length tic-width))
21471 (/ X-length tic-width)
21472 (1+ (/ X-length tic-width)))))
21476 (print-X-axis-tic-line
21477 tic-number leading-spaces X-tic)
21479 (print-X-axis-numbered-line
21480 tic-number leading-spaces horizontal-step)))
21487 (numbers-list &optional vertical-step horizontal-step)
21488 "Print labeled bar graph of the NUMBERS-LIST.
21489 The numbers-list consists of the Y-axis values.
21493 Optionally, VERTICAL-STEP, a positive integer,
21494 specifies how much a Y axis label increments for
21495 each line. For example, a step of 5 means that
21496 each row is five units.
21500 Optionally, HORIZONTAL-STEP, a positive integer,
21501 specifies how much an X axis label increments for
21503 (let* ((symbol-width (length graph-blank))
21504 ;; @code{height} @r{is both the largest number}
21505 ;; @r{and the number with the most digits.}
21506 (height (apply 'max numbers-list))
21509 (height-of-top-line
21510 (if (zerop (% height Y-axis-label-spacing))
21513 (* (1+ (/ height Y-axis-label-spacing))
21514 Y-axis-label-spacing)))
21517 (vertical-step (or vertical-step 1))
21518 (full-Y-label-width
21522 (* height-of-top-line vertical-step))
21527 height-of-top-line full-Y-label-width vertical-step)
21529 numbers-list height-of-top-line symbol-width)
21530 (print-X-axis numbers-list horizontal-step)))
21537 Graphing Definitions Re-listed
21540 Here are all the graphing definitions in their final form:
21544 (defvar top-of-ranges
21547 110 120 130 140 150
21548 160 170 180 190 200
21549 210 220 230 240 250)
21550 "List specifying ranges for `defuns-per-range'.")
21554 (defvar graph-symbol "*"
21555 "String used as symbol in graph, usually an asterisk.")
21559 (defvar graph-blank " "
21560 "String used as blank in graph, usually a blank space.
21561 graph-blank must be the same number of columns wide
21566 (defvar Y-axis-tic " - "
21567 "String that follows number in a Y axis label.")
21571 (defvar Y-axis-label-spacing 5
21572 "Number of lines from one Y axis label to next.")
21576 (defvar X-axis-tic-symbol "|"
21577 "String to insert to point to a column in X axis.")
21581 (defvar X-axis-label-spacing
21582 (if (boundp 'graph-blank)
21583 (* 5 (length graph-blank)) 5)
21584 "Number of units from one X axis label to next.")
21590 (defun count-words-in-defun ()
21591 "Return the number of words and symbols in a defun."
21592 (beginning-of-defun)
21594 (end (save-excursion (end-of-defun) (point))))
21599 (and (< (point) end)
21601 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21603 (setq count (1+ count)))
21610 (defun lengths-list-file (filename)
21611 "Return list of definitions' lengths within FILE.
21612 The returned list is a list of numbers.
21613 Each number is the number of words or
21614 symbols in one function definition."
21618 (message "Working on `%s' ... " filename)
21620 (let ((buffer (find-file-noselect filename))
21622 (set-buffer buffer)
21623 (setq buffer-read-only t)
21625 (goto-char (point-min))
21629 (while (re-search-forward "^(defun" nil t)
21631 (cons (count-words-in-defun) lengths-list)))
21632 (kill-buffer buffer)
21639 (defun lengths-list-many-files (list-of-files)
21640 "Return list of lengths of defuns in LIST-OF-FILES."
21641 (let (lengths-list)
21642 ;;; @r{true-or-false-test}
21643 (while list-of-files
21649 ;;; @r{Generate a lengths' list.}
21651 (expand-file-name (car list-of-files)))))
21652 ;;; @r{Make files' list shorter.}
21653 (setq list-of-files (cdr list-of-files)))
21654 ;;; @r{Return final value of lengths' list.}
21661 (defun defuns-per-range (sorted-lengths top-of-ranges)
21662 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21663 (let ((top-of-range (car top-of-ranges))
21664 (number-within-range 0)
21665 defuns-per-range-list)
21670 (while top-of-ranges
21674 ;; @r{Need number for numeric test.}
21675 (car sorted-lengths)
21676 (< (car sorted-lengths) top-of-range))
21678 ;; @r{Count number of definitions within current range.}
21679 (setq number-within-range (1+ number-within-range))
21680 (setq sorted-lengths (cdr sorted-lengths)))
21684 ;; @r{Exit inner loop but remain within outer loop.}
21686 (setq defuns-per-range-list
21687 (cons number-within-range defuns-per-range-list))
21688 (setq number-within-range 0) ; @r{Reset count to zero.}
21690 ;; @r{Move to next range.}
21691 (setq top-of-ranges (cdr top-of-ranges))
21692 ;; @r{Specify next top of range value.}
21693 (setq top-of-range (car top-of-ranges)))
21697 ;; @r{Exit outer loop and count the number of defuns larger than}
21698 ;; @r{ the largest top-of-range value.}
21699 (setq defuns-per-range-list
21701 (length sorted-lengths)
21702 defuns-per-range-list))
21704 ;; @r{Return a list of the number of definitions within each range,}
21705 ;; @r{ smallest to largest.}
21706 (nreverse defuns-per-range-list)))
21712 (defun column-of-graph (max-graph-height actual-height)
21713 "Return list of MAX-GRAPH-HEIGHT strings;
21714 ACTUAL-HEIGHT are graph-symbols.
21715 The graph-symbols are contiguous entries at the end
21717 The list will be inserted as one column of a graph.
21718 The strings are either graph-blank or graph-symbol."
21722 (let ((insert-list nil)
21723 (number-of-top-blanks
21724 (- max-graph-height actual-height)))
21726 ;; @r{Fill in @code{graph-symbols}.}
21727 (while (> actual-height 0)
21728 (setq insert-list (cons graph-symbol insert-list))
21729 (setq actual-height (1- actual-height)))
21733 ;; @r{Fill in @code{graph-blanks}.}
21734 (while (> number-of-top-blanks 0)
21735 (setq insert-list (cons graph-blank insert-list))
21736 (setq number-of-top-blanks
21737 (1- number-of-top-blanks)))
21739 ;; @r{Return whole list.}
21746 (defun Y-axis-element (number full-Y-label-width)
21747 "Construct a NUMBERed label element.
21748 A numbered element looks like this ` 5 - ',
21749 and is padded as needed so all line up with
21750 the element for the largest number."
21753 (let* ((leading-spaces
21754 (- full-Y-label-width
21756 (concat (number-to-string number)
21761 (make-string leading-spaces ? )
21762 (number-to-string number)
21769 (defun print-Y-axis
21770 (height full-Y-label-width &optional vertical-step)
21771 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21772 Height must be the maximum height of the graph.
21773 Full width is the width of the highest label element.
21774 Optionally, print according to VERTICAL-STEP."
21777 ;; Value of height and full-Y-label-width
21778 ;; are passed by `print-graph'.
21779 (let ((start (point)))
21781 (Y-axis-column height full-Y-label-width vertical-step))
21784 ;; @r{Place point ready for inserting graph.}
21786 ;; @r{Move point forward by value of} full-Y-label-width
21787 (forward-char full-Y-label-width)))
21793 (defun print-X-axis-tic-line
21794 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21795 "Print ticks for X axis."
21796 (insert X-axis-leading-spaces)
21797 (insert X-axis-tic-symbol) ; @r{Under first column.}
21800 ;; @r{Insert second tic in the right spot.}
21803 (- (* symbol-width X-axis-label-spacing)
21804 ;; @r{Insert white space up to second tic symbol.}
21805 (* 2 (length X-axis-tic-symbol)))
21807 X-axis-tic-symbol))
21810 ;; @r{Insert remaining ticks.}
21811 (while (> number-of-X-tics 1)
21812 (insert X-axis-tic-element)
21813 (setq number-of-X-tics (1- number-of-X-tics))))
21819 (defun X-axis-element (number)
21820 "Construct a numbered X axis element."
21821 (let ((leading-spaces
21822 (- (* symbol-width X-axis-label-spacing)
21823 (length (number-to-string number)))))
21824 (concat (make-string leading-spaces ? )
21825 (number-to-string number))))
21831 (defun graph-body-print (numbers-list height symbol-width)
21832 "Print a bar graph of the NUMBERS-LIST.
21833 The numbers-list consists of the Y-axis values.
21834 HEIGHT is maximum height of graph.
21835 SYMBOL-WIDTH is number of each column."
21838 (let (from-position)
21839 (while numbers-list
21840 (setq from-position (point))
21842 (column-of-graph height (car numbers-list)))
21843 (goto-char from-position)
21844 (forward-char symbol-width)
21847 ;; @r{Draw graph column by column.}
21849 (setq numbers-list (cdr numbers-list)))
21850 ;; @r{Place point for X axis labels.}
21851 (forward-line height)
21858 (defun Y-axis-column
21859 (height width-of-label &optional vertical-step)
21860 "Construct list of labels for Y axis.
21861 HEIGHT is maximum height of graph.
21862 WIDTH-OF-LABEL is maximum width of label.
21865 VERTICAL-STEP, an option, is a positive integer
21866 that specifies how much a Y axis label increments
21867 for each line. For example, a step of 5 means
21868 that each line is five units of the graph."
21870 (number-per-line (or vertical-step 1)))
21873 (while (> height 1)
21874 (if (zerop (% height Y-axis-label-spacing))
21875 ;; @r{Insert label.}
21879 (* height number-per-line)
21884 ;; @r{Else, insert blanks.}
21887 (make-string width-of-label ? )
21889 (setq height (1- height)))
21892 ;; @r{Insert base line.}
21893 (setq Y-axis (cons (Y-axis-element
21894 (or vertical-step 1)
21897 (nreverse Y-axis)))
21903 (defun print-X-axis-numbered-line
21904 (number-of-X-tics X-axis-leading-spaces
21905 &optional horizontal-step)
21906 "Print line of X-axis numbers"
21907 (let ((number X-axis-label-spacing)
21908 (horizontal-step (or horizontal-step 1)))
21911 (insert X-axis-leading-spaces)
21913 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21916 ;; @r{Insert white space up to next number.}
21917 (- (* symbol-width X-axis-label-spacing)
21918 (1- (length (number-to-string horizontal-step)))
21921 (number-to-string (* number horizontal-step))))
21924 ;; @r{Insert remaining numbers.}
21925 (setq number (+ number X-axis-label-spacing))
21926 (while (> number-of-X-tics 1)
21927 (insert (X-axis-element (* number horizontal-step)))
21928 (setq number (+ number X-axis-label-spacing))
21929 (setq number-of-X-tics (1- number-of-X-tics)))))
21935 (defun print-X-axis (numbers-list horizontal-step)
21936 "Print X axis labels to length of NUMBERS-LIST.
21937 Optionally, HORIZONTAL-STEP, a positive integer,
21938 specifies how much an X axis label increments for
21942 ;; Value of symbol-width and full-Y-label-width
21943 ;; are passed by `print-graph'.
21944 (let* ((leading-spaces
21945 (make-string full-Y-label-width ? ))
21946 ;; symbol-width @r{is provided by} graph-body-print
21947 (tic-width (* symbol-width X-axis-label-spacing))
21948 (X-length (length numbers-list))
21954 ;; @r{Make a string of blanks.}
21955 (- (* symbol-width X-axis-label-spacing)
21956 (length X-axis-tic-symbol))
21960 ;; @r{Concatenate blanks with tic symbol.}
21961 X-axis-tic-symbol))
21963 (if (zerop (% X-length tic-width))
21964 (/ X-length tic-width)
21965 (1+ (/ X-length tic-width)))))
21969 (print-X-axis-tic-line
21970 tic-number leading-spaces X-tic)
21972 (print-X-axis-numbered-line
21973 tic-number leading-spaces horizontal-step)))
21979 (defun one-fiftieth (full-range)
21980 "Return list, each number of which is 1/50th previous."
21981 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21988 (numbers-list &optional vertical-step horizontal-step)
21989 "Print labeled bar graph of the NUMBERS-LIST.
21990 The numbers-list consists of the Y-axis values.
21994 Optionally, VERTICAL-STEP, a positive integer,
21995 specifies how much a Y axis label increments for
21996 each line. For example, a step of 5 means that
21997 each row is five units.
22001 Optionally, HORIZONTAL-STEP, a positive integer,
22002 specifies how much an X axis label increments for
22004 (let* ((symbol-width (length graph-blank))
22005 ;; @code{height} @r{is both the largest number}
22006 ;; @r{and the number with the most digits.}
22007 (height (apply 'max numbers-list))
22010 (height-of-top-line
22011 (if (zerop (% height Y-axis-label-spacing))
22014 (* (1+ (/ height Y-axis-label-spacing))
22015 Y-axis-label-spacing)))
22018 (vertical-step (or vertical-step 1))
22019 (full-Y-label-width
22023 (* height-of-top-line vertical-step))
22029 height-of-top-line full-Y-label-width vertical-step)
22031 numbers-list height-of-top-line symbol-width)
22032 (print-X-axis numbers-list horizontal-step)))
22039 @node Final printed graph, , Another Bug, Print Whole Graph
22040 @appendixsubsec The Printed Graph
22042 When made and installed, you can call the @code{print-graph} command
22048 (print-graph fiftieth-list-for-graph 50 10)
22078 50 - ***************** * *
22080 10 50 100 150 200 250 300 350
22087 The largest group of functions contain 10 -- 19 words and symbols each.
22089 @node Free Software and Free Manuals, GNU Free Documentation License, Full Graph, Top
22090 @appendix Free Software and Free Manuals
22092 @strong{by Richard M. Stallman}
22095 The biggest deficiency in free operating systems is not in the
22096 software---it is the lack of good free manuals that we can include in
22097 these systems. Many of our most important programs do not come with
22098 full manuals. Documentation is an essential part of any software
22099 package; when an important free software package does not come with a
22100 free manual, that is a major gap. We have many such gaps today.
22102 Once upon a time, many years ago, I thought I would learn Perl. I got
22103 a copy of a free manual, but I found it hard to read. When I asked
22104 Perl users about alternatives, they told me that there were better
22105 introductory manuals---but those were not free.
22107 Why was this? The authors of the good manuals had written them for
22108 O'Reilly Associates, which published them with restrictive terms---no
22109 copying, no modification, source files not available---which exclude
22110 them from the free software community.
22112 That wasn't the first time this sort of thing has happened, and (to
22113 our community's great loss) it was far from the last. Proprietary
22114 manual publishers have enticed a great many authors to restrict their
22115 manuals since then. Many times I have heard a GNU user eagerly tell me
22116 about a manual that he is writing, with which he expects to help the
22117 GNU project---and then had my hopes dashed, as he proceeded to explain
22118 that he had signed a contract with a publisher that would restrict it
22119 so that we cannot use it.
22121 Given that writing good English is a rare skill among programmers, we
22122 can ill afford to lose manuals this way.
22124 Free documentation, like free software, is a matter of freedom, not
22125 price. The problem with these manuals was not that O'Reilly Associates
22126 charged a price for printed copies---that in itself is fine. The Free
22127 Software Foundation @uref{http://shop.fsf.org, sells printed copies} of
22128 free @uref{http://www.gnu.org/doc/doc.html, GNU manuals}, too.
22129 But GNU manuals are available in source code form, while these manuals
22130 are available only on paper. GNU manuals come with permission to copy
22131 and modify; the Perl manuals do not. These restrictions are the
22134 The criterion for a free manual is pretty much the same as for free
22135 software: it is a matter of giving all users certain
22136 freedoms. Redistribution (including commercial redistribution) must be
22137 permitted, so that the manual can accompany every copy of the program,
22138 on-line or on paper. Permission for modification is crucial too.
22140 As a general rule, I don't believe that it is essential for people to
22141 have permission to modify all sorts of articles and books. The issues
22142 for writings are not necessarily the same as those for software. For
22143 example, I don't think you or I are obliged to give permission to
22144 modify articles like this one, which describe our actions and our
22147 But there is a particular reason why the freedom to modify is crucial
22148 for documentation for free software. When people exercise their right
22149 to modify the software, and add or change its features, if they are
22150 conscientious they will change the manual too---so they can provide
22151 accurate and usable documentation with the modified program. A manual
22152 which forbids programmers to be conscientious and finish the job, or
22153 more precisely requires them to write a new manual from scratch if
22154 they change the program, does not fill our community's needs.
22156 While a blanket prohibition on modification is unacceptable, some
22157 kinds of limits on the method of modification pose no problem. For
22158 example, requirements to preserve the original author's copyright
22159 notice, the distribution terms, or the list of authors, are ok. It is
22160 also no problem to require modified versions to include notice that
22161 they were modified, even to have entire sections that may not be
22162 deleted or changed, as long as these sections deal with nontechnical
22163 topics. (Some GNU manuals have them.)
22165 These kinds of restrictions are not a problem because, as a practical
22166 matter, they don't stop the conscientious programmer from adapting the
22167 manual to fit the modified program. In other words, they don't block
22168 the free software community from making full use of the manual.
22170 However, it must be possible to modify all the technical content of
22171 the manual, and then distribute the result in all the usual media,
22172 through all the usual channels; otherwise, the restrictions do block
22173 the community, the manual is not free, and so we need another manual.
22175 Unfortunately, it is often hard to find someone to write another
22176 manual when a proprietary manual exists. The obstacle is that many
22177 users think that a proprietary manual is good enough---so they don't
22178 see the need to write a free manual. They do not see that the free
22179 operating system has a gap that needs filling.
22181 Why do users think that proprietary manuals are good enough? Some have
22182 not considered the issue. I hope this article will do something to
22185 Other users consider proprietary manuals acceptable for the same
22186 reason so many people consider proprietary software acceptable: they
22187 judge in purely practical terms, not using freedom as a
22188 criterion. These people are entitled to their opinions, but since
22189 those opinions spring from values which do not include freedom, they
22190 are no guide for those of us who do value freedom.
22192 Please spread the word about this issue. We continue to lose manuals
22193 to proprietary publishing. If we spread the word that proprietary
22194 manuals are not sufficient, perhaps the next person who wants to help
22195 GNU by writing documentation will realize, before it is too late, that
22196 he must above all make it free.
22198 We can also encourage commercial publishers to sell free, copylefted
22199 manuals instead of proprietary ones. One way you can help this is to
22200 check the distribution terms of a manual before you buy it, and prefer
22201 copylefted manuals to non-copylefted ones.
22205 Note: The Free Software Foundation maintains a page on its Web site
22206 that lists free books available from other publishers:@*
22207 @uref{http://www.gnu.org/doc/other-free-books.html}
22209 @node GNU Free Documentation License, Index, Free Software and Free Manuals, Top
22210 @appendix GNU Free Documentation License
22212 @cindex FDL, GNU Free Documentation License
22213 @include doclicense.texi
22215 @node Index, About the Author, GNU Free Documentation License, Top
22216 @comment node-name, next, previous, up
22220 MENU ENTRY: NODE NAME.
22226 @c Place biographical information on right-hand (verso) page
22229 \par\vfill\supereject
22231 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22232 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22235 % \par\vfill\supereject
22236 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22237 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22238 %\page\hbox{}%\page
22239 %\page\hbox{}%\page
22246 @c ================ Biographical information ================
22250 @center About the Author
22255 @node About the Author, , Index, Top
22256 @unnumbered About the Author
22260 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22261 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22262 world on software freedom. Chassell was a founding Director and
22263 Treasurer of the Free Software Foundation, Inc. He is co-author of
22264 the @cite{Texinfo} manual, and has edited more than a dozen other
22265 books. He graduated from Cambridge University, in England. He has an
22266 abiding interest in social and economic history and flies his own
22273 @c @c Prevent page number on blank verso, so eject it first.
22275 @c \par\vfill\supereject
22280 @c @evenheading @thispage @| @| @thistitle
22281 @c @oddheading @| @| @thispage