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 ----------------------------------------------------
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, 1991, 1992, 1993, 1994, 1995, 1997, 2001,
232 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
233 Free Software Foundation, Inc.
239 GNU Press, @hfill @uref{http://www.gnupress.org}@*
240 a division of the @hfill General: @email{press@@gnu.org}@*
241 Free Software Foundation, Inc. @hfill Orders:@w{ } @email{sales@@gnu.org}@*
242 51 Franklin Street, Fifth Floor @hfill Tel: +1 (617) 542-5942@*
243 Boston, MA 02110-1301 USA @hfill Fax: +1 (617) 542-2652@*
250 GNU Press, Website: http://www.gnupress.org
251 a division of the General: press@@gnu.org
252 Free Software Foundation, Inc. Orders: sales@@gnu.org
253 51 Franklin Street, Fifth Floor Tel: +1 (617) 542-5942
254 Boston, MA 02110-1301 USA Fax: +1 (617) 542-2652
259 @c Printed copies are available for $30 each.@*
262 Permission is granted to copy, distribute and/or modify this document
263 under the terms of the GNU Free Documentation License, Version 1.3 or
264 any later version published by the Free Software Foundation; there
265 being no Invariant Section, with the Front-Cover Texts being ``A GNU
266 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
267 the license is included in the section entitled ``GNU Free
268 Documentation License''.
270 (a) The FSF's Back-Cover Text is: ``You have the freedom to
271 copy and modify this GNU manual. Buying copies from the FSF
272 supports it in developing GNU and promoting software freedom.''
275 @c half title; two lines here, so do not use `shorttitlepage'
278 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
280 {\begingroup\hbox{}\vskip 0.25in \chaprm%
281 \centerline{Programming in Emacs Lisp}%
282 \endgroup\page\hbox{}\page}
287 @center @titlefont{An Introduction to}
289 @center @titlefont{Programming in Emacs Lisp}
291 @center Revised Third Edition
293 @center by Robert J. Chassell
296 @vskip 0pt plus 1filll
302 @evenheading @thispage @| @| @thischapter
303 @oddheading @thissection @| @| @thispage
307 @c Keep T.O.C. short by tightening up for largebook
310 \global\parskip 2pt plus 1pt
311 \global\advance\baselineskip by -1pt
320 @node Top, Preface, (dir), (dir)
321 @top An Introduction to Programming in Emacs Lisp
325 This master menu first lists each chapter and index; then it lists
326 every node in every chapter.
329 @c >>>> Set pageno appropriately <<<<
331 @c The first page of the Preface is a roman numeral; it is the first
332 @c right handed page after the Table of Contents; hence the following
333 @c setting must be for an odd negative number.
336 @c global@pageno = -11
340 * Preface:: What to look for.
341 * List Processing:: What is Lisp?
342 * Practicing Evaluation:: Running several programs.
343 * Writing Defuns:: How to write function definitions.
344 * Buffer Walk Through:: Exploring a few buffer-related functions.
345 * More Complex:: A few, even more complex functions.
346 * Narrowing & Widening:: Restricting your and Emacs attention to
348 * car cdr & cons:: Fundamental functions in Lisp.
349 * Cutting & Storing Text:: Removing text and saving it.
350 * List Implementation:: How lists are implemented in the computer.
351 * Yanking:: Pasting stored text.
352 * Loops & Recursion:: How to repeat a process.
353 * Regexp Search:: Regular expression searches.
354 * Counting Words:: A review of repetition and regexps.
355 * Words in a defun:: Counting words in a @code{defun}.
356 * Readying a Graph:: A prototype graph printing function.
357 * Emacs Initialization:: How to write a @file{.emacs} file.
358 * Debugging:: How to run the Emacs Lisp debuggers.
359 * Conclusion:: Now you have the basics.
360 * the-the:: An appendix: how to find reduplicated words.
361 * Kill Ring:: An appendix: how the kill ring works.
362 * Full Graph:: How to create a graph with labelled axes.
363 * Free Software and Free Manuals::
364 * GNU Free Documentation License::
369 --- The Detailed Node Listing ---
373 * Why:: Why learn Emacs Lisp?
374 * On Reading this Text:: Read, gain familiarity, pick up habits....
375 * Who You Are:: For whom this is written.
377 * Note for Novices:: You can read this as a novice.
382 * Lisp Lists:: What are lists?
383 * Run a Program:: Any list in Lisp is a program ready to run.
384 * Making Errors:: Generating an error message.
385 * Names & Definitions:: Names of symbols and function definitions.
386 * Lisp Interpreter:: What the Lisp interpreter does.
387 * Evaluation:: Running a program.
388 * Variables:: Returning a value from a variable.
389 * Arguments:: Passing information to a function.
390 * set & setq:: Setting the value of a variable.
391 * Summary:: The major points.
392 * Error Message Exercises::
396 * Numbers Lists:: List have numbers, other lists, in them.
397 * Lisp Atoms:: Elemental entities.
398 * Whitespace in Lists:: Formatting lists to be readable.
399 * Typing Lists:: How GNU Emacs helps you type lists.
403 * Complications:: Variables, Special forms, Lists within.
404 * Byte Compiling:: Specially processing code for speed.
408 * How the Interpreter Acts:: Returns and Side Effects...
409 * Evaluating Inner Lists:: Lists within lists...
413 * fill-column Example::
414 * Void Function:: The error message for a symbol
416 * Void Variable:: The error message for a symbol without a value.
420 * Data types:: Types of data passed to a function.
421 * Args as Variable or List:: An argument can be the value
422 of a variable or list.
423 * Variable Number of Arguments:: Some functions may take a
424 variable number of arguments.
425 * Wrong Type of Argument:: Passing an argument of the wrong type
427 * message:: A useful function for sending messages.
429 Setting the Value of a Variable
431 * Using set:: Setting values.
432 * Using setq:: Setting a quoted value.
433 * Counting:: Using @code{setq} to count.
435 Practicing Evaluation
437 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
439 * Buffer Names:: Buffers and files are different.
440 * Getting Buffers:: Getting a buffer itself, not merely its name.
441 * Switching Buffers:: How to change to another buffer.
442 * Buffer Size & Locations:: Where point is located and the size of
444 * Evaluation Exercise::
446 How To Write Function Definitions
448 * Primitive Functions::
449 * defun:: The @code{defun} special form.
450 * Install:: Install a function definition.
451 * Interactive:: Making a function interactive.
452 * Interactive Options:: Different options for @code{interactive}.
453 * Permanent Installation:: Installing code permanently.
454 * let:: Creating and initializing local variables.
456 * else:: If--then--else expressions.
457 * Truth & Falsehood:: What Lisp considers false and true.
458 * save-excursion:: Keeping track of point, mark, and buffer.
462 Install a Function Definition
464 * Effect of installation::
465 * Change a defun:: How to change a function definition.
467 Make a Function Interactive
469 * Interactive multiply-by-seven:: An overview.
470 * multiply-by-seven in detail:: The interactive version.
474 * Prevent confusion::
475 * Parts of let Expression::
476 * Sample let Expression::
477 * Uninitialized let Variables::
479 The @code{if} Special Form
481 * if in more detail::
482 * type-of-animal in detail:: An example of an @code{if} expression.
484 Truth and Falsehood in Emacs Lisp
486 * nil explained:: @code{nil} has two meanings.
488 @code{save-excursion}
490 * Point and mark:: A review of various locations.
491 * Template for save-excursion::
493 A Few Buffer--Related Functions
495 * Finding More:: How to find more information.
496 * simplified-beginning-of-buffer:: Shows @code{goto-char},
497 @code{point-min}, and @code{push-mark}.
498 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
499 * append-to-buffer:: Uses @code{save-excursion} and
500 @code{insert-buffer-substring}.
501 * Buffer Related Review:: Review.
504 The Definition of @code{mark-whole-buffer}
506 * mark-whole-buffer overview::
507 * Body of mark-whole-buffer:: Only three lines of code.
509 The Definition of @code{append-to-buffer}
511 * append-to-buffer overview::
512 * append interactive:: A two part interactive expression.
513 * append-to-buffer body:: Incorporates a @code{let} expression.
514 * append save-excursion:: How the @code{save-excursion} works.
516 A Few More Complex Functions
518 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
519 * insert-buffer:: Read-only, and with @code{or}.
520 * beginning-of-buffer:: Shows @code{goto-char},
521 @code{point-min}, and @code{push-mark}.
522 * Second Buffer Related Review::
523 * optional Exercise::
525 The Definition of @code{insert-buffer}
527 * insert-buffer code::
528 * insert-buffer interactive:: When you can read, but not write.
529 * insert-buffer body:: The body has an @code{or} and a @code{let}.
530 * if & or:: Using an @code{if} instead of an @code{or}.
531 * Insert or:: How the @code{or} expression works.
532 * Insert let:: Two @code{save-excursion} expressions.
533 * New insert-buffer::
535 The Interactive Expression in @code{insert-buffer}
537 * Read-only buffer:: When a buffer cannot be modified.
538 * b for interactive:: An existing buffer or else its name.
540 Complete Definition of @code{beginning-of-buffer}
542 * Optional Arguments::
543 * beginning-of-buffer opt arg:: Example with optional argument.
544 * beginning-of-buffer complete::
546 @code{beginning-of-buffer} with an Argument
548 * Disentangle beginning-of-buffer::
549 * Large buffer case::
550 * Small buffer case::
552 Narrowing and Widening
554 * Narrowing advantages:: The advantages of narrowing
555 * save-restriction:: The @code{save-restriction} special form.
556 * what-line:: The number of the line that point is on.
559 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
561 * Strange Names:: An historical aside: why the strange names?
562 * car & cdr:: Functions for extracting part of a list.
563 * cons:: Constructing a list.
564 * nthcdr:: Calling @code{cdr} repeatedly.
566 * setcar:: Changing the first element of a list.
567 * setcdr:: Changing the rest of a list.
573 * length:: How to find the length of a list.
575 Cutting and Storing Text
577 * Storing Text:: Text is stored in a list.
578 * zap-to-char:: Cutting out text up to a character.
579 * kill-region:: Cutting text out of a region.
580 * copy-region-as-kill:: A definition for copying text.
581 * Digression into C:: Minor note on C programming language macros.
582 * defvar:: How to give a variable an initial value.
583 * cons & search-fwd Review::
588 * Complete zap-to-char:: The complete implementation.
589 * zap-to-char interactive:: A three part interactive expression.
590 * zap-to-char body:: A short overview.
591 * search-forward:: How to search for a string.
592 * progn:: The @code{progn} special form.
593 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
597 * Complete kill-region:: The function definition.
598 * condition-case:: Dealing with a problem.
601 @code{copy-region-as-kill}
603 * Complete copy-region-as-kill:: The complete function definition.
604 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
606 The Body of @code{copy-region-as-kill}
608 * last-command & this-command::
609 * kill-append function::
610 * kill-new function::
612 Initializing a Variable with @code{defvar}
614 * See variable current value::
615 * defvar and asterisk::
617 How Lists are Implemented
620 * Symbols as Chest:: Exploring a powerful metaphor.
625 * Kill Ring Overview::
626 * kill-ring-yank-pointer:: The kill ring is a list.
627 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
631 * while:: Causing a stretch of code to repeat.
633 * Recursion:: Causing a function to call itself.
638 * Looping with while:: Repeat so long as test returns true.
639 * Loop Example:: A @code{while} loop that uses a list.
640 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
641 * Incrementing Loop:: A loop with an incrementing counter.
642 * Incrementing Loop Details::
643 * Decrementing Loop:: A loop with a decrementing counter.
645 Details of an Incrementing Loop
647 * Incrementing Example:: Counting pebbles in a triangle.
648 * Inc Example parts:: The parts of the function definition.
649 * Inc Example altogether:: Putting the function definition together.
651 Loop with a Decrementing Counter
653 * Decrementing Example:: More pebbles on the beach.
654 * Dec Example parts:: The parts of the function definition.
655 * Dec Example altogether:: Putting the function definition together.
657 Save your time: @code{dolist} and @code{dotimes}
664 * Building Robots:: Same model, different serial number ...
665 * Recursive Definition Parts:: Walk until you stop ...
666 * Recursion with list:: Using a list as the test whether to recurse.
667 * Recursive triangle function::
668 * Recursion with cond::
669 * Recursive Patterns:: Often used templates.
670 * No Deferment:: Don't store up work ...
671 * No deferment solution::
673 Recursion in Place of a Counter
675 * Recursive Example arg of 1 or 2::
676 * Recursive Example arg of 3 or 4::
684 Regular Expression Searches
686 * sentence-end:: The regular expression for @code{sentence-end}.
687 * re-search-forward:: Very similar to @code{search-forward}.
688 * forward-sentence:: A straightforward example of regexp search.
689 * forward-paragraph:: A somewhat complex example.
690 * etags:: How to create your own @file{TAGS} table.
692 * re-search Exercises::
694 @code{forward-sentence}
696 * Complete forward-sentence::
697 * fwd-sentence while loops:: Two @code{while} loops.
698 * fwd-sentence re-search:: A regular expression search.
700 @code{forward-paragraph}: a Goldmine of Functions
702 * forward-paragraph in brief:: Key parts of the function definition.
703 * fwd-para let:: The @code{let*} expression.
704 * fwd-para while:: The forward motion @code{while} loop.
706 Counting: Repetition and Regexps
707 @set COUNT-WORDS count-words-example
708 @c Length of variable name chosen so that things still line up when expanded.
711 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
712 * recursive-count-words:: Start with case of no words in region.
713 * Counting Exercise::
715 The @code{@value{COUNT-WORDS}} Function
717 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
718 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
720 Counting Words in a @code{defun}
722 * Divide and Conquer::
723 * Words and Symbols:: What to count?
724 * Syntax:: What constitutes a word or symbol?
725 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
726 * Several defuns:: Counting several defuns in a file.
727 * Find a File:: Do you want to look at a file?
728 * lengths-list-file:: A list of the lengths of many definitions.
729 * Several files:: Counting in definitions in different files.
730 * Several files recursively:: Recursively counting in different files.
731 * Prepare the data:: Prepare the data for display in a graph.
733 Count Words in @code{defuns} in Different Files
735 * lengths-list-many-files:: Return a list of the lengths of defuns.
736 * append:: Attach one list to another.
738 Prepare the Data for Display in a Graph
740 * Data for Display in Detail::
741 * Sorting:: Sorting lists.
742 * Files List:: Making a list of files.
743 * Counting function definitions::
747 * Columns of a graph::
748 * graph-body-print:: How to print the body of a graph.
749 * recursive-graph-body-print::
751 * Line Graph Exercise::
753 Your @file{.emacs} File
755 * Default Configuration::
756 * Site-wide Init:: You can write site-wide init files.
757 * defcustom:: Emacs will write code for you.
758 * Beginning a .emacs File:: How to write a @code{.emacs file}.
759 * Text and Auto-fill:: Automatically wrap lines.
760 * Mail Aliases:: Use abbreviations for email addresses.
761 * Indent Tabs Mode:: Don't use tabs with @TeX{}
762 * Keybindings:: Create some personal keybindings.
763 * Keymaps:: More about key binding.
764 * Loading Files:: Load (i.e., evaluate) files automatically.
765 * Autoload:: Make functions available.
766 * Simple Extension:: Define a function; bind it to a key.
767 * X11 Colors:: Colors in X.
769 * Mode Line:: How to customize your mode line.
773 * debug:: How to use the built-in debugger.
774 * debug-on-entry:: Start debugging when you call a function.
775 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
776 * edebug:: How to use Edebug, a source level debugger.
777 * Debugging Exercises::
779 Handling the Kill Ring
781 * What the Kill Ring Does::
783 * yank:: Paste a copy of a clipped element.
784 * yank-pop:: Insert element pointed to.
787 The @code{current-kill} Function
789 * Code for current-kill::
790 * Understanding current-kill::
792 @code{current-kill} in Outline
794 * Body of current-kill::
795 * Digression concerning error:: How to mislead humans, but not computers.
796 * Determining the Element::
798 A Graph with Labelled Axes
801 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
802 * print-Y-axis:: Print a label for the vertical axis.
803 * print-X-axis:: Print a horizontal label.
804 * Print Whole Graph:: The function to print a complete graph.
806 The @code{print-Y-axis} Function
808 * print-Y-axis in Detail::
809 * Height of label:: What height for the Y axis?
810 * Compute a Remainder:: How to compute the remainder of a division.
811 * Y Axis Element:: Construct a line for the Y axis.
812 * Y-axis-column:: Generate a list of Y axis labels.
813 * print-Y-axis Penultimate:: A not quite final version.
815 The @code{print-X-axis} Function
817 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
818 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
820 Printing the Whole Graph
822 * The final version:: A few changes.
823 * Test print-graph:: Run a short test.
824 * Graphing words in defuns:: Executing the final code.
825 * lambda:: How to write an anonymous function.
826 * mapcar:: Apply a function to elements of a list.
827 * Another Bug:: Yet another bug @dots{} most insidious.
828 * Final printed graph:: The graph itself!
833 @node Preface, List Processing, Top, Top
834 @comment node-name, next, previous, up
837 Most of the GNU Emacs integrated environment is written in the programming
838 language called Emacs Lisp. The code written in this programming
839 language is the software---the sets of instructions---that tell the
840 computer what to do when you give it commands. Emacs is designed so
841 that you can write new code in Emacs Lisp and easily install it as an
842 extension to the editor.
844 (GNU Emacs is sometimes called an ``extensible editor'', but it does
845 much more than provide editing capabilities. It is better to refer to
846 Emacs as an ``extensible computing environment''. However, that
847 phrase is quite a mouthful. It is easier to refer to Emacs simply as
848 an editor. Moreover, everything you do in Emacs---find the Mayan date
849 and phases of the moon, simplify polynomials, debug code, manage
850 files, read letters, write books---all these activities are kinds of
851 editing in the most general sense of the word.)
854 * Why:: Why learn Emacs Lisp?
855 * On Reading this Text:: Read, gain familiarity, pick up habits....
856 * Who You Are:: For whom this is written.
858 * Note for Novices:: You can read this as a novice.
862 @node Why, On Reading this Text, Preface, Preface
864 @unnumberedsec Why Study Emacs Lisp?
867 Although Emacs Lisp is usually thought of in association only with Emacs,
868 it is a full computer programming language. You can use Emacs Lisp as
869 you would any other programming language.
871 Perhaps you want to understand programming; perhaps you want to extend
872 Emacs; or perhaps you want to become a programmer. This introduction to
873 Emacs Lisp is designed to get you started: to guide you in learning the
874 fundamentals of programming, and more importantly, to show you how you
875 can teach yourself to go further.
877 @node On Reading this Text, Who You Are, Why, Preface
878 @comment node-name, next, previous, up
879 @unnumberedsec On Reading this Text
881 All through this document, you will see little sample programs you can
882 run inside of Emacs. If you read this document in Info inside of GNU
883 Emacs, you can run the programs as they appear. (This is easy to do and
884 is explained when the examples are presented.) Alternatively, you can
885 read this introduction as a printed book while sitting beside a computer
886 running Emacs. (This is what I like to do; I like printed books.) If
887 you don't have a running Emacs beside you, you can still read this book,
888 but in this case, it is best to treat it as a novel or as a travel guide
889 to a country not yet visited: interesting, but not the same as being
892 Much of this introduction is dedicated to walk-throughs or guided tours
893 of code used in GNU Emacs. These tours are designed for two purposes:
894 first, to give you familiarity with real, working code (code you use
895 every day); and, second, to give you familiarity with the way Emacs
896 works. It is interesting to see how a working environment is
899 hope that you will pick up the habit of browsing through source code.
900 You can learn from it and mine it for ideas. Having GNU Emacs is like
901 having a dragon's cave of treasures.
903 In addition to learning about Emacs as an editor and Emacs Lisp as a
904 programming language, the examples and guided tours will give you an
905 opportunity to get acquainted with Emacs as a Lisp programming
906 environment. GNU Emacs supports programming and provides tools that
907 you will want to become comfortable using, such as @kbd{M-.} (the key
908 which invokes the @code{find-tag} command). You will also learn about
909 buffers and other objects that are part of the environment.
910 Learning about these features of Emacs is like learning new routes
911 around your home town.
914 In addition, I have written several programs as extended examples.
915 Although these are examples, the programs are real. I use them.
916 Other people use them. You may use them. Beyond the fragments of
917 programs used for illustrations, there is very little in here that is
918 `just for teaching purposes'; what you see is used. This is a great
919 advantage of Emacs Lisp: it is easy to learn to use it for work.
922 Finally, I hope to convey some of the skills for using Emacs to
923 learn aspects of programming that you don't know. You can often use
924 Emacs to help you understand what puzzles you or to find out how to do
925 something new. This self-reliance is not only a pleasure, but an
928 @node Who You Are, Lisp History, On Reading this Text, Preface
929 @comment node-name, next, previous, up
930 @unnumberedsec For Whom This is Written
932 This text is written as an elementary introduction for people who are
933 not programmers. If you are a programmer, you may not be satisfied with
934 this primer. The reason is that you may have become expert at reading
935 reference manuals and be put off by the way this text is organized.
937 An expert programmer who reviewed this text said to me:
940 @i{I prefer to learn from reference manuals. I ``dive into'' each
941 paragraph, and ``come up for air'' between paragraphs.}
943 @i{When I get to the end of a paragraph, I assume that that subject is
944 done, finished, that I know everything I need (with the
945 possible exception of the case when the next paragraph starts talking
946 about it in more detail). I expect that a well written reference manual
947 will not have a lot of redundancy, and that it will have excellent
948 pointers to the (one) place where the information I want is.}
951 This introduction is not written for this person!
953 Firstly, I try to say everything at least three times: first, to
954 introduce it; second, to show it in context; and third, to show it in a
955 different context, or to review it.
957 Secondly, I hardly ever put all the information about a subject in one
958 place, much less in one paragraph. To my way of thinking, that imposes
959 too heavy a burden on the reader. Instead I try to explain only what
960 you need to know at the time. (Sometimes I include a little extra
961 information so you won't be surprised later when the additional
962 information is formally introduced.)
964 When you read this text, you are not expected to learn everything the
965 first time. Frequently, you need only make, as it were, a `nodding
966 acquaintance' with some of the items mentioned. My hope is that I have
967 structured the text and given you enough hints that you will be alert to
968 what is important, and concentrate on it.
970 You will need to ``dive into'' some paragraphs; there is no other way
971 to read them. But I have tried to keep down the number of such
972 paragraphs. This book is intended as an approachable hill, rather than
973 as a daunting mountain.
975 This introduction to @cite{Programming in Emacs Lisp} has a companion
978 @cite{The GNU Emacs Lisp Reference Manual}.
981 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
982 Emacs Lisp Reference Manual}.
984 The reference manual has more detail than this introduction. In the
985 reference manual, all the information about one topic is concentrated
986 in one place. You should turn to it if you are like the programmer
987 quoted above. And, of course, after you have read this
988 @cite{Introduction}, you will find the @cite{Reference Manual} useful
989 when you are writing your own programs.
991 @node Lisp History, Note for Novices, Who You Are, Preface
992 @unnumberedsec Lisp History
995 Lisp was first developed in the late 1950s at the Massachusetts
996 Institute of Technology for research in artificial intelligence. The
997 great power of the Lisp language makes it superior for other purposes as
998 well, such as writing editor commands and integrated environments.
1002 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
1003 in the 1960s. It is somewhat inspired by Common Lisp, which became a
1004 standard in the 1980s. However, Emacs Lisp is much simpler than Common
1005 Lisp. (The standard Emacs distribution contains an optional extensions
1006 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
1008 @node Note for Novices, Thank You, Lisp History, Preface
1009 @comment node-name, next, previous, up
1010 @unnumberedsec A Note for Novices
1012 If you don't know GNU Emacs, you can still read this document
1013 profitably. However, I recommend you learn Emacs, if only to learn to
1014 move around your computer screen. You can teach yourself how to use
1015 Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This
1016 means you press and release the @key{CTRL} key and the @kbd{h} at the
1017 same time, and then press and release @kbd{t}.)
1019 Also, I often refer to one of Emacs' standard commands by listing the
1020 keys which you press to invoke the command and then giving the name of
1021 the command in parentheses, like this: @kbd{M-C-\}
1022 (@code{indent-region}). What this means is that the
1023 @code{indent-region} command is customarily invoked by typing
1024 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
1025 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
1026 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
1027 @key{META} key, @key{CTRL} key and @key{\} key all at the same time.
1028 (On many modern keyboards the @key{META} key is labelled
1030 Sometimes a combination like this is called a keychord, since it is
1031 similar to the way you play a chord on a piano. If your keyboard does
1032 not have a @key{META} key, the @key{ESC} key prefix is used in place
1033 of it. In this case, @kbd{M-C-\} means that you press and release your
1034 @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at
1035 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
1036 along with the key that is labelled @key{ALT} and, at the same time,
1037 press the @key{\} key.
1039 In addition to typing a lone keychord, you can prefix what you type
1040 with @kbd{C-u}, which is called the `universal argument'. The
1041 @kbd{C-u} keychord passes an argument to the subsequent command.
1042 Thus, to indent a region of plain text by 6 spaces, mark the region,
1043 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
1044 Emacs either passes the number 4 to the command or otherwise runs the
1045 command differently than it would otherwise.) @xref{Arguments, ,
1046 Numeric Arguments, emacs, The GNU Emacs Manual}.
1048 If you are reading this in Info using GNU Emacs, you can read through
1049 this whole document just by pressing the space bar, @key{SPC}.
1050 (To learn about Info, type @kbd{C-h i} and then select Info.)
1052 A note on terminology: when I use the word Lisp alone, I often am
1053 referring to the various dialects of Lisp in general, but when I speak
1054 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
1056 @node Thank You, , Note for Novices, Preface
1057 @comment node-name, next, previous, up
1058 @unnumberedsec Thank You
1060 My thanks to all who helped me with this book. My especial thanks to
1061 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
1062 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.@:
1063 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
1064 @w{Philip Johnson} and @w{David Stampe} for their patient
1065 encouragement. My mistakes are my own.
1069 @email{bob@@gnu.org}
1072 @c ================ Beginning of main text ================
1074 @c Start main text on right-hand (verso) page
1077 \par\vfill\supereject
1080 \par\vfill\supereject
1082 \par\vfill\supereject
1084 \par\vfill\supereject
1090 @evenheading @thispage @| @| @thischapter
1091 @oddheading @thissection @| @| @thispage
1095 @node List Processing, Practicing Evaluation, Preface, Top
1096 @comment node-name, next, previous, up
1097 @chapter List Processing
1099 To the untutored eye, Lisp is a strange programming language. In Lisp
1100 code there are parentheses everywhere. Some people even claim that
1101 the name stands for `Lots of Isolated Silly Parentheses'. But the
1102 claim is unwarranted. Lisp stands for LISt Processing, and the
1103 programming language handles @emph{lists} (and lists of lists) by
1104 putting them between parentheses. The parentheses mark the boundaries
1105 of the list. Sometimes a list is preceded by a single apostrophe or
1106 quotation mark, @samp{'}@footnote{The single apostrophe or quotation
1107 mark is an abbreviation for the function @code{quote}; you need not
1108 think about functions now; functions are defined in @ref{Making
1109 Errors, , Generate an Error Message}.} Lists are the basis of Lisp.
1112 * Lisp Lists:: What are lists?
1113 * Run a Program:: Any list in Lisp is a program ready to run.
1114 * Making Errors:: Generating an error message.
1115 * Names & Definitions:: Names of symbols and function definitions.
1116 * Lisp Interpreter:: What the Lisp interpreter does.
1117 * Evaluation:: Running a program.
1118 * Variables:: Returning a value from a variable.
1119 * Arguments:: Passing information to a function.
1120 * set & setq:: Setting the value of a variable.
1121 * Summary:: The major points.
1122 * Error Message Exercises::
1125 @node Lisp Lists, Run a Program, List Processing, List Processing
1126 @comment node-name, next, previous, up
1130 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1131 This list is preceded by a single apostrophe. It could just as well be
1132 written as follows, which looks more like the kind of list you are likely
1133 to be familiar with:
1145 The elements of this list are the names of the four different flowers,
1146 separated from each other by whitespace and surrounded by parentheses,
1147 like flowers in a field with a stone wall around them.
1148 @cindex Flowers in a field
1151 * Numbers Lists:: List have numbers, other lists, in them.
1152 * Lisp Atoms:: Elemental entities.
1153 * Whitespace in Lists:: Formatting lists to be readable.
1154 * Typing Lists:: How GNU Emacs helps you type lists.
1157 @node Numbers Lists, Lisp Atoms, Lisp Lists, Lisp Lists
1159 @unnumberedsubsec Numbers, Lists inside of Lists
1162 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1163 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1164 separated by whitespace.
1166 In Lisp, both data and programs are represented the same way; that is,
1167 they are both lists of words, numbers, or other lists, separated by
1168 whitespace and surrounded by parentheses. (Since a program looks like
1169 data, one program may easily serve as data for another; this is a very
1170 powerful feature of Lisp.) (Incidentally, these two parenthetical
1171 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1172 @samp{.} as punctuation marks.)
1175 Here is another list, this time with a list inside of it:
1178 '(this list has (a list inside of it))
1181 The components of this list are the words @samp{this}, @samp{list},
1182 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1183 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1184 @samp{of}, @samp{it}.
1186 @node Lisp Atoms, Whitespace in Lists, Numbers Lists, Lisp Lists
1187 @comment node-name, next, previous, up
1188 @subsection Lisp Atoms
1191 In Lisp, what we have been calling words are called @dfn{atoms}. This
1192 term comes from the historical meaning of the word atom, which means
1193 `indivisible'. As far as Lisp is concerned, the words we have been
1194 using in the lists cannot be divided into any smaller parts and still
1195 mean the same thing as part of a program; likewise with numbers and
1196 single character symbols like @samp{+}. On the other hand, unlike an
1197 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1198 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1200 In a list, atoms are separated from each other by whitespace. They can be
1201 right next to a parenthesis.
1203 @cindex @samp{empty list} defined
1204 Technically speaking, a list in Lisp consists of parentheses surrounding
1205 atoms separated by whitespace or surrounding other lists or surrounding
1206 both atoms and other lists. A list can have just one atom in it or
1207 have nothing in it at all. A list with nothing in it looks like this:
1208 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1209 empty list is considered both an atom and a list at the same time.
1211 @cindex Symbolic expressions, introduced
1212 @cindex @samp{expression} defined
1213 @cindex @samp{form} defined
1214 The printed representation of both atoms and lists are called
1215 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1216 The word @dfn{expression} by itself can refer to either the printed
1217 representation, or to the atom or list as it is held internally in the
1218 computer. Often, people use the term @dfn{expression}
1219 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1220 as a synonym for expression.)
1222 Incidentally, the atoms that make up our universe were named such when
1223 they were thought to be indivisible; but it has been found that physical
1224 atoms are not indivisible. Parts can split off an atom or it can
1225 fission into two parts of roughly equal size. Physical atoms were named
1226 prematurely, before their truer nature was found. In Lisp, certain
1227 kinds of atom, such as an array, can be separated into parts; but the
1228 mechanism for doing this is different from the mechanism for splitting a
1229 list. As far as list operations are concerned, the atoms of a list are
1232 As in English, the meanings of the component letters of a Lisp atom
1233 are different from the meaning the letters make as a word. For
1234 example, the word for the South American sloth, the @samp{ai}, is
1235 completely different from the two words, @samp{a}, and @samp{i}.
1237 There are many kinds of atom in nature but only a few in Lisp: for
1238 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1239 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1240 listed in the examples above are all symbols. In everyday Lisp
1241 conversation, the word ``atom'' is not often used, because programmers
1242 usually try to be more specific about what kind of atom they are dealing
1243 with. Lisp programming is mostly about symbols (and sometimes numbers)
1244 within lists. (Incidentally, the preceding three word parenthetical
1245 remark is a proper list in Lisp, since it consists of atoms, which in
1246 this case are symbols, separated by whitespace and enclosed by
1247 parentheses, without any non-Lisp punctuation.)
1250 Text between double quotation marks---even sentences or
1251 paragraphs---is also an atom. Here is an example:
1252 @cindex Text between double quotation marks
1255 '(this list includes "text between quotation marks.")
1258 @cindex @samp{string} defined
1260 In Lisp, all of the quoted text including the punctuation mark and the
1261 blank spaces is a single atom. This kind of atom is called a
1262 @dfn{string} (for `string of characters') and is the sort of thing that
1263 is used for messages that a computer can print for a human to read.
1264 Strings are a different kind of atom than numbers or symbols and are
1267 @node Whitespace in Lists, Typing Lists, Lisp Atoms, Lisp Lists
1268 @comment node-name, next, previous, up
1269 @subsection Whitespace in Lists
1270 @cindex Whitespace in lists
1273 The amount of whitespace in a list does not matter. From the point of view
1274 of the Lisp language,
1285 is exactly the same as this:
1288 '(this list looks like this)
1291 Both examples show what to Lisp is the same list, the list made up of
1292 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1293 @samp{this} in that order.
1295 Extra whitespace and newlines are designed to make a list more readable
1296 by humans. When Lisp reads the expression, it gets rid of all the extra
1297 whitespace (but it needs to have at least one space between atoms in
1298 order to tell them apart.)
1300 Odd as it seems, the examples we have seen cover almost all of what Lisp
1301 lists look like! Every other list in Lisp looks more or less like one
1302 of these examples, except that the list may be longer and more complex.
1303 In brief, a list is between parentheses, a string is between quotation
1304 marks, a symbol looks like a word, and a number looks like a number.
1305 (For certain situations, square brackets, dots and a few other special
1306 characters may be used; however, we will go quite far without them.)
1308 @node Typing Lists, , Whitespace in Lists, Lisp Lists
1309 @comment node-name, next, previous, up
1310 @subsection GNU Emacs Helps You Type Lists
1311 @cindex Help typing lists
1312 @cindex Formatting help
1314 When you type a Lisp expression in GNU Emacs using either Lisp
1315 Interaction mode or Emacs Lisp mode, you have available to you several
1316 commands to format the Lisp expression so it is easy to read. For
1317 example, pressing the @key{TAB} key automatically indents the line the
1318 cursor is on by the right amount. A command to properly indent the
1319 code in a region is customarily bound to @kbd{M-C-\}. Indentation is
1320 designed so that you can see which elements of a list belong to which
1321 list---elements of a sub-list are indented more than the elements of
1324 In addition, when you type a closing parenthesis, Emacs momentarily
1325 jumps the cursor back to the matching opening parenthesis, so you can
1326 see which one it is. This is very useful, since every list you type
1327 in Lisp must have its closing parenthesis match its opening
1328 parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs
1329 Manual}, for more information about Emacs' modes.)
1331 @node Run a Program, Making Errors, Lisp Lists, List Processing
1332 @comment node-name, next, previous, up
1333 @section Run a Program
1334 @cindex Run a program
1335 @cindex Program, running one
1337 @cindex @samp{evaluate} defined
1338 A list in Lisp---any list---is a program ready to run. If you run it
1339 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1340 of three things: do nothing except return to you the list itself; send
1341 you an error message; or, treat the first symbol in the list as a
1342 command to do something. (Usually, of course, it is the last of these
1343 three things that you really want!)
1345 @c use code for the single apostrophe, not samp.
1346 The single apostrophe, @code{'}, that I put in front of some of the
1347 example lists in preceding sections is called a @dfn{quote}; when it
1348 precedes a list, it tells Lisp to do nothing with the list, other than
1349 take it as it is written. But if there is no quote preceding a list,
1350 the first item of the list is special: it is a command for the computer
1351 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1352 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1353 understands that the @code{+} is an instruction to do something with the
1354 rest of the list: add the numbers that follow.
1357 If you are reading this inside of GNU Emacs in Info, here is how you can
1358 evaluate such a list: place your cursor immediately after the right
1359 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1365 @c use code for the number four, not samp.
1367 You will see the number @code{4} appear in the echo area. (In the
1368 jargon, what you have just done is ``evaluate the list.'' The echo area
1369 is the line at the bottom of the screen that displays or ``echoes''
1370 text.) Now try the same thing with a quoted list: place the cursor
1371 right after the following list and type @kbd{C-x C-e}:
1374 '(this is a quoted list)
1378 You will see @code{(this is a quoted list)} appear in the echo area.
1380 @cindex Lisp interpreter, explained
1381 @cindex Interpreter, Lisp, explained
1382 In both cases, what you are doing is giving a command to the program
1383 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1384 interpreter a command to evaluate the expression. The name of the Lisp
1385 interpreter comes from the word for the task done by a human who comes
1386 up with the meaning of an expression---who ``interprets'' it.
1388 You can also evaluate an atom that is not part of a list---one that is
1389 not surrounded by parentheses; again, the Lisp interpreter translates
1390 from the humanly readable expression to the language of the computer.
1391 But before discussing this (@pxref{Variables}), we will discuss what the
1392 Lisp interpreter does when you make an error.
1394 @node Making Errors, Names & Definitions, Run a Program, List Processing
1395 @comment node-name, next, previous, up
1396 @section Generate an Error Message
1397 @cindex Generate an error message
1398 @cindex Error message generation
1400 Partly so you won't worry if you do it accidentally, we will now give
1401 a command to the Lisp interpreter that generates an error message.
1402 This is a harmless activity; and indeed, we will often try to generate
1403 error messages intentionally. Once you understand the jargon, error
1404 messages can be informative. Instead of being called ``error''
1405 messages, they should be called ``help'' messages. They are like
1406 signposts to a traveller in a strange country; deciphering them can be
1407 hard, but once understood, they can point the way.
1409 The error message is generated by a built-in GNU Emacs debugger. We
1410 will `enter the debugger'. You get out of the debugger by typing @code{q}.
1412 What we will do is evaluate a list that is not quoted and does not
1413 have a meaningful command as its first element. Here is a list almost
1414 exactly the same as the one we just used, but without the single-quote
1415 in front of it. Position the cursor right after it and type @kbd{C-x
1419 (this is an unquoted list)
1423 What you see depends on which version of Emacs you are running. GNU
1424 Emacs version 22 provides more information than version 20 and before.
1425 First, the more recent result of generating an error; then the
1426 earlier, version 20 result.
1430 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1431 you will see the following in it:
1435 ---------- Buffer: *Backtrace* ----------
1436 Debugger entered--Lisp error: (void-function this)
1437 (this is an unquoted list)
1438 eval((this is an unquoted list))
1439 eval-last-sexp-1(nil)
1441 call-interactively(eval-last-sexp)
1442 ---------- Buffer: *Backtrace* ----------
1448 Your cursor will be in this window (you may have to wait a few seconds
1449 before it becomes visible). To quit the debugger and make the
1450 debugger window go away, type:
1457 Please type @kbd{q} right now, so you become confident that you can
1458 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1461 @cindex @samp{function} defined
1462 Based on what we already know, we can almost read this error message.
1464 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1465 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1466 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1467 an abbreviation for `evaluate' and @code{sexp} is an abbreviation for
1468 `symbolic expression'. The command means `evaluate last symbolic
1469 expression', which is the expression just before your cursor.
1471 Each line above tells you what the Lisp interpreter evaluated next.
1472 The most recent action is at the top. The buffer is called the
1473 @file{*Backtrace*} buffer because it enables you to track Emacs
1477 At the top of the @file{*Backtrace*} buffer, you see the line:
1480 Debugger entered--Lisp error: (void-function this)
1484 The Lisp interpreter tried to evaluate the first atom of the list, the
1485 word @samp{this}. It is this action that generated the error message
1486 @samp{void-function this}.
1488 The message contains the words @samp{void-function} and @samp{this}.
1490 @cindex @samp{function} defined
1491 The word @samp{function} was mentioned once before. It is a very
1492 important word. For our purposes, we can define it by saying that a
1493 @dfn{function} is a set of instructions to the computer that tell the
1494 computer to do something.
1496 Now we can begin to understand the error message: @samp{void-function
1497 this}. The function (that is, the word @samp{this}) does not have a
1498 definition of any set of instructions for the computer to carry out.
1500 The slightly odd word, @samp{void-function}, is designed to cover the
1501 way Emacs Lisp is implemented, which is that when a symbol does not
1502 have a function definition attached to it, the place that should
1503 contain the instructions is `void'.
1505 On the other hand, since we were able to add 2 plus 2 successfully, by
1506 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1507 have a set of instructions for the computer to obey and those
1508 instructions must be to add the numbers that follow the @code{+}.
1511 In GNU Emacs version 20, and in earlier versions, you will see only
1512 one line of error message; it will appear in the echo area and look
1516 Symbol's function definition is void:@: this
1520 (Also, your terminal may beep at you---some do, some don't; and others
1521 blink. This is just a device to get your attention.) The message goes
1522 away as soon as you type another key, even just to move the cursor.
1524 We know the meaning of the word @samp{Symbol}. It refers to the first
1525 atom of the list, the word @samp{this}. The word @samp{function}
1526 refers to the instructions that tell the computer what to do.
1527 (Technically, the symbol tells the computer where to find the
1528 instructions, but this is a complication we can ignore for the
1531 The error message can be understood: @samp{Symbol's function
1532 definition is void:@: this}. The symbol (that is, the word
1533 @samp{this}) lacks instructions for the computer to carry out.
1535 @node Names & Definitions, Lisp Interpreter, Making Errors, List Processing
1536 @comment node-name, next, previous, up
1537 @section Symbol Names and Function Definitions
1538 @cindex Symbol names
1540 We can articulate another characteristic of Lisp based on what we have
1541 discussed so far---an important characteristic: a symbol, like
1542 @code{+}, is not itself the set of instructions for the computer to
1543 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1544 of locating the definition or set of instructions. What we see is the
1545 name through which the instructions can be found. Names of people
1546 work the same way. I can be referred to as @samp{Bob}; however, I am
1547 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1548 consciousness consistently associated with a particular life-form.
1549 The name is not me, but it can be used to refer to me.
1551 In Lisp, one set of instructions can be attached to several names.
1552 For example, the computer instructions for adding numbers can be
1553 linked to the symbol @code{plus} as well as to the symbol @code{+}
1554 (and are in some dialects of Lisp). Among humans, I can be referred
1555 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1557 On the other hand, a symbol can have only one function definition
1558 attached to it at a time. Otherwise, the computer would be confused as
1559 to which definition to use. If this were the case among people, only
1560 one person in the world could be named @samp{Bob}. However, the function
1561 definition to which the name refers can be changed readily.
1562 (@xref{Install, , Install a Function Definition}.)
1564 Since Emacs Lisp is large, it is customary to name symbols in a way
1565 that identifies the part of Emacs to which the function belongs.
1566 Thus, all the names for functions that deal with Texinfo start with
1567 @samp{texinfo-} and those for functions that deal with reading mail
1568 start with @samp{rmail-}.
1570 @node Lisp Interpreter, Evaluation, Names & Definitions, List Processing
1571 @comment node-name, next, previous, up
1572 @section The Lisp Interpreter
1573 @cindex Lisp interpreter, what it does
1574 @cindex Interpreter, what it does
1576 Based on what we have seen, we can now start to figure out what the
1577 Lisp interpreter does when we command it to evaluate a list.
1578 First, it looks to see whether there is a quote before the list; if
1579 there is, the interpreter just gives us the list. On the other
1580 hand, if there is no quote, the interpreter looks at the first element
1581 in the list and sees whether it has a function definition. If it does,
1582 the interpreter carries out the instructions in the function definition.
1583 Otherwise, the interpreter prints an error message.
1585 This is how Lisp works. Simple. There are added complications which we
1586 will get to in a minute, but these are the fundamentals. Of course, to
1587 write Lisp programs, you need to know how to write function definitions
1588 and attach them to names, and how to do this without confusing either
1589 yourself or the computer.
1592 * Complications:: Variables, Special forms, Lists within.
1593 * Byte Compiling:: Specially processing code for speed.
1596 @node Complications, Byte Compiling, Lisp Interpreter, Lisp Interpreter
1598 @unnumberedsubsec Complications
1601 Now, for the first complication. In addition to lists, the Lisp
1602 interpreter can evaluate a symbol that is not quoted and does not have
1603 parentheses around it. The Lisp interpreter will attempt to determine
1604 the symbol's value as a @dfn{variable}. This situation is described
1605 in the section on variables. (@xref{Variables}.)
1607 @cindex Special form
1608 The second complication occurs because some functions are unusual and do
1609 not work in the usual manner. Those that don't are called @dfn{special
1610 forms}. They are used for special jobs, like defining a function, and
1611 there are not many of them. In the next few chapters, you will be
1612 introduced to several of the more important special forms.
1614 The third and final complication is this: if the function that the
1615 Lisp interpreter is looking at is not a special form, and if it is part
1616 of a list, the Lisp interpreter looks to see whether the list has a list
1617 inside of it. If there is an inner list, the Lisp interpreter first
1618 figures out what it should do with the inside list, and then it works on
1619 the outside list. If there is yet another list embedded inside the
1620 inner list, it works on that one first, and so on. It always works on
1621 the innermost list first. The interpreter works on the innermost list
1622 first, to evaluate the result of that list. The result may be
1623 used by the enclosing expression.
1625 Otherwise, the interpreter works left to right, from one expression to
1628 @node Byte Compiling, , Complications, Lisp Interpreter
1629 @subsection Byte Compiling
1630 @cindex Byte compiling
1632 One other aspect of interpreting: the Lisp interpreter is able to
1633 interpret two kinds of entity: humanly readable code, on which we will
1634 focus exclusively, and specially processed code, called @dfn{byte
1635 compiled} code, which is not humanly readable. Byte compiled code
1636 runs faster than humanly readable code.
1638 You can transform humanly readable code into byte compiled code by
1639 running one of the compile commands such as @code{byte-compile-file}.
1640 Byte compiled code is usually stored in a file that ends with a
1641 @file{.elc} extension rather than a @file{.el} extension. You will
1642 see both kinds of file in the @file{emacs/lisp} directory; the files
1643 to read are those with @file{.el} extensions.
1645 As a practical matter, for most things you might do to customize or
1646 extend Emacs, you do not need to byte compile; and I will not discuss
1647 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1648 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1651 @node Evaluation, Variables, Lisp Interpreter, List Processing
1652 @comment node-name, next, previous, up
1656 When the Lisp interpreter works on an expression, the term for the
1657 activity is called @dfn{evaluation}. We say that the interpreter
1658 `evaluates the expression'. I've used this term several times before.
1659 The word comes from its use in everyday language, `to ascertain the
1660 value or amount of; to appraise', according to @cite{Webster's New
1661 Collegiate Dictionary}.
1664 * How the Interpreter Acts:: Returns and Side Effects...
1665 * Evaluating Inner Lists:: Lists within lists...
1668 @node How the Interpreter Acts, Evaluating Inner Lists, Evaluation, Evaluation
1670 @unnumberedsubsec How the Lisp Interpreter Acts
1673 @cindex @samp{returned value} explained
1674 After evaluating an expression, the Lisp interpreter will most likely
1675 @dfn{return} the value that the computer produces by carrying out the
1676 instructions it found in the function definition, or perhaps it will
1677 give up on that function and produce an error message. (The interpreter
1678 may also find itself tossed, so to speak, to a different function or it
1679 may attempt to repeat continually what it is doing for ever and ever in
1680 what is called an `infinite loop'. These actions are less common; and
1681 we can ignore them.) Most frequently, the interpreter returns a value.
1683 @cindex @samp{side effect} defined
1684 At the same time the interpreter returns a value, it may do something
1685 else as well, such as move a cursor or copy a file; this other kind of
1686 action is called a @dfn{side effect}. Actions that we humans think are
1687 important, such as printing results, are often ``side effects'' to the
1688 Lisp interpreter. The jargon can sound peculiar, but it turns out that
1689 it is fairly easy to learn to use side effects.
1691 In summary, evaluating a symbolic expression most commonly causes the
1692 Lisp interpreter to return a value and perhaps carry out a side effect;
1693 or else produce an error.
1695 @node Evaluating Inner Lists, , How the Interpreter Acts, Evaluation
1696 @comment node-name, next, previous, up
1697 @subsection Evaluating Inner Lists
1698 @cindex Inner list evaluation
1699 @cindex Evaluating inner lists
1701 If evaluation applies to a list that is inside another list, the outer
1702 list may use the value returned by the first evaluation as information
1703 when the outer list is evaluated. This explains why inner expressions
1704 are evaluated first: the values they return are used by the outer
1708 We can investigate this process by evaluating another addition example.
1709 Place your cursor after the following expression and type @kbd{C-x C-e}:
1716 The number 8 will appear in the echo area.
1718 What happens is that the Lisp interpreter first evaluates the inner
1719 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1720 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1721 returns the value 8. Since there are no more enclosing expressions to
1722 evaluate, the interpreter prints that value in the echo area.
1724 Now it is easy to understand the name of the command invoked by the
1725 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1726 letters @code{sexp} are an abbreviation for `symbolic expression', and
1727 @code{eval} is an abbreviation for `evaluate'. The command means
1728 `evaluate last symbolic expression'.
1730 As an experiment, you can try evaluating the expression by putting the
1731 cursor at the beginning of the next line immediately following the
1732 expression, or inside the expression.
1735 Here is another copy of the expression:
1742 If you place the cursor at the beginning of the blank line that
1743 immediately follows the expression and type @kbd{C-x C-e}, you will
1744 still get the value 8 printed in the echo area. Now try putting the
1745 cursor inside the expression. If you put it right after the next to
1746 last parenthesis (so it appears to sit on top of the last parenthesis),
1747 you will get a 6 printed in the echo area! This is because the command
1748 evaluates the expression @code{(+ 3 3)}.
1750 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1751 you will get the number itself. In Lisp, if you evaluate a number, you
1752 get the number itself---this is how numbers differ from symbols. If you
1753 evaluate a list starting with a symbol like @code{+}, you will get a
1754 value returned that is the result of the computer carrying out the
1755 instructions in the function definition attached to that name. If a
1756 symbol by itself is evaluated, something different happens, as we will
1757 see in the next section.
1759 @node Variables, Arguments, Evaluation, List Processing
1760 @comment node-name, next, previous, up
1764 In Emacs Lisp, a symbol can have a value attached to it just as it can
1765 have a function definition attached to it. The two are different.
1766 The function definition is a set of instructions that a computer will
1767 obey. A value, on the other hand, is something, such as number or a
1768 name, that can vary (which is why such a symbol is called a variable).
1769 The value of a symbol can be any expression in Lisp, such as a symbol,
1770 number, list, or string. A symbol that has a value is often called a
1773 A symbol can have both a function definition and a value attached to
1774 it at the same time. Or it can have just one or the other.
1775 The two are separate. This is somewhat similar
1776 to the way the name Cambridge can refer to the city in Massachusetts
1777 and have some information attached to the name as well, such as
1778 ``great programming center''.
1781 (Incidentally, in Emacs Lisp, a symbol can have two
1782 other things attached to it, too: a property list and a documentation
1783 string; these are discussed later.)
1786 Another way to think about this is to imagine a symbol as being a chest
1787 of drawers. The function definition is put in one drawer, the value in
1788 another, and so on. What is put in the drawer holding the value can be
1789 changed without affecting the contents of the drawer holding the
1790 function definition, and vice-verse.
1793 * fill-column Example::
1794 * Void Function:: The error message for a symbol
1796 * Void Variable:: The error message for a symbol without a value.
1799 @node fill-column Example, Void Function, Variables, Variables
1801 @unnumberedsubsec @code{fill-column}, an Example Variable
1804 @findex fill-column, @r{an example variable}
1805 @cindex Example variable, @code{fill-column}
1806 @cindex Variable, example of, @code{fill-column}
1807 The variable @code{fill-column} illustrates a symbol with a value
1808 attached to it: in every GNU Emacs buffer, this symbol is set to some
1809 value, usually 72 or 70, but sometimes to some other value. To find the
1810 value of this symbol, evaluate it by itself. If you are reading this in
1811 Info inside of GNU Emacs, you can do this by putting the cursor after
1812 the symbol and typing @kbd{C-x C-e}:
1819 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1820 area. This is the value for which @code{fill-column} is set for me as I
1821 write this. It may be different for you in your Info buffer. Notice
1822 that the value returned as a variable is printed in exactly the same way
1823 as the value returned by a function carrying out its instructions. From
1824 the point of view of the Lisp interpreter, a value returned is a value
1825 returned. What kind of expression it came from ceases to matter once
1828 A symbol can have any value attached to it or, to use the jargon, we can
1829 @dfn{bind} the variable to a value: to a number, such as 72; to a
1830 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1831 oak)}; we can even bind a variable to a function definition.
1833 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1834 Setting the Value of a Variable}, for information about one way to do
1837 @node Void Function, Void Variable, fill-column Example, Variables
1838 @comment node-name, next, previous, up
1839 @subsection Error Message for a Symbol Without a Function
1840 @cindex Symbol without function error
1841 @cindex Error for symbol without function
1843 When we evaluated @code{fill-column} to find its value as a variable,
1844 we did not place parentheses around the word. This is because we did
1845 not intend to use it as a function name.
1847 If @code{fill-column} were the first or only element of a list, the
1848 Lisp interpreter would attempt to find the function definition
1849 attached to it. But @code{fill-column} has no function definition.
1850 Try evaluating this:
1858 In GNU Emacs version 22, you will create a @file{*Backtrace*} buffer
1863 ---------- Buffer: *Backtrace* ----------
1864 Debugger entered--Lisp error: (void-function fill-column)
1867 eval-last-sexp-1(nil)
1869 call-interactively(eval-last-sexp)
1870 ---------- Buffer: *Backtrace* ----------
1875 (Remember, to quit the debugger and make the debugger window go away,
1876 type @kbd{q} in the @file{*Backtrace*} buffer.)
1880 In GNU Emacs 20 and before, you will produce an error message that says:
1883 Symbol's function definition is void:@: fill-column
1887 (The message will go away as soon as you move the cursor or type
1891 @node Void Variable, , Void Function, Variables
1892 @comment node-name, next, previous, up
1893 @subsection Error Message for a Symbol Without a Value
1894 @cindex Symbol without value error
1895 @cindex Error for symbol without value
1897 If you attempt to evaluate a symbol that does not have a value bound to
1898 it, you will receive an error message. You can see this by
1899 experimenting with our 2 plus 2 addition. In the following expression,
1900 put your cursor right after the @code{+}, before the first number 2,
1909 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1914 ---------- Buffer: *Backtrace* ----------
1915 Debugger entered--Lisp error: (void-variable +)
1917 eval-last-sexp-1(nil)
1919 call-interactively(eval-last-sexp)
1920 ---------- Buffer: *Backtrace* ----------
1925 (As with the other times we entered the debugger, you can quit by
1926 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1928 This backtrace is different from the very first error message we saw,
1929 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1930 In this case, the function does not have a value as a variable; while
1931 in the other error message, the function (the word `this') did not
1934 In this experiment with the @code{+}, what we did was cause the Lisp
1935 interpreter to evaluate the @code{+} and look for the value of the
1936 variable instead of the function definition. We did this by placing the
1937 cursor right after the symbol rather than after the parenthesis of the
1938 enclosing list as we did before. As a consequence, the Lisp interpreter
1939 evaluated the preceding s-expression, which in this case was the
1942 Since @code{+} does not have a value bound to it, just the function
1943 definition, the error message reported that the symbol's value as a
1948 In GNU Emacs version 20 and before, your error message will say:
1951 Symbol's value as variable is void:@: +
1955 The meaning is the same as in GNU Emacs 22.
1958 @node Arguments, set & setq, Variables, List Processing
1959 @comment node-name, next, previous, up
1962 @cindex Passing information to functions
1964 To see how information is passed to functions, let's look again at
1965 our old standby, the addition of two plus two. In Lisp, this is written
1972 If you evaluate this expression, the number 4 will appear in your echo
1973 area. What the Lisp interpreter does is add the numbers that follow
1976 @cindex @samp{argument} defined
1977 The numbers added by @code{+} are called the @dfn{arguments} of the
1978 function @code{+}. These numbers are the information that is given to
1979 or @dfn{passed} to the function.
1981 The word `argument' comes from the way it is used in mathematics and
1982 does not refer to a disputation between two people; instead it refers to
1983 the information presented to the function, in this case, to the
1984 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1985 that follow the function. The values returned by the evaluation of
1986 these atoms or lists are passed to the function. Different functions
1987 require different numbers of arguments; some functions require none at
1988 all.@footnote{It is curious to track the path by which the word `argument'
1989 came to have two different meanings, one in mathematics and the other in
1990 everyday English. According to the @cite{Oxford English Dictionary},
1991 the word derives from the Latin for @samp{to make clear, prove}; thus it
1992 came to mean, by one thread of derivation, `the evidence offered as
1993 proof', which is to say, `the information offered', which led to its
1994 meaning in Lisp. But in the other thread of derivation, it came to mean
1995 `to assert in a manner against which others may make counter
1996 assertions', which led to the meaning of the word as a disputation.
1997 (Note here that the English word has two different definitions attached
1998 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
1999 have two different function definitions at the same time.)}
2002 * Data types:: Types of data passed to a function.
2003 * Args as Variable or List:: An argument can be the value
2004 of a variable or list.
2005 * Variable Number of Arguments:: Some functions may take a
2006 variable number of arguments.
2007 * Wrong Type of Argument:: Passing an argument of the wrong type
2009 * message:: A useful function for sending messages.
2012 @node Data types, Args as Variable or List, Arguments, Arguments
2013 @comment node-name, next, previous, up
2014 @subsection Arguments' Data Types
2016 @cindex Types of data
2017 @cindex Arguments' data types
2019 The type of data that should be passed to a function depends on what
2020 kind of information it uses. The arguments to a function such as
2021 @code{+} must have values that are numbers, since @code{+} adds numbers.
2022 Other functions use different kinds of data for their arguments.
2026 For example, the @code{concat} function links together or unites two or
2027 more strings of text to produce a string. The arguments are strings.
2028 Concatenating the two character strings @code{abc}, @code{def} produces
2029 the single string @code{abcdef}. This can be seen by evaluating the
2033 (concat "abc" "def")
2037 The value produced by evaluating this expression is @code{"abcdef"}.
2039 A function such as @code{substring} uses both a string and numbers as
2040 arguments. The function returns a part of the string, a substring of
2041 the first argument. This function takes three arguments. Its first
2042 argument is the string of characters, the second and third arguments are
2043 numbers that indicate the beginning and end of the substring. The
2044 numbers are a count of the number of characters (including spaces and
2045 punctuations) from the beginning of the string.
2048 For example, if you evaluate the following:
2051 (substring "The quick brown fox jumped." 16 19)
2055 you will see @code{"fox"} appear in the echo area. The arguments are the
2056 string and the two numbers.
2058 Note that the string passed to @code{substring} is a single atom even
2059 though it is made up of several words separated by spaces. Lisp counts
2060 everything between the two quotation marks as part of the string,
2061 including the spaces. You can think of the @code{substring} function as
2062 a kind of `atom smasher' since it takes an otherwise indivisible atom
2063 and extracts a part. However, @code{substring} is only able to extract
2064 a substring from an argument that is a string, not from another type of
2065 atom such as a number or symbol.
2067 @node Args as Variable or List, Variable Number of Arguments, Data types, Arguments
2068 @comment node-name, next, previous, up
2069 @subsection An Argument as the Value of a Variable or List
2071 An argument can be a symbol that returns a value when it is evaluated.
2072 For example, when the symbol @code{fill-column} by itself is evaluated,
2073 it returns a number. This number can be used in an addition.
2076 Position the cursor after the following expression and type @kbd{C-x
2084 The value will be a number two more than what you get by evaluating
2085 @code{fill-column} alone. For me, this is 74, because my value of
2086 @code{fill-column} is 72.
2088 As we have just seen, an argument can be a symbol that returns a value
2089 when evaluated. In addition, an argument can be a list that returns a
2090 value when it is evaluated. For example, in the following expression,
2091 the arguments to the function @code{concat} are the strings
2092 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2093 @code{(number-to-string (+ 2 fill-column))}.
2095 @c For GNU Emacs 22, need number-to-string
2097 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2101 If you evaluate this expression---and if, as with my Emacs,
2102 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2103 appear in the echo area. (Note that you must put spaces after the
2104 word @samp{The} and before the word @samp{red} so they will appear in
2105 the final string. The function @code{number-to-string} converts the
2106 integer that the addition function returns to a string.
2107 @code{number-to-string} is also known as @code{int-to-string}.)
2109 @node Variable Number of Arguments, Wrong Type of Argument, Args as Variable or List, Arguments
2110 @comment node-name, next, previous, up
2111 @subsection Variable Number of Arguments
2112 @cindex Variable number of arguments
2113 @cindex Arguments, variable number of
2115 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2116 number of arguments. (The @code{*} is the symbol for multiplication.)
2117 This can be seen by evaluating each of the following expressions in
2118 the usual way. What you will see in the echo area is printed in this
2119 text after @samp{@result{}}, which you may read as `evaluates to'.
2122 In the first set, the functions have no arguments:
2133 In this set, the functions have one argument each:
2144 In this set, the functions have three arguments each:
2148 (+ 3 4 5) @result{} 12
2150 (* 3 4 5) @result{} 60
2154 @node Wrong Type of Argument, message, Variable Number of Arguments, Arguments
2155 @comment node-name, next, previous, up
2156 @subsection Using the Wrong Type Object as an Argument
2157 @cindex Wrong type of argument
2158 @cindex Argument, wrong type of
2160 When a function is passed an argument of the wrong type, the Lisp
2161 interpreter produces an error message. For example, the @code{+}
2162 function expects the values of its arguments to be numbers. As an
2163 experiment we can pass it the quoted symbol @code{hello} instead of a
2164 number. Position the cursor after the following expression and type
2172 When you do this you will generate an error message. What has happened
2173 is that @code{+} has tried to add the 2 to the value returned by
2174 @code{'hello}, but the value returned by @code{'hello} is the symbol
2175 @code{hello}, not a number. Only numbers can be added. So @code{+}
2176 could not carry out its addition.
2179 In GNU Emacs version 22, you will create and enter a
2180 @file{*Backtrace*} buffer that says:
2185 ---------- Buffer: *Backtrace* ----------
2186 Debugger entered--Lisp error:
2187 (wrong-type-argument number-or-marker-p hello)
2189 eval((+ 2 (quote hello)))
2190 eval-last-sexp-1(nil)
2192 call-interactively(eval-last-sexp)
2193 ---------- Buffer: *Backtrace* ----------
2198 As usual, the error message tries to be helpful and makes sense after you
2199 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2200 the abbreviation @code{'hello}.}
2202 The first part of the error message is straightforward; it says
2203 @samp{wrong type argument}. Next comes the mysterious jargon word
2204 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2205 kind of argument the @code{+} expected.
2207 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2208 trying to determine whether the information presented it (the value of
2209 the argument) is a number or a marker (a special object representing a
2210 buffer position). What it does is test to see whether the @code{+} is
2211 being given numbers to add. It also tests to see whether the
2212 argument is something called a marker, which is a specific feature of
2213 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2214 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2215 its position is kept as a marker. The mark can be considered a
2216 number---the number of characters the location is from the beginning
2217 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2218 numeric value of marker positions as numbers.
2220 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2221 practice started in the early days of Lisp programming. The @samp{p}
2222 stands for `predicate'. In the jargon used by the early Lisp
2223 researchers, a predicate refers to a function to determine whether some
2224 property is true or false. So the @samp{p} tells us that
2225 @code{number-or-marker-p} is the name of a function that determines
2226 whether it is true or false that the argument supplied is a number or
2227 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2228 a function that tests whether its argument has the value of zero, and
2229 @code{listp}, a function that tests whether its argument is a list.
2231 Finally, the last part of the error message is the symbol @code{hello}.
2232 This is the value of the argument that was passed to @code{+}. If the
2233 addition had been passed the correct type of object, the value passed
2234 would have been a number, such as 37, rather than a symbol like
2235 @code{hello}. But then you would not have got the error message.
2239 In GNU Emacs version 20 and before, the echo area displays an error
2243 Wrong type argument:@: number-or-marker-p, hello
2246 This says, in different words, the same as the top line of the
2247 @file{*Backtrace*} buffer.
2250 @node message, , Wrong Type of Argument, Arguments
2251 @comment node-name, next, previous, up
2252 @subsection The @code{message} Function
2255 Like @code{+}, the @code{message} function takes a variable number of
2256 arguments. It is used to send messages to the user and is so useful
2257 that we will describe it here.
2260 A message is printed in the echo area. For example, you can print a
2261 message in your echo area by evaluating the following list:
2264 (message "This message appears in the echo area!")
2267 The whole string between double quotation marks is a single argument
2268 and is printed @i{in toto}. (Note that in this example, the message
2269 itself will appear in the echo area within double quotes; that is
2270 because you see the value returned by the @code{message} function. In
2271 most uses of @code{message} in programs that you write, the text will
2272 be printed in the echo area as a side-effect, without the quotes.
2273 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2274 detail}, for an example of this.)
2276 However, if there is a @samp{%s} in the quoted string of characters, the
2277 @code{message} function does not print the @samp{%s} as such, but looks
2278 to the argument that follows the string. It evaluates the second
2279 argument and prints the value at the location in the string where the
2283 You can see this by positioning the cursor after the following
2284 expression and typing @kbd{C-x C-e}:
2287 (message "The name of this buffer is: %s." (buffer-name))
2291 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2292 echo area. The function @code{buffer-name} returns the name of the
2293 buffer as a string, which the @code{message} function inserts in place
2296 To print a value as an integer, use @samp{%d} in the same way as
2297 @samp{%s}. For example, to print a message in the echo area that
2298 states the value of the @code{fill-column}, evaluate the following:
2301 (message "The value of fill-column is %d." fill-column)
2305 On my system, when I evaluate this list, @code{"The value of
2306 fill-column is 72."} appears in my echo area@footnote{Actually, you
2307 can use @code{%s} to print a number. It is non-specific. @code{%d}
2308 prints only the part of a number left of a decimal point, and not
2309 anything that is not a number.}.
2311 If there is more than one @samp{%s} in the quoted string, the value of
2312 the first argument following the quoted string is printed at the
2313 location of the first @samp{%s} and the value of the second argument is
2314 printed at the location of the second @samp{%s}, and so on.
2317 For example, if you evaluate the following,
2321 (message "There are %d %s in the office!"
2322 (- fill-column 14) "pink elephants")
2327 a rather whimsical message will appear in your echo area. On my system
2328 it says, @code{"There are 58 pink elephants in the office!"}.
2330 The expression @code{(- fill-column 14)} is evaluated and the resulting
2331 number is inserted in place of the @samp{%d}; and the string in double
2332 quotes, @code{"pink elephants"}, is treated as a single argument and
2333 inserted in place of the @samp{%s}. (That is to say, a string between
2334 double quotes evaluates to itself, like a number.)
2336 Finally, here is a somewhat complex example that not only illustrates
2337 the computation of a number, but also shows how you can use an
2338 expression within an expression to generate the text that is substituted
2343 (message "He saw %d %s"
2347 "The quick brown foxes jumped." 16 21)
2352 In this example, @code{message} has three arguments: the string,
2353 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2354 the expression beginning with the function @code{concat}. The value
2355 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2356 in place of the @samp{%d}; and the value returned by the expression
2357 beginning with @code{concat} is inserted in place of the @samp{%s}.
2359 When your fill column is 70 and you evaluate the expression, the
2360 message @code{"He saw 38 red foxes leaping."} appears in your echo
2363 @node set & setq, Summary, Arguments, List Processing
2364 @comment node-name, next, previous, up
2365 @section Setting the Value of a Variable
2366 @cindex Variable, setting value
2367 @cindex Setting value of variable
2369 @cindex @samp{bind} defined
2370 There are several ways by which a variable can be given a value. One of
2371 the ways is to use either the function @code{set} or the function
2372 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2373 jargon for this process is to @dfn{bind} a variable to a value.)
2375 The following sections not only describe how @code{set} and @code{setq}
2376 work but also illustrate how arguments are passed.
2379 * Using set:: Setting values.
2380 * Using setq:: Setting a quoted value.
2381 * Counting:: Using @code{setq} to count.
2384 @node Using set, Using setq, set & setq, set & setq
2385 @comment node-name, next, previous, up
2386 @subsection Using @code{set}
2389 To set the value of the symbol @code{flowers} to the list @code{'(rose
2390 violet daisy buttercup)}, evaluate the following expression by
2391 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2394 (set 'flowers '(rose violet daisy buttercup))
2398 The list @code{(rose violet daisy buttercup)} will appear in the echo
2399 area. This is what is @emph{returned} by the @code{set} function. As a
2400 side effect, the symbol @code{flowers} is bound to the list; that is,
2401 the symbol @code{flowers}, which can be viewed as a variable, is given
2402 the list as its value. (This process, by the way, illustrates how a
2403 side effect to the Lisp interpreter, setting the value, can be the
2404 primary effect that we humans are interested in. This is because every
2405 Lisp function must return a value if it does not get an error, but it
2406 will only have a side effect if it is designed to have one.)
2408 After evaluating the @code{set} expression, you can evaluate the symbol
2409 @code{flowers} and it will return the value you just set. Here is the
2410 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2417 When you evaluate @code{flowers}, the list
2418 @code{(rose violet daisy buttercup)} appears in the echo area.
2420 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2421 in front of it, what you will see in the echo area is the symbol itself,
2422 @code{flowers}. Here is the quoted symbol, so you can try this:
2428 Note also, that when you use @code{set}, you need to quote both
2429 arguments to @code{set}, unless you want them evaluated. Since we do
2430 not want either argument evaluated, neither the variable
2431 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2432 are quoted. (When you use @code{set} without quoting its first
2433 argument, the first argument is evaluated before anything else is
2434 done. If you did this and @code{flowers} did not have a value
2435 already, you would get an error message that the @samp{Symbol's value
2436 as variable is void}; on the other hand, if @code{flowers} did return
2437 a value after it was evaluated, the @code{set} would attempt to set
2438 the value that was returned. There are situations where this is the
2439 right thing for the function to do; but such situations are rare.)
2441 @node Using setq, Counting, Using set, set & setq
2442 @comment node-name, next, previous, up
2443 @subsection Using @code{setq}
2446 As a practical matter, you almost always quote the first argument to
2447 @code{set}. The combination of @code{set} and a quoted first argument
2448 is so common that it has its own name: the special form @code{setq}.
2449 This special form is just like @code{set} except that the first argument
2450 is quoted automatically, so you don't need to type the quote mark
2451 yourself. Also, as an added convenience, @code{setq} permits you to set
2452 several different variables to different values, all in one expression.
2454 To set the value of the variable @code{carnivores} to the list
2455 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2459 (setq carnivores '(lion tiger leopard))
2463 This is exactly the same as using @code{set} except the first argument
2464 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2465 means @code{quote}.)
2468 With @code{set}, the expression would look like this:
2471 (set 'carnivores '(lion tiger leopard))
2474 Also, @code{setq} can be used to assign different values to
2475 different variables. The first argument is bound to the value
2476 of the second argument, the third argument is bound to the value of the
2477 fourth argument, and so on. For example, you could use the following to
2478 assign a list of trees to the symbol @code{trees} and a list of herbivores
2479 to the symbol @code{herbivores}:
2483 (setq trees '(pine fir oak maple)
2484 herbivores '(gazelle antelope zebra))
2489 (The expression could just as well have been on one line, but it might
2490 not have fit on a page; and humans find it easier to read nicely
2493 Although I have been using the term `assign', there is another way of
2494 thinking about the workings of @code{set} and @code{setq}; and that is to
2495 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2496 list. This latter way of thinking is very common and in forthcoming
2497 chapters we shall come upon at least one symbol that has `pointer' as
2498 part of its name. The name is chosen because the symbol has a value,
2499 specifically a list, attached to it; or, expressed another way,
2500 the symbol is set to ``point'' to the list.
2502 @node Counting, , Using setq, set & setq
2503 @comment node-name, next, previous, up
2504 @subsection Counting
2507 Here is an example that shows how to use @code{setq} in a counter. You
2508 might use this to count how many times a part of your program repeats
2509 itself. First set a variable to zero; then add one to the number each
2510 time the program repeats itself. To do this, you need a variable that
2511 serves as a counter, and two expressions: an initial @code{setq}
2512 expression that sets the counter variable to zero; and a second
2513 @code{setq} expression that increments the counter each time it is
2518 (setq counter 0) ; @r{Let's call this the initializer.}
2520 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2522 counter ; @r{This is the counter.}
2527 (The text following the @samp{;} are comments. @xref{Change a
2528 defun, , Change a Function Definition}.)
2530 If you evaluate the first of these expressions, the initializer,
2531 @code{(setq counter 0)}, and then evaluate the third expression,
2532 @code{counter}, the number @code{0} will appear in the echo area. If
2533 you then evaluate the second expression, the incrementer, @code{(setq
2534 counter (+ counter 1))}, the counter will get the value 1. So if you
2535 again evaluate @code{counter}, the number @code{1} will appear in the
2536 echo area. Each time you evaluate the second expression, the value of
2537 the counter will be incremented.
2539 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2540 the Lisp interpreter first evaluates the innermost list; this is the
2541 addition. In order to evaluate this list, it must evaluate the variable
2542 @code{counter} and the number @code{1}. When it evaluates the variable
2543 @code{counter}, it receives its current value. It passes this value and
2544 the number @code{1} to the @code{+} which adds them together. The sum
2545 is then returned as the value of the inner list and passed to the
2546 @code{setq} which sets the variable @code{counter} to this new value.
2547 Thus, the value of the variable, @code{counter}, is changed.
2549 @node Summary, Error Message Exercises, set & setq, List Processing
2550 @comment node-name, next, previous, up
2553 Learning Lisp is like climbing a hill in which the first part is the
2554 steepest. You have now climbed the most difficult part; what remains
2555 becomes easier as you progress onwards.
2563 Lisp programs are made up of expressions, which are lists or single atoms.
2566 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2567 surrounded by parentheses. A list can be empty.
2570 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2571 character symbols like @code{+}, strings of characters between double
2572 quotation marks, or numbers.
2575 A number evaluates to itself.
2578 A string between double quotes also evaluates to itself.
2581 When you evaluate a symbol by itself, its value is returned.
2584 When you evaluate a list, the Lisp interpreter looks at the first symbol
2585 in the list and then at the function definition bound to that symbol.
2586 Then the instructions in the function definition are carried out.
2589 A single quotation mark,
2596 , tells the Lisp interpreter that it should
2597 return the following expression as written, and not evaluate it as it
2598 would if the quote were not there.
2601 Arguments are the information passed to a function. The arguments to a
2602 function are computed by evaluating the rest of the elements of the list
2603 of which the function is the first element.
2606 A function always returns a value when it is evaluated (unless it gets
2607 an error); in addition, it may also carry out some action called a
2608 ``side effect''. In many cases, a function's primary purpose is to
2609 create a side effect.
2612 @node Error Message Exercises, , Summary, List Processing
2613 @comment node-name, next, previous, up
2616 A few simple exercises:
2620 Generate an error message by evaluating an appropriate symbol that is
2621 not within parentheses.
2624 Generate an error message by evaluating an appropriate symbol that is
2625 between parentheses.
2628 Create a counter that increments by two rather than one.
2631 Write an expression that prints a message in the echo area when
2635 @node Practicing Evaluation, Writing Defuns, List Processing, Top
2636 @comment node-name, next, previous, up
2637 @chapter Practicing Evaluation
2638 @cindex Practicing evaluation
2639 @cindex Evaluation practice
2641 Before learning how to write a function definition in Emacs Lisp, it is
2642 useful to spend a little time evaluating various expressions that have
2643 already been written. These expressions will be lists with the
2644 functions as their first (and often only) element. Since some of the
2645 functions associated with buffers are both simple and interesting, we
2646 will start with those. In this section, we will evaluate a few of
2647 these. In another section, we will study the code of several other
2648 buffer-related functions, to see how they were written.
2651 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2653 * Buffer Names:: Buffers and files are different.
2654 * Getting Buffers:: Getting a buffer itself, not merely its name.
2655 * Switching Buffers:: How to change to another buffer.
2656 * Buffer Size & Locations:: Where point is located and the size of
2658 * Evaluation Exercise::
2661 @node How to Evaluate, Buffer Names, Practicing Evaluation, Practicing Evaluation
2663 @unnumberedsec How to Evaluate
2666 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2667 command to move the cursor or to scroll the screen, @i{you are evaluating
2668 an expression,} the first element of which is a function. @i{This is
2671 @cindex @samp{interactive function} defined
2672 @cindex @samp{command} defined
2673 When you type keys, you cause the Lisp interpreter to evaluate an
2674 expression and that is how you get your results. Even typing plain text
2675 involves evaluating an Emacs Lisp function, in this case, one that uses
2676 @code{self-insert-command}, which simply inserts the character you
2677 typed. The functions you evaluate by typing keystrokes are called
2678 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2679 interactive will be illustrated in the chapter on how to write function
2680 definitions. @xref{Interactive, , Making a Function Interactive}.
2682 In addition to typing keyboard commands, we have seen a second way to
2683 evaluate an expression: by positioning the cursor after a list and
2684 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2685 section. There are other ways to evaluate an expression as well; these
2686 will be described as we come to them.
2688 Besides being used for practicing evaluation, the functions shown in the
2689 next few sections are important in their own right. A study of these
2690 functions makes clear the distinction between buffers and files, how to
2691 switch to a buffer, and how to determine a location within it.
2693 @node Buffer Names, Getting Buffers, How to Evaluate, Practicing Evaluation
2694 @comment node-name, next, previous, up
2695 @section Buffer Names
2697 @findex buffer-file-name
2699 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2700 the difference between a file and a buffer. When you evaluate the
2701 following expression, @code{(buffer-name)}, the name of the buffer
2702 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2703 the name of the file to which the buffer refers appears in the echo
2704 area. Usually, the name returned by @code{(buffer-name)} is the same as
2705 the name of the file to which it refers, and the name returned by
2706 @code{(buffer-file-name)} is the full path-name of the file.
2708 A file and a buffer are two different entities. A file is information
2709 recorded permanently in the computer (unless you delete it). A buffer,
2710 on the other hand, is information inside of Emacs that will vanish at
2711 the end of the editing session (or when you kill the buffer). Usually,
2712 a buffer contains information that you have copied from a file; we say
2713 the buffer is @dfn{visiting} that file. This copy is what you work on
2714 and modify. Changes to the buffer do not change the file, until you
2715 save the buffer. When you save the buffer, the buffer is copied to the file
2716 and is thus saved permanently.
2719 If you are reading this in Info inside of GNU Emacs, you can evaluate
2720 each of the following expressions by positioning the cursor after it and
2721 typing @kbd{C-x C-e}.
2732 When I do this in Info, the value returned by evaluating
2733 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2734 evaluating @code{(buffer-file-name)} is @file{nil}.
2736 On the other hand, while I am writing this document, the value
2737 returned by evaluating @code{(buffer-name)} is
2738 @file{"introduction.texinfo"}, and the value returned by evaluating
2739 @code{(buffer-file-name)} is
2740 @file{"/gnu/work/intro/introduction.texinfo"}.
2742 @cindex @code{nil}, history of word
2743 The former is the name of the buffer and the latter is the name of the
2744 file. In Info, the buffer name is @file{"*info*"}. Info does not
2745 point to any file, so the result of evaluating
2746 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2747 from the Latin word for `nothing'; in this case, it means that the
2748 buffer is not associated with any file. (In Lisp, @code{nil} is also
2749 used to mean `false' and is a synonym for the empty list, @code{()}.)
2751 When I am writing, the name of my buffer is
2752 @file{"introduction.texinfo"}. The name of the file to which it
2753 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2755 (In the expressions, the parentheses tell the Lisp interpreter to
2756 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2757 functions; without the parentheses, the interpreter would attempt to
2758 evaluate the symbols as variables. @xref{Variables}.)
2760 In spite of the distinction between files and buffers, you will often
2761 find that people refer to a file when they mean a buffer and vice-verse.
2762 Indeed, most people say, ``I am editing a file,'' rather than saying,
2763 ``I am editing a buffer which I will soon save to a file.'' It is
2764 almost always clear from context what people mean. When dealing with
2765 computer programs, however, it is important to keep the distinction in mind,
2766 since the computer is not as smart as a person.
2768 @cindex Buffer, history of word
2769 The word `buffer', by the way, comes from the meaning of the word as a
2770 cushion that deadens the force of a collision. In early computers, a
2771 buffer cushioned the interaction between files and the computer's
2772 central processing unit. The drums or tapes that held a file and the
2773 central processing unit were pieces of equipment that were very
2774 different from each other, working at their own speeds, in spurts. The
2775 buffer made it possible for them to work together effectively.
2776 Eventually, the buffer grew from being an intermediary, a temporary
2777 holding place, to being the place where work is done. This
2778 transformation is rather like that of a small seaport that grew into a
2779 great city: once it was merely the place where cargo was warehoused
2780 temporarily before being loaded onto ships; then it became a business
2781 and cultural center in its own right.
2783 Not all buffers are associated with files. For example, a
2784 @file{*scratch*} buffer does not visit any file. Similarly, a
2785 @file{*Help*} buffer is not associated with any file.
2787 In the old days, when you lacked a @file{~/.emacs} file and started an
2788 Emacs session by typing the command @code{emacs} alone, without naming
2789 any files, Emacs started with the @file{*scratch*} buffer visible.
2790 Nowadays, you will see a splash screen. You can follow one of the
2791 commands suggested on the splash screen, visit a file, or press the
2792 spacebar to reach the @file{*scratch*} buffer.
2794 If you switch to the @file{*scratch*} buffer, type
2795 @code{(buffer-name)}, position the cursor after it, and then type
2796 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2797 will be returned and will appear in the echo area. @code{"*scratch*"}
2798 is the name of the buffer. When you type @code{(buffer-file-name)} in
2799 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2800 in the echo area, just as it does when you evaluate
2801 @code{(buffer-file-name)} in Info.
2803 Incidentally, if you are in the @file{*scratch*} buffer and want the
2804 value returned by an expression to appear in the @file{*scratch*}
2805 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2806 instead of @kbd{C-x C-e}. This causes the value returned to appear
2807 after the expression. The buffer will look like this:
2810 (buffer-name)"*scratch*"
2814 You cannot do this in Info since Info is read-only and it will not allow
2815 you to change the contents of the buffer. But you can do this in any
2816 buffer you can edit; and when you write code or documentation (such as
2817 this book), this feature is very useful.
2819 @node Getting Buffers, Switching Buffers, Buffer Names, Practicing Evaluation
2820 @comment node-name, next, previous, up
2821 @section Getting Buffers
2822 @findex current-buffer
2823 @findex other-buffer
2824 @cindex Getting a buffer
2826 The @code{buffer-name} function returns the @emph{name} of the buffer;
2827 to get the buffer @emph{itself}, a different function is needed: the
2828 @code{current-buffer} function. If you use this function in code, what
2829 you get is the buffer itself.
2831 A name and the object or entity to which the name refers are different
2832 from each other. You are not your name. You are a person to whom
2833 others refer by name. If you ask to speak to George and someone hands you
2834 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2835 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2836 not be satisfied. You do not want to speak to the name, but to the
2837 person to whom the name refers. A buffer is similar: the name of the
2838 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2839 get a buffer itself, you need to use a function such as
2840 @code{current-buffer}.
2842 However, there is a slight complication: if you evaluate
2843 @code{current-buffer} in an expression on its own, as we will do here,
2844 what you see is a printed representation of the name of the buffer
2845 without the contents of the buffer. Emacs works this way for two
2846 reasons: the buffer may be thousands of lines long---too long to be
2847 conveniently displayed; and, another buffer may have the same contents
2848 but a different name, and it is important to distinguish between them.
2851 Here is an expression containing the function:
2858 If you evaluate this expression in Info in Emacs in the usual way,
2859 @file{#<buffer *info*>} will appear in the echo area. The special
2860 format indicates that the buffer itself is being returned, rather than
2863 Incidentally, while you can type a number or symbol into a program, you
2864 cannot do that with the printed representation of a buffer: the only way
2865 to get a buffer itself is with a function such as @code{current-buffer}.
2867 A related function is @code{other-buffer}. This returns the most
2868 recently selected buffer other than the one you are in currently, not
2869 a printed representation of its name. If you have recently switched
2870 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2871 will return that buffer.
2874 You can see this by evaluating the expression:
2881 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2882 the name of whatever other buffer you switched back from most
2883 recently@footnote{Actually, by default, if the buffer from which you
2884 just switched is visible to you in another window, @code{other-buffer}
2885 will choose the most recent buffer that you cannot see; this is a
2886 subtlety that I often forget.}.
2888 @node Switching Buffers, Buffer Size & Locations, Getting Buffers, Practicing Evaluation
2889 @comment node-name, next, previous, up
2890 @section Switching Buffers
2891 @findex switch-to-buffer
2893 @cindex Switching to a buffer
2895 The @code{other-buffer} function actually provides a buffer when it is
2896 used as an argument to a function that requires one. We can see this
2897 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2900 But first, a brief introduction to the @code{switch-to-buffer}
2901 function. When you switched back and forth from Info to the
2902 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2903 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2904 rather, to save typing, you probably only typed @kbd{RET} if the
2905 default buffer was @file{*scratch*}, or if it was different, then you
2906 typed just part of the name, such as @code{*sc}, pressed your
2907 @kbd{TAB} key to cause it to expand to the full name, and then typed
2908 your @kbd{RET} key.} when prompted in the minibuffer for the name of
2909 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2910 b}, cause the Lisp interpreter to evaluate the interactive function
2911 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2912 different keystrokes call or run different functions. For example,
2913 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2914 @code{forward-sentence}, and so on.
2916 By writing @code{switch-to-buffer} in an expression, and giving it a
2917 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2921 Here is the Lisp expression:
2924 (switch-to-buffer (other-buffer))
2928 The symbol @code{switch-to-buffer} is the first element of the list,
2929 so the Lisp interpreter will treat it as a function and carry out the
2930 instructions that are attached to it. But before doing that, the
2931 interpreter will note that @code{other-buffer} is inside parentheses
2932 and work on that symbol first. @code{other-buffer} is the first (and
2933 in this case, the only) element of this list, so the Lisp interpreter
2934 calls or runs the function. It returns another buffer. Next, the
2935 interpreter runs @code{switch-to-buffer}, passing to it, as an
2936 argument, the other buffer, which is what Emacs will switch to. If
2937 you are reading this in Info, try this now. Evaluate the expression.
2938 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2939 expression will move you to your most recent other buffer that you
2940 cannot see. If you really want to go to your most recently selected
2941 buffer, even if you can still see it, you need to evaluate the
2942 following more complex expression:
2945 (switch-to-buffer (other-buffer (current-buffer) t))
2949 In this case, the first argument to @code{other-buffer} tells it which
2950 buffer to skip---the current one---and the second argument tells
2951 @code{other-buffer} it is OK to switch to a visible buffer.
2952 In regular use, @code{switch-to-buffer} takes you to an invisible
2953 window since you would most likely use @kbd{C-x o} (@code{other-window})
2954 to go to another visible buffer.}
2956 In the programming examples in later sections of this document, you will
2957 see the function @code{set-buffer} more often than
2958 @code{switch-to-buffer}. This is because of a difference between
2959 computer programs and humans: humans have eyes and expect to see the
2960 buffer on which they are working on their computer terminals. This is
2961 so obvious, it almost goes without saying. However, programs do not
2962 have eyes. When a computer program works on a buffer, that buffer does
2963 not need to be visible on the screen.
2965 @code{switch-to-buffer} is designed for humans and does two different
2966 things: it switches the buffer to which Emacs' attention is directed; and
2967 it switches the buffer displayed in the window to the new buffer.
2968 @code{set-buffer}, on the other hand, does only one thing: it switches
2969 the attention of the computer program to a different buffer. The buffer
2970 on the screen remains unchanged (of course, normally nothing happens
2971 there until the command finishes running).
2973 @cindex @samp{call} defined
2974 Also, we have just introduced another jargon term, the word @dfn{call}.
2975 When you evaluate a list in which the first symbol is a function, you
2976 are calling that function. The use of the term comes from the notion of
2977 the function as an entity that can do something for you if you `call'
2978 it---just as a plumber is an entity who can fix a leak if you call him
2981 @node Buffer Size & Locations, Evaluation Exercise, Switching Buffers, Practicing Evaluation
2982 @comment node-name, next, previous, up
2983 @section Buffer Size and the Location of Point
2984 @cindex Size of buffer
2986 @cindex Point location
2987 @cindex Location of point
2989 Finally, let's look at several rather simple functions,
2990 @code{buffer-size}, @code{point}, @code{point-min}, and
2991 @code{point-max}. These give information about the size of a buffer and
2992 the location of point within it.
2994 The function @code{buffer-size} tells you the size of the current
2995 buffer; that is, the function returns a count of the number of
2996 characters in the buffer.
3003 You can evaluate this in the usual way, by positioning the
3004 cursor after the expression and typing @kbd{C-x C-e}.
3006 @cindex @samp{point} defined
3007 In Emacs, the current position of the cursor is called @dfn{point}.
3008 The expression @code{(point)} returns a number that tells you where the
3009 cursor is located as a count of the number of characters from the
3010 beginning of the buffer up to point.
3013 You can see the character count for point in this buffer by evaluating
3014 the following expression in the usual way:
3021 As I write this, the value of @code{point} is 65724. The @code{point}
3022 function is frequently used in some of the examples later in this
3026 The value of point depends, of course, on its location within the
3027 buffer. If you evaluate point in this spot, the number will be larger:
3034 For me, the value of point in this location is 66043, which means that
3035 there are 319 characters (including spaces) between the two
3036 expressions. (Doubtless, you will see different numbers, since I will
3037 have edited this since I first evaluated point.)
3039 @cindex @samp{narrowing} defined
3040 The function @code{point-min} is somewhat similar to @code{point}, but
3041 it returns the value of the minimum permissible value of point in the
3042 current buffer. This is the number 1 unless @dfn{narrowing} is in
3043 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3044 or a program, to operations on just a part of a buffer.
3045 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3046 function @code{point-max} returns the value of the maximum permissible
3047 value of point in the current buffer.
3049 @node Evaluation Exercise, , Buffer Size & Locations, Practicing Evaluation
3052 Find a file with which you are working and move towards its middle.
3053 Find its buffer name, file name, length, and your position in the file.
3055 @node Writing Defuns, Buffer Walk Through, Practicing Evaluation, Top
3056 @comment node-name, next, previous, up
3057 @chapter How To Write Function Definitions
3058 @cindex Definition writing
3059 @cindex Function definition writing
3060 @cindex Writing a function definition
3062 When the Lisp interpreter evaluates a list, it looks to see whether the
3063 first symbol on the list has a function definition attached to it; or,
3064 put another way, whether the symbol points to a function definition. If
3065 it does, the computer carries out the instructions in the definition. A
3066 symbol that has a function definition is called, simply, a function
3067 (although, properly speaking, the definition is the function and the
3068 symbol refers to it.)
3071 * Primitive Functions::
3072 * defun:: The @code{defun} special form.
3073 * Install:: Install a function definition.
3074 * Interactive:: Making a function interactive.
3075 * Interactive Options:: Different options for @code{interactive}.
3076 * Permanent Installation:: Installing code permanently.
3077 * let:: Creating and initializing local variables.
3079 * else:: If--then--else expressions.
3080 * Truth & Falsehood:: What Lisp considers false and true.
3081 * save-excursion:: Keeping track of point, mark, and buffer.
3086 @node Primitive Functions, defun, Writing Defuns, Writing Defuns
3088 @unnumberedsec An Aside about Primitive Functions
3090 @cindex Primitive functions
3091 @cindex Functions, primitive
3093 @cindex C language primitives
3094 @cindex Primitives written in C
3095 All functions are defined in terms of other functions, except for a few
3096 @dfn{primitive} functions that are written in the C programming
3097 language. When you write functions' definitions, you will write them in
3098 Emacs Lisp and use other functions as your building blocks. Some of the
3099 functions you will use will themselves be written in Emacs Lisp (perhaps
3100 by you) and some will be primitives written in C. The primitive
3101 functions are used exactly like those written in Emacs Lisp and behave
3102 like them. They are written in C so we can easily run GNU Emacs on any
3103 computer that has sufficient power and can run C.
3105 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3106 distinguish between the use of functions written in C and the use of
3107 functions written in Emacs Lisp. The difference is irrelevant. I
3108 mention the distinction only because it is interesting to know. Indeed,
3109 unless you investigate, you won't know whether an already-written
3110 function is written in Emacs Lisp or C.
3112 @node defun, Install, Primitive Functions, Writing Defuns
3113 @comment node-name, next, previous, up
3114 @section The @code{defun} Special Form
3116 @cindex Special form of @code{defun}
3118 @cindex @samp{function definition} defined
3119 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3120 it that tells the computer what to do when the function is called.
3121 This code is called the @dfn{function definition} and is created by
3122 evaluating a Lisp expression that starts with the symbol @code{defun}
3123 (which is an abbreviation for @emph{define function}). Because
3124 @code{defun} does not evaluate its arguments in the usual way, it is
3125 called a @dfn{special form}.
3127 In subsequent sections, we will look at function definitions from the
3128 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3129 we will describe a simple function definition so you can see how it
3130 looks. This function definition uses arithmetic because it makes for a
3131 simple example. Some people dislike examples using arithmetic; however,
3132 if you are such a person, do not despair. Hardly any of the code we
3133 will study in the remainder of this introduction involves arithmetic or
3134 mathematics. The examples mostly involve text in one way or another.
3136 A function definition has up to five parts following the word
3141 The name of the symbol to which the function definition should be
3145 A list of the arguments that will be passed to the function. If no
3146 arguments will be passed to the function, this is an empty list,
3150 Documentation describing the function. (Technically optional, but
3151 strongly recommended.)
3154 Optionally, an expression to make the function interactive so you can
3155 use it by typing @kbd{M-x} and then the name of the function; or by
3156 typing an appropriate key or keychord.
3158 @cindex @samp{body} defined
3160 The code that instructs the computer what to do: the @dfn{body} of the
3161 function definition.
3164 It is helpful to think of the five parts of a function definition as
3165 being organized in a template, with slots for each part:
3169 (defun @var{function-name} (@var{arguments}@dots{})
3170 "@var{optional-documentation}@dots{}"
3171 (interactive @var{argument-passing-info}) ; @r{optional}
3176 As an example, here is the code for a function that multiplies its
3177 argument by 7. (This example is not interactive. @xref{Interactive,
3178 , Making a Function Interactive}, for that information.)
3182 (defun multiply-by-seven (number)
3183 "Multiply NUMBER by seven."
3188 This definition begins with a parenthesis and the symbol @code{defun},
3189 followed by the name of the function.
3191 @cindex @samp{argument list} defined
3192 The name of the function is followed by a list that contains the
3193 arguments that will be passed to the function. This list is called
3194 the @dfn{argument list}. In this example, the list has only one
3195 element, the symbol, @code{number}. When the function is used, the
3196 symbol will be bound to the value that is used as the argument to the
3199 Instead of choosing the word @code{number} for the name of the argument,
3200 I could have picked any other name. For example, I could have chosen
3201 the word @code{multiplicand}. I picked the word `number' because it
3202 tells what kind of value is intended for this slot; but I could just as
3203 well have chosen the word `multiplicand' to indicate the role that the
3204 value placed in this slot will play in the workings of the function. I
3205 could have called it @code{foogle}, but that would have been a bad
3206 choice because it would not tell humans what it means. The choice of
3207 name is up to the programmer and should be chosen to make the meaning of
3210 Indeed, you can choose any name you wish for a symbol in an argument
3211 list, even the name of a symbol used in some other function: the name
3212 you use in an argument list is private to that particular definition.
3213 In that definition, the name refers to a different entity than any use
3214 of the same name outside the function definition. Suppose you have a
3215 nick-name `Shorty' in your family; when your family members refer to
3216 `Shorty', they mean you. But outside your family, in a movie, for
3217 example, the name `Shorty' refers to someone else. Because a name in an
3218 argument list is private to the function definition, you can change the
3219 value of such a symbol inside the body of a function without changing
3220 its value outside the function. The effect is similar to that produced
3221 by a @code{let} expression. (@xref{let, , @code{let}}.)
3224 Note also that we discuss the word `number' in two different ways: as a
3225 symbol that appears in the code, and as the name of something that will
3226 be replaced by a something else during the evaluation of the function.
3227 In the first case, @code{number} is a symbol, not a number; it happens
3228 that within the function, it is a variable who value is the number in
3229 question, but our primary interest in it is as a symbol. On the other
3230 hand, when we are talking about the function, our interest is that we
3231 will substitute a number for the word @var{number}. To keep this
3232 distinction clear, we use different typography for the two
3233 circumstances. When we talk about this function, or about how it works,
3234 we refer to this number by writing @var{number}. In the function
3235 itself, we refer to it by writing @code{number}.
3238 The argument list is followed by the documentation string that
3239 describes the function. This is what you see when you type
3240 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3241 write a documentation string like this, you should make the first line
3242 a complete sentence since some commands, such as @code{apropos}, print
3243 only the first line of a multi-line documentation string. Also, you
3244 should not indent the second line of a documentation string, if you
3245 have one, because that looks odd when you use @kbd{C-h f}
3246 (@code{describe-function}). The documentation string is optional, but
3247 it is so useful, it should be included in almost every function you
3250 @findex * @r{(multiplication)}
3251 The third line of the example consists of the body of the function
3252 definition. (Most functions' definitions, of course, are longer than
3253 this.) In this function, the body is the list, @code{(* 7 number)}, which
3254 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3255 @code{*} is the function for multiplication, just as @code{+} is the
3256 function for addition.)
3258 When you use the @code{multiply-by-seven} function, the argument
3259 @code{number} evaluates to the actual number you want used. Here is an
3260 example that shows how @code{multiply-by-seven} is used; but don't try
3261 to evaluate this yet!
3264 (multiply-by-seven 3)
3268 The symbol @code{number}, specified in the function definition in the
3269 next section, is given or ``bound to'' the value 3 in the actual use of
3270 the function. Note that although @code{number} was inside parentheses
3271 in the function definition, the argument passed to the
3272 @code{multiply-by-seven} function is not in parentheses. The
3273 parentheses are written in the function definition so the computer can
3274 figure out where the argument list ends and the rest of the function
3277 If you evaluate this example, you are likely to get an error message.
3278 (Go ahead, try it!) This is because we have written the function
3279 definition, but not yet told the computer about the definition---we have
3280 not yet installed (or `loaded') the function definition in Emacs.
3281 Installing a function is the process that tells the Lisp interpreter the
3282 definition of the function. Installation is described in the next
3285 @node Install, Interactive, defun, Writing Defuns
3286 @comment node-name, next, previous, up
3287 @section Install a Function Definition
3288 @cindex Install a Function Definition
3289 @cindex Definition installation
3290 @cindex Function definition installation
3292 If you are reading this inside of Info in Emacs, you can try out the
3293 @code{multiply-by-seven} function by first evaluating the function
3294 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3295 the function definition follows. Place the cursor after the last
3296 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3297 do this, @code{multiply-by-seven} will appear in the echo area. (What
3298 this means is that when a function definition is evaluated, the value it
3299 returns is the name of the defined function.) At the same time, this
3300 action installs the function definition.
3304 (defun multiply-by-seven (number)
3305 "Multiply NUMBER by seven."
3311 By evaluating this @code{defun}, you have just installed
3312 @code{multiply-by-seven} in Emacs. The function is now just as much a
3313 part of Emacs as @code{forward-word} or any other editing function you
3314 use. (@code{multiply-by-seven} will stay installed until you quit
3315 Emacs. To reload code automatically whenever you start Emacs, see
3316 @ref{Permanent Installation, , Installing Code Permanently}.)
3319 * Effect of installation::
3320 * Change a defun:: How to change a function definition.
3323 @node Effect of installation, Change a defun, Install, Install
3325 @unnumberedsubsec The effect of installation
3328 You can see the effect of installing @code{multiply-by-seven} by
3329 evaluating the following sample. Place the cursor after the following
3330 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3334 (multiply-by-seven 3)
3337 If you wish, you can read the documentation for the function by typing
3338 @kbd{C-h f} (@code{describe-function}) and then the name of the
3339 function, @code{multiply-by-seven}. When you do this, a
3340 @file{*Help*} window will appear on your screen that says:
3344 multiply-by-seven is a Lisp function.
3345 (multiply-by-seven NUMBER)
3347 Multiply NUMBER by seven.
3352 (To return to a single window on your screen, type @kbd{C-x 1}.)
3354 @node Change a defun, , Effect of installation, Install
3355 @comment node-name, next, previous, up
3356 @subsection Change a Function Definition
3357 @cindex Changing a function definition
3358 @cindex Function definition, how to change
3359 @cindex Definition, how to change
3361 If you want to change the code in @code{multiply-by-seven}, just rewrite
3362 it. To install the new version in place of the old one, evaluate the
3363 function definition again. This is how you modify code in Emacs. It is
3366 As an example, you can change the @code{multiply-by-seven} function to
3367 add the number to itself seven times instead of multiplying the number
3368 by seven. It produces the same answer, but by a different path. At
3369 the same time, we will add a comment to the code; a comment is text
3370 that the Lisp interpreter ignores, but that a human reader may find
3371 useful or enlightening. The comment is that this is the ``second
3376 (defun multiply-by-seven (number) ; @r{Second version.}
3377 "Multiply NUMBER by seven."
3378 (+ number number number number number number number))
3382 @cindex Comments in Lisp code
3383 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3384 line that follows a semicolon is a comment. The end of the line is the
3385 end of the comment. To stretch a comment over two or more lines, begin
3386 each line with a semicolon.
3388 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3389 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3390 Reference Manual}, for more about comments.
3392 You can install this version of the @code{multiply-by-seven} function by
3393 evaluating it in the same way you evaluated the first function: place
3394 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3396 In summary, this is how you write code in Emacs Lisp: you write a
3397 function; install it; test it; and then make fixes or enhancements and
3400 @node Interactive, Interactive Options, Install, Writing Defuns
3401 @comment node-name, next, previous, up
3402 @section Make a Function Interactive
3403 @cindex Interactive functions
3406 You make a function interactive by placing a list that begins with
3407 the special form @code{interactive} immediately after the
3408 documentation. A user can invoke an interactive function by typing
3409 @kbd{M-x} and then the name of the function; or by typing the keys to
3410 which it is bound, for example, by typing @kbd{C-n} for
3411 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3413 Interestingly, when you call an interactive function interactively,
3414 the value returned is not automatically displayed in the echo area.
3415 This is because you often call an interactive function for its side
3416 effects, such as moving forward by a word or line, and not for the
3417 value returned. If the returned value were displayed in the echo area
3418 each time you typed a key, it would be very distracting.
3421 * Interactive multiply-by-seven:: An overview.
3422 * multiply-by-seven in detail:: The interactive version.
3425 @node Interactive multiply-by-seven, multiply-by-seven in detail, Interactive, Interactive
3427 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3430 Both the use of the special form @code{interactive} and one way to
3431 display a value in the echo area can be illustrated by creating an
3432 interactive version of @code{multiply-by-seven}.
3439 (defun multiply-by-seven (number) ; @r{Interactive version.}
3440 "Multiply NUMBER by seven."
3442 (message "The result is %d" (* 7 number)))
3447 You can install this code by placing your cursor after it and typing
3448 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3449 Then, you can use this code by typing @kbd{C-u} and a number and then
3450 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3451 @samp{The result is @dots{}} followed by the product will appear in the
3454 Speaking more generally, you invoke a function like this in either of two
3459 By typing a prefix argument that contains the number to be passed, and
3460 then typing @kbd{M-x} and the name of the function, as with
3461 @kbd{C-u 3 M-x forward-sentence}; or,
3464 By typing whatever key or keychord the function is bound to, as with
3469 Both the examples just mentioned work identically to move point forward
3470 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3471 it could not be used as an example of key binding.)
3473 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3476 A prefix argument is passed to an interactive function by typing the
3477 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3478 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3479 type @kbd{C-u} without a number, it defaults to 4).
3481 @node multiply-by-seven in detail, , Interactive multiply-by-seven, Interactive
3482 @comment node-name, next, previous, up
3483 @subsection An Interactive @code{multiply-by-seven}
3485 Let's look at the use of the special form @code{interactive} and then at
3486 the function @code{message} in the interactive version of
3487 @code{multiply-by-seven}. You will recall that the function definition
3492 (defun multiply-by-seven (number) ; @r{Interactive version.}
3493 "Multiply NUMBER by seven."
3495 (message "The result is %d" (* 7 number)))
3499 In this function, the expression, @code{(interactive "p")}, is a list of
3500 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3501 the function and use its value for the argument of the function.
3504 The argument will be a number. This means that the symbol
3505 @code{number} will be bound to a number in the line:
3508 (message "The result is %d" (* 7 number))
3513 For example, if your prefix argument is 5, the Lisp interpreter will
3514 evaluate the line as if it were:
3517 (message "The result is %d" (* 7 5))
3521 (If you are reading this in GNU Emacs, you can evaluate this expression
3522 yourself.) First, the interpreter will evaluate the inner list, which
3523 is @code{(* 7 5)}. This returns a value of 35. Next, it
3524 will evaluate the outer list, passing the values of the second and
3525 subsequent elements of the list to the function @code{message}.
3527 As we have seen, @code{message} is an Emacs Lisp function especially
3528 designed for sending a one line message to a user. (@xref{message, ,
3529 The @code{message} function}.) In summary, the @code{message}
3530 function prints its first argument in the echo area as is, except for
3531 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3532 which we have not mentioned). When it sees a control sequence, the
3533 function looks to the second or subsequent arguments and prints the
3534 value of the argument in the location in the string where the control
3535 sequence is located.
3537 In the interactive @code{multiply-by-seven} function, the control string
3538 is @samp{%d}, which requires a number, and the value returned by
3539 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3540 is printed in place of the @samp{%d} and the message is @samp{The result
3543 (Note that when you call the function @code{multiply-by-seven}, the
3544 message is printed without quotes, but when you call @code{message}, the
3545 text is printed in double quotes. This is because the value returned by
3546 @code{message} is what appears in the echo area when you evaluate an
3547 expression whose first element is @code{message}; but when embedded in a
3548 function, @code{message} prints the text as a side effect without
3551 @node Interactive Options, Permanent Installation, Interactive, Writing Defuns
3552 @comment node-name, next, previous, up
3553 @section Different Options for @code{interactive}
3554 @cindex Options for @code{interactive}
3555 @cindex Interactive options
3557 In the example, @code{multiply-by-seven} used @code{"p"} as the
3558 argument to @code{interactive}. This argument told Emacs to interpret
3559 your typing either @kbd{C-u} followed by a number or @key{META}
3560 followed by a number as a command to pass that number to the function
3561 as its argument. Emacs has more than twenty characters predefined for
3562 use with @code{interactive}. In almost every case, one of these
3563 options will enable you to pass the right information interactively to
3564 a function. (@xref{Interactive Codes, , Code Characters for
3565 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3568 Consider the function @code{zap-to-char}. Its interactive expression
3572 (interactive "p\ncZap to char: ")
3575 The first part of the argument to @code{interactive} is @samp{p}, with
3576 which you are already familiar. This argument tells Emacs to
3577 interpret a `prefix', as a number to be passed to the function. You
3578 can specify a prefix either by typing @kbd{C-u} followed by a number
3579 or by typing @key{META} followed by a number. The prefix is the
3580 number of specified characters. Thus, if your prefix is three and the
3581 specified character is @samp{x}, then you will delete all the text up
3582 to and including the third next @samp{x}. If you do not set a prefix,
3583 then you delete all the text up to and including the specified
3584 character, but no more.
3586 The @samp{c} tells the function the name of the character to which to delete.
3588 More formally, a function with two or more arguments can have
3589 information passed to each argument by adding parts to the string that
3590 follows @code{interactive}. When you do this, the information is
3591 passed to each argument in the same order it is specified in the
3592 @code{interactive} list. In the string, each part is separated from
3593 the next part by a @samp{\n}, which is a newline. For example, you
3594 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3595 This causes Emacs to pass the value of the prefix argument (if there
3596 is one) and the character.
3598 In this case, the function definition looks like the following, where
3599 @code{arg} and @code{char} are the symbols to which @code{interactive}
3600 binds the prefix argument and the specified character:
3604 (defun @var{name-of-function} (arg char)
3605 "@var{documentation}@dots{}"
3606 (interactive "p\ncZap to char: ")
3607 @var{body-of-function}@dots{})
3612 (The space after the colon in the prompt makes it look better when you
3613 are prompted. @xref{copy-to-buffer, , The Definition of
3614 @code{copy-to-buffer}}, for an example.)
3616 When a function does not take arguments, @code{interactive} does not
3617 require any. Such a function contains the simple expression
3618 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3621 Alternatively, if the special letter-codes are not right for your
3622 application, you can pass your own arguments to @code{interactive} as
3625 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3626 for an example. @xref{Using Interactive, , Using @code{Interactive},
3627 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3628 explanation about this technique.
3630 @node Permanent Installation, let, Interactive Options, Writing Defuns
3631 @comment node-name, next, previous, up
3632 @section Install Code Permanently
3633 @cindex Install code permanently
3634 @cindex Permanent code installation
3635 @cindex Code installation
3637 When you install a function definition by evaluating it, it will stay
3638 installed until you quit Emacs. The next time you start a new session
3639 of Emacs, the function will not be installed unless you evaluate the
3640 function definition again.
3642 At some point, you may want to have code installed automatically
3643 whenever you start a new session of Emacs. There are several ways of
3648 If you have code that is just for yourself, you can put the code for the
3649 function definition in your @file{.emacs} initialization file. When you
3650 start Emacs, your @file{.emacs} file is automatically evaluated and all
3651 the function definitions within it are installed.
3652 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3655 Alternatively, you can put the function definitions that you want
3656 installed in one or more files of their own and use the @code{load}
3657 function to cause Emacs to evaluate and thereby install each of the
3658 functions in the files.
3659 @xref{Loading Files, , Loading Files}.
3662 Thirdly, if you have code that your whole site will use, it is usual
3663 to put it in a file called @file{site-init.el} that is loaded when
3664 Emacs is built. This makes the code available to everyone who uses
3665 your machine. (See the @file{INSTALL} file that is part of the Emacs
3669 Finally, if you have code that everyone who uses Emacs may want, you
3670 can post it on a computer network or send a copy to the Free Software
3671 Foundation. (When you do this, please license the code and its
3672 documentation under a license that permits other people to run, copy,
3673 study, modify, and redistribute the code and which protects you from
3674 having your work taken from you.) If you send a copy of your code to
3675 the Free Software Foundation, and properly protect yourself and
3676 others, it may be included in the next release of Emacs. In large
3677 part, this is how Emacs has grown over the past years, by donations.
3679 @node let, if, Permanent Installation, Writing Defuns
3680 @comment node-name, next, previous, up
3684 The @code{let} expression is a special form in Lisp that you will need
3685 to use in most function definitions.
3687 @code{let} is used to attach or bind a symbol to a value in such a way
3688 that the Lisp interpreter will not confuse the variable with a
3689 variable of the same name that is not part of the function.
3691 To understand why the @code{let} special form is necessary, consider
3692 the situation in which you own a home that you generally refer to as
3693 `the house', as in the sentence, ``The house needs painting.'' If you
3694 are visiting a friend and your host refers to `the house', he is
3695 likely to be referring to @emph{his} house, not yours, that is, to a
3698 If your friend is referring to his house and you think he is referring
3699 to your house, you may be in for some confusion. The same thing could
3700 happen in Lisp if a variable that is used inside of one function has
3701 the same name as a variable that is used inside of another function,
3702 and the two are not intended to refer to the same value. The
3703 @code{let} special form prevents this kind of confusion.
3706 * Prevent confusion::
3707 * Parts of let Expression::
3708 * Sample let Expression::
3709 * Uninitialized let Variables::
3712 @node Prevent confusion, Parts of let Expression, let, let
3714 @unnumberedsubsec @code{let} Prevents Confusion
3717 @cindex @samp{local variable} defined
3718 @cindex @samp{variable, local}, defined
3719 The @code{let} special form prevents confusion. @code{let} creates a
3720 name for a @dfn{local variable} that overshadows any use of the same
3721 name outside the @code{let} expression. This is like understanding
3722 that whenever your host refers to `the house', he means his house, not
3723 yours. (Symbols used in argument lists work the same way.
3724 @xref{defun, , The @code{defun} Special Form}.)
3726 Local variables created by a @code{let} expression retain their value
3727 @emph{only} within the @code{let} expression itself (and within
3728 expressions called within the @code{let} expression); the local
3729 variables have no effect outside the @code{let} expression.
3731 Another way to think about @code{let} is that it is like a @code{setq}
3732 that is temporary and local. The values set by @code{let} are
3733 automatically undone when the @code{let} is finished. The setting
3734 only affects expressions that are inside the bounds of the @code{let}
3735 expression. In computer science jargon, we would say ``the binding of
3736 a symbol is visible only in functions called in the @code{let} form;
3737 in Emacs Lisp, scoping is dynamic, not lexical.''
3739 @code{let} can create more than one variable at once. Also,
3740 @code{let} gives each variable it creates an initial value, either a
3741 value specified by you, or @code{nil}. (In the jargon, this is called
3742 `binding the variable to the value'.) After @code{let} has created
3743 and bound the variables, it executes the code in the body of the
3744 @code{let}, and returns the value of the last expression in the body,
3745 as the value of the whole @code{let} expression. (`Execute' is a jargon
3746 term that means to evaluate a list; it comes from the use of the word
3747 meaning `to give practical effect to' (@cite{Oxford English
3748 Dictionary}). Since you evaluate an expression to perform an action,
3749 `execute' has evolved as a synonym to `evaluate'.)
3751 @node Parts of let Expression, Sample let Expression, Prevent confusion, let
3752 @comment node-name, next, previous, up
3753 @subsection The Parts of a @code{let} Expression
3754 @cindex @code{let} expression, parts of
3755 @cindex Parts of @code{let} expression
3757 @cindex @samp{varlist} defined
3758 A @code{let} expression is a list of three parts. The first part is
3759 the symbol @code{let}. The second part is a list, called a
3760 @dfn{varlist}, each element of which is either a symbol by itself or a
3761 two-element list, the first element of which is a symbol. The third
3762 part of the @code{let} expression is the body of the @code{let}. The
3763 body usually consists of one or more lists.
3766 A template for a @code{let} expression looks like this:
3769 (let @var{varlist} @var{body}@dots{})
3773 The symbols in the varlist are the variables that are given initial
3774 values by the @code{let} special form. Symbols by themselves are given
3775 the initial value of @code{nil}; and each symbol that is the first
3776 element of a two-element list is bound to the value that is returned
3777 when the Lisp interpreter evaluates the second element.
3779 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3780 this case, in a @code{let} expression, Emacs binds the symbol
3781 @code{thread} to an initial value of @code{nil}, and binds the symbol
3782 @code{needles} to an initial value of 3.
3784 When you write a @code{let} expression, what you do is put the
3785 appropriate expressions in the slots of the @code{let} expression
3788 If the varlist is composed of two-element lists, as is often the case,
3789 the template for the @code{let} expression looks like this:
3793 (let ((@var{variable} @var{value})
3794 (@var{variable} @var{value})
3800 @node Sample let Expression, Uninitialized let Variables, Parts of let Expression, let
3801 @comment node-name, next, previous, up
3802 @subsection Sample @code{let} Expression
3803 @cindex Sample @code{let} expression
3804 @cindex @code{let} expression sample
3806 The following expression creates and gives initial values
3807 to the two variables @code{zebra} and @code{tiger}. The body of the
3808 @code{let} expression is a list which calls the @code{message} function.
3812 (let ((zebra 'stripes)
3814 (message "One kind of animal has %s and another is %s."
3819 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3821 The two variables are @code{zebra} and @code{tiger}. Each variable is
3822 the first element of a two-element list and each value is the second
3823 element of its two-element list. In the varlist, Emacs binds the
3824 variable @code{zebra} to the value @code{stripes}@footnote{According
3825 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3826 become impossibly dangerous as they grow older'' but the claim here is
3827 that they do not become fierce like a tiger. (1997, W. W. Norton and
3828 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3829 variable @code{tiger} to the value @code{fierce}. In this example,
3830 both values are symbols preceded by a quote. The values could just as
3831 well have been another list or a string. The body of the @code{let}
3832 follows after the list holding the variables. In this example, the
3833 body is a list that uses the @code{message} function to print a string
3837 You may evaluate the example in the usual fashion, by placing the
3838 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3839 this, the following will appear in the echo area:
3842 "One kind of animal has stripes and another is fierce."
3845 As we have seen before, the @code{message} function prints its first
3846 argument, except for @samp{%s}. In this example, the value of the variable
3847 @code{zebra} is printed at the location of the first @samp{%s} and the
3848 value of the variable @code{tiger} is printed at the location of the
3851 @node Uninitialized let Variables, , Sample let Expression, let
3852 @comment node-name, next, previous, up
3853 @subsection Uninitialized Variables in a @code{let} Statement
3854 @cindex Uninitialized @code{let} variables
3855 @cindex @code{let} variables uninitialized
3857 If you do not bind the variables in a @code{let} statement to specific
3858 initial values, they will automatically be bound to an initial value of
3859 @code{nil}, as in the following expression:
3868 "Here are %d variables with %s, %s, and %s value."
3869 birch pine fir oak))
3874 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3877 If you evaluate this expression in the usual way, the following will
3878 appear in your echo area:
3881 "Here are 3 variables with nil, nil, and some value."
3885 In this example, Emacs binds the symbol @code{birch} to the number 3,
3886 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3887 the symbol @code{oak} to the value @code{some}.
3889 Note that in the first part of the @code{let}, the variables @code{pine}
3890 and @code{fir} stand alone as atoms that are not surrounded by
3891 parentheses; this is because they are being bound to @code{nil}, the
3892 empty list. But @code{oak} is bound to @code{some} and so is a part of
3893 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3894 number 3 and so is in a list with that number. (Since a number
3895 evaluates to itself, the number does not need to be quoted. Also, the
3896 number is printed in the message using a @samp{%d} rather than a
3897 @samp{%s}.) The four variables as a group are put into a list to
3898 delimit them from the body of the @code{let}.
3900 @node if, else, let, Writing Defuns
3901 @comment node-name, next, previous, up
3902 @section The @code{if} Special Form
3904 @cindex Conditional with @code{if}
3906 A third special form, in addition to @code{defun} and @code{let}, is the
3907 conditional @code{if}. This form is used to instruct the computer to
3908 make decisions. You can write function definitions without using
3909 @code{if}, but it is used often enough, and is important enough, to be
3910 included here. It is used, for example, in the code for the
3911 function @code{beginning-of-buffer}.
3913 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3914 @emph{then} an expression is evaluated.'' If the test is not true, the
3915 expression is not evaluated. For example, you might make a decision
3916 such as, ``if it is warm and sunny, then go to the beach!''
3919 * if in more detail::
3920 * type-of-animal in detail:: An example of an @code{if} expression.
3923 @node if in more detail, type-of-animal in detail, if, if
3925 @unnumberedsubsec @code{if} in more detail
3928 @cindex @samp{if-part} defined
3929 @cindex @samp{then-part} defined
3930 An @code{if} expression written in Lisp does not use the word `then';
3931 the test and the action are the second and third elements of the list
3932 whose first element is @code{if}. Nonetheless, the test part of an
3933 @code{if} expression is often called the @dfn{if-part} and the second
3934 argument is often called the @dfn{then-part}.
3936 Also, when an @code{if} expression is written, the true-or-false-test
3937 is usually written on the same line as the symbol @code{if}, but the
3938 action to carry out if the test is true, the ``then-part'', is written
3939 on the second and subsequent lines. This makes the @code{if}
3940 expression easier to read.
3944 (if @var{true-or-false-test}
3945 @var{action-to-carry-out-if-test-is-true})
3950 The true-or-false-test will be an expression that
3951 is evaluated by the Lisp interpreter.
3953 Here is an example that you can evaluate in the usual manner. The test
3954 is whether the number 5 is greater than the number 4. Since it is, the
3955 message @samp{5 is greater than 4!} will be printed.
3959 (if (> 5 4) ; @r{if-part}
3960 (message "5 is greater than 4!")) ; @r{then-part}
3965 (The function @code{>} tests whether its first argument is greater than
3966 its second argument and returns true if it is.)
3967 @findex > (greater than)
3969 Of course, in actual use, the test in an @code{if} expression will not
3970 be fixed for all time as it is by the expression @code{(> 5 4)}.
3971 Instead, at least one of the variables used in the test will be bound to
3972 a value that is not known ahead of time. (If the value were known ahead
3973 of time, we would not need to run the test!)
3975 For example, the value may be bound to an argument of a function
3976 definition. In the following function definition, the character of the
3977 animal is a value that is passed to the function. If the value bound to
3978 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3979 tiger!} will be printed; otherwise, @code{nil} will be returned.
3983 (defun type-of-animal (characteristic)
3984 "Print message in echo area depending on CHARACTERISTIC.
3985 If the CHARACTERISTIC is the symbol `fierce',
3986 then warn of a tiger."
3987 (if (equal characteristic 'fierce)
3988 (message "It's a tiger!")))
3994 If you are reading this inside of GNU Emacs, you can evaluate the
3995 function definition in the usual way to install it in Emacs, and then you
3996 can evaluate the following two expressions to see the results:
4000 (type-of-animal 'fierce)
4002 (type-of-animal 'zebra)
4007 @c Following sentences rewritten to prevent overfull hbox.
4009 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4010 following message printed in the echo area: @code{"It's a tiger!"}; and
4011 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
4012 printed in the echo area.
4014 @node type-of-animal in detail, , if in more detail, if
4015 @comment node-name, next, previous, up
4016 @subsection The @code{type-of-animal} Function in Detail
4018 Let's look at the @code{type-of-animal} function in detail.
4020 The function definition for @code{type-of-animal} was written by filling
4021 the slots of two templates, one for a function definition as a whole, and
4022 a second for an @code{if} expression.
4025 The template for every function that is not interactive is:
4029 (defun @var{name-of-function} (@var{argument-list})
4030 "@var{documentation}@dots{}"
4036 The parts of the function that match this template look like this:
4040 (defun type-of-animal (characteristic)
4041 "Print message in echo area depending on CHARACTERISTIC.
4042 If the CHARACTERISTIC is the symbol `fierce',
4043 then warn of a tiger."
4044 @var{body: the} @code{if} @var{expression})
4048 The name of function is @code{type-of-animal}; it is passed the value
4049 of one argument. The argument list is followed by a multi-line
4050 documentation string. The documentation string is included in the
4051 example because it is a good habit to write documentation string for
4052 every function definition. The body of the function definition
4053 consists of the @code{if} expression.
4056 The template for an @code{if} expression looks like this:
4060 (if @var{true-or-false-test}
4061 @var{action-to-carry-out-if-the-test-returns-true})
4066 In the @code{type-of-animal} function, the code for the @code{if}
4071 (if (equal characteristic 'fierce)
4072 (message "It's a tiger!")))
4077 Here, the true-or-false-test is the expression:
4080 (equal characteristic 'fierce)
4084 In Lisp, @code{equal} is a function that determines whether its first
4085 argument is equal to its second argument. The second argument is the
4086 quoted symbol @code{'fierce} and the first argument is the value of the
4087 symbol @code{characteristic}---in other words, the argument passed to
4090 In the first exercise of @code{type-of-animal}, the argument
4091 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4092 is equal to @code{fierce}, the expression, @code{(equal characteristic
4093 'fierce)}, returns a value of true. When this happens, the @code{if}
4094 evaluates the second argument or then-part of the @code{if}:
4095 @code{(message "It's tiger!")}.
4097 On the other hand, in the second exercise of @code{type-of-animal}, the
4098 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4099 is not equal to @code{fierce}, so the then-part is not evaluated and
4100 @code{nil} is returned by the @code{if} expression.
4102 @node else, Truth & Falsehood, if, Writing Defuns
4103 @comment node-name, next, previous, up
4104 @section If--then--else Expressions
4107 An @code{if} expression may have an optional third argument, called
4108 the @dfn{else-part}, for the case when the true-or-false-test returns
4109 false. When this happens, the second argument or then-part of the
4110 overall @code{if} expression is @emph{not} evaluated, but the third or
4111 else-part @emph{is} evaluated. You might think of this as the cloudy
4112 day alternative for the decision ``if it is warm and sunny, then go to
4113 the beach, else read a book!''.
4115 The word ``else'' is not written in the Lisp code; the else-part of an
4116 @code{if} expression comes after the then-part. In the written Lisp, the
4117 else-part is usually written to start on a line of its own and is
4118 indented less than the then-part:
4122 (if @var{true-or-false-test}
4123 @var{action-to-carry-out-if-the-test-returns-true}
4124 @var{action-to-carry-out-if-the-test-returns-false})
4128 For example, the following @code{if} expression prints the message @samp{4
4129 is not greater than 5!} when you evaluate it in the usual way:
4133 (if (> 4 5) ; @r{if-part}
4134 (message "4 falsely greater than 5!") ; @r{then-part}
4135 (message "4 is not greater than 5!")) ; @r{else-part}
4140 Note that the different levels of indentation make it easy to
4141 distinguish the then-part from the else-part. (GNU Emacs has several
4142 commands that automatically indent @code{if} expressions correctly.
4143 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4145 We can extend the @code{type-of-animal} function to include an
4146 else-part by simply incorporating an additional part to the @code{if}
4150 You can see the consequences of doing this if you evaluate the following
4151 version of the @code{type-of-animal} function definition to install it
4152 and then evaluate the two subsequent expressions to pass different
4153 arguments to the function.
4157 (defun type-of-animal (characteristic) ; @r{Second version.}
4158 "Print message in echo area depending on CHARACTERISTIC.
4159 If the CHARACTERISTIC is the symbol `fierce',
4160 then warn of a tiger;
4161 else say it's not fierce."
4162 (if (equal characteristic 'fierce)
4163 (message "It's a tiger!")
4164 (message "It's not fierce!")))
4171 (type-of-animal 'fierce)
4173 (type-of-animal 'zebra)
4178 @c Following sentence rewritten to prevent overfull hbox.
4180 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4181 following message printed in the echo area: @code{"It's a tiger!"}; but
4182 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4183 @code{"It's not fierce!"}.
4185 (Of course, if the @var{characteristic} were @code{ferocious}, the
4186 message @code{"It's not fierce!"} would be printed; and it would be
4187 misleading! When you write code, you need to take into account the
4188 possibility that some such argument will be tested by the @code{if}
4189 and write your program accordingly.)
4191 @node Truth & Falsehood, save-excursion, else, Writing Defuns
4192 @comment node-name, next, previous, up
4193 @section Truth and Falsehood in Emacs Lisp
4194 @cindex Truth and falsehood in Emacs Lisp
4195 @cindex Falsehood and truth in Emacs Lisp
4198 There is an important aspect to the truth test in an @code{if}
4199 expression. So far, we have spoken of `true' and `false' as values of
4200 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4201 `false' is just our old friend @code{nil}. Anything else---anything
4204 The expression that tests for truth is interpreted as @dfn{true}
4205 if the result of evaluating it is a value that is not @code{nil}. In
4206 other words, the result of the test is considered true if the value
4207 returned is a number such as 47, a string such as @code{"hello"}, or a
4208 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4209 long as it is not empty), or even a buffer!
4212 * nil explained:: @code{nil} has two meanings.
4215 @node nil explained, , Truth & Falsehood, Truth & Falsehood
4217 @unnumberedsubsec An explanation of @code{nil}
4220 Before illustrating a test for truth, we need an explanation of @code{nil}.
4222 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4223 empty list. Second, it means false and is the value returned when a
4224 true-or-false-test tests false. @code{nil} can be written as an empty
4225 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4226 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4227 to use @code{nil} for false and @code{()} for the empty list.
4229 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4230 list---is considered true. This means that if an evaluation returns
4231 something that is not an empty list, an @code{if} expression will test
4232 true. For example, if a number is put in the slot for the test, it
4233 will be evaluated and will return itself, since that is what numbers
4234 do when evaluated. In this conditional, the @code{if} expression will
4235 test true. The expression tests false only when @code{nil}, an empty
4236 list, is returned by evaluating the expression.
4238 You can see this by evaluating the two expressions in the following examples.
4240 In the first example, the number 4 is evaluated as the test in the
4241 @code{if} expression and returns itself; consequently, the then-part
4242 of the expression is evaluated and returned: @samp{true} appears in
4243 the echo area. In the second example, the @code{nil} indicates false;
4244 consequently, the else-part of the expression is evaluated and
4245 returned: @samp{false} appears in the echo area.
4262 Incidentally, if some other useful value is not available for a test that
4263 returns true, then the Lisp interpreter will return the symbol @code{t}
4264 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4265 when evaluated, as you can see by evaluating it in the usual way:
4273 On the other hand, this function returns @code{nil} if the test is false.
4279 @node save-excursion, Review, Truth & Falsehood, Writing Defuns
4280 @comment node-name, next, previous, up
4281 @section @code{save-excursion}
4282 @findex save-excursion
4283 @cindex Region, what it is
4284 @cindex Preserving point, mark, and buffer
4285 @cindex Point, mark, buffer preservation
4289 The @code{save-excursion} function is the fourth and final special form
4290 that we will discuss in this chapter.
4292 In Emacs Lisp programs used for editing, the @code{save-excursion}
4293 function is very common. It saves the location of point and mark,
4294 executes the body of the function, and then restores point and mark to
4295 their previous positions if their locations were changed. Its primary
4296 purpose is to keep the user from being surprised and disturbed by
4297 unexpected movement of point or mark.
4300 * Point and mark:: A review of various locations.
4301 * Template for save-excursion::
4304 @node Point and mark, Template for save-excursion, save-excursion, save-excursion
4306 @unnumberedsubsec Point and Mark
4309 Before discussing @code{save-excursion}, however, it may be useful
4310 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4311 the current location of the cursor. Wherever the cursor
4312 is, that is point. More precisely, on terminals where the cursor
4313 appears to be on top of a character, point is immediately before the
4314 character. In Emacs Lisp, point is an integer. The first character in
4315 a buffer is number one, the second is number two, and so on. The
4316 function @code{point} returns the current position of the cursor as a
4317 number. Each buffer has its own value for point.
4319 The @dfn{mark} is another position in the buffer; its value can be set
4320 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4321 a mark has been set, you can use the command @kbd{C-x C-x}
4322 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4323 and set the mark to be the previous position of point. In addition, if
4324 you set another mark, the position of the previous mark is saved in the
4325 mark ring. Many mark positions can be saved this way. You can jump the
4326 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4329 The part of the buffer between point and mark is called @dfn{the
4330 region}. Numerous commands work on the region, including
4331 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4332 @code{print-region}.
4334 The @code{save-excursion} special form saves the locations of point and
4335 mark and restores those positions after the code within the body of the
4336 special form is evaluated by the Lisp interpreter. Thus, if point were
4337 in the beginning of a piece of text and some code moved point to the end
4338 of the buffer, the @code{save-excursion} would put point back to where
4339 it was before, after the expressions in the body of the function were
4342 In Emacs, a function frequently moves point as part of its internal
4343 workings even though a user would not expect this. For example,
4344 @code{count-lines-region} moves point. To prevent the user from being
4345 bothered by jumps that are both unexpected and (from the user's point of
4346 view) unnecessary, @code{save-excursion} is often used to keep point and
4347 mark in the location expected by the user. The use of
4348 @code{save-excursion} is good housekeeping.
4350 To make sure the house stays clean, @code{save-excursion} restores the
4351 values of point and mark even if something goes wrong in the code inside
4352 of it (or, to be more precise and to use the proper jargon, ``in case of
4353 abnormal exit''). This feature is very helpful.
4355 In addition to recording the values of point and mark,
4356 @code{save-excursion} keeps track of the current buffer, and restores
4357 it, too. This means you can write code that will change the buffer and
4358 have @code{save-excursion} switch you back to the original buffer.
4359 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4360 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4362 @node Template for save-excursion, , Point and mark, save-excursion
4363 @comment node-name, next, previous, up
4364 @subsection Template for a @code{save-excursion} Expression
4367 The template for code using @code{save-excursion} is simple:
4377 The body of the function is one or more expressions that will be
4378 evaluated in sequence by the Lisp interpreter. If there is more than
4379 one expression in the body, the value of the last one will be returned
4380 as the value of the @code{save-excursion} function. The other
4381 expressions in the body are evaluated only for their side effects; and
4382 @code{save-excursion} itself is used only for its side effect (which
4383 is restoring the positions of point and mark).
4386 In more detail, the template for a @code{save-excursion} expression
4392 @var{first-expression-in-body}
4393 @var{second-expression-in-body}
4394 @var{third-expression-in-body}
4396 @var{last-expression-in-body})
4401 An expression, of course, may be a symbol on its own or a list.
4403 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4404 within the body of a @code{let} expression. It looks like this:
4414 @node Review, defun Exercises, save-excursion, Writing Defuns
4415 @comment node-name, next, previous, up
4418 In the last few chapters we have introduced a fair number of functions
4419 and special forms. Here they are described in brief, along with a few
4420 similar functions that have not been mentioned yet.
4423 @item eval-last-sexp
4424 Evaluate the last symbolic expression before the current location of
4425 point. The value is printed in the echo area unless the function is
4426 invoked with an argument; in that case, the output is printed in the
4427 current buffer. This command is normally bound to @kbd{C-x C-e}.
4430 Define function. This special form has up to five parts: the name,
4431 a template for the arguments that will be passed to the function,
4432 documentation, an optional interactive declaration, and the body of the
4436 For example, in an early version of Emacs, the function definition was
4437 as follows. (It is slightly more complex now that it seeks the first
4438 non-whitespace character rather than the first visible character.)
4442 (defun back-to-indentation ()
4443 "Move point to first visible character on line."
4445 (beginning-of-line 1)
4446 (skip-chars-forward " \t"))
4453 (defun backward-to-indentation (&optional arg)
4454 "Move backward ARG lines and position at first nonblank character."
4456 (forward-line (- (or arg 1)))
4457 (skip-chars-forward " \t"))
4459 (defun back-to-indentation ()
4460 "Move point to the first non-whitespace character on this line."
4462 (beginning-of-line 1)
4463 (skip-syntax-forward " " (line-end-position))
4464 ;; Move back over chars that have whitespace syntax but have the p flag.
4465 (backward-prefix-chars))
4469 Declare to the interpreter that the function can be used
4470 interactively. This special form may be followed by a string with one
4471 or more parts that pass the information to the arguments of the
4472 function, in sequence. These parts may also tell the interpreter to
4473 prompt for information. Parts of the string are separated by
4474 newlines, @samp{\n}.
4477 Common code characters are:
4481 The name of an existing buffer.
4484 The name of an existing file.
4487 The numeric prefix argument. (Note that this `p' is lower case.)
4490 Point and the mark, as two numeric arguments, smallest first. This
4491 is the only code letter that specifies two successive arguments
4495 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4496 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4500 Declare that a list of variables is for use within the body of the
4501 @code{let} and give them an initial value, either @code{nil} or a
4502 specified value; then evaluate the rest of the expressions in the body
4503 of the @code{let} and return the value of the last one. Inside the
4504 body of the @code{let}, the Lisp interpreter does not see the values of
4505 the variables of the same names that are bound outside of the
4513 (let ((foo (buffer-name))
4514 (bar (buffer-size)))
4516 "This buffer is %s and has %d characters."
4521 @item save-excursion
4522 Record the values of point and mark and the current buffer before
4523 evaluating the body of this special form. Restore the values of point
4524 and mark and buffer afterward.
4531 (message "We are %d characters into this buffer."
4534 (goto-char (point-min)) (point))))
4539 Evaluate the first argument to the function; if it is true, evaluate
4540 the second argument; else evaluate the third argument, if there is one.
4542 The @code{if} special form is called a @dfn{conditional}. There are
4543 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4551 (if (= 22 emacs-major-version)
4552 (message "This is version 22 Emacs")
4553 (message "This is not version 22 Emacs"))
4562 The @code{<} function tests whether its first argument is smaller than
4563 its second argument. A corresponding function, @code{>}, tests whether
4564 the first argument is greater than the second. Likewise, @code{<=}
4565 tests whether the first argument is less than or equal to the second and
4566 @code{>=} tests whether the first argument is greater than or equal to
4567 the second. In all cases, both arguments must be numbers or markers
4568 (markers indicate positions in buffers).
4572 The @code{=} function tests whether two arguments, both numbers or
4578 Test whether two objects are the same. @code{equal} uses one meaning
4579 of the word `same' and @code{eq} uses another: @code{equal} returns
4580 true if the two objects have a similar structure and contents, such as
4581 two copies of the same book. On the other hand, @code{eq}, returns
4582 true if both arguments are actually the same object.
4591 The @code{string-lessp} function tests whether its first argument is
4592 smaller than the second argument. A shorter, alternative name for the
4593 same function (a @code{defalias}) is @code{string<}.
4595 The arguments to @code{string-lessp} must be strings or symbols; the
4596 ordering is lexicographic, so case is significant. The print names of
4597 symbols are used instead of the symbols themselves.
4599 @cindex @samp{empty string} defined
4600 An empty string, @samp{""}, a string with no characters in it, is
4601 smaller than any string of characters.
4603 @code{string-equal} provides the corresponding test for equality. Its
4604 shorter, alternative name is @code{string=}. There are no string test
4605 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4608 Print a message in the echo area. The first argument is a string that
4609 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4610 arguments that follow the string. The argument used by @samp{%s} must
4611 be a string or a symbol; the argument used by @samp{%d} must be a
4612 number. The argument used by @samp{%c} must be an @sc{ascii} code
4613 number; it will be printed as the character with that @sc{ascii} code.
4614 (Various other %-sequences have not been mentioned.)
4618 The @code{setq} function sets the value of its first argument to the
4619 value of the second argument. The first argument is automatically
4620 quoted by @code{setq}. It does the same for succeeding pairs of
4621 arguments. Another function, @code{set}, takes only two arguments and
4622 evaluates both of them before setting the value returned by its first
4623 argument to the value returned by its second argument.
4626 Without an argument, return the name of the buffer, as a string.
4628 @itemx buffer-file-name
4629 Without an argument, return the name of the file the buffer is
4632 @item current-buffer
4633 Return the buffer in which Emacs is active; it may not be
4634 the buffer that is visible on the screen.
4637 Return the most recently selected buffer (other than the buffer passed
4638 to @code{other-buffer} as an argument and other than the current
4641 @item switch-to-buffer
4642 Select a buffer for Emacs to be active in and display it in the current
4643 window so users can look at it. Usually bound to @kbd{C-x b}.
4646 Switch Emacs' attention to a buffer on which programs will run. Don't
4647 alter what the window is showing.
4650 Return the number of characters in the current buffer.
4653 Return the value of the current position of the cursor, as an
4654 integer counting the number of characters from the beginning of the
4658 Return the minimum permissible value of point in
4659 the current buffer. This is 1, unless narrowing is in effect.
4662 Return the value of the maximum permissible value of point in the
4663 current buffer. This is the end of the buffer, unless narrowing is in
4668 @node defun Exercises, , Review, Writing Defuns
4673 Write a non-interactive function that doubles the value of its
4674 argument, a number. Make that function interactive.
4677 Write a function that tests whether the current value of
4678 @code{fill-column} is greater than the argument passed to the function,
4679 and if so, prints an appropriate message.
4682 @node Buffer Walk Through, More Complex, Writing Defuns, Top
4683 @comment node-name, next, previous, up
4684 @chapter A Few Buffer--Related Functions
4686 In this chapter we study in detail several of the functions used in GNU
4687 Emacs. This is called a ``walk-through''. These functions are used as
4688 examples of Lisp code, but are not imaginary examples; with the
4689 exception of the first, simplified function definition, these functions
4690 show the actual code used in GNU Emacs. You can learn a great deal from
4691 these definitions. The functions described here are all related to
4692 buffers. Later, we will study other functions.
4695 * Finding More:: How to find more information.
4696 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4697 @code{point-min}, and @code{push-mark}.
4698 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4699 * append-to-buffer:: Uses @code{save-excursion} and
4700 @code{insert-buffer-substring}.
4701 * Buffer Related Review:: Review.
4702 * Buffer Exercises::
4705 @node Finding More, simplified-beginning-of-buffer, Buffer Walk Through, Buffer Walk Through
4706 @section Finding More Information
4708 @findex describe-function, @r{introduced}
4709 @cindex Find function documentation
4710 In this walk-through, I will describe each new function as we come to
4711 it, sometimes in detail and sometimes briefly. If you are interested,
4712 you can get the full documentation of any Emacs Lisp function at any
4713 time by typing @kbd{C-h f} and then the name of the function (and then
4714 @key{RET}). Similarly, you can get the full documentation for a
4715 variable by typing @kbd{C-h v} and then the name of the variable (and
4718 @cindex Find source of function
4719 @c In version 22, tells location both of C and of Emacs Lisp
4720 Also, @code{describe-function} will tell you the location of the
4721 function definition.
4723 Put point into the name of the file that contains the function and
4724 press the @key{RET} key. In this case, @key{RET} means
4725 @code{push-button} rather than `return' or `enter'. Emacs will take
4726 you directly to the function definition.
4731 If you move point over the file name and press
4732 the @key{RET} key, which in this case means @code{help-follow} rather
4733 than `return' or `enter', Emacs will take you directly to the function
4737 More generally, if you want to see a function in its original source
4738 file, you can use the @code{find-tag} function to jump to it.
4739 @code{find-tag} works with a wide variety of languages, not just
4740 Lisp, and C, and it works with non-programming text as well. For
4741 example, @code{find-tag} will jump to the various nodes in the
4742 Texinfo source file of this document.
4743 The @code{find-tag} function depends on `tags tables' that record
4744 the locations of the functions, variables, and other items to which
4745 @code{find-tag} jumps.
4747 To use the @code{find-tag} command, type @kbd{M-.} (i.e., press the
4748 period key while holding down the @key{META} key, or else type the
4749 @key{ESC} key and then type the period key), and then, at the prompt,
4750 type in the name of the function whose source code you want to see,
4751 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4752 switch buffers and display the source code for the function on your
4753 screen. To switch back to your current buffer, type @kbd{C-x b
4754 @key{RET}}. (On some keyboards, the @key{META} key is labelled
4757 @c !!! 22.1.1 tags table location in this paragraph
4758 @cindex TAGS table, specifying
4760 Depending on how the initial default values of your copy of Emacs are
4761 set, you may also need to specify the location of your `tags table',
4762 which is a file called @file{TAGS}. For example, if you are
4763 interested in Emacs sources, the tags table you will most likely want,
4764 if it has already been created for you, will be in a subdirectory of
4765 the @file{/usr/local/share/emacs/} directory; thus you would use the
4766 @code{M-x visit-tags-table} command and specify a pathname such as
4767 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4768 has not already been created, you will have to create it yourself. It
4769 will be in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4772 To create a @file{TAGS} file in a specific directory, switch to that
4773 directory in Emacs using @kbd{M-x cd} command, or list the directory
4774 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4775 @w{@code{etags *.el}} as the command to execute:
4778 M-x compile RET etags *.el RET
4781 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4783 After you become more familiar with Emacs Lisp, you will find that you will
4784 frequently use @code{find-tag} to navigate your way around source code;
4785 and you will create your own @file{TAGS} tables.
4787 @cindex Library, as term for `file'
4788 Incidentally, the files that contain Lisp code are conventionally
4789 called @dfn{libraries}. The metaphor is derived from that of a
4790 specialized library, such as a law library or an engineering library,
4791 rather than a general library. Each library, or file, contains
4792 functions that relate to a particular topic or activity, such as
4793 @file{abbrev.el} for handling abbreviations and other typing
4794 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4795 libraries provide code for a single activity, as the various
4796 @file{rmail@dots{}} files provide code for reading electronic mail.)
4797 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4798 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4799 by topic keywords.''
4801 @node simplified-beginning-of-buffer, mark-whole-buffer, Finding More, Buffer Walk Through
4802 @comment node-name, next, previous, up
4803 @section A Simplified @code{beginning-of-buffer} Definition
4804 @findex simplified-beginning-of-buffer
4806 The @code{beginning-of-buffer} command is a good function to start with
4807 since you are likely to be familiar with it and it is easy to
4808 understand. Used as an interactive command, @code{beginning-of-buffer}
4809 moves the cursor to the beginning of the buffer, leaving the mark at the
4810 previous position. It is generally bound to @kbd{M-<}.
4812 In this section, we will discuss a shortened version of the function
4813 that shows how it is most frequently used. This shortened function
4814 works as written, but it does not contain the code for a complex option.
4815 In another section, we will describe the entire function.
4816 (@xref{beginning-of-buffer, , Complete Definition of
4817 @code{beginning-of-buffer}}.)
4819 Before looking at the code, let's consider what the function
4820 definition has to contain: it must include an expression that makes
4821 the function interactive so it can be called by typing @kbd{M-x
4822 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4823 must include code to leave a mark at the original position in the
4824 buffer; and it must include code to move the cursor to the beginning
4828 Here is the complete text of the shortened version of the function:
4832 (defun simplified-beginning-of-buffer ()
4833 "Move point to the beginning of the buffer;
4834 leave mark at previous position."
4837 (goto-char (point-min)))
4841 Like all function definitions, this definition has five parts following
4842 the special form @code{defun}:
4846 The name: in this example, @code{simplified-beginning-of-buffer}.
4849 A list of the arguments: in this example, an empty list, @code{()},
4852 The documentation string.
4855 The interactive expression.
4862 In this function definition, the argument list is empty; this means that
4863 this function does not require any arguments. (When we look at the
4864 definition for the complete function, we will see that it may be passed
4865 an optional argument.)
4867 The interactive expression tells Emacs that the function is intended to
4868 be used interactively. In this example, @code{interactive} does not have
4869 an argument because @code{simplified-beginning-of-buffer} does not
4873 The body of the function consists of the two lines:
4878 (goto-char (point-min))
4882 The first of these lines is the expression, @code{(push-mark)}. When
4883 this expression is evaluated by the Lisp interpreter, it sets a mark at
4884 the current position of the cursor, wherever that may be. The position
4885 of this mark is saved in the mark ring.
4887 The next line is @code{(goto-char (point-min))}. This expression
4888 jumps the cursor to the minimum point in the buffer, that is, to the
4889 beginning of the buffer (or to the beginning of the accessible portion
4890 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4891 Narrowing and Widening}.)
4893 The @code{push-mark} command sets a mark at the place where the cursor
4894 was located before it was moved to the beginning of the buffer by the
4895 @code{(goto-char (point-min))} expression. Consequently, you can, if
4896 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4898 That is all there is to the function definition!
4900 @findex describe-function
4901 When you are reading code such as this and come upon an unfamiliar
4902 function, such as @code{goto-char}, you can find out what it does by
4903 using the @code{describe-function} command. To use this command, type
4904 @kbd{C-h f} and then type in the name of the function and press
4905 @key{RET}. The @code{describe-function} command will print the
4906 function's documentation string in a @file{*Help*} window. For
4907 example, the documentation for @code{goto-char} is:
4911 Set point to POSITION, a number or marker.
4912 Beginning of buffer is position (point-min), end is (point-max).
4917 The function's one argument is the desired position.
4920 (The prompt for @code{describe-function} will offer you the symbol
4921 under or preceding the cursor, so you can save typing by positioning
4922 the cursor right over or after the function and then typing @kbd{C-h f
4925 The @code{end-of-buffer} function definition is written in the same way as
4926 the @code{beginning-of-buffer} definition except that the body of the
4927 function contains the expression @code{(goto-char (point-max))} in place
4928 of @code{(goto-char (point-min))}.
4930 @node mark-whole-buffer, append-to-buffer, simplified-beginning-of-buffer, Buffer Walk Through
4931 @comment node-name, next, previous, up
4932 @section The Definition of @code{mark-whole-buffer}
4933 @findex mark-whole-buffer
4935 The @code{mark-whole-buffer} function is no harder to understand than the
4936 @code{simplified-beginning-of-buffer} function. In this case, however,
4937 we will look at the complete function, not a shortened version.
4939 The @code{mark-whole-buffer} function is not as commonly used as the
4940 @code{beginning-of-buffer} function, but is useful nonetheless: it
4941 marks a whole buffer as a region by putting point at the beginning and
4942 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4946 * mark-whole-buffer overview::
4947 * Body of mark-whole-buffer:: Only three lines of code.
4950 @node mark-whole-buffer overview, Body of mark-whole-buffer, mark-whole-buffer, mark-whole-buffer
4952 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4956 In GNU Emacs 22, the code for the complete function looks like this:
4960 (defun mark-whole-buffer ()
4961 "Put point at beginning and mark at end of buffer.
4962 You probably should not use this function in Lisp programs;
4963 it is usually a mistake for a Lisp function to use any subroutine
4964 that uses or sets the mark."
4967 (push-mark (point-max) nil t)
4968 (goto-char (point-min)))
4973 Like all other functions, the @code{mark-whole-buffer} function fits
4974 into the template for a function definition. The template looks like
4979 (defun @var{name-of-function} (@var{argument-list})
4980 "@var{documentation}@dots{}"
4981 (@var{interactive-expression}@dots{})
4986 Here is how the function works: the name of the function is
4987 @code{mark-whole-buffer}; it is followed by an empty argument list,
4988 @samp{()}, which means that the function does not require arguments.
4989 The documentation comes next.
4991 The next line is an @code{(interactive)} expression that tells Emacs
4992 that the function will be used interactively. These details are similar
4993 to the @code{simplified-beginning-of-buffer} function described in the
4997 @node Body of mark-whole-buffer, , mark-whole-buffer overview, mark-whole-buffer
4998 @comment node-name, next, previous, up
4999 @subsection Body of @code{mark-whole-buffer}
5001 The body of the @code{mark-whole-buffer} function consists of three
5008 (push-mark (point-max) nil t)
5009 (goto-char (point-min))
5013 The first of these lines is the expression, @code{(push-mark (point))}.
5015 This line does exactly the same job as the first line of the body of
5016 the @code{simplified-beginning-of-buffer} function, which is written
5017 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
5018 at the current position of the cursor.
5020 I don't know why the expression in @code{mark-whole-buffer} is written
5021 @code{(push-mark (point))} and the expression in
5022 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
5023 whoever wrote the code did not know that the arguments for
5024 @code{push-mark} are optional and that if @code{push-mark} is not
5025 passed an argument, the function automatically sets mark at the
5026 location of point by default. Or perhaps the expression was written
5027 so as to parallel the structure of the next line. In any case, the
5028 line causes Emacs to determine the position of point and set a mark
5031 In earlier versions of GNU Emacs, the next line of
5032 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
5033 expression sets a mark at the point in the buffer that has the highest
5034 number. This will be the end of the buffer (or, if the buffer is
5035 narrowed, the end of the accessible portion of the buffer.
5036 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
5037 narrowing.) After this mark has been set, the previous mark, the one
5038 set at point, is no longer set, but Emacs remembers its position, just
5039 as all other recent marks are always remembered. This means that you
5040 can, if you wish, go back to that position by typing @kbd{C-u
5044 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5048 (push-mark (point-max) nil t)
5052 The expression works nearly the same as before. It sets a mark at the
5053 highest numbered place in the buffer that it can. However, in this
5054 version, @code{push-mark} has two additional arguments. The second
5055 argument to @code{push-mark} is @code{nil}. This tells the function
5056 it @emph{should} display a message that says `Mark set' when it pushes
5057 the mark. The third argument is @code{t}. This tells
5058 @code{push-mark} to activate the mark when Transient Mark mode is
5059 turned on. Transient Mark mode highlights the currently active
5060 region. It is often turned off.
5062 Finally, the last line of the function is @code{(goto-char
5063 (point-min)))}. This is written exactly the same way as it is written
5064 in @code{beginning-of-buffer}. The expression moves the cursor to
5065 the minimum point in the buffer, that is, to the beginning of the buffer
5066 (or to the beginning of the accessible portion of the buffer). As a
5067 result of this, point is placed at the beginning of the buffer and mark
5068 is set at the end of the buffer. The whole buffer is, therefore, the
5071 @node append-to-buffer, Buffer Related Review, mark-whole-buffer, Buffer Walk Through
5072 @comment node-name, next, previous, up
5073 @section The Definition of @code{append-to-buffer}
5074 @findex append-to-buffer
5076 The @code{append-to-buffer} command is more complex than the
5077 @code{mark-whole-buffer} command. What it does is copy the region
5078 (that is, the part of the buffer between point and mark) from the
5079 current buffer to a specified buffer.
5082 * append-to-buffer overview::
5083 * append interactive:: A two part interactive expression.
5084 * append-to-buffer body:: Incorporates a @code{let} expression.
5085 * append save-excursion:: How the @code{save-excursion} works.
5088 @node append-to-buffer overview, append interactive, append-to-buffer, append-to-buffer
5090 @unnumberedsubsec An Overview of @code{append-to-buffer}
5093 @findex insert-buffer-substring
5094 The @code{append-to-buffer} command uses the
5095 @code{insert-buffer-substring} function to copy the region.
5096 @code{insert-buffer-substring} is described by its name: it takes a
5097 string of characters from part of a buffer, a ``substring'', and
5098 inserts them into another buffer.
5100 Most of @code{append-to-buffer} is
5101 concerned with setting up the conditions for
5102 @code{insert-buffer-substring} to work: the code must specify both the
5103 buffer to which the text will go, the window it comes from and goes
5104 to, and the region that will be copied.
5107 Here is the complete text of the function:
5111 (defun append-to-buffer (buffer start end)
5112 "Append to specified buffer the text of the region.
5113 It is inserted into that buffer before its point.
5117 When calling from a program, give three arguments:
5118 BUFFER (or buffer name), START and END.
5119 START and END specify the portion of the current buffer to be copied."
5121 (list (read-buffer "Append to buffer: " (other-buffer
5122 (current-buffer) t))
5123 (region-beginning) (region-end)))
5126 (let ((oldbuf (current-buffer)))
5128 (let* ((append-to (get-buffer-create buffer))
5129 (windows (get-buffer-window-list append-to t t))
5131 (set-buffer append-to)
5132 (setq point (point))
5133 (barf-if-buffer-read-only)
5134 (insert-buffer-substring oldbuf start end)
5135 (dolist (window windows)
5136 (when (= (window-point window) point)
5137 (set-window-point window (point))))))))
5141 The function can be understood by looking at it as a series of
5142 filled-in templates.
5144 The outermost template is for the function definition. In this
5145 function, it looks like this (with several slots filled in):
5149 (defun append-to-buffer (buffer start end)
5150 "@var{documentation}@dots{}"
5151 (interactive @dots{})
5156 The first line of the function includes its name and three arguments.
5157 The arguments are the @code{buffer} to which the text will be copied, and
5158 the @code{start} and @code{end} of the region in the current buffer that
5161 The next part of the function is the documentation, which is clear and
5162 complete. As is conventional, the three arguments are written in
5163 upper case so you will notice them easily. Even better, they are
5164 described in the same order as in the argument list.
5166 Note that the documentation distinguishes between a buffer and its
5167 name. (The function can handle either.)
5169 @node append interactive, append-to-buffer body, append-to-buffer overview, append-to-buffer
5170 @comment node-name, next, previous, up
5171 @subsection The @code{append-to-buffer} Interactive Expression
5173 Since the @code{append-to-buffer} function will be used interactively,
5174 the function must have an @code{interactive} expression. (For a
5175 review of @code{interactive}, see @ref{Interactive, , Making a
5176 Function Interactive}.) The expression reads as follows:
5182 "Append to buffer: "
5183 (other-buffer (current-buffer) t))
5190 This expression is not one with letters standing for parts, as
5191 described earlier. Instead, it starts a list with these parts:
5193 The first part of the list is an expression to read the name of a
5194 buffer and return it as a string. That is @code{read-buffer}. The
5195 function requires a prompt as its first argument, @samp{"Append to
5196 buffer: "}. Its second argument tells the command what value to
5197 provide if you don't specify anything.
5199 In this case that second argument is an expression containing the
5200 function @code{other-buffer}, an exception, and a @samp{t}, standing
5203 The first argument to @code{other-buffer}, the exception, is yet
5204 another function, @code{current-buffer}. That is not going to be
5205 returned. The second argument is the symbol for true, @code{t}. that
5206 tells @code{other-buffer} that it may show visible buffers (except in
5207 this case, it will not show the current buffer, which makes sense).
5210 The expression looks like this:
5213 (other-buffer (current-buffer) t)
5216 The second and third arguments to the @code{list} expression are
5217 @code{(region-beginning)} and @code{(region-end)}. These two
5218 functions specify the beginning and end of the text to be appended.
5221 Originally, the command used the letters @samp{B} and @samp{r}.
5222 The whole @code{interactive} expression looked like this:
5225 (interactive "BAppend to buffer:@: \nr")
5229 But when that was done, the default value of the buffer switched to
5230 was invisible. That was not wanted.
5232 (The prompt was separated from the second argument with a newline,
5233 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5234 two arguments that follow the symbol @code{buffer} in the function's
5235 argument list (that is, @code{start} and @code{end}) to the values of
5236 point and mark. That argument worked fine.)
5238 @node append-to-buffer body, append save-excursion, append interactive, append-to-buffer
5239 @comment node-name, next, previous, up
5240 @subsection The Body of @code{append-to-buffer}
5243 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5245 (defun append-to-buffer (buffer start end)
5246 "Append to specified buffer the text of the region.
5247 It is inserted into that buffer before its point.
5249 When calling from a program, give three arguments:
5250 BUFFER (or buffer name), START and END.
5251 START and END specify the portion of the current buffer to be copied."
5253 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5254 (region-beginning) (region-end)))
5255 (let ((oldbuf (current-buffer)))
5257 (let* ((append-to (get-buffer-create buffer))
5258 (windows (get-buffer-window-list append-to t t))
5260 (set-buffer append-to)
5261 (setq point (point))
5262 (barf-if-buffer-read-only)
5263 (insert-buffer-substring oldbuf start end)
5264 (dolist (window windows)
5265 (when (= (window-point window) point)
5266 (set-window-point window (point))))))))
5269 The body of the @code{append-to-buffer} function begins with @code{let}.
5271 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5272 @code{let} expression is to create and give initial values to one or
5273 more variables that will only be used within the body of the
5274 @code{let}. This means that such a variable will not be confused with
5275 any variable of the same name outside the @code{let} expression.
5277 We can see how the @code{let} expression fits into the function as a
5278 whole by showing a template for @code{append-to-buffer} with the
5279 @code{let} expression in outline:
5283 (defun append-to-buffer (buffer start end)
5284 "@var{documentation}@dots{}"
5285 (interactive @dots{})
5286 (let ((@var{variable} @var{value}))
5291 The @code{let} expression has three elements:
5295 The symbol @code{let};
5298 A varlist containing, in this case, a single two-element list,
5299 @code{(@var{variable} @var{value})};
5302 The body of the @code{let} expression.
5306 In the @code{append-to-buffer} function, the varlist looks like this:
5309 (oldbuf (current-buffer))
5313 In this part of the @code{let} expression, the one variable,
5314 @code{oldbuf}, is bound to the value returned by the
5315 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5316 used to keep track of the buffer in which you are working and from
5317 which you will copy.
5319 The element or elements of a varlist are surrounded by a set of
5320 parentheses so the Lisp interpreter can distinguish the varlist from
5321 the body of the @code{let}. As a consequence, the two-element list
5322 within the varlist is surrounded by a circumscribing set of parentheses.
5323 The line looks like this:
5327 (let ((oldbuf (current-buffer)))
5333 The two parentheses before @code{oldbuf} might surprise you if you did
5334 not realize that the first parenthesis before @code{oldbuf} marks the
5335 boundary of the varlist and the second parenthesis marks the beginning
5336 of the two-element list, @code{(oldbuf (current-buffer))}.
5338 @node append save-excursion, , append-to-buffer body, append-to-buffer
5339 @comment node-name, next, previous, up
5340 @subsection @code{save-excursion} in @code{append-to-buffer}
5342 The body of the @code{let} expression in @code{append-to-buffer}
5343 consists of a @code{save-excursion} expression.
5345 The @code{save-excursion} function saves the locations of point and
5346 mark, and restores them to those positions after the expressions in the
5347 body of the @code{save-excursion} complete execution. In addition,
5348 @code{save-excursion} keeps track of the original buffer, and
5349 restores it. This is how @code{save-excursion} is used in
5350 @code{append-to-buffer}.
5353 @cindex Indentation for formatting
5354 @cindex Formatting convention
5355 Incidentally, it is worth noting here that a Lisp function is normally
5356 formatted so that everything that is enclosed in a multi-line spread is
5357 indented more to the right than the first symbol. In this function
5358 definition, the @code{let} is indented more than the @code{defun}, and
5359 the @code{save-excursion} is indented more than the @code{let}, like
5375 This formatting convention makes it easy to see that the lines in
5376 the body of the @code{save-excursion} are enclosed by the parentheses
5377 associated with @code{save-excursion}, just as the
5378 @code{save-excursion} itself is enclosed by the parentheses associated
5379 with the @code{let}:
5383 (let ((oldbuf (current-buffer)))
5386 (set-buffer @dots{})
5387 (insert-buffer-substring oldbuf start end)
5393 The use of the @code{save-excursion} function can be viewed as a process
5394 of filling in the slots of a template:
5399 @var{first-expression-in-body}
5400 @var{second-expression-in-body}
5402 @var{last-expression-in-body})
5408 In this function, the body of the @code{save-excursion} contains only
5409 one expression, the @code{let*} expression. You know about a
5410 @code{let} function. The @code{let*} function is different. It has a
5411 @samp{*} in its name. It enables Emacs to set each variable in its
5412 varlist in sequence, one after another.
5414 Its critical feature is that variables later in the varlist can make
5415 use of the values to which Emacs set variables earlier in the varlist.
5416 @xref{fwd-para let, , The @code{let*} expression}.
5418 We will skip functions like @code{let*} and focus on two: the
5419 @code{set-buffer} function and the @code{insert-buffer-substring}
5423 In the old days, the @code{set-buffer} expression was simply
5426 (set-buffer (get-buffer-create buffer))
5434 (set-buffer append-to)
5438 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5439 on in the @code{let*} expression. That extra binding would not be
5440 necessary except for that @code{append-to} is used later in the
5441 varlist as an argument to @code{get-buffer-window-list}.
5446 (let ((oldbuf (current-buffer)))
5448 (let* ((append-to (get-buffer-create buffer))
5449 (windows (get-buffer-window-list append-to t t))
5451 (set-buffer append-to)
5452 (setq point (point))
5453 (barf-if-buffer-read-only)
5454 (insert-buffer-substring oldbuf start end)
5455 (dolist (window windows)
5456 (when (= (window-point window) point)
5457 (set-window-point window (point))))))))
5460 The @code{append-to-buffer} function definition inserts text from the
5461 buffer in which you are currently to a named buffer. It happens that
5462 @code{insert-buffer-substring} copies text from another buffer to the
5463 current buffer, just the reverse---that is why the
5464 @code{append-to-buffer} definition starts out with a @code{let} that
5465 binds the local symbol @code{oldbuf} to the value returned by
5466 @code{current-buffer}.
5469 The @code{insert-buffer-substring} expression looks like this:
5472 (insert-buffer-substring oldbuf start end)
5476 The @code{insert-buffer-substring} function copies a string
5477 @emph{from} the buffer specified as its first argument and inserts the
5478 string into the present buffer. In this case, the argument to
5479 @code{insert-buffer-substring} is the value of the variable created
5480 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5481 was the current buffer when you gave the @code{append-to-buffer}
5484 After @code{insert-buffer-substring} has done its work,
5485 @code{save-excursion} will restore the action to the original buffer
5486 and @code{append-to-buffer} will have done its job.
5489 Written in skeletal form, the workings of the body look like this:
5493 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5494 (save-excursion ; @r{Keep track of buffer.}
5496 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5498 @var{change-back-to-original-buffer-when-finished}
5499 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5503 In summary, @code{append-to-buffer} works as follows: it saves the
5504 value of the current buffer in the variable called @code{oldbuf}. It
5505 gets the new buffer (creating one if need be) and switches Emacs'
5506 attention to it. Using the value of @code{oldbuf}, it inserts the
5507 region of text from the old buffer into the new buffer; and then using
5508 @code{save-excursion}, it brings you back to your original buffer.
5510 In looking at @code{append-to-buffer}, you have explored a fairly
5511 complex function. It shows how to use @code{let} and
5512 @code{save-excursion}, and how to change to and come back from another
5513 buffer. Many function definitions use @code{let},
5514 @code{save-excursion}, and @code{set-buffer} this way.
5516 @node Buffer Related Review, Buffer Exercises, append-to-buffer, Buffer Walk Through
5517 @comment node-name, next, previous, up
5520 Here is a brief summary of the various functions discussed in this chapter.
5523 @item describe-function
5524 @itemx describe-variable
5525 Print the documentation for a function or variable.
5526 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5529 Find the file containing the source for a function or variable and
5530 switch buffers to it, positioning point at the beginning of the item.
5531 Conventionally bound to @kbd{M-.} (that's a period following the
5534 @item save-excursion
5535 Save the location of point and mark and restore their values after the
5536 arguments to @code{save-excursion} have been evaluated. Also, remember
5537 the current buffer and return to it.
5540 Set mark at a location and record the value of the previous mark on the
5541 mark ring. The mark is a location in the buffer that will keep its
5542 relative position even if text is added to or removed from the buffer.
5545 Set point to the location specified by the value of the argument, which
5546 can be a number, a marker, or an expression that returns the number of
5547 a position, such as @code{(point-min)}.
5549 @item insert-buffer-substring
5550 Copy a region of text from a buffer that is passed to the function as
5551 an argument and insert the region into the current buffer.
5553 @item mark-whole-buffer
5554 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5557 Switch the attention of Emacs to another buffer, but do not change the
5558 window being displayed. Used when the program rather than a human is
5559 to work on a different buffer.
5561 @item get-buffer-create
5563 Find a named buffer or create one if a buffer of that name does not
5564 exist. The @code{get-buffer} function returns @code{nil} if the named
5565 buffer does not exist.
5569 @node Buffer Exercises, , Buffer Related Review, Buffer Walk Through
5574 Write your own @code{simplified-end-of-buffer} function definition;
5575 then test it to see whether it works.
5578 Use @code{if} and @code{get-buffer} to write a function that prints a
5579 message telling you whether a buffer exists.
5582 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5586 @node More Complex, Narrowing & Widening, Buffer Walk Through, Top
5587 @comment node-name, next, previous, up
5588 @chapter A Few More Complex Functions
5590 In this chapter, we build on what we have learned in previous chapters
5591 by looking at more complex functions. The @code{copy-to-buffer}
5592 function illustrates use of two @code{save-excursion} expressions in
5593 one definition, while the @code{insert-buffer} function illustrates
5594 use of an asterisk in an @code{interactive} expression, use of
5595 @code{or}, and the important distinction between a name and the object
5596 to which the name refers.
5599 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5600 * insert-buffer:: Read-only, and with @code{or}.
5601 * beginning-of-buffer:: Shows @code{goto-char},
5602 @code{point-min}, and @code{push-mark}.
5603 * Second Buffer Related Review::
5604 * optional Exercise::
5607 @node copy-to-buffer, insert-buffer, More Complex, More Complex
5608 @comment node-name, next, previous, up
5609 @section The Definition of @code{copy-to-buffer}
5610 @findex copy-to-buffer
5612 After understanding how @code{append-to-buffer} works, it is easy to
5613 understand @code{copy-to-buffer}. This function copies text into a
5614 buffer, but instead of adding to the second buffer, it replaces all the
5615 previous text in the second buffer.
5618 The body of @code{copy-to-buffer} looks like this,
5623 (interactive "BCopy to buffer: \nr")
5624 (let ((oldbuf (current-buffer)))
5625 (with-current-buffer (get-buffer-create buffer)
5626 (barf-if-buffer-read-only)
5629 (insert-buffer-substring oldbuf start end)))))
5633 The @code{copy-to-buffer} function has a simpler @code{interactive}
5634 expression than @code{append-to-buffer}.
5637 The definition then says
5640 (with-current-buffer (get-buffer-create buffer) @dots{}
5643 First, look at the earliest inner expression; that is evaluated first.
5644 That expression starts with @code{get-buffer-create buffer}. The
5645 function tells the computer to use the buffer with the name specified
5646 as the one to which you are copying, or if such a buffer does not
5647 exist, to create it. Then, the @code{with-current-buffer} function
5648 evaluates its body with that buffer temporarily current.
5650 (This demonstrates another way to shift the computer's attention but
5651 not the user's. The @code{append-to-buffer} function showed how to do
5652 the same with @code{save-excursion} and @code{set-buffer}.
5653 @code{with-current-buffer} is a newer, and arguably easier,
5656 The @code{barf-if-buffer-read-only} function sends you an error
5657 message saying the buffer is read-only if you cannot modify it.
5659 The next line has the @code{erase-buffer} function as its sole
5660 contents. That function erases the buffer.
5662 Finally, the last two lines contain the @code{save-excursion}
5663 expression with @code{insert-buffer-substring} as its body.
5664 The @code{insert-buffer-substring} expression copies the text from
5665 the buffer you are in (and you have not seen the computer shift its
5666 attention, so you don't know that that buffer is now called
5669 Incidentally, this is what is meant by `replacement'. To replace text,
5670 Emacs erases the previous text and then inserts new text.
5673 In outline, the body of @code{copy-to-buffer} looks like this:
5677 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5678 (@var{with-the-buffer-you-are-copying-to}
5679 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5682 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5686 @node insert-buffer, beginning-of-buffer, copy-to-buffer, More Complex
5687 @comment node-name, next, previous, up
5688 @section The Definition of @code{insert-buffer}
5689 @findex insert-buffer
5691 @code{insert-buffer} is yet another buffer-related function. This
5692 command copies another buffer @emph{into} the current buffer. It is the
5693 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5694 copy a region of text @emph{from} the current buffer to another buffer.
5696 Here is a discussion based on the original code. The code was
5697 simplified in 2003 and is harder to understand.
5699 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5700 a discussion of the new body.)
5702 In addition, this code illustrates the use of @code{interactive} with a
5703 buffer that might be @dfn{read-only} and the important distinction
5704 between the name of an object and the object actually referred to.
5707 * insert-buffer code::
5708 * insert-buffer interactive:: When you can read, but not write.
5709 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5710 * if & or:: Using an @code{if} instead of an @code{or}.
5711 * Insert or:: How the @code{or} expression works.
5712 * Insert let:: Two @code{save-excursion} expressions.
5713 * New insert-buffer::
5716 @node insert-buffer code, insert-buffer interactive, insert-buffer, insert-buffer
5718 @unnumberedsubsec The Code for @code{insert-buffer}
5722 Here is the earlier code:
5726 (defun insert-buffer (buffer)
5727 "Insert after point the contents of BUFFER.
5728 Puts mark after the inserted text.
5729 BUFFER may be a buffer or a buffer name."
5730 (interactive "*bInsert buffer:@: ")
5733 (or (bufferp buffer)
5734 (setq buffer (get-buffer buffer)))
5735 (let (start end newmark)
5739 (setq start (point-min) end (point-max)))
5742 (insert-buffer-substring buffer start end)
5743 (setq newmark (point)))
5744 (push-mark newmark)))
5749 As with other function definitions, you can use a template to see an
5750 outline of the function:
5754 (defun insert-buffer (buffer)
5755 "@var{documentation}@dots{}"
5756 (interactive "*bInsert buffer:@: ")
5761 @node insert-buffer interactive, insert-buffer body, insert-buffer code, insert-buffer
5762 @comment node-name, next, previous, up
5763 @subsection The Interactive Expression in @code{insert-buffer}
5764 @findex interactive, @r{example use of}
5766 In @code{insert-buffer}, the argument to the @code{interactive}
5767 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5771 * Read-only buffer:: When a buffer cannot be modified.
5772 * b for interactive:: An existing buffer or else its name.
5775 @node Read-only buffer, b for interactive, insert-buffer interactive, insert-buffer interactive
5776 @comment node-name, next, previous, up
5777 @unnumberedsubsubsec A Read-only Buffer
5778 @cindex Read-only buffer
5779 @cindex Asterisk for read-only buffer
5780 @findex * @r{for read-only buffer}
5782 The asterisk is for the situation when the current buffer is a
5783 read-only buffer---a buffer that cannot be modified. If
5784 @code{insert-buffer} is called when the current buffer is read-only, a
5785 message to this effect is printed in the echo area and the terminal
5786 may beep or blink at you; you will not be permitted to insert anything
5787 into current buffer. The asterisk does not need to be followed by a
5788 newline to separate it from the next argument.
5790 @node b for interactive, , Read-only buffer, insert-buffer interactive
5791 @comment node-name, next, previous, up
5792 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5794 The next argument in the interactive expression starts with a lower
5795 case @samp{b}. (This is different from the code for
5796 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5797 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5798 The lower-case @samp{b} tells the Lisp interpreter that the argument
5799 for @code{insert-buffer} should be an existing buffer or else its
5800 name. (The upper-case @samp{B} option provides for the possibility
5801 that the buffer does not exist.) Emacs will prompt you for the name
5802 of the buffer, offering you a default buffer, with name completion
5803 enabled. If the buffer does not exist, you receive a message that
5804 says ``No match''; your terminal may beep at you as well.
5806 The new and simplified code generates a list for @code{interactive}.
5807 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5808 functions with which we are already familiar and the @code{progn}
5809 special form with which we are not. (It will be described later.)
5811 @node insert-buffer body, if & or, insert-buffer interactive, insert-buffer
5812 @comment node-name, next, previous, up
5813 @subsection The Body of the @code{insert-buffer} Function
5815 The body of the @code{insert-buffer} function has two major parts: an
5816 @code{or} expression and a @code{let} expression. The purpose of the
5817 @code{or} expression is to ensure that the argument @code{buffer} is
5818 bound to a buffer and not just the name of a buffer. The body of the
5819 @code{let} expression contains the code which copies the other buffer
5820 into the current buffer.
5823 In outline, the two expressions fit into the @code{insert-buffer}
5828 (defun insert-buffer (buffer)
5829 "@var{documentation}@dots{}"
5830 (interactive "*bInsert buffer:@: ")
5835 (let (@var{varlist})
5836 @var{body-of-}@code{let}@dots{} )
5840 To understand how the @code{or} expression ensures that the argument
5841 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5842 is first necessary to understand the @code{or} function.
5844 Before doing this, let me rewrite this part of the function using
5845 @code{if} so that you can see what is done in a manner that will be familiar.
5847 @node if & or, Insert or, insert-buffer body, insert-buffer
5848 @comment node-name, next, previous, up
5849 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5851 The job to be done is to make sure the value of @code{buffer} is a
5852 buffer itself and not the name of a buffer. If the value is the name,
5853 then the buffer itself must be got.
5855 You can imagine yourself at a conference where an usher is wandering
5856 around holding a list with your name on it and looking for you: the
5857 usher is ``bound'' to your name, not to you; but when the usher finds
5858 you and takes your arm, the usher becomes ``bound'' to you.
5861 In Lisp, you might describe this situation like this:
5865 (if (not (holding-on-to-guest))
5866 (find-and-take-arm-of-guest))
5870 We want to do the same thing with a buffer---if we do not have the
5871 buffer itself, we want to get it.
5874 Using a predicate called @code{bufferp} that tells us whether we have a
5875 buffer (rather than its name), we can write the code like this:
5879 (if (not (bufferp buffer)) ; @r{if-part}
5880 (setq buffer (get-buffer buffer))) ; @r{then-part}
5885 Here, the true-or-false-test of the @code{if} expression is
5886 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5887 @w{@code{(setq buffer (get-buffer buffer))}}.
5889 In the test, the function @code{bufferp} returns true if its argument is
5890 a buffer---but false if its argument is the name of the buffer. (The
5891 last character of the function name @code{bufferp} is the character
5892 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5893 indicates that the function is a predicate, which is a term that means
5894 that the function will determine whether some property is true or false.
5895 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5899 The function @code{not} precedes the expression @code{(bufferp buffer)},
5900 so the true-or-false-test looks like this:
5903 (not (bufferp buffer))
5907 @code{not} is a function that returns true if its argument is false
5908 and false if its argument is true. So if @code{(bufferp buffer)}
5909 returns true, the @code{not} expression returns false and vice-verse:
5910 what is ``not true'' is false and what is ``not false'' is true.
5912 Using this test, the @code{if} expression works as follows: when the
5913 value of the variable @code{buffer} is actually a buffer rather than
5914 its name, the true-or-false-test returns false and the @code{if}
5915 expression does not evaluate the then-part. This is fine, since we do
5916 not need to do anything to the variable @code{buffer} if it really is
5919 On the other hand, when the value of @code{buffer} is not a buffer
5920 itself, but the name of a buffer, the true-or-false-test returns true
5921 and the then-part of the expression is evaluated. In this case, the
5922 then-part is @code{(setq buffer (get-buffer buffer))}. This
5923 expression uses the @code{get-buffer} function to return an actual
5924 buffer itself, given its name. The @code{setq} then sets the variable
5925 @code{buffer} to the value of the buffer itself, replacing its previous
5926 value (which was the name of the buffer).
5928 @node Insert or, Insert let, if & or, insert-buffer
5929 @comment node-name, next, previous, up
5930 @subsection The @code{or} in the Body
5932 The purpose of the @code{or} expression in the @code{insert-buffer}
5933 function is to ensure that the argument @code{buffer} is bound to a
5934 buffer and not just to the name of a buffer. The previous section shows
5935 how the job could have been done using an @code{if} expression.
5936 However, the @code{insert-buffer} function actually uses @code{or}.
5937 To understand this, it is necessary to understand how @code{or} works.
5940 An @code{or} function can have any number of arguments. It evaluates
5941 each argument in turn and returns the value of the first of its
5942 arguments that is not @code{nil}. Also, and this is a crucial feature
5943 of @code{or}, it does not evaluate any subsequent arguments after
5944 returning the first non-@code{nil} value.
5947 The @code{or} expression looks like this:
5951 (or (bufferp buffer)
5952 (setq buffer (get-buffer buffer)))
5957 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5958 This expression returns true (a non-@code{nil} value) if the buffer is
5959 actually a buffer, and not just the name of a buffer. In the @code{or}
5960 expression, if this is the case, the @code{or} expression returns this
5961 true value and does not evaluate the next expression---and this is fine
5962 with us, since we do not want to do anything to the value of
5963 @code{buffer} if it really is a buffer.
5965 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5966 which it will be if the value of @code{buffer} is the name of a buffer,
5967 the Lisp interpreter evaluates the next element of the @code{or}
5968 expression. This is the expression @code{(setq buffer (get-buffer
5969 buffer))}. This expression returns a non-@code{nil} value, which
5970 is the value to which it sets the variable @code{buffer}---and this
5971 value is a buffer itself, not the name of a buffer.
5973 The result of all this is that the symbol @code{buffer} is always
5974 bound to a buffer itself rather than to the name of a buffer. All
5975 this is necessary because the @code{set-buffer} function in a
5976 following line only works with a buffer itself, not with the name to a
5980 Incidentally, using @code{or}, the situation with the usher would be
5984 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5987 @node Insert let, New insert-buffer, Insert or, insert-buffer
5988 @comment node-name, next, previous, up
5989 @subsection The @code{let} Expression in @code{insert-buffer}
5991 After ensuring that the variable @code{buffer} refers to a buffer itself
5992 and not just to the name of a buffer, the @code{insert-buffer function}
5993 continues with a @code{let} expression. This specifies three local
5994 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5995 to the initial value @code{nil}. These variables are used inside the
5996 remainder of the @code{let} and temporarily hide any other occurrence of
5997 variables of the same name in Emacs until the end of the @code{let}.
6000 The body of the @code{let} contains two @code{save-excursion}
6001 expressions. First, we will look at the inner @code{save-excursion}
6002 expression in detail. The expression looks like this:
6008 (setq start (point-min) end (point-max)))
6013 The expression @code{(set-buffer buffer)} changes Emacs' attention
6014 from the current buffer to the one from which the text will copied.
6015 In that buffer, the variables @code{start} and @code{end} are set to
6016 the beginning and end of the buffer, using the commands
6017 @code{point-min} and @code{point-max}. Note that we have here an
6018 illustration of how @code{setq} is able to set two variables in the
6019 same expression. The first argument of @code{setq} is set to the
6020 value of its second, and its third argument is set to the value of its
6023 After the body of the inner @code{save-excursion} is evaluated, the
6024 @code{save-excursion} restores the original buffer, but @code{start} and
6025 @code{end} remain set to the values of the beginning and end of the
6026 buffer from which the text will be copied.
6029 The outer @code{save-excursion} expression looks like this:
6034 (@var{inner-}@code{save-excursion}@var{-expression}
6035 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
6036 (insert-buffer-substring buffer start end)
6037 (setq newmark (point)))
6042 The @code{insert-buffer-substring} function copies the text
6043 @emph{into} the current buffer @emph{from} the region indicated by
6044 @code{start} and @code{end} in @code{buffer}. Since the whole of the
6045 second buffer lies between @code{start} and @code{end}, the whole of
6046 the second buffer is copied into the buffer you are editing. Next,
6047 the value of point, which will be at the end of the inserted text, is
6048 recorded in the variable @code{newmark}.
6050 After the body of the outer @code{save-excursion} is evaluated, point
6051 and mark are relocated to their original places.
6053 However, it is convenient to locate a mark at the end of the newly
6054 inserted text and locate point at its beginning. The @code{newmark}
6055 variable records the end of the inserted text. In the last line of
6056 the @code{let} expression, the @code{(push-mark newmark)} expression
6057 function sets a mark to this location. (The previous location of the
6058 mark is still accessible; it is recorded on the mark ring and you can
6059 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6060 located at the beginning of the inserted text, which is where it was
6061 before you called the insert function, the position of which was saved
6062 by the first @code{save-excursion}.
6065 The whole @code{let} expression looks like this:
6069 (let (start end newmark)
6073 (setq start (point-min) end (point-max)))
6074 (insert-buffer-substring buffer start end)
6075 (setq newmark (point)))
6076 (push-mark newmark))
6080 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6081 function uses @code{let}, @code{save-excursion}, and
6082 @code{set-buffer}. In addition, the function illustrates one way to
6083 use @code{or}. All these functions are building blocks that we will
6084 find and use again and again.
6086 @node New insert-buffer, , Insert let, insert-buffer
6087 @comment node-name, next, previous, up
6088 @subsection New Body for @code{insert-buffer}
6089 @findex insert-buffer, new version body
6090 @findex new version body for insert-buffer
6092 The body in the GNU Emacs 22 version is more confusing than the original.
6095 It consists of two expressions,
6101 (insert-buffer-substring (get-buffer buffer))
6109 except, and this is what confuses novices, very important work is done
6110 inside the @code{push-mark} expression.
6112 The @code{get-buffer} function returns a buffer with the name
6113 provided. You will note that the function is @emph{not} called
6114 @code{get-buffer-create}; it does not create a buffer if one does not
6115 already exist. The buffer returned by @code{get-buffer}, an existing
6116 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6117 whole of the buffer (since you did not specify anything else).
6119 The location into which the buffer is inserted is recorded by
6120 @code{push-mark}. Then the function returns @code{nil}, the value of
6121 its last command. Put another way, the @code{insert-buffer} function
6122 exists only to produce a side effect, inserting another buffer, not to
6125 @node beginning-of-buffer, Second Buffer Related Review, insert-buffer, More Complex
6126 @comment node-name, next, previous, up
6127 @section Complete Definition of @code{beginning-of-buffer}
6128 @findex beginning-of-buffer
6130 The basic structure of the @code{beginning-of-buffer} function has
6131 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6132 Simplified @code{beginning-of-buffer} Definition}.)
6133 This section describes the complex part of the definition.
6135 As previously described, when invoked without an argument,
6136 @code{beginning-of-buffer} moves the cursor to the beginning of the
6137 buffer (in truth, the beginning of the accessible portion of the
6138 buffer), leaving the mark at the previous position. However, when the
6139 command is invoked with a number between one and ten, the function
6140 considers that number to be a fraction of the length of the buffer,
6141 measured in tenths, and Emacs moves the cursor that fraction of the
6142 way from the beginning of the buffer. Thus, you can either call this
6143 function with the key command @kbd{M-<}, which will move the cursor to
6144 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6145 M-<} which will move the cursor to a point 70% of the way through the
6146 buffer. If a number bigger than ten is used for the argument, it
6147 moves to the end of the buffer.
6149 The @code{beginning-of-buffer} function can be called with or without an
6150 argument. The use of the argument is optional.
6153 * Optional Arguments::
6154 * beginning-of-buffer opt arg:: Example with optional argument.
6155 * beginning-of-buffer complete::
6158 @node Optional Arguments, beginning-of-buffer opt arg, beginning-of-buffer, beginning-of-buffer
6159 @subsection Optional Arguments
6161 Unless told otherwise, Lisp expects that a function with an argument in
6162 its function definition will be called with a value for that argument.
6163 If that does not happen, you get an error and a message that says
6164 @samp{Wrong number of arguments}.
6166 @cindex Optional arguments
6169 However, optional arguments are a feature of Lisp: a particular
6170 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6171 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6172 @samp{optional} is part of the keyword.) In a function definition, if
6173 an argument follows the keyword @code{&optional}, no value need be
6174 passed to that argument when the function is called.
6177 The first line of the function definition of @code{beginning-of-buffer}
6178 therefore looks like this:
6181 (defun beginning-of-buffer (&optional arg)
6185 In outline, the whole function looks like this:
6189 (defun beginning-of-buffer (&optional arg)
6190 "@var{documentation}@dots{}"
6192 (or (@var{is-the-argument-a-cons-cell} arg)
6193 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6195 (let (@var{determine-size-and-set-it})
6197 (@var{if-there-is-an-argument}
6198 @var{figure-out-where-to-go}
6205 The function is similar to the @code{simplified-beginning-of-buffer}
6206 function except that the @code{interactive} expression has @code{"P"}
6207 as an argument and the @code{goto-char} function is followed by an
6208 if-then-else expression that figures out where to put the cursor if
6209 there is an argument that is not a cons cell.
6211 (Since I do not explain a cons cell for many more chapters, please
6212 consider ignoring the function @code{consp}. @xref{List
6213 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6214 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6217 The @code{"P"} in the @code{interactive} expression tells Emacs to
6218 pass a prefix argument, if there is one, to the function in raw form.
6219 A prefix argument is made by typing the @key{META} key followed by a
6220 number, or by typing @kbd{C-u} and then a number. (If you don't type
6221 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6222 @code{"p"} in the @code{interactive} expression causes the function to
6223 convert a prefix arg to a number.)
6225 The true-or-false-test of the @code{if} expression looks complex, but
6226 it is not: it checks whether @code{arg} has a value that is not
6227 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6228 does; it checks whether its argument is a cons cell.) If @code{arg}
6229 has a value that is not @code{nil} (and is not a cons cell), which
6230 will be the case if @code{beginning-of-buffer} is called with a
6231 numeric argument, then this true-or-false-test will return true and
6232 the then-part of the @code{if} expression will be evaluated. On the
6233 other hand, if @code{beginning-of-buffer} is not called with an
6234 argument, the value of @code{arg} will be @code{nil} and the else-part
6235 of the @code{if} expression will be evaluated. The else-part is
6236 simply @code{point-min}, and when this is the outcome, the whole
6237 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6238 is how we saw the @code{beginning-of-buffer} function in its
6241 @node beginning-of-buffer opt arg, beginning-of-buffer complete, Optional Arguments, beginning-of-buffer
6242 @subsection @code{beginning-of-buffer} with an Argument
6244 When @code{beginning-of-buffer} is called with an argument, an
6245 expression is evaluated which calculates what value to pass to
6246 @code{goto-char}. This expression is rather complicated at first sight.
6247 It includes an inner @code{if} expression and much arithmetic. It looks
6252 (if (> (buffer-size) 10000)
6253 ;; @r{Avoid overflow for large buffer sizes!}
6254 (* (prefix-numeric-value arg)
6259 size (prefix-numeric-value arg))) 10)))
6264 * Disentangle beginning-of-buffer::
6265 * Large buffer case::
6266 * Small buffer case::
6269 @node Disentangle beginning-of-buffer, Large buffer case, beginning-of-buffer opt arg, beginning-of-buffer opt arg
6271 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6274 Like other complex-looking expressions, the conditional expression
6275 within @code{beginning-of-buffer} can be disentangled by looking at it
6276 as parts of a template, in this case, the template for an if-then-else
6277 expression. In skeletal form, the expression looks like this:
6281 (if (@var{buffer-is-large}
6282 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6283 @var{else-use-alternate-calculation}
6287 The true-or-false-test of this inner @code{if} expression checks the
6288 size of the buffer. The reason for this is that the old version 18
6289 Emacs used numbers that are no bigger than eight million or so and in
6290 the computation that followed, the programmer feared that Emacs might
6291 try to use over-large numbers if the buffer were large. The term
6292 `overflow', mentioned in the comment, means numbers that are over
6293 large. More recent versions of Emacs use larger numbers, but this
6294 code has not been touched, if only because people now look at buffers
6295 that are far, far larger than ever before.
6297 There are two cases: if the buffer is large and if it is not.
6299 @node Large buffer case, Small buffer case, Disentangle beginning-of-buffer, beginning-of-buffer opt arg
6300 @comment node-name, next, previous, up
6301 @unnumberedsubsubsec What happens in a large buffer
6303 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6304 whether the size of the buffer is greater than 10,000 characters. To do
6305 this, it uses the @code{>} function and the computation of @code{size}
6306 that comes from the let expression.
6308 In the old days, the function @code{buffer-size} was used. Not only
6309 was that function called several times, it gave the size of the whole
6310 buffer, not the accessible part. The computation makes much more
6311 sense when it handles just the accessible part. (@xref{Narrowing &
6312 Widening, , Narrowing and Widening}, for more information on focusing
6313 attention to an `accessible' part.)
6316 The line looks like this:
6324 When the buffer is large, the then-part of the @code{if} expression is
6325 evaluated. It reads like this (after formatting for easy reading):
6330 (prefix-numeric-value arg)
6336 This expression is a multiplication, with two arguments to the function
6339 The first argument is @code{(prefix-numeric-value arg)}. When
6340 @code{"P"} is used as the argument for @code{interactive}, the value
6341 passed to the function as its argument is passed a ``raw prefix
6342 argument'', and not a number. (It is a number in a list.) To perform
6343 the arithmetic, a conversion is necessary, and
6344 @code{prefix-numeric-value} does the job.
6346 @findex / @r{(division)}
6348 The second argument is @code{(/ size 10)}. This expression divides
6349 the numeric value by ten --- the numeric value of the size of the
6350 accessible portion of the buffer. This produces a number that tells
6351 how many characters make up one tenth of the buffer size. (In Lisp,
6352 @code{/} is used for division, just as @code{*} is used for
6356 In the multiplication expression as a whole, this amount is multiplied
6357 by the value of the prefix argument---the multiplication looks like this:
6361 (* @var{numeric-value-of-prefix-arg}
6362 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6367 If, for example, the prefix argument is @samp{7}, the one-tenth value
6368 will be multiplied by 7 to give a position 70% of the way through.
6371 The result of all this is that if the accessible portion of the buffer
6372 is large, the @code{goto-char} expression reads like this:
6376 (goto-char (* (prefix-numeric-value arg)
6381 This puts the cursor where we want it.
6383 @node Small buffer case, , Large buffer case, beginning-of-buffer opt arg
6384 @comment node-name, next, previous, up
6385 @unnumberedsubsubsec What happens in a small buffer
6387 If the buffer contains fewer than 10,000 characters, a slightly
6388 different computation is performed. You might think this is not
6389 necessary, since the first computation could do the job. However, in
6390 a small buffer, the first method may not put the cursor on exactly the
6391 desired line; the second method does a better job.
6394 The code looks like this:
6396 @c Keep this on one line.
6398 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6403 This is code in which you figure out what happens by discovering how the
6404 functions are embedded in parentheses. It is easier to read if you
6405 reformat it with each expression indented more deeply than its
6406 enclosing expression:
6414 (prefix-numeric-value arg)))
6421 Looking at parentheses, we see that the innermost operation is
6422 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6423 a number. In the following expression, this number is multiplied by
6424 the size of the accessible portion of the buffer:
6427 (* size (prefix-numeric-value arg))
6431 This multiplication creates a number that may be larger than the size of
6432 the buffer---seven times larger if the argument is 7, for example. Ten
6433 is then added to this number and finally the large number is divided by
6434 ten to provide a value that is one character larger than the percentage
6435 position in the buffer.
6437 The number that results from all this is passed to @code{goto-char} and
6438 the cursor is moved to that point.
6441 @node beginning-of-buffer complete, , beginning-of-buffer opt arg, beginning-of-buffer
6442 @comment node-name, next, previous, up
6443 @subsection The Complete @code{beginning-of-buffer}
6446 Here is the complete text of the @code{beginning-of-buffer} function:
6452 (defun beginning-of-buffer (&optional arg)
6453 "Move point to the beginning of the buffer;
6454 leave mark at previous position.
6455 With \\[universal-argument] prefix,
6456 do not set mark at previous position.
6458 put point N/10 of the way from the beginning.
6460 If the buffer is narrowed,
6461 this command uses the beginning and size
6462 of the accessible part of the buffer.
6466 Don't use this command in Lisp programs!
6467 \(goto-char (point-min)) is faster
6468 and avoids clobbering the mark."
6471 (and transient-mark-mode mark-active)
6475 (let ((size (- (point-max) (point-min))))
6476 (goto-char (if (and arg (not (consp arg)))
6479 ;; Avoid overflow for large buffer sizes!
6480 (* (prefix-numeric-value arg)
6482 (/ (+ 10 (* size (prefix-numeric-value arg)))
6485 (if arg (forward-line 1)))
6490 From before GNU Emacs 22
6493 (defun beginning-of-buffer (&optional arg)
6494 "Move point to the beginning of the buffer;
6495 leave mark at previous position.
6496 With arg N, put point N/10 of the way
6497 from the true beginning.
6500 Don't use this in Lisp programs!
6501 \(goto-char (point-min)) is faster
6502 and does not set the mark."
6509 (if (> (buffer-size) 10000)
6510 ;; @r{Avoid overflow for large buffer sizes!}
6511 (* (prefix-numeric-value arg)
6512 (/ (buffer-size) 10))
6515 (/ (+ 10 (* (buffer-size)
6516 (prefix-numeric-value arg)))
6519 (if arg (forward-line 1)))
6525 Except for two small points, the previous discussion shows how this
6526 function works. The first point deals with a detail in the
6527 documentation string, and the second point concerns the last line of
6531 In the documentation string, there is reference to an expression:
6534 \\[universal-argument]
6538 A @samp{\\} is used before the first square bracket of this
6539 expression. This @samp{\\} tells the Lisp interpreter to substitute
6540 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6541 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6542 be different. (@xref{Documentation Tips, , Tips for Documentation
6543 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6547 Finally, the last line of the @code{beginning-of-buffer} command says
6548 to move point to the beginning of the next line if the command is
6549 invoked with an argument:
6552 (if arg (forward-line 1)))
6556 This puts the cursor at the beginning of the first line after the
6557 appropriate tenths position in the buffer. This is a flourish that
6558 means that the cursor is always located @emph{at least} the requested
6559 tenths of the way through the buffer, which is a nicety that is,
6560 perhaps, not necessary, but which, if it did not occur, would be sure
6563 On the other hand, it also means that if you specify the command with
6564 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6565 argument' is simply a cons cell, then the command puts you at the
6566 beginning of the second line @dots{} I don't know whether this is
6567 intended or whether no one has dealt with the code to avoid this
6570 @node Second Buffer Related Review, optional Exercise, beginning-of-buffer, More Complex
6571 @comment node-name, next, previous, up
6574 Here is a brief summary of some of the topics covered in this chapter.
6578 Evaluate each argument in sequence, and return the value of the first
6579 argument that is not @code{nil}; if none return a value that is not
6580 @code{nil}, return @code{nil}. In brief, return the first true value
6581 of the arguments; return a true value if one @emph{or} any of the
6585 Evaluate each argument in sequence, and if any are @code{nil}, return
6586 @code{nil}; if none are @code{nil}, return the value of the last
6587 argument. In brief, return a true value only if all the arguments are
6588 true; return a true value if one @emph{and} each of the others is
6592 A keyword used to indicate that an argument to a function definition
6593 is optional; this means that the function can be evaluated without the
6594 argument, if desired.
6596 @item prefix-numeric-value
6597 Convert the `raw prefix argument' produced by @code{(interactive
6598 "P")} to a numeric value.
6601 Move point forward to the beginning of the next line, or if the argument
6602 is greater than one, forward that many lines. If it can't move as far
6603 forward as it is supposed to, @code{forward-line} goes forward as far as
6604 it can and then returns a count of the number of additional lines it was
6605 supposed to move but couldn't.
6608 Delete the entire contents of the current buffer.
6611 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6614 @node optional Exercise, , Second Buffer Related Review, More Complex
6615 @section @code{optional} Argument Exercise
6617 Write an interactive function with an optional argument that tests
6618 whether its argument, a number, is greater than or equal to, or else,
6619 less than the value of @code{fill-column}, and tells you which, in a
6620 message. However, if you do not pass an argument to the function, use
6621 56 as a default value.
6623 @node Narrowing & Widening, car cdr & cons, More Complex, Top
6624 @comment node-name, next, previous, up
6625 @chapter Narrowing and Widening
6626 @cindex Focusing attention (narrowing)
6630 Narrowing is a feature of Emacs that makes it possible for you to focus
6631 on a specific part of a buffer, and work without accidentally changing
6632 other parts. Narrowing is normally disabled since it can confuse
6636 * Narrowing advantages:: The advantages of narrowing
6637 * save-restriction:: The @code{save-restriction} special form.
6638 * what-line:: The number of the line that point is on.
6642 @node Narrowing advantages, save-restriction, Narrowing & Widening, Narrowing & Widening
6644 @unnumberedsec The Advantages of Narrowing
6647 With narrowing, the rest of a buffer is made invisible, as if it weren't
6648 there. This is an advantage if, for example, you want to replace a word
6649 in one part of a buffer but not in another: you narrow to the part you want
6650 and the replacement is carried out only in that section, not in the rest
6651 of the buffer. Searches will only work within a narrowed region, not
6652 outside of one, so if you are fixing a part of a document, you can keep
6653 yourself from accidentally finding parts you do not need to fix by
6654 narrowing just to the region you want.
6655 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6657 However, narrowing does make the rest of the buffer invisible, which
6658 can scare people who inadvertently invoke narrowing and think they
6659 have deleted a part of their file. Moreover, the @code{undo} command
6660 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6661 (nor should it), so people can become quite desperate if they do not
6662 know that they can return the rest of a buffer to visibility with the
6663 @code{widen} command.
6664 (The key binding for @code{widen} is @kbd{C-x n w}.)
6666 Narrowing is just as useful to the Lisp interpreter as to a human.
6667 Often, an Emacs Lisp function is designed to work on just part of a
6668 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6669 buffer that has been narrowed. The @code{what-line} function, for
6670 example, removes the narrowing from a buffer, if it has any narrowing
6671 and when it has finished its job, restores the narrowing to what it was.
6672 On the other hand, the @code{count-lines} function, which is called by
6673 @code{what-line}, uses narrowing to restrict itself to just that portion
6674 of the buffer in which it is interested and then restores the previous
6677 @node save-restriction, what-line, Narrowing advantages, Narrowing & Widening
6678 @comment node-name, next, previous, up
6679 @section The @code{save-restriction} Special Form
6680 @findex save-restriction
6682 In Emacs Lisp, you can use the @code{save-restriction} special form to
6683 keep track of whatever narrowing is in effect, if any. When the Lisp
6684 interpreter meets with @code{save-restriction}, it executes the code
6685 in the body of the @code{save-restriction} expression, and then undoes
6686 any changes to narrowing that the code caused. If, for example, the
6687 buffer is narrowed and the code that follows @code{save-restriction}
6688 gets rid of the narrowing, @code{save-restriction} returns the buffer
6689 to its narrowed region afterwards. In the @code{what-line} command,
6690 any narrowing the buffer may have is undone by the @code{widen}
6691 command that immediately follows the @code{save-restriction} command.
6692 Any original narrowing is restored just before the completion of the
6696 The template for a @code{save-restriction} expression is simple:
6706 The body of the @code{save-restriction} is one or more expressions that
6707 will be evaluated in sequence by the Lisp interpreter.
6709 Finally, a point to note: when you use both @code{save-excursion} and
6710 @code{save-restriction}, one right after the other, you should use
6711 @code{save-excursion} outermost. If you write them in reverse order,
6712 you may fail to record narrowing in the buffer to which Emacs switches
6713 after calling @code{save-excursion}. Thus, when written together,
6714 @code{save-excursion} and @code{save-restriction} should be written
6725 In other circumstances, when not written together, the
6726 @code{save-excursion} and @code{save-restriction} special forms must
6727 be written in the order appropriate to the function.
6743 /usr/local/src/emacs/lisp/simple.el
6746 "Print the current buffer line number and narrowed line number of point."
6748 (let ((start (point-min))
6749 (n (line-number-at-pos)))
6751 (message "Line %d" n)
6755 (message "line %d (narrowed line %d)"
6756 (+ n (line-number-at-pos start) -1) n))))))
6758 (defun line-number-at-pos (&optional pos)
6759 "Return (narrowed) buffer line number at position POS.
6760 If POS is nil, use current buffer location.
6761 Counting starts at (point-min), so the value refers
6762 to the contents of the accessible portion of the buffer."
6763 (let ((opoint (or pos (point))) start)
6765 (goto-char (point-min))
6766 (setq start (point))
6769 (1+ (count-lines start (point))))))
6771 (defun count-lines (start end)
6772 "Return number of lines between START and END.
6773 This is usually the number of newlines between them,
6774 but can be one more if START is not equal to END
6775 and the greater of them is not at the start of a line."
6778 (narrow-to-region start end)
6779 (goto-char (point-min))
6780 (if (eq selective-display t)
6783 (while (re-search-forward "[\n\C-m]" nil t 40)
6784 (setq done (+ 40 done)))
6785 (while (re-search-forward "[\n\C-m]" nil t 1)
6786 (setq done (+ 1 done)))
6787 (goto-char (point-max))
6788 (if (and (/= start end)
6792 (- (buffer-size) (forward-line (buffer-size)))))))
6795 @node what-line, narrow Exercise, save-restriction, Narrowing & Widening
6796 @comment node-name, next, previous, up
6797 @section @code{what-line}
6799 @cindex Widening, example of
6801 The @code{what-line} command tells you the number of the line in which
6802 the cursor is located. The function illustrates the use of the
6803 @code{save-restriction} and @code{save-excursion} commands. Here is the
6804 original text of the function:
6809 "Print the current line number (in the buffer) of point."
6816 (1+ (count-lines 1 (point)))))))
6820 (In recent versions of GNU Emacs, the @code{what-line} function has
6821 been expanded to tell you your line number in a narrowed buffer as
6822 well as your line number in a widened buffer. The recent version is
6823 more complex than the version shown here. If you feel adventurous,
6824 you might want to look at it after figuring out how this version
6825 works. You will probably need to use @kbd{C-h f}
6826 (@code{describe-function}). The newer version uses a conditional to
6827 determine whether the buffer has been narrowed.
6829 (Also, it uses @code{line-number-at-pos}, which among other simple
6830 expressions, such as @code{(goto-char (point-min))}, moves point to
6831 the beginning of the current line with @code{(forward-line 0)} rather
6832 than @code{beginning-of-line}.)
6834 The @code{what-line} function as shown here has a documentation line
6835 and is interactive, as you would expect. The next two lines use the
6836 functions @code{save-restriction} and @code{widen}.
6838 The @code{save-restriction} special form notes whatever narrowing is in
6839 effect, if any, in the current buffer and restores that narrowing after
6840 the code in the body of the @code{save-restriction} has been evaluated.
6842 The @code{save-restriction} special form is followed by @code{widen}.
6843 This function undoes any narrowing the current buffer may have had
6844 when @code{what-line} was called. (The narrowing that was there is
6845 the narrowing that @code{save-restriction} remembers.) This widening
6846 makes it possible for the line counting commands to count from the
6847 beginning of the buffer. Otherwise, they would have been limited to
6848 counting within the accessible region. Any original narrowing is
6849 restored just before the completion of the function by the
6850 @code{save-restriction} special form.
6852 The call to @code{widen} is followed by @code{save-excursion}, which
6853 saves the location of the cursor (i.e., of point) and of the mark, and
6854 restores them after the code in the body of the @code{save-excursion}
6855 uses the @code{beginning-of-line} function to move point.
6857 (Note that the @code{(widen)} expression comes between the
6858 @code{save-restriction} and @code{save-excursion} special forms. When
6859 you write the two @code{save- @dots{}} expressions in sequence, write
6860 @code{save-excursion} outermost.)
6863 The last two lines of the @code{what-line} function are functions to
6864 count the number of lines in the buffer and then print the number in the
6870 (1+ (count-lines 1 (point)))))))
6874 The @code{message} function prints a one-line message at the bottom of
6875 the Emacs screen. The first argument is inside of quotation marks and
6876 is printed as a string of characters. However, it may contain a
6877 @samp{%d} expression to print a following argument. @samp{%d} prints
6878 the argument as a decimal, so the message will say something such as
6882 The number that is printed in place of the @samp{%d} is computed by the
6883 last line of the function:
6886 (1+ (count-lines 1 (point)))
6892 (defun count-lines (start end)
6893 "Return number of lines between START and END.
6894 This is usually the number of newlines between them,
6895 but can be one more if START is not equal to END
6896 and the greater of them is not at the start of a line."
6899 (narrow-to-region start end)
6900 (goto-char (point-min))
6901 (if (eq selective-display t)
6904 (while (re-search-forward "[\n\C-m]" nil t 40)
6905 (setq done (+ 40 done)))
6906 (while (re-search-forward "[\n\C-m]" nil t 1)
6907 (setq done (+ 1 done)))
6908 (goto-char (point-max))
6909 (if (and (/= start end)
6913 (- (buffer-size) (forward-line (buffer-size)))))))
6917 What this does is count the lines from the first position of the
6918 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6919 one to that number. (The @code{1+} function adds one to its
6920 argument.) We add one to it because line 2 has only one line before
6921 it, and @code{count-lines} counts only the lines @emph{before} the
6924 After @code{count-lines} has done its job, and the message has been
6925 printed in the echo area, the @code{save-excursion} restores point and
6926 mark to their original positions; and @code{save-restriction} restores
6927 the original narrowing, if any.
6929 @node narrow Exercise, , what-line, Narrowing & Widening
6930 @section Exercise with Narrowing
6932 Write a function that will display the first 60 characters of the
6933 current buffer, even if you have narrowed the buffer to its latter
6934 half so that the first line is inaccessible. Restore point, mark, and
6935 narrowing. For this exercise, you need to use a whole potpourri of
6936 functions, including @code{save-restriction}, @code{widen},
6937 @code{goto-char}, @code{point-min}, @code{message}, and
6938 @code{buffer-substring}.
6940 @cindex Properties, mention of @code{buffer-substring-no-properties}
6941 (@code{buffer-substring} is a previously unmentioned function you will
6942 have to investigate yourself; or perhaps you will have to use
6943 @code{buffer-substring-no-properties} or
6944 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6945 properties are a feature otherwise not discussed here. @xref{Text
6946 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6949 Additionally, do you really need @code{goto-char} or @code{point-min}?
6950 Or can you write the function without them?
6952 @node car cdr & cons, Cutting & Storing Text, Narrowing & Widening, Top
6953 @comment node-name, next, previous, up
6954 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6955 @findex car, @r{introduced}
6956 @findex cdr, @r{introduced}
6958 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6959 functions. The @code{cons} function is used to construct lists, and
6960 the @code{car} and @code{cdr} functions are used to take them apart.
6962 In the walk through of the @code{copy-region-as-kill} function, we
6963 will see @code{cons} as well as two variants on @code{cdr},
6964 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6967 * Strange Names:: An historical aside: why the strange names?
6968 * car & cdr:: Functions for extracting part of a list.
6969 * cons:: Constructing a list.
6970 * nthcdr:: Calling @code{cdr} repeatedly.
6972 * setcar:: Changing the first element of a list.
6973 * setcdr:: Changing the rest of a list.
6977 @node Strange Names, car & cdr, car cdr & cons, car cdr & cons
6979 @unnumberedsec Strange Names
6982 The name of the @code{cons} function is not unreasonable: it is an
6983 abbreviation of the word `construct'. The origins of the names for
6984 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6985 is an acronym from the phrase `Contents of the Address part of the
6986 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6987 the phrase `Contents of the Decrement part of the Register'. These
6988 phrases refer to specific pieces of hardware on the very early
6989 computer on which the original Lisp was developed. Besides being
6990 obsolete, the phrases have been completely irrelevant for more than 25
6991 years to anyone thinking about Lisp. Nonetheless, although a few
6992 brave scholars have begun to use more reasonable names for these
6993 functions, the old terms are still in use. In particular, since the
6994 terms are used in the Emacs Lisp source code, we will use them in this
6997 @node car & cdr, cons, Strange Names, car cdr & cons
6998 @comment node-name, next, previous, up
6999 @section @code{car} and @code{cdr}
7001 The @sc{car} of a list is, quite simply, the first item in the list.
7002 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
7006 If you are reading this in Info in GNU Emacs, you can see this by
7007 evaluating the following:
7010 (car '(rose violet daisy buttercup))
7014 After evaluating the expression, @code{rose} will appear in the echo
7017 Clearly, a more reasonable name for the @code{car} function would be
7018 @code{first} and this is often suggested.
7020 @code{car} does not remove the first item from the list; it only reports
7021 what it is. After @code{car} has been applied to a list, the list is
7022 still the same as it was. In the jargon, @code{car} is
7023 `non-destructive'. This feature turns out to be important.
7025 The @sc{cdr} of a list is the rest of the list, that is, the
7026 @code{cdr} function returns the part of the list that follows the
7027 first item. Thus, while the @sc{car} of the list @code{'(rose violet
7028 daisy buttercup)} is @code{rose}, the rest of the list, the value
7029 returned by the @code{cdr} function, is @code{(violet daisy
7033 You can see this by evaluating the following in the usual way:
7036 (cdr '(rose violet daisy buttercup))
7040 When you evaluate this, @code{(violet daisy buttercup)} will appear in
7043 Like @code{car}, @code{cdr} does not remove any elements from the
7044 list---it just returns a report of what the second and subsequent
7047 Incidentally, in the example, the list of flowers is quoted. If it were
7048 not, the Lisp interpreter would try to evaluate the list by calling
7049 @code{rose} as a function. In this example, we do not want to do that.
7051 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
7053 (There is a lesson here: when you name new functions, consider very
7054 carefully what you are doing, since you may be stuck with the names
7055 for far longer than you expect. The reason this document perpetuates
7056 these names is that the Emacs Lisp source code uses them, and if I did
7057 not use them, you would have a hard time reading the code; but do,
7058 please, try to avoid using these terms yourself. The people who come
7059 after you will be grateful to you.)
7061 When @code{car} and @code{cdr} are applied to a list made up of symbols,
7062 such as the list @code{(pine fir oak maple)}, the element of the list
7063 returned by the function @code{car} is the symbol @code{pine} without
7064 any parentheses around it. @code{pine} is the first element in the
7065 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7066 oak maple)}, as you can see by evaluating the following expressions in
7071 (car '(pine fir oak maple))
7073 (cdr '(pine fir oak maple))
7077 On the other hand, in a list of lists, the first element is itself a
7078 list. @code{car} returns this first element as a list. For example,
7079 the following list contains three sub-lists, a list of carnivores, a
7080 list of herbivores and a list of sea mammals:
7084 (car '((lion tiger cheetah)
7085 (gazelle antelope zebra)
7086 (whale dolphin seal)))
7091 In this example, the first element or @sc{car} of the list is the list of
7092 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7093 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7097 (cdr '((lion tiger cheetah)
7098 (gazelle antelope zebra)
7099 (whale dolphin seal)))
7103 It is worth saying again that @code{car} and @code{cdr} are
7104 non-destructive---that is, they do not modify or change lists to which
7105 they are applied. This is very important for how they are used.
7107 Also, in the first chapter, in the discussion about atoms, I said that
7108 in Lisp, ``certain kinds of atom, such as an array, can be separated
7109 into parts; but the mechanism for doing this is different from the
7110 mechanism for splitting a list. As far as Lisp is concerned, the
7111 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7112 @code{car} and @code{cdr} functions are used for splitting lists and
7113 are considered fundamental to Lisp. Since they cannot split or gain
7114 access to the parts of an array, an array is considered an atom.
7115 Conversely, the other fundamental function, @code{cons}, can put
7116 together or construct a list, but not an array. (Arrays are handled
7117 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7118 Emacs Lisp Reference Manual}.)
7120 @node cons, nthcdr, car & cdr, car cdr & cons
7121 @comment node-name, next, previous, up
7122 @section @code{cons}
7123 @findex cons, @r{introduced}
7125 The @code{cons} function constructs lists; it is the inverse of
7126 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7127 a four element list from the three element list, @code{(fir oak maple)}:
7130 (cons 'pine '(fir oak maple))
7135 After evaluating this list, you will see
7138 (pine fir oak maple)
7142 appear in the echo area. @code{cons} causes the creation of a new
7143 list in which the element is followed by the elements of the original
7146 We often say that `@code{cons} puts a new element at the beginning of
7147 a list; it attaches or pushes elements onto the list', but this
7148 phrasing can be misleading, since @code{cons} does not change an
7149 existing list, but creates a new one.
7151 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7155 * length:: How to find the length of a list.
7158 @node Build a list, length, cons, cons
7160 @unnumberedsubsec Build a list
7163 @code{cons} must have a list to attach to.@footnote{Actually, you can
7164 @code{cons} an element to an atom to produce a dotted pair. Dotted
7165 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7166 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7167 cannot start from absolutely nothing. If you are building a list, you
7168 need to provide at least an empty list at the beginning. Here is a
7169 series of @code{cons} expressions that build up a list of flowers. If
7170 you are reading this in Info in GNU Emacs, you can evaluate each of
7171 the expressions in the usual way; the value is printed in this text
7172 after @samp{@result{}}, which you may read as `evaluates to'.
7176 (cons 'buttercup ())
7177 @result{} (buttercup)
7181 (cons 'daisy '(buttercup))
7182 @result{} (daisy buttercup)
7186 (cons 'violet '(daisy buttercup))
7187 @result{} (violet daisy buttercup)
7191 (cons 'rose '(violet daisy buttercup))
7192 @result{} (rose violet daisy buttercup)
7197 In the first example, the empty list is shown as @code{()} and a list
7198 made up of @code{buttercup} followed by the empty list is constructed.
7199 As you can see, the empty list is not shown in the list that was
7200 constructed. All that you see is @code{(buttercup)}. The empty list is
7201 not counted as an element of a list because there is nothing in an empty
7202 list. Generally speaking, an empty list is invisible.
7204 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7205 two element list by putting @code{daisy} in front of @code{buttercup};
7206 and the third example constructs a three element list by putting
7207 @code{violet} in front of @code{daisy} and @code{buttercup}.
7209 @node length, , Build a list, cons
7210 @comment node-name, next, previous, up
7211 @subsection Find the Length of a List: @code{length}
7214 You can find out how many elements there are in a list by using the Lisp
7215 function @code{length}, as in the following examples:
7219 (length '(buttercup))
7224 (length '(daisy buttercup))
7229 (length (cons 'violet '(daisy buttercup)))
7235 In the third example, the @code{cons} function is used to construct a
7236 three element list which is then passed to the @code{length} function as
7240 We can also use @code{length} to count the number of elements in an
7251 As you would expect, the number of elements in an empty list is zero.
7253 An interesting experiment is to find out what happens if you try to find
7254 the length of no list at all; that is, if you try to call @code{length}
7255 without giving it an argument, not even an empty list:
7263 What you see, if you evaluate this, is the error message
7266 Lisp error: (wrong-number-of-arguments length 0)
7270 This means that the function receives the wrong number of
7271 arguments, zero, when it expects some other number of arguments. In
7272 this case, one argument is expected, the argument being a list whose
7273 length the function is measuring. (Note that @emph{one} list is
7274 @emph{one} argument, even if the list has many elements inside it.)
7276 The part of the error message that says @samp{length} is the name of
7280 @code{length} is still a subroutine, but you need C-h f to discover that.
7282 In an earlier version:
7283 This is written with a special notation, @samp{#<subr},
7284 that indicates that the function @code{length} is one of the primitive
7285 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7286 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7287 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7291 @node nthcdr, nth, cons, car cdr & cons
7292 @comment node-name, next, previous, up
7293 @section @code{nthcdr}
7296 The @code{nthcdr} function is associated with the @code{cdr} function.
7297 What it does is take the @sc{cdr} of a list repeatedly.
7299 If you take the @sc{cdr} of the list @code{(pine fir
7300 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7301 repeat this on what was returned, you will be returned the list
7302 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7303 list will just give you the original @sc{cdr} since the function does
7304 not change the list. You need to evaluate the @sc{cdr} of the
7305 @sc{cdr} and so on.) If you continue this, eventually you will be
7306 returned an empty list, which in this case, instead of being shown as
7307 @code{()} is shown as @code{nil}.
7310 For review, here is a series of repeated @sc{cdr}s, the text following
7311 the @samp{@result{}} shows what is returned.
7315 (cdr '(pine fir oak maple))
7316 @result{}(fir oak maple)
7320 (cdr '(fir oak maple))
7321 @result{} (oak maple)
7346 You can also do several @sc{cdr}s without printing the values in
7351 (cdr (cdr '(pine fir oak maple)))
7352 @result{} (oak maple)
7357 In this example, the Lisp interpreter evaluates the innermost list first.
7358 The innermost list is quoted, so it just passes the list as it is to the
7359 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7360 second and subsequent elements of the list to the outermost @code{cdr},
7361 which produces a list composed of the third and subsequent elements of
7362 the original list. In this example, the @code{cdr} function is repeated
7363 and returns a list that consists of the original list without its
7366 The @code{nthcdr} function does the same as repeating the call to
7367 @code{cdr}. In the following example, the argument 2 is passed to the
7368 function @code{nthcdr}, along with the list, and the value returned is
7369 the list without its first two items, which is exactly the same
7370 as repeating @code{cdr} twice on the list:
7374 (nthcdr 2 '(pine fir oak maple))
7375 @result{} (oak maple)
7380 Using the original four element list, we can see what happens when
7381 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7386 ;; @r{Leave the list as it was.}
7387 (nthcdr 0 '(pine fir oak maple))
7388 @result{} (pine fir oak maple)
7392 ;; @r{Return a copy without the first element.}
7393 (nthcdr 1 '(pine fir oak maple))
7394 @result{} (fir oak maple)
7398 ;; @r{Return a copy of the list without three elements.}
7399 (nthcdr 3 '(pine fir oak maple))
7404 ;; @r{Return a copy lacking all four elements.}
7405 (nthcdr 4 '(pine fir oak maple))
7410 ;; @r{Return a copy lacking all elements.}
7411 (nthcdr 5 '(pine fir oak maple))
7416 @node nth, setcar, nthcdr, car cdr & cons
7417 @comment node-name, next, previous, up
7421 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7422 The @code{nth} function takes the @sc{car} of the result returned by
7423 @code{nthcdr}. It returns the Nth element of the list.
7426 Thus, if it were not defined in C for speed, the definition of
7427 @code{nth} would be:
7432 "Returns the Nth element of LIST.
7433 N counts from zero. If LIST is not that long, nil is returned."
7434 (car (nthcdr n list)))
7439 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7440 but its definition was redone in C in the 1980s.)
7442 The @code{nth} function returns a single element of a list.
7443 This can be very convenient.
7445 Note that the elements are numbered from zero, not one. That is to
7446 say, the first element of a list, its @sc{car} is the zeroth element.
7447 This is called `zero-based' counting and often bothers people who
7448 are accustomed to the first element in a list being number one, which
7456 (nth 0 '("one" "two" "three"))
7459 (nth 1 '("one" "two" "three"))
7464 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7465 @code{cdr}, does not change the original list---the function is
7466 non-destructive. This is in sharp contrast to the @code{setcar} and
7467 @code{setcdr} functions.
7469 @node setcar, setcdr, nth, car cdr & cons
7470 @comment node-name, next, previous, up
7471 @section @code{setcar}
7474 As you might guess from their names, the @code{setcar} and @code{setcdr}
7475 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7476 They actually change the original list, unlike @code{car} and @code{cdr}
7477 which leave the original list as it was. One way to find out how this
7478 works is to experiment. We will start with the @code{setcar} function.
7481 First, we can make a list and then set the value of a variable to the
7482 list, using the @code{setq} function. Here is a list of animals:
7485 (setq animals '(antelope giraffe lion tiger))
7489 If you are reading this in Info inside of GNU Emacs, you can evaluate
7490 this expression in the usual fashion, by positioning the cursor after
7491 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7492 as I write this. This is one of the advantages of having the
7493 interpreter built into the computing environment. Incidentally, when
7494 there is nothing on the line after the final parentheses, such as a
7495 comment, point can be on the next line. Thus, if your cursor is in
7496 the first column of the next line, you do not need to move it.
7497 Indeed, Emacs permits any amount of white space after the final
7501 When we evaluate the variable @code{animals}, we see that it is bound to
7502 the list @code{(antelope giraffe lion tiger)}:
7507 @result{} (antelope giraffe lion tiger)
7512 Put another way, the variable @code{animals} points to the list
7513 @code{(antelope giraffe lion tiger)}.
7515 Next, evaluate the function @code{setcar} while passing it two
7516 arguments, the variable @code{animals} and the quoted symbol
7517 @code{hippopotamus}; this is done by writing the three element list
7518 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7522 (setcar animals 'hippopotamus)
7527 After evaluating this expression, evaluate the variable @code{animals}
7528 again. You will see that the list of animals has changed:
7533 @result{} (hippopotamus giraffe lion tiger)
7538 The first element on the list, @code{antelope} is replaced by
7539 @code{hippopotamus}.
7541 So we can see that @code{setcar} did not add a new element to the list
7542 as @code{cons} would have; it replaced @code{antelope} with
7543 @code{hippopotamus}; it @emph{changed} the list.
7545 @node setcdr, cons Exercise, setcar, car cdr & cons
7546 @comment node-name, next, previous, up
7547 @section @code{setcdr}
7550 The @code{setcdr} function is similar to the @code{setcar} function,
7551 except that the function replaces the second and subsequent elements of
7552 a list rather than the first element.
7554 (To see how to change the last element of a list, look ahead to
7555 @ref{kill-new function, , The @code{kill-new} function}, which uses
7556 the @code{nthcdr} and @code{setcdr} functions.)
7559 To see how this works, set the value of the variable to a list of
7560 domesticated animals by evaluating the following expression:
7563 (setq domesticated-animals '(horse cow sheep goat))
7568 If you now evaluate the list, you will be returned the list
7569 @code{(horse cow sheep goat)}:
7573 domesticated-animals
7574 @result{} (horse cow sheep goat)
7579 Next, evaluate @code{setcdr} with two arguments, the name of the
7580 variable which has a list as its value, and the list to which the
7581 @sc{cdr} of the first list will be set;
7584 (setcdr domesticated-animals '(cat dog))
7588 If you evaluate this expression, the list @code{(cat dog)} will appear
7589 in the echo area. This is the value returned by the function. The
7590 result we are interested in is the ``side effect'', which we can see by
7591 evaluating the variable @code{domesticated-animals}:
7595 domesticated-animals
7596 @result{} (horse cat dog)
7601 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7602 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7603 @code{(cow sheep goat)} to @code{(cat dog)}.
7605 @node cons Exercise, , setcdr, car cdr & cons
7608 Construct a list of four birds by evaluating several expressions with
7609 @code{cons}. Find out what happens when you @code{cons} a list onto
7610 itself. Replace the first element of the list of four birds with a
7611 fish. Replace the rest of that list with a list of other fish.
7613 @node Cutting & Storing Text, List Implementation, car cdr & cons, Top
7614 @comment node-name, next, previous, up
7615 @chapter Cutting and Storing Text
7616 @cindex Cutting and storing text
7617 @cindex Storing and cutting text
7618 @cindex Killing text
7619 @cindex Clipping text
7620 @cindex Erasing text
7621 @cindex Deleting text
7623 Whenever you cut or clip text out of a buffer with a `kill' command in
7624 GNU Emacs, it is stored in a list and you can bring it back with a
7627 (The use of the word `kill' in Emacs for processes which specifically
7628 @emph{do not} destroy the values of the entities is an unfortunate
7629 historical accident. A much more appropriate word would be `clip' since
7630 that is what the kill commands do; they clip text out of a buffer and
7631 put it into storage from which it can be brought back. I have often
7632 been tempted to replace globally all occurrences of `kill' in the Emacs
7633 sources with `clip' and all occurrences of `killed' with `clipped'.)
7636 * Storing Text:: Text is stored in a list.
7637 * zap-to-char:: Cutting out text up to a character.
7638 * kill-region:: Cutting text out of a region.
7639 * copy-region-as-kill:: A definition for copying text.
7640 * Digression into C:: Minor note on C programming language macros.
7641 * defvar:: How to give a variable an initial value.
7642 * cons & search-fwd Review::
7643 * search Exercises::
7646 @node Storing Text, zap-to-char, Cutting & Storing Text, Cutting & Storing Text
7648 @unnumberedsec Storing Text in a List
7651 When text is cut out of a buffer, it is stored on a list. Successive
7652 pieces of text are stored on the list successively, so the list might
7656 ("a piece of text" "previous piece")
7661 The function @code{cons} can be used to create a new list from a piece
7662 of text (an `atom', to use the jargon) and an existing list, like
7667 (cons "another piece"
7668 '("a piece of text" "previous piece"))
7674 If you evaluate this expression, a list of three elements will appear in
7678 ("another piece" "a piece of text" "previous piece")
7681 With the @code{car} and @code{nthcdr} functions, you can retrieve
7682 whichever piece of text you want. For example, in the following code,
7683 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7684 and the @code{car} returns the first element of that remainder---the
7685 second element of the original list:
7689 (car (nthcdr 1 '("another piece"
7692 @result{} "a piece of text"
7696 The actual functions in Emacs are more complex than this, of course.
7697 The code for cutting and retrieving text has to be written so that
7698 Emacs can figure out which element in the list you want---the first,
7699 second, third, or whatever. In addition, when you get to the end of
7700 the list, Emacs should give you the first element of the list, rather
7701 than nothing at all.
7703 The list that holds the pieces of text is called the @dfn{kill ring}.
7704 This chapter leads up to a description of the kill ring and how it is
7705 used by first tracing how the @code{zap-to-char} function works. This
7706 function uses (or `calls') a function that invokes a function that
7707 manipulates the kill ring. Thus, before reaching the mountains, we
7708 climb the foothills.
7710 A subsequent chapter describes how text that is cut from the buffer is
7711 retrieved. @xref{Yanking, , Yanking Text Back}.
7713 @node zap-to-char, kill-region, Storing Text, Cutting & Storing Text
7714 @comment node-name, next, previous, up
7715 @section @code{zap-to-char}
7718 The @code{zap-to-char} function changed little between GNU Emacs
7719 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7720 calls another function, @code{kill-region}, which enjoyed a major
7723 The @code{kill-region} function in Emacs 19 is complex, but does not
7724 use code that is important at this time. We will skip it.
7726 The @code{kill-region} function in Emacs 22 is easier to read than the
7727 same function in Emacs 19 and introduces a very important concept,
7728 that of error handling. We will walk through the function.
7730 But first, let us look at the interactive @code{zap-to-char} function.
7733 * Complete zap-to-char:: The complete implementation.
7734 * zap-to-char interactive:: A three part interactive expression.
7735 * zap-to-char body:: A short overview.
7736 * search-forward:: How to search for a string.
7737 * progn:: The @code{progn} special form.
7738 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7741 @node Complete zap-to-char, zap-to-char interactive, zap-to-char, zap-to-char
7743 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7746 The @code{zap-to-char} function removes the text in the region between
7747 the location of the cursor (i.e., of point) up to and including the
7748 next occurrence of a specified character. The text that
7749 @code{zap-to-char} removes is put in the kill ring; and it can be
7750 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7751 the command is given an argument, it removes text through that number
7752 of occurrences. Thus, if the cursor were at the beginning of this
7753 sentence and the character were @samp{s}, @samp{Thus} would be
7754 removed. If the argument were two, @samp{Thus, if the curs} would be
7755 removed, up to and including the @samp{s} in @samp{cursor}.
7757 If the specified character is not found, @code{zap-to-char} will say
7758 ``Search failed'', tell you the character you typed, and not remove
7761 In order to determine how much text to remove, @code{zap-to-char} uses
7762 a search function. Searches are used extensively in code that
7763 manipulates text, and we will focus attention on them as well as on the
7767 @c GNU Emacs version 19
7768 (defun zap-to-char (arg char) ; version 19 implementation
7769 "Kill up to and including ARG'th occurrence of CHAR.
7770 Goes backward if ARG is negative; error if CHAR not found."
7771 (interactive "*p\ncZap to char: ")
7772 (kill-region (point)
7775 (char-to-string char) nil nil arg)
7780 Here is the complete text of the version 22 implementation of the function:
7785 (defun zap-to-char (arg char)
7786 "Kill up to and including ARG'th occurrence of CHAR.
7787 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7788 Goes backward if ARG is negative; error if CHAR not found."
7789 (interactive "p\ncZap to char: ")
7790 (if (char-table-p translation-table-for-input)
7791 (setq char (or (aref translation-table-for-input char) char)))
7792 (kill-region (point) (progn
7793 (search-forward (char-to-string char)
7799 The documentation is thorough. You do need to know the jargon meaning
7802 @node zap-to-char interactive, zap-to-char body, Complete zap-to-char, zap-to-char
7803 @comment node-name, next, previous, up
7804 @subsection The @code{interactive} Expression
7807 The interactive expression in the @code{zap-to-char} command looks like
7811 (interactive "p\ncZap to char: ")
7814 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7815 two different things. First, and most simply, is the @samp{p}.
7816 This part is separated from the next part by a newline, @samp{\n}.
7817 The @samp{p} means that the first argument to the function will be
7818 passed the value of a `processed prefix'. The prefix argument is
7819 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7820 the function is called interactively without a prefix, 1 is passed to
7823 The second part of @code{"p\ncZap to char:@: "} is
7824 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7825 indicates that @code{interactive} expects a prompt and that the
7826 argument will be a character. The prompt follows the @samp{c} and is
7827 the string @samp{Zap to char:@: } (with a space after the colon to
7830 What all this does is prepare the arguments to @code{zap-to-char} so they
7831 are of the right type, and give the user a prompt.
7833 In a read-only buffer, the @code{zap-to-char} function copies the text
7834 to the kill ring, but does not remove it. The echo area displays a
7835 message saying that the buffer is read-only. Also, the terminal may
7836 beep or blink at you.
7838 @node zap-to-char body, search-forward, zap-to-char interactive, zap-to-char
7839 @comment node-name, next, previous, up
7840 @subsection The Body of @code{zap-to-char}
7842 The body of the @code{zap-to-char} function contains the code that
7843 kills (that is, removes) the text in the region from the current
7844 position of the cursor up to and including the specified character.
7846 The first part of the code looks like this:
7849 (if (char-table-p translation-table-for-input)
7850 (setq char (or (aref translation-table-for-input char) char)))
7851 (kill-region (point) (progn
7852 (search-forward (char-to-string char) nil nil arg)
7857 @code{char-table-p} is an hitherto unseen function. It determines
7858 whether its argument is a character table. When it is, it sets the
7859 character passed to @code{zap-to-char} to one of them, if that
7860 character exists, or to the character itself. (This becomes important
7861 for certain characters in non-European languages. The @code{aref}
7862 function extracts an element from an array. It is an array-specific
7863 function that is not described in this document. @xref{Arrays, ,
7864 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7867 @code{(point)} is the current position of the cursor.
7869 The next part of the code is an expression using @code{progn}. The body
7870 of the @code{progn} consists of calls to @code{search-forward} and
7873 It is easier to understand how @code{progn} works after learning about
7874 @code{search-forward}, so we will look at @code{search-forward} and
7875 then at @code{progn}.
7877 @node search-forward, progn, zap-to-char body, zap-to-char
7878 @comment node-name, next, previous, up
7879 @subsection The @code{search-forward} Function
7880 @findex search-forward
7882 The @code{search-forward} function is used to locate the
7883 zapped-for-character in @code{zap-to-char}. If the search is
7884 successful, @code{search-forward} leaves point immediately after the
7885 last character in the target string. (In @code{zap-to-char}, the
7886 target string is just one character long. @code{zap-to-char} uses the
7887 function @code{char-to-string} to ensure that the computer treats that
7888 character as a string.) If the search is backwards,
7889 @code{search-forward} leaves point just before the first character in
7890 the target. Also, @code{search-forward} returns @code{t} for true.
7891 (Moving point is therefore a `side effect'.)
7894 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7897 (search-forward (char-to-string char) nil nil arg)
7900 The @code{search-forward} function takes four arguments:
7904 The first argument is the target, what is searched for. This must be a
7905 string, such as @samp{"z"}.
7907 As it happens, the argument passed to @code{zap-to-char} is a single
7908 character. Because of the way computers are built, the Lisp
7909 interpreter may treat a single character as being different from a
7910 string of characters. Inside the computer, a single character has a
7911 different electronic format than a string of one character. (A single
7912 character can often be recorded in the computer using exactly one
7913 byte; but a string may be longer, and the computer needs to be ready
7914 for this.) Since the @code{search-forward} function searches for a
7915 string, the character that the @code{zap-to-char} function receives as
7916 its argument must be converted inside the computer from one format to
7917 the other; otherwise the @code{search-forward} function will fail.
7918 The @code{char-to-string} function is used to make this conversion.
7921 The second argument bounds the search; it is specified as a position in
7922 the buffer. In this case, the search can go to the end of the buffer,
7923 so no bound is set and the second argument is @code{nil}.
7926 The third argument tells the function what it should do if the search
7927 fails---it can signal an error (and print a message) or it can return
7928 @code{nil}. A @code{nil} as the third argument causes the function to
7929 signal an error when the search fails.
7932 The fourth argument to @code{search-forward} is the repeat count---how
7933 many occurrences of the string to look for. This argument is optional
7934 and if the function is called without a repeat count, this argument is
7935 passed the value 1. If this argument is negative, the search goes
7940 In template form, a @code{search-forward} expression looks like this:
7944 (search-forward "@var{target-string}"
7945 @var{limit-of-search}
7946 @var{what-to-do-if-search-fails}
7951 We will look at @code{progn} next.
7953 @node progn, Summing up zap-to-char, search-forward, zap-to-char
7954 @comment node-name, next, previous, up
7955 @subsection The @code{progn} Special Form
7958 @code{progn} is a special form that causes each of its arguments to be
7959 evaluated in sequence and then returns the value of the last one. The
7960 preceding expressions are evaluated only for the side effects they
7961 perform. The values produced by them are discarded.
7964 The template for a @code{progn} expression is very simple:
7973 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7974 put point in exactly the right position; and return the location of
7975 point so that @code{kill-region} will know how far to kill to.
7977 The first argument to the @code{progn} is @code{search-forward}. When
7978 @code{search-forward} finds the string, the function leaves point
7979 immediately after the last character in the target string. (In this
7980 case the target string is just one character long.) If the search is
7981 backwards, @code{search-forward} leaves point just before the first
7982 character in the target. The movement of point is a side effect.
7984 The second and last argument to @code{progn} is the expression
7985 @code{(point)}. This expression returns the value of point, which in
7986 this case will be the location to which it has been moved by
7987 @code{search-forward}. (In the source, a line that tells the function
7988 to go to the previous character, if it is going forward, was commented
7989 out in 1999; I don't remember whether that feature or mis-feature was
7990 ever a part of the distributed source.) The value of @code{point} is
7991 returned by the @code{progn} expression and is passed to
7992 @code{kill-region} as @code{kill-region}'s second argument.
7994 @node Summing up zap-to-char, , progn, zap-to-char
7995 @comment node-name, next, previous, up
7996 @subsection Summing up @code{zap-to-char}
7998 Now that we have seen how @code{search-forward} and @code{progn} work,
7999 we can see how the @code{zap-to-char} function works as a whole.
8001 The first argument to @code{kill-region} is the position of the cursor
8002 when the @code{zap-to-char} command is given---the value of point at
8003 that time. Within the @code{progn}, the search function then moves
8004 point to just after the zapped-to-character and @code{point} returns the
8005 value of this location. The @code{kill-region} function puts together
8006 these two values of point, the first one as the beginning of the region
8007 and the second one as the end of the region, and removes the region.
8009 The @code{progn} special form is necessary because the
8010 @code{kill-region} command takes two arguments; and it would fail if
8011 @code{search-forward} and @code{point} expressions were written in
8012 sequence as two additional arguments. The @code{progn} expression is
8013 a single argument to @code{kill-region} and returns the one value that
8014 @code{kill-region} needs for its second argument.
8016 @node kill-region, copy-region-as-kill, zap-to-char, Cutting & Storing Text
8017 @comment node-name, next, previous, up
8018 @section @code{kill-region}
8021 The @code{zap-to-char} function uses the @code{kill-region} function.
8022 This function clips text from a region and copies that text to
8023 the kill ring, from which it may be retrieved.
8028 (defun kill-region (beg end &optional yank-handler)
8029 "Kill (\"cut\") text between point and mark.
8030 This deletes the text from the buffer and saves it in the kill ring.
8031 The command \\[yank] can retrieve it from there.
8032 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
8034 If you want to append the killed region to the last killed text,
8035 use \\[append-next-kill] before \\[kill-region].
8037 If the buffer is read-only, Emacs will beep and refrain from deleting
8038 the text, but put the text in the kill ring anyway. This means that
8039 you can use the killing commands to copy text from a read-only buffer.
8041 This is the primitive for programs to kill text (as opposed to deleting it).
8042 Supply two arguments, character positions indicating the stretch of text
8044 Any command that calls this function is a \"kill command\".
8045 If the previous command was also a kill command,
8046 the text killed this time appends to the text killed last time
8047 to make one entry in the kill ring.
8049 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
8050 specifies the yank-handler text property to be set on the killed
8051 text. See `insert-for-yank'."
8052 ;; Pass point first, then mark, because the order matters
8053 ;; when calling kill-append.
8054 (interactive (list (point) (mark)))
8055 (unless (and beg end)
8056 (error "The mark is not set now, so there is no region"))
8058 (let ((string (filter-buffer-substring beg end t)))
8059 (when string ;STRING is nil if BEG = END
8060 ;; Add that string to the kill ring, one way or another.
8061 (if (eq last-command 'kill-region)
8062 (kill-append string (< end beg) yank-handler)
8063 (kill-new string nil yank-handler)))
8064 (when (or string (eq last-command 'kill-region))
8065 (setq this-command 'kill-region))
8067 ((buffer-read-only text-read-only)
8068 ;; The code above failed because the buffer, or some of the characters
8069 ;; in the region, are read-only.
8070 ;; We should beep, in case the user just isn't aware of this.
8071 ;; However, there's no harm in putting
8072 ;; the region's text in the kill ring, anyway.
8073 (copy-region-as-kill beg end)
8074 ;; Set this-command now, so it will be set even if we get an error.
8075 (setq this-command 'kill-region)
8076 ;; This should barf, if appropriate, and give us the correct error.
8077 (if kill-read-only-ok
8078 (progn (message "Read only text copied to kill ring") nil)
8079 ;; Signal an error if the buffer is read-only.
8080 (barf-if-buffer-read-only)
8081 ;; If the buffer isn't read-only, the text is.
8082 (signal 'text-read-only (list (current-buffer)))))))
8085 The Emacs 22 version of that function uses @code{condition-case} and
8086 @code{copy-region-as-kill}, both of which we will explain.
8087 @code{condition-case} is an important special form.
8089 In essence, the @code{kill-region} function calls
8090 @code{condition-case}, which takes three arguments. In this function,
8091 the first argument does nothing. The second argument contains the
8092 code that does the work when all goes well. The third argument
8093 contains the code that is called in the event of an error.
8096 * Complete kill-region:: The function definition.
8097 * condition-case:: Dealing with a problem.
8101 @node Complete kill-region, condition-case, kill-region, kill-region
8103 @unnumberedsubsec The Complete @code{kill-region} Definition
8107 We will go through the @code{condition-case} code in a moment. First,
8108 let us look at the definition of @code{kill-region}, with comments
8114 (defun kill-region (beg end)
8115 "Kill (\"cut\") text between point and mark.
8116 This deletes the text from the buffer and saves it in the kill ring.
8117 The command \\[yank] can retrieve it from there. @dots{} "
8121 ;; @bullet{} Since order matters, pass point first.
8122 (interactive (list (point) (mark)))
8123 ;; @bullet{} And tell us if we cannot cut the text.
8124 ;; `unless' is an `if' without a then-part.
8125 (unless (and beg end)
8126 (error "The mark is not set now, so there is no region"))
8130 ;; @bullet{} `condition-case' takes three arguments.
8131 ;; If the first argument is nil, as it is here,
8132 ;; information about the error signal is not
8133 ;; stored for use by another function.
8138 ;; @bullet{} The second argument to `condition-case' tells the
8139 ;; Lisp interpreter what to do when all goes well.
8143 ;; It starts with a `let' function that extracts the string
8144 ;; and tests whether it exists. If so (that is what the
8145 ;; `when' checks), it calls an `if' function that determines
8146 ;; whether the previous command was another call to
8147 ;; `kill-region'; if it was, then the new text is appended to
8148 ;; the previous text; if not, then a different function,
8149 ;; `kill-new', is called.
8153 ;; The `kill-append' function concatenates the new string and
8154 ;; the old. The `kill-new' function inserts text into a new
8155 ;; item in the kill ring.
8159 ;; `when' is an `if' without an else-part. The second `when'
8160 ;; again checks whether the current string exists; in
8161 ;; addition, it checks whether the previous command was
8162 ;; another call to `kill-region'. If one or the other
8163 ;; condition is true, then it sets the current command to
8164 ;; be `kill-region'.
8167 (let ((string (filter-buffer-substring beg end t)))
8168 (when string ;STRING is nil if BEG = END
8169 ;; Add that string to the kill ring, one way or another.
8170 (if (eq last-command 'kill-region)
8173 ;; @minus{} `yank-handler' is an optional argument to
8174 ;; `kill-region' that tells the `kill-append' and
8175 ;; `kill-new' functions how deal with properties
8176 ;; added to the text, such as `bold' or `italics'.
8177 (kill-append string (< end beg) yank-handler)
8178 (kill-new string nil yank-handler)))
8179 (when (or string (eq last-command 'kill-region))
8180 (setq this-command 'kill-region))
8185 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8186 ;; what to do with an error.
8189 ;; The third argument has a conditions part and a body part.
8190 ;; If the conditions are met (in this case,
8191 ;; if text or buffer are read-only)
8192 ;; then the body is executed.
8195 ;; The first part of the third argument is the following:
8196 ((buffer-read-only text-read-only) ;; the if-part
8197 ;; @dots{} the then-part
8198 (copy-region-as-kill beg end)
8201 ;; Next, also as part of the then-part, set this-command, so
8202 ;; it will be set in an error
8203 (setq this-command 'kill-region)
8204 ;; Finally, in the then-part, send a message if you may copy
8205 ;; the text to the kill ring without signally an error, but
8206 ;; don't if you may not.
8209 (if kill-read-only-ok
8210 (progn (message "Read only text copied to kill ring") nil)
8211 (barf-if-buffer-read-only)
8212 ;; If the buffer isn't read-only, the text is.
8213 (signal 'text-read-only (list (current-buffer)))))
8221 (defun kill-region (beg end)
8222 "Kill between point and mark.
8223 The text is deleted but saved in the kill ring."
8228 ;; 1. `condition-case' takes three arguments.
8229 ;; If the first argument is nil, as it is here,
8230 ;; information about the error signal is not
8231 ;; stored for use by another function.
8236 ;; 2. The second argument to `condition-case'
8237 ;; tells the Lisp interpreter what to do when all goes well.
8241 ;; The `delete-and-extract-region' function usually does the
8242 ;; work. If the beginning and ending of the region are both
8243 ;; the same, then the variable `string' will be empty, or nil
8244 (let ((string (delete-and-extract-region beg end)))
8248 ;; `when' is an `if' clause that cannot take an `else-part'.
8249 ;; Emacs normally sets the value of `last-command' to the
8250 ;; previous command.
8253 ;; `kill-append' concatenates the new string and the old.
8254 ;; `kill-new' inserts text into a new item in the kill ring.
8256 (if (eq last-command 'kill-region)
8257 ;; if true, prepend string
8258 (kill-append string (< end beg))
8260 (setq this-command 'kill-region))
8264 ;; 3. The third argument to `condition-case' tells the interpreter
8265 ;; what to do with an error.
8268 ;; The third argument has a conditions part and a body part.
8269 ;; If the conditions are met (in this case,
8270 ;; if text or buffer are read-only)
8271 ;; then the body is executed.
8274 ((buffer-read-only text-read-only) ;; this is the if-part
8276 (copy-region-as-kill beg end)
8279 (if kill-read-only-ok ;; usually this variable is nil
8280 (message "Read only text copied to kill ring")
8281 ;; or else, signal an error if the buffer is read-only;
8282 (barf-if-buffer-read-only)
8283 ;; and, in any case, signal that the text is read-only.
8284 (signal 'text-read-only (list (current-buffer)))))))
8289 @node condition-case, Lisp macro, Complete kill-region, kill-region
8290 @comment node-name, next, previous, up
8291 @subsection @code{condition-case}
8292 @findex condition-case
8294 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8295 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8296 expression, it provides you with help; in the jargon, this is called
8297 ``signaling an error''. Usually, the computer stops the program and
8298 shows you a message.
8300 However, some programs undertake complicated actions. They should not
8301 simply stop on an error. In the @code{kill-region} function, the most
8302 likely error is that you will try to kill text that is read-only and
8303 cannot be removed. So the @code{kill-region} function contains code
8304 to handle this circumstance. This code, which makes up the body of
8305 the @code{kill-region} function, is inside of a @code{condition-case}
8309 The template for @code{condition-case} looks like this:
8316 @var{error-handler}@dots{})
8320 The second argument, @var{bodyform}, is straightforward. The
8321 @code{condition-case} special form causes the Lisp interpreter to
8322 evaluate the code in @var{bodyform}. If no error occurs, the special
8323 form returns the code's value and produces the side-effects, if any.
8325 In short, the @var{bodyform} part of a @code{condition-case}
8326 expression determines what should happen when everything works
8329 However, if an error occurs, among its other actions, the function
8330 generating the error signal will define one or more error condition
8333 An error handler is the third argument to @code{condition case}.
8334 An error handler has two parts, a @var{condition-name} and a
8335 @var{body}. If the @var{condition-name} part of an error handler
8336 matches a condition name generated by an error, then the @var{body}
8337 part of the error handler is run.
8339 As you will expect, the @var{condition-name} part of an error handler
8340 may be either a single condition name or a list of condition names.
8342 Also, a complete @code{condition-case} expression may contain more
8343 than one error handler. When an error occurs, the first applicable
8346 Lastly, the first argument to the @code{condition-case} expression,
8347 the @var{var} argument, is sometimes bound to a variable that
8348 contains information about the error. However, if that argument is
8349 nil, as is the case in @code{kill-region}, that information is
8353 In brief, in the @code{kill-region} function, the code
8354 @code{condition-case} works like this:
8358 @var{If no errors}, @var{run only this code}
8359 @var{but}, @var{if errors}, @var{run this other code}.
8366 copy-region-as-kill is short, 12 lines, and uses
8367 filter-buffer-substring, which is longer, 39 lines
8368 and has delete-and-extract-region in it.
8369 delete-and-extract-region is written in C.
8371 see Initializing a Variable with @code{defvar}
8373 Initializing a Variable with @code{defvar} includes line 8350
8376 @node Lisp macro, , condition-case, kill-region
8377 @comment node-name, next, previous, up
8378 @subsection Lisp macro
8382 The part of the @code{condition-case} expression that is evaluated in
8383 the expectation that all goes well has a @code{when}. The code uses
8384 @code{when} to determine whether the @code{string} variable points to
8387 A @code{when} expression is simply a programmers' convenience. It is
8388 an @code{if} without the possibility of an else clause. In your mind,
8389 you can replace @code{when} with @code{if} and understand what goes
8390 on. That is what the Lisp interpreter does.
8392 Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro}
8393 enables you to define new control constructs and other language
8394 features. It tells the interpreter how to compute another Lisp
8395 expression which will in turn compute the value. In this case, the
8396 `other expression' is an @code{if} expression.
8398 The @code{kill-region} function definition also has an @code{unless}
8399 macro; it is the converse of @code{when}. The @code{unless} macro is
8400 an @code{if} without a then clause
8402 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8403 Emacs Lisp Reference Manual}. The C programming language also
8404 provides macros. These are different, but also useful.
8407 We will briefly look at C macros in
8408 @ref{Digression into C}.
8412 Regarding the @code{when} macro, in the @code{condition-case}
8413 expression, when the string has content, then another conditional
8414 expression is executed. This is an @code{if} with both a then-part
8419 (if (eq last-command 'kill-region)
8420 (kill-append string (< end beg) yank-handler)
8421 (kill-new string nil yank-handler))
8425 The then-part is evaluated if the previous command was another call to
8426 @code{kill-region}; if not, the else-part is evaluated.
8428 @code{yank-handler} is an optional argument to @code{kill-region} that
8429 tells the @code{kill-append} and @code{kill-new} functions how deal
8430 with properties added to the text, such as `bold' or `italics'.
8432 @code{last-command} is a variable that comes with Emacs that we have
8433 not seen before. Normally, whenever a function is executed, Emacs
8434 sets the value of @code{last-command} to the previous command.
8437 In this segment of the definition, the @code{if} expression checks
8438 whether the previous command was @code{kill-region}. If it was,
8441 (kill-append string (< end beg) yank-handler)
8445 concatenates a copy of the newly clipped text to the just previously
8446 clipped text in the kill ring.
8448 @node copy-region-as-kill, Digression into C, kill-region, Cutting & Storing Text
8449 @comment node-name, next, previous, up
8450 @section @code{copy-region-as-kill}
8451 @findex copy-region-as-kill
8454 The @code{copy-region-as-kill} function copies a region of text from a
8455 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8456 in the @code{kill-ring}.
8458 If you call @code{copy-region-as-kill} immediately after a
8459 @code{kill-region} command, Emacs appends the newly copied text to the
8460 previously copied text. This means that if you yank back the text, you
8461 get it all, from both this and the previous operation. On the other
8462 hand, if some other command precedes the @code{copy-region-as-kill},
8463 the function copies the text into a separate entry in the kill ring.
8466 * Complete copy-region-as-kill:: The complete function definition.
8467 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8470 @node Complete copy-region-as-kill, copy-region-as-kill body, copy-region-as-kill, copy-region-as-kill
8472 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8476 Here is the complete text of the version 22 @code{copy-region-as-kill}
8481 (defun copy-region-as-kill (beg end)
8482 "Save the region as if killed, but don't kill it.
8483 In Transient Mark mode, deactivate the mark.
8484 If `interprogram-cut-function' is non-nil, also save the text for a window
8485 system cut and paste."
8489 (if (eq last-command 'kill-region)
8490 (kill-append (filter-buffer-substring beg end) (< end beg))
8491 (kill-new (filter-buffer-substring beg end)))
8494 (if transient-mark-mode
8495 (setq deactivate-mark t))
8501 As usual, this function can be divided into its component parts:
8505 (defun copy-region-as-kill (@var{argument-list})
8506 "@var{documentation}@dots{}"
8512 The arguments are @code{beg} and @code{end} and the function is
8513 interactive with @code{"r"}, so the two arguments must refer to the
8514 beginning and end of the region. If you have been reading though this
8515 document from the beginning, understanding these parts of a function is
8516 almost becoming routine.
8518 The documentation is somewhat confusing unless you remember that the
8519 word `kill' has a meaning different from usual. The `Transient Mark'
8520 and @code{interprogram-cut-function} comments explain certain
8523 After you once set a mark, a buffer always contains a region. If you
8524 wish, you can use Transient Mark mode to highlight the region
8525 temporarily. (No one wants to highlight the region all the time, so
8526 Transient Mark mode highlights it only at appropriate times. Many
8527 people turn off Transient Mark mode, so the region is never
8530 Also, a windowing system allows you to copy, cut, and paste among
8531 different programs. In the X windowing system, for example, the
8532 @code{interprogram-cut-function} function is @code{x-select-text},
8533 which works with the windowing system's equivalent of the Emacs kill
8536 The body of the @code{copy-region-as-kill} function starts with an
8537 @code{if} clause. What this clause does is distinguish between two
8538 different situations: whether or not this command is executed
8539 immediately after a previous @code{kill-region} command. In the first
8540 case, the new region is appended to the previously copied text.
8541 Otherwise, it is inserted into the beginning of the kill ring as a
8542 separate piece of text from the previous piece.
8544 The last two lines of the function prevent the region from lighting up
8545 if Transient Mark mode is turned on.
8547 The body of @code{copy-region-as-kill} merits discussion in detail.
8549 @node copy-region-as-kill body, , Complete copy-region-as-kill, copy-region-as-kill
8550 @comment node-name, next, previous, up
8551 @subsection The Body of @code{copy-region-as-kill}
8553 The @code{copy-region-as-kill} function works in much the same way as
8554 the @code{kill-region} function. Both are written so that two or more
8555 kills in a row combine their text into a single entry. If you yank
8556 back the text from the kill ring, you get it all in one piece.
8557 Moreover, kills that kill forward from the current position of the
8558 cursor are added to the end of the previously copied text and commands
8559 that copy text backwards add it to the beginning of the previously
8560 copied text. This way, the words in the text stay in the proper
8563 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8564 use of the @code{last-command} variable that keeps track of the
8565 previous Emacs command.
8568 * last-command & this-command::
8569 * kill-append function::
8570 * kill-new function::
8573 @node last-command & this-command, kill-append function, copy-region-as-kill body, copy-region-as-kill body
8575 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8578 Normally, whenever a function is executed, Emacs sets the value of
8579 @code{this-command} to the function being executed (which in this case
8580 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8581 the value of @code{last-command} to the previous value of
8582 @code{this-command}.
8584 In the first part of the body of the @code{copy-region-as-kill}
8585 function, an @code{if} expression determines whether the value of
8586 @code{last-command} is @code{kill-region}. If so, the then-part of
8587 the @code{if} expression is evaluated; it uses the @code{kill-append}
8588 function to concatenate the text copied at this call to the function
8589 with the text already in the first element (the @sc{car}) of the kill
8590 ring. On the other hand, if the value of @code{last-command} is not
8591 @code{kill-region}, then the @code{copy-region-as-kill} function
8592 attaches a new element to the kill ring using the @code{kill-new}
8596 The @code{if} expression reads as follows; it uses @code{eq}:
8600 (if (eq last-command 'kill-region)
8602 (kill-append (filter-buffer-substring beg end) (< end beg))
8604 (kill-new (filter-buffer-substring beg end)))
8608 @findex filter-buffer-substring
8609 (The @code{filter-buffer-substring} function returns a filtered
8610 substring of the buffer, if any. Optionally---the arguments are not
8611 here, so neither is done---the function may delete the initial text or
8612 return the text without its properties; this function is a replacement
8613 for the older @code{buffer-substring} function, which came before text
8614 properties were implemented.)
8616 @findex eq @r{(example of use)}
8618 The @code{eq} function tests whether its first argument is the same Lisp
8619 object as its second argument. The @code{eq} function is similar to the
8620 @code{equal} function in that it is used to test for equality, but
8621 differs in that it determines whether two representations are actually
8622 the same object inside the computer, but with different names.
8623 @code{equal} determines whether the structure and contents of two
8624 expressions are the same.
8626 If the previous command was @code{kill-region}, then the Emacs Lisp
8627 interpreter calls the @code{kill-append} function
8629 @node kill-append function, kill-new function, last-command & this-command, copy-region-as-kill body
8630 @unnumberedsubsubsec The @code{kill-append} function
8634 The @code{kill-append} function looks like this:
8639 (defun kill-append (string before-p &optional yank-handler)
8640 "Append STRING to the end of the latest kill in the kill ring.
8641 If BEFORE-P is non-nil, prepend STRING to the kill.
8643 (let* ((cur (car kill-ring)))
8644 (kill-new (if before-p (concat string cur) (concat cur string))
8645 (or (= (length cur) 0)
8647 (get-text-property 0 'yank-handler cur)))
8654 (defun kill-append (string before-p)
8655 "Append STRING to the end of the latest kill in the kill ring.
8656 If BEFORE-P is non-nil, prepend STRING to the kill.
8657 If `interprogram-cut-function' is set, pass the resulting kill to
8659 (kill-new (if before-p
8660 (concat string (car kill-ring))
8661 (concat (car kill-ring) string))
8666 The @code{kill-append} function is fairly straightforward. It uses
8667 the @code{kill-new} function, which we will discuss in more detail in
8670 (Also, the function provides an optional argument called
8671 @code{yank-handler}; when invoked, this argument tells the function
8672 how to deal with properties added to the text, such as `bold' or
8675 @c !!! bug in GNU Emacs 22 version of kill-append ?
8676 It has a @code{let*} function to set the value of the first element of
8677 the kill ring to @code{cur}. (I do not know why the function does not
8678 use @code{let} instead; only one value is set in the expression.
8679 Perhaps this is a bug that produces no problems?)
8681 Consider the conditional that is one of the two arguments to
8682 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8683 the @sc{car} of the kill ring. Whether it prepends or appends the
8684 text depends on the results of an @code{if} expression:
8688 (if before-p ; @r{if-part}
8689 (concat string cur) ; @r{then-part}
8690 (concat cur string)) ; @r{else-part}
8695 If the region being killed is before the region that was killed in the
8696 last command, then it should be prepended before the material that was
8697 saved in the previous kill; and conversely, if the killed text follows
8698 what was just killed, it should be appended after the previous text.
8699 The @code{if} expression depends on the predicate @code{before-p} to
8700 decide whether the newly saved text should be put before or after the
8701 previously saved text.
8703 The symbol @code{before-p} is the name of one of the arguments to
8704 @code{kill-append}. When the @code{kill-append} function is
8705 evaluated, it is bound to the value returned by evaluating the actual
8706 argument. In this case, this is the expression @code{(< end beg)}.
8707 This expression does not directly determine whether the killed text in
8708 this command is located before or after the kill text of the last
8709 command; what it does is determine whether the value of the variable
8710 @code{end} is less than the value of the variable @code{beg}. If it
8711 is, it means that the user is most likely heading towards the
8712 beginning of the buffer. Also, the result of evaluating the predicate
8713 expression, @code{(< end beg)}, will be true and the text will be
8714 prepended before the previous text. On the other hand, if the value of
8715 the variable @code{end} is greater than the value of the variable
8716 @code{beg}, the text will be appended after the previous text.
8719 When the newly saved text will be prepended, then the string with the new
8720 text will be concatenated before the old text:
8728 But if the text will be appended, it will be concatenated
8732 (concat cur string))
8735 To understand how this works, we first need to review the
8736 @code{concat} function. The @code{concat} function links together or
8737 unites two strings of text. The result is a string. For example:
8741 (concat "abc" "def")
8747 (car '("first element" "second element")))
8748 @result{} "new first element"
8751 '("first element" "second element")) " modified")
8752 @result{} "first element modified"
8756 We can now make sense of @code{kill-append}: it modifies the contents
8757 of the kill ring. The kill ring is a list, each element of which is
8758 saved text. The @code{kill-append} function uses the @code{kill-new}
8759 function which in turn uses the @code{setcar} function.
8761 @node kill-new function, , kill-append function, copy-region-as-kill body
8762 @unnumberedsubsubsec The @code{kill-new} function
8765 @c in GNU Emacs 22, additional documentation to kill-new:
8767 Optional third arguments YANK-HANDLER controls how the STRING is later
8768 inserted into a buffer; see `insert-for-yank' for details.
8769 When a yank handler is specified, STRING must be non-empty (the yank
8770 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8772 When the yank handler has a non-nil PARAM element, the original STRING
8773 argument is not used by `insert-for-yank'. However, since Lisp code
8774 may access and use elements from the kill ring directly, the STRING
8775 argument should still be a \"useful\" string for such uses."
8778 The @code{kill-new} function looks like this:
8782 (defun kill-new (string &optional replace yank-handler)
8783 "Make STRING the latest kill in the kill ring.
8784 Set `kill-ring-yank-pointer' to point to it.
8786 If `interprogram-cut-function' is non-nil, apply it to STRING.
8787 Optional second argument REPLACE non-nil means that STRING will replace
8788 the front of the kill ring, rather than being added to the list.
8792 (if (> (length string) 0)
8794 (put-text-property 0 (length string)
8795 'yank-handler yank-handler string))
8797 (signal 'args-out-of-range
8798 (list string "yank-handler specified for empty string"))))
8801 (if (fboundp 'menu-bar-update-yank-menu)
8802 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8805 (if (and replace kill-ring)
8806 (setcar kill-ring string)
8807 (push string kill-ring)
8808 (if (> (length kill-ring) kill-ring-max)
8809 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8812 (setq kill-ring-yank-pointer kill-ring)
8813 (if interprogram-cut-function
8814 (funcall interprogram-cut-function string (not replace))))
8819 (defun kill-new (string &optional replace)
8820 "Make STRING the latest kill in the kill ring.
8821 Set the kill-ring-yank pointer to point to it.
8822 If `interprogram-cut-function' is non-nil, apply it to STRING.
8823 Optional second argument REPLACE non-nil means that STRING will replace
8824 the front of the kill ring, rather than being added to the list."
8825 (and (fboundp 'menu-bar-update-yank-menu)
8826 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8827 (if (and replace kill-ring)
8828 (setcar kill-ring string)
8829 (setq kill-ring (cons string kill-ring))
8830 (if (> (length kill-ring) kill-ring-max)
8831 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8832 (setq kill-ring-yank-pointer kill-ring)
8833 (if interprogram-cut-function
8834 (funcall interprogram-cut-function string (not replace))))
8837 (Notice that the function is not interactive.)
8839 As usual, we can look at this function in parts.
8841 The function definition has an optional @code{yank-handler} argument,
8842 which when invoked tells the function how to deal with properties
8843 added to the text, such as `bold' or `italics'. We will skip that.
8846 The first line of the documentation makes sense:
8849 Make STRING the latest kill in the kill ring.
8853 Let's skip over the rest of the documentation for the moment.
8856 Also, let's skip over the initial @code{if} expression and those lines
8857 of code involving @code{menu-bar-update-yank-menu}. We will explain
8861 The critical lines are these:
8865 (if (and replace kill-ring)
8867 (setcar kill-ring string)
8871 (push string kill-ring)
8874 (setq kill-ring (cons string kill-ring))
8875 (if (> (length kill-ring) kill-ring-max)
8876 ;; @r{avoid overly long kill ring}
8877 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8880 (setq kill-ring-yank-pointer kill-ring)
8881 (if interprogram-cut-function
8882 (funcall interprogram-cut-function string (not replace))))
8886 The conditional test is @w{@code{(and replace kill-ring)}}.
8887 This will be true when two conditions are met: the kill ring has
8888 something in it, and the @code{replace} variable is true.
8891 When the @code{kill-append} function sets @code{replace} to be true
8892 and when the kill ring has at least one item in it, the @code{setcar}
8893 expression is executed:
8896 (setcar kill-ring string)
8899 The @code{setcar} function actually changes the first element of the
8900 @code{kill-ring} list to the value of @code{string}. It replaces the
8904 On the other hand, if the kill ring is empty, or replace is false, the
8905 else-part of the condition is executed:
8908 (push string kill-ring)
8913 @code{push} puts its first argument onto the second. It is similar to
8917 (setq kill-ring (cons string kill-ring))
8925 (add-to-list kill-ring string)
8929 When it is false, the expression first constructs a new version of the
8930 kill ring by prepending @code{string} to the existing kill ring as a
8931 new element (that is what the @code{push} does). Then it executes a
8932 second @code{if} clause. This second @code{if} clause keeps the kill
8933 ring from growing too long.
8935 Let's look at these two expressions in order.
8937 The @code{push} line of the else-part sets the new value of the kill
8938 ring to what results from adding the string being killed to the old
8941 We can see how this works with an example.
8947 (setq example-list '("here is a clause" "another clause"))
8952 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8953 @code{example-list} and see what it returns:
8958 @result{} ("here is a clause" "another clause")
8964 Now, we can add a new element on to this list by evaluating the
8965 following expression:
8966 @findex push, @r{example}
8969 (push "a third clause" example-list)
8974 When we evaluate @code{example-list}, we find its value is:
8979 @result{} ("a third clause" "here is a clause" "another clause")
8984 Thus, the third clause is added to the list by @code{push}.
8987 Now for the second part of the @code{if} clause. This expression
8988 keeps the kill ring from growing too long. It looks like this:
8992 (if (> (length kill-ring) kill-ring-max)
8993 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8997 The code checks whether the length of the kill ring is greater than
8998 the maximum permitted length. This is the value of
8999 @code{kill-ring-max} (which is 60, by default). If the length of the
9000 kill ring is too long, then this code sets the last element of the
9001 kill ring to @code{nil}. It does this by using two functions,
9002 @code{nthcdr} and @code{setcdr}.
9004 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
9005 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
9006 @sc{car} of a list. In this case, however, @code{setcdr} will not be
9007 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
9008 function is used to cause it to set the @sc{cdr} of the next to last
9009 element of the kill ring---this means that since the @sc{cdr} of the
9010 next to last element is the last element of the kill ring, it will set
9011 the last element of the kill ring.
9013 @findex nthcdr, @r{example}
9014 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
9015 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
9016 @dots{} It does this @var{N} times and returns the results.
9017 (@xref{nthcdr, , @code{nthcdr}}.)
9019 @findex setcdr, @r{example}
9020 Thus, if we had a four element list that was supposed to be three
9021 elements long, we could set the @sc{cdr} of the next to last element
9022 to @code{nil}, and thereby shorten the list. (If you set the last
9023 element to some other value than @code{nil}, which you could do, then
9024 you would not have shortened the list. @xref{setcdr, ,
9027 You can see shortening by evaluating the following three expressions
9028 in turn. First set the value of @code{trees} to @code{(maple oak pine
9029 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
9030 and then find the value of @code{trees}:
9034 (setq trees '(maple oak pine birch))
9035 @result{} (maple oak pine birch)
9039 (setcdr (nthcdr 2 trees) nil)
9043 @result{} (maple oak pine)
9048 (The value returned by the @code{setcdr} expression is @code{nil} since
9049 that is what the @sc{cdr} is set to.)
9051 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
9052 @sc{cdr} a number of times that is one less than the maximum permitted
9053 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
9054 element (which will be the rest of the elements in the kill ring) to
9055 @code{nil}. This prevents the kill ring from growing too long.
9058 The next to last expression in the @code{kill-new} function is
9061 (setq kill-ring-yank-pointer kill-ring)
9064 The @code{kill-ring-yank-pointer} is a global variable that is set to be
9065 the @code{kill-ring}.
9067 Even though the @code{kill-ring-yank-pointer} is called a
9068 @samp{pointer}, it is a variable just like the kill ring. However, the
9069 name has been chosen to help humans understand how the variable is used.
9072 Now, to return to an early expression in the body of the function:
9076 (if (fboundp 'menu-bar-update-yank-menu)
9077 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9082 It starts with an @code{if} expression
9084 In this case, the expression tests first to see whether
9085 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9086 calls it. The @code{fboundp} function returns true if the symbol it
9087 is testing has a function definition that `is not void'. If the
9088 symbol's function definition were void, we would receive an error
9089 message, as we did when we created errors intentionally (@pxref{Making
9090 Errors, , Generate an Error Message}).
9093 The then-part contains an expression whose first element is the
9094 function @code{and}.
9097 The @code{and} special form evaluates each of its arguments until one
9098 of the arguments returns a value of @code{nil}, in which case the
9099 @code{and} expression returns @code{nil}; however, if none of the
9100 arguments returns a value of @code{nil}, the value resulting from
9101 evaluating the last argument is returned. (Since such a value is not
9102 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9103 @code{and} expression returns a true value only if all its arguments
9104 are true. (@xref{Second Buffer Related Review}.)
9106 The expression determines whether the second argument to
9107 @code{menu-bar-update-yank-menu} is true or not.
9109 ;; If we're supposed to be extending an existing string, and that
9110 ;; string really is at the front of the menu, then update it in place.
9113 @code{menu-bar-update-yank-menu} is one of the functions that make it
9114 possible to use the `Select and Paste' menu in the Edit item of a menu
9115 bar; using a mouse, you can look at the various pieces of text you
9116 have saved and select one piece to paste.
9118 The last expression in the @code{kill-new} function adds the newly
9119 copied string to whatever facility exists for copying and pasting
9120 among different programs running in a windowing system. In the X
9121 Windowing system, for example, the @code{x-select-text} function takes
9122 the string and stores it in memory operated by X. You can paste the
9123 string in another program, such as an Xterm.
9126 The expression looks like this:
9130 (if interprogram-cut-function
9131 (funcall interprogram-cut-function string (not replace))))
9135 If an @code{interprogram-cut-function} exists, then Emacs executes
9136 @code{funcall}, which in turn calls its first argument as a function
9137 and passes the remaining arguments to it. (Incidentally, as far as I
9138 can see, this @code{if} expression could be replaced by an @code{and}
9139 expression similar to the one in the first part of the function.)
9141 We are not going to discuss windowing systems and other programs
9142 further, but merely note that this is a mechanism that enables GNU
9143 Emacs to work easily and well with other programs.
9145 This code for placing text in the kill ring, either concatenated with
9146 an existing element or as a new element, leads us to the code for
9147 bringing back text that has been cut out of the buffer---the yank
9148 commands. However, before discussing the yank commands, it is better
9149 to learn how lists are implemented in a computer. This will make
9150 clear such mysteries as the use of the term `pointer'. But before
9151 that, we will digress into C.
9154 @c is this true in Emacs 22? Does not seems to be
9156 (If the @w{@code{(< end beg))}}
9157 expression is true, @code{kill-append} prepends the string to the just
9158 previously clipped text. For a detailed discussion, see
9159 @ref{kill-append function, , The @code{kill-append} function}.)
9161 If you then yank back the text, i.e., `paste' it, you get both
9162 pieces of text at once. That way, if you delete two words in a row,
9163 and then yank them back, you get both words, in their proper order,
9164 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9167 On the other hand, if the previous command is not @code{kill-region},
9168 then the @code{kill-new} function is called, which adds the text to
9169 the kill ring as the latest item, and sets the
9170 @code{kill-ring-yank-pointer} variable to point to it.
9174 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9175 @c use the delete-and-extract-region function
9177 2006 Oct 26, the Digression into C is now OK but should come after
9178 copy-region-as-kill and filter-buffer-substring
9182 copy-region-as-kill is short, 12 lines, and uses
9183 filter-buffer-substring, which is longer, 39 lines
9184 and has delete-and-extract-region in it.
9185 delete-and-extract-region is written in C.
9187 see Initializing a Variable with @code{defvar}
9190 @node Digression into C, defvar, copy-region-as-kill, Cutting & Storing Text
9191 @comment node-name, next, previous, up
9192 @section Digression into C
9193 @findex delete-and-extract-region
9194 @cindex C, a digression into
9195 @cindex Digression into C
9197 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9198 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9199 function, which in turn uses the @code{delete-and-extract-region}
9200 function. It removes the contents of a region and you cannot get them
9203 Unlike the other code discussed here, the
9204 @code{delete-and-extract-region} function is not written in Emacs
9205 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9206 system. Since it is very simple, I will digress briefly from Lisp and
9209 @c GNU Emacs 22 in /usr/local/src/emacs/src/editfns.c
9210 @c the DEFUN for buffer-substring-no-properties
9213 Like many of the other Emacs primitives,
9214 @code{delete-and-extract-region} is written as an instance of a C
9215 macro, a macro being a template for code. The complete macro looks
9220 DEFUN ("buffer-substring-no-properties", Fbuffer_substring_no_properties,
9221 Sbuffer_substring_no_properties, 2, 2, 0,
9222 doc: /* Return the characters of part of the buffer,
9223 without the text properties.
9224 The two arguments START and END are character positions;
9225 they can be in either order. */)
9227 Lisp_Object start, end;
9231 validate_region (&start, &end);
9235 return make_buffer_string (b, e, 0);
9240 Without going into the details of the macro writing process, let me
9241 point out that this macro starts with the word @code{DEFUN}. The word
9242 @code{DEFUN} was chosen since the code serves the same purpose as
9243 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9244 @file{emacs/src/lisp.h}.)
9246 The word @code{DEFUN} is followed by seven parts inside of
9251 The first part is the name given to the function in Lisp,
9252 @code{delete-and-extract-region}.
9255 The second part is the name of the function in C,
9256 @code{Fdelete_and_extract_region}. By convention, it starts with
9257 @samp{F}. Since C does not use hyphens in names, underscores are used
9261 The third part is the name for the C constant structure that records
9262 information on this function for internal use. It is the name of the
9263 function in C but begins with an @samp{S} instead of an @samp{F}.
9266 The fourth and fifth parts specify the minimum and maximum number of
9267 arguments the function can have. This function demands exactly 2
9271 The sixth part is nearly like the argument that follows the
9272 @code{interactive} declaration in a function written in Lisp: a letter
9273 followed, perhaps, by a prompt. The only difference from the Lisp is
9274 when the macro is called with no arguments. Then you write a @code{0}
9275 (which is a `null string'), as in this macro.
9277 If you were to specify arguments, you would place them between
9278 quotation marks. The C macro for @code{goto-char} includes
9279 @code{"NGoto char: "} in this position to indicate that the function
9280 expects a raw prefix, in this case, a numerical location in a buffer,
9281 and provides a prompt.
9284 The seventh part is a documentation string, just like the one for a
9285 function written in Emacs Lisp, except that every newline must be
9286 written explicitly as @samp{\n} followed by a backslash and carriage
9290 Thus, the first two lines of documentation for @code{goto-char} are
9295 "Set point to POSITION, a number or marker.\n\
9296 Beginning of buffer is position (point-min), end is (point-max)."
9302 In a C macro, the formal parameters come next, with a statement of
9303 what kind of object they are, followed by what might be called the `body'
9304 of the macro. For @code{delete-and-extract-region} the `body'
9305 consists of the following four lines:
9309 validate_region (&start, &end);
9310 if (XINT (start) == XINT (end))
9311 return build_string ("");
9312 return del_range_1 (XINT (start), XINT (end), 1, 1);
9316 The @code{validate_region} function checks whether the values
9317 passed as the beginning and end of the region are the proper type and
9318 are within range. If the beginning and end positions are the same,
9319 then return and empty string.
9321 The @code{del_range_1} function actually deletes the text. It is a
9322 complex function we will not look into. It updates the buffer and
9323 does other things. However, it is worth looking at the two arguments
9324 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9325 @w{@code{XINT (end)}}.
9327 As far as the C language is concerned, @code{start} and @code{end} are
9328 two integers that mark the beginning and end of the region to be
9329 deleted@footnote{More precisely, and requiring more expert knowledge
9330 to understand, the two integers are of type `Lisp_Object', which can
9331 also be a C union instead of an integer type.}.
9333 In early versions of Emacs, these two numbers were thirty-two bits
9334 long, but the code is slowly being generalized to handle other
9335 lengths. Three of the available bits are used to specify the type of
9336 information; the remaining bits are used as `content'.
9338 @samp{XINT} is a C macro that extracts the relevant number from the
9339 longer collection of bits; the three other bits are discarded.
9342 The command in @code{delete-and-extract-region} looks like this:
9345 del_range_1 (XINT (start), XINT (end), 1, 1);
9349 It deletes the region between the beginning position, @code{start},
9350 and the ending position, @code{end}.
9352 From the point of view of the person writing Lisp, Emacs is all very
9353 simple; but hidden underneath is a great deal of complexity to make it
9356 @node defvar, cons & search-fwd Review, Digression into C, Cutting & Storing Text
9357 @comment node-name, next, previous, up
9358 @section Initializing a Variable with @code{defvar}
9360 @cindex Initializing a variable
9361 @cindex Variable initialization
9366 copy-region-as-kill is short, 12 lines, and uses
9367 filter-buffer-substring, which is longer, 39 lines
9368 and has delete-and-extract-region in it.
9369 delete-and-extract-region is written in C.
9371 see Initializing a Variable with @code{defvar}
9375 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9376 functions within it, @code{kill-append} and @code{kill-new}, copy a
9377 region in a buffer and save it in a variable called the
9378 @code{kill-ring}. This section describes how the @code{kill-ring}
9379 variable is created and initialized using the @code{defvar} special
9382 (Again we note that the term @code{kill-ring} is a misnomer. The text
9383 that is clipped out of the buffer can be brought back; it is not a ring
9384 of corpses, but a ring of resurrectable text.)
9386 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9387 given an initial value by using the @code{defvar} special form. The
9388 name comes from ``define variable''.
9390 The @code{defvar} special form is similar to @code{setq} in that it sets
9391 the value of a variable. It is unlike @code{setq} in two ways: first,
9392 it only sets the value of the variable if the variable does not already
9393 have a value. If the variable already has a value, @code{defvar} does
9394 not override the existing value. Second, @code{defvar} has a
9395 documentation string.
9397 (Another special form, @code{defcustom}, is designed for variables
9398 that people customize. It has more features than @code{defvar}.
9399 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9402 * See variable current value::
9403 * defvar and asterisk::
9406 @node See variable current value, defvar and asterisk, defvar, defvar
9408 @unnumberedsubsec Seeing the Current Value of a Variable
9411 You can see the current value of a variable, any variable, by using
9412 the @code{describe-variable} function, which is usually invoked by
9413 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9414 (followed by @key{RET}) when prompted, you will see what is in your
9415 current kill ring---this may be quite a lot! Conversely, if you have
9416 been doing nothing this Emacs session except read this document, you
9417 may have nothing in it. Also, you will see the documentation for
9423 List of killed text sequences.
9424 Since the kill ring is supposed to interact nicely with cut-and-paste
9425 facilities offered by window systems, use of this variable should
9428 interact nicely with `interprogram-cut-function' and
9429 `interprogram-paste-function'. The functions `kill-new',
9430 `kill-append', and `current-kill' are supposed to implement this
9431 interaction; you may want to use them instead of manipulating the kill
9437 The kill ring is defined by a @code{defvar} in the following way:
9441 (defvar kill-ring nil
9442 "List of killed text sequences.
9448 In this variable definition, the variable is given an initial value of
9449 @code{nil}, which makes sense, since if you have saved nothing, you want
9450 nothing back if you give a @code{yank} command. The documentation
9451 string is written just like the documentation string of a @code{defun}.
9452 As with the documentation string of the @code{defun}, the first line of
9453 the documentation should be a complete sentence, since some commands,
9454 like @code{apropos}, print only the first line of documentation.
9455 Succeeding lines should not be indented; otherwise they look odd when
9456 you use @kbd{C-h v} (@code{describe-variable}).
9458 @node defvar and asterisk, , See variable current value, defvar
9459 @subsection @code{defvar} and an asterisk
9460 @findex defvar @r{for a user customizable variable}
9461 @findex defvar @r{with an asterisk}
9463 In the past, Emacs used the @code{defvar} special form both for
9464 internal variables that you would not expect a user to change and for
9465 variables that you do expect a user to change. Although you can still
9466 use @code{defvar} for user customizable variables, please use
9467 @code{defcustom} instead, since that special form provides a path into
9468 the Customization commands. (@xref{defcustom, , Specifying Variables
9469 using @code{defcustom}}.)
9471 When you specified a variable using the @code{defvar} special form,
9472 you could distinguish a variable that a user might want to change from
9473 others by typing an asterisk, @samp{*}, in the first column of its
9474 documentation string. For example:
9478 (defvar shell-command-default-error-buffer nil
9479 "*Buffer name for `shell-command' @dots{} error output.
9484 @findex set-variable
9486 You could (and still can) use the @code{set-variable} command to
9487 change the value of @code{shell-command-default-error-buffer}
9488 temporarily. However, options set using @code{set-variable} are set
9489 only for the duration of your editing session. The new values are not
9490 saved between sessions. Each time Emacs starts, it reads the original
9491 value, unless you change the value within your @file{.emacs} file,
9492 either by setting it manually or by using @code{customize}.
9493 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9495 For me, the major use of the @code{set-variable} command is to suggest
9496 variables that I might want to set in my @file{.emacs} file. There
9497 are now more than 700 such variables --- far too many to remember
9498 readily. Fortunately, you can press @key{TAB} after calling the
9499 @code{M-x set-variable} command to see the list of variables.
9500 (@xref{Examining, , Examining and Setting Variables, emacs,
9501 The GNU Emacs Manual}.)
9504 @node cons & search-fwd Review, search Exercises, defvar, Cutting & Storing Text
9505 @comment node-name, next, previous, up
9508 Here is a brief summary of some recently introduced functions.
9513 @code{car} returns the first element of a list; @code{cdr} returns the
9514 second and subsequent elements of a list.
9521 (car '(1 2 3 4 5 6 7))
9523 (cdr '(1 2 3 4 5 6 7))
9524 @result{} (2 3 4 5 6 7)
9529 @code{cons} constructs a list by prepending its first argument to its
9543 @code{funcall} evaluates its first argument as a function. It passes
9544 its remaining arguments to its first argument.
9547 Return the result of taking @sc{cdr} `n' times on a list.
9555 The `rest of the rest', as it were.
9562 (nthcdr 3 '(1 2 3 4 5 6 7))
9569 @code{setcar} changes the first element of a list; @code{setcdr}
9570 changes the second and subsequent elements of a list.
9577 (setq triple '(1 2 3))
9584 (setcdr triple '("foo" "bar"))
9587 @result{} (37 "foo" "bar")
9592 Evaluate each argument in sequence and then return the value of the
9605 @item save-restriction
9606 Record whatever narrowing is in effect in the current buffer, if any,
9607 and restore that narrowing after evaluating the arguments.
9609 @item search-forward
9610 Search for a string, and if the string is found, move point. With a
9611 regular expression, use the similar @code{re-search-forward}.
9612 (@xref{Regexp Search, , Regular Expression Searches}, for an
9613 explanation of regular expression patterns and searches.)
9617 @code{search-forward} and @code{re-search-forward} take four
9622 The string or regular expression to search for.
9625 Optionally, the limit of the search.
9628 Optionally, what to do if the search fails, return @code{nil} or an
9632 Optionally, how many times to repeat the search; if negative, the
9633 search goes backwards.
9637 @itemx delete-and-extract-region
9638 @itemx copy-region-as-kill
9640 @code{kill-region} cuts the text between point and mark from the
9641 buffer and stores that text in the kill ring, so you can get it back
9644 @code{copy-region-as-kill} copies the text between point and mark into
9645 the kill ring, from which you can get it by yanking. The function
9646 does not cut or remove the text from the buffer.
9649 @code{delete-and-extract-region} removes the text between point and
9650 mark from the buffer and throws it away. You cannot get it back.
9651 (This is not an interactive command.)
9654 @node search Exercises, , cons & search-fwd Review, Cutting & Storing Text
9655 @section Searching Exercises
9659 Write an interactive function that searches for a string. If the
9660 search finds the string, leave point after it and display a message
9661 that says ``Found!''. (Do not use @code{search-forward} for the name
9662 of this function; if you do, you will overwrite the existing version of
9663 @code{search-forward} that comes with Emacs. Use a name such as
9664 @code{test-search} instead.)
9667 Write a function that prints the third element of the kill ring in the
9668 echo area, if any; if the kill ring does not contain a third element,
9669 print an appropriate message.
9672 @node List Implementation, Yanking, Cutting & Storing Text, Top
9673 @comment node-name, next, previous, up
9674 @chapter How Lists are Implemented
9675 @cindex Lists in a computer
9677 In Lisp, atoms are recorded in a straightforward fashion; if the
9678 implementation is not straightforward in practice, it is, nonetheless,
9679 straightforward in theory. The atom @samp{rose}, for example, is
9680 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9681 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9682 is equally simple, but it takes a moment to get used to the idea. A
9683 list is kept using a series of pairs of pointers. In the series, the
9684 first pointer in each pair points to an atom or to another list, and the
9685 second pointer in each pair points to the next pair, or to the symbol
9686 @code{nil}, which marks the end of the list.
9688 A pointer itself is quite simply the electronic address of what is
9689 pointed to. Hence, a list is kept as a series of electronic addresses.
9692 * Lists diagrammed::
9693 * Symbols as Chest:: Exploring a powerful metaphor.
9697 @node Lists diagrammed, Symbols as Chest, List Implementation, List Implementation
9699 @unnumberedsec Lists diagrammed
9702 For example, the list @code{(rose violet buttercup)} has three elements,
9703 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9704 electronic address of @samp{rose} is recorded in a segment of computer
9705 memory along with the address that gives the electronic address of where
9706 the atom @samp{violet} is located; and that address (the one that tells
9707 where @samp{violet} is located) is kept along with an address that tells
9708 where the address for the atom @samp{buttercup} is located.
9711 This sounds more complicated than it is and is easier seen in a diagram:
9713 @c clear print-postscript-figures
9714 @c !!! cons-cell-diagram #1
9718 ___ ___ ___ ___ ___ ___
9719 |___|___|--> |___|___|--> |___|___|--> nil
9722 --> rose --> violet --> buttercup
9726 @ifset print-postscript-figures
9729 @center @image{cons-1}
9730 %%%% old method of including an image
9731 % \input /usr/local/lib/tex/inputs/psfig.tex
9732 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9737 @ifclear print-postscript-figures
9741 ___ ___ ___ ___ ___ ___
9742 |___|___|--> |___|___|--> |___|___|--> nil
9745 --> rose --> violet --> buttercup
9752 In the diagram, each box represents a word of computer memory that
9753 holds a Lisp object, usually in the form of a memory address. The boxes,
9754 i.e.@: the addresses, are in pairs. Each arrow points to what the address
9755 is the address of, either an atom or another pair of addresses. The
9756 first box is the electronic address of @samp{rose} and the arrow points
9757 to @samp{rose}; the second box is the address of the next pair of boxes,
9758 the first part of which is the address of @samp{violet} and the second
9759 part of which is the address of the next pair. The very last box
9760 points to the symbol @code{nil}, which marks the end of the list.
9763 When a variable is set to a list with a function such as @code{setq},
9764 it stores the address of the first box in the variable. Thus,
9765 evaluation of the expression
9768 (setq bouquet '(rose violet buttercup))
9773 creates a situation like this:
9775 @c cons-cell-diagram #2
9781 | ___ ___ ___ ___ ___ ___
9782 --> |___|___|--> |___|___|--> |___|___|--> nil
9785 --> rose --> violet --> buttercup
9789 @ifset print-postscript-figures
9792 @center @image{cons-2}
9793 %%%% old method of including an image
9794 % \input /usr/local/lib/tex/inputs/psfig.tex
9795 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9800 @ifclear print-postscript-figures
9806 | ___ ___ ___ ___ ___ ___
9807 --> |___|___|--> |___|___|--> |___|___|--> nil
9810 --> rose --> violet --> buttercup
9817 In this example, the symbol @code{bouquet} holds the address of the first
9821 This same list can be illustrated in a different sort of box notation
9824 @c cons-cell-diagram #2a
9830 | -------------- --------------- ----------------
9831 | | car | cdr | | car | cdr | | car | cdr |
9832 -->| rose | o------->| violet | o------->| butter- | nil |
9833 | | | | | | | cup | |
9834 -------------- --------------- ----------------
9838 @ifset print-postscript-figures
9841 @center @image{cons-2a}
9842 %%%% old method of including an image
9843 % \input /usr/local/lib/tex/inputs/psfig.tex
9844 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9849 @ifclear print-postscript-figures
9855 | -------------- --------------- ----------------
9856 | | car | cdr | | car | cdr | | car | cdr |
9857 -->| rose | o------->| violet | o------->| butter- | nil |
9858 | | | | | | | cup | |
9859 -------------- --------------- ----------------
9865 (Symbols consist of more than pairs of addresses, but the structure of
9866 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9867 consists of a group of address-boxes, one of which is the address of
9868 the printed word @samp{bouquet}, a second of which is the address of a
9869 function definition attached to the symbol, if any, a third of which
9870 is the address of the first pair of address-boxes for the list
9871 @code{(rose violet buttercup)}, and so on. Here we are showing that
9872 the symbol's third address-box points to the first pair of
9873 address-boxes for the list.)
9875 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9876 changed; the symbol simply has an address further down the list. (In
9877 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9878 evaluation of the following expression
9881 (setq flowers (cdr bouquet))
9888 @c cons-cell-diagram #3
9895 | ___ ___ | ___ ___ ___ ___
9896 --> | | | --> | | | | | |
9897 |___|___|----> |___|___|--> |___|___|--> nil
9900 --> rose --> violet --> buttercup
9905 @ifset print-postscript-figures
9908 @center @image{cons-3}
9909 %%%% old method of including an image
9910 % \input /usr/local/lib/tex/inputs/psfig.tex
9911 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9916 @ifclear print-postscript-figures
9923 | ___ ___ | ___ ___ ___ ___
9924 --> | | | --> | | | | | |
9925 |___|___|----> |___|___|--> |___|___|--> nil
9928 --> rose --> violet --> buttercup
9936 The value of @code{flowers} is @code{(violet buttercup)}, which is
9937 to say, the symbol @code{flowers} holds the address of the pair of
9938 address-boxes, the first of which holds the address of @code{violet},
9939 and the second of which holds the address of @code{buttercup}.
9941 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9942 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9943 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9944 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9945 information about cons cells and dotted pairs.
9948 The function @code{cons} adds a new pair of addresses to the front of
9949 a series of addresses like that shown above. For example, evaluating
9953 (setq bouquet (cons 'lily bouquet))
9960 @c cons-cell-diagram #4
9967 | ___ ___ ___ ___ | ___ ___ ___ ___
9968 --> | | | | | | --> | | | | | |
9969 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9972 --> lily --> rose --> violet --> buttercup
9977 @ifset print-postscript-figures
9980 @center @image{cons-4}
9981 %%%% old method of including an image
9982 % \input /usr/local/lib/tex/inputs/psfig.tex
9983 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9988 @ifclear print-postscript-figures
9995 | ___ ___ ___ ___ | ___ ___ ___ ___
9996 --> | | | | | | --> | | | | | |
9997 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
10000 --> lily --> rose --> violet --> buttercup
10009 However, this does not change the value of the symbol
10010 @code{flowers}, as you can see by evaluating the following,
10013 (eq (cdr (cdr bouquet)) flowers)
10017 which returns @code{t} for true.
10019 Until it is reset, @code{flowers} still has the value
10020 @code{(violet buttercup)}; that is, it has the address of the cons
10021 cell whose first address is of @code{violet}. Also, this does not
10022 alter any of the pre-existing cons cells; they are all still there.
10024 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
10025 of the next cons cell in the series; to get the @sc{car} of a list,
10026 you get the address of the first element of the list; to @code{cons} a
10027 new element on a list, you add a new cons cell to the front of the list.
10028 That is all there is to it! The underlying structure of Lisp is
10029 brilliantly simple!
10031 And what does the last address in a series of cons cells refer to? It
10032 is the address of the empty list, of @code{nil}.
10034 In summary, when a Lisp variable is set to a value, it is provided with
10035 the address of the list to which the variable refers.
10037 @node Symbols as Chest, List Exercise, Lists diagrammed, List Implementation
10038 @section Symbols as a Chest of Drawers
10039 @cindex Symbols as a Chest of Drawers
10040 @cindex Chest of Drawers, metaphor for a symbol
10041 @cindex Drawers, Chest of, metaphor for a symbol
10043 In an earlier section, I suggested that you might imagine a symbol as
10044 being a chest of drawers. The function definition is put in one
10045 drawer, the value in another, and so on. What is put in the drawer
10046 holding the value can be changed without affecting the contents of the
10047 drawer holding the function definition, and vice-verse.
10049 Actually, what is put in each drawer is the address of the value or
10050 function definition. It is as if you found an old chest in the attic,
10051 and in one of its drawers you found a map giving you directions to
10052 where the buried treasure lies.
10054 (In addition to its name, symbol definition, and variable value, a
10055 symbol has a `drawer' for a @dfn{property list} which can be used to
10056 record other information. Property lists are not discussed here; see
10057 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
10058 Reference Manual}.)
10061 Here is a fanciful representation:
10063 @c chest-of-drawers diagram
10068 Chest of Drawers Contents of Drawers
10072 ---------------------
10073 | directions to | [map to]
10074 | symbol name | bouquet
10076 +---------------------+
10078 | symbol definition | [none]
10080 +---------------------+
10081 | directions to | [map to]
10082 | variable value | (rose violet buttercup)
10084 +---------------------+
10086 | property list | [not described here]
10088 +---------------------+
10094 @ifset print-postscript-figures
10097 @center @image{drawers}
10098 %%%% old method of including an image
10099 % \input /usr/local/lib/tex/inputs/psfig.tex
10100 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10105 @ifclear print-postscript-figures
10110 Chest of Drawers Contents of Drawers
10114 ---------------------
10115 | directions to | [map to]
10116 | symbol name | bouquet
10118 +---------------------+
10120 | symbol definition | [none]
10122 +---------------------+
10123 | directions to | [map to]
10124 | variable value | (rose violet buttercup)
10126 +---------------------+
10128 | property list | [not described here]
10130 +---------------------+
10138 @node List Exercise, , Symbols as Chest, List Implementation
10141 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10142 more flowers on to this list and set this new list to
10143 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10144 What does the @code{more-flowers} list now contain?
10146 @node Yanking, Loops & Recursion, List Implementation, Top
10147 @comment node-name, next, previous, up
10148 @chapter Yanking Text Back
10150 @cindex Text retrieval
10151 @cindex Retrieving text
10152 @cindex Pasting text
10154 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10155 you can bring it back with a `yank' command. The text that is cut out of
10156 the buffer is put in the kill ring and the yank commands insert the
10157 appropriate contents of the kill ring back into a buffer (not necessarily
10158 the original buffer).
10160 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10161 the kill ring into the current buffer. If the @kbd{C-y} command is
10162 followed immediately by @kbd{M-y}, the first element is replaced by
10163 the second element. Successive @kbd{M-y} commands replace the second
10164 element with the third, fourth, or fifth element, and so on. When the
10165 last element in the kill ring is reached, it is replaced by the first
10166 element and the cycle is repeated. (Thus the kill ring is called a
10167 `ring' rather than just a `list'. However, the actual data structure
10168 that holds the text is a list.
10169 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10170 list is handled as a ring.)
10173 * Kill Ring Overview::
10174 * kill-ring-yank-pointer:: The kill ring is a list.
10175 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
10178 @node Kill Ring Overview, kill-ring-yank-pointer, Yanking, Yanking
10179 @comment node-name, next, previous, up
10180 @section Kill Ring Overview
10181 @cindex Kill ring overview
10183 The kill ring is a list of textual strings. This is what it looks like:
10186 ("some text" "a different piece of text" "yet more text")
10189 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10190 string of characters saying @samp{some text} would be inserted in this
10191 buffer where my cursor is located.
10193 The @code{yank} command is also used for duplicating text by copying it.
10194 The copied text is not cut from the buffer, but a copy of it is put on the
10195 kill ring and is inserted by yanking it back.
10197 Three functions are used for bringing text back from the kill ring:
10198 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10199 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10200 which is used by the two other functions.
10202 These functions refer to the kill ring through a variable called the
10203 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10204 @code{yank} and @code{yank-pop} functions is:
10207 (insert (car kill-ring-yank-pointer))
10211 (Well, no more. In GNU Emacs 22, the function has been replaced by
10212 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10213 repetitively for each @code{yank-handler} segment. In turn,
10214 @code{insert-for-yank-1} strips text properties from the inserted text
10215 according to @code{yank-excluded-properties}. Otherwise, it is just
10216 like @code{insert}. We will stick with plain @code{insert} since it
10217 is easier to understand.)
10219 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10220 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10222 @node kill-ring-yank-pointer, yank nthcdr Exercises, Kill Ring Overview, Yanking
10223 @comment node-name, next, previous, up
10224 @section The @code{kill-ring-yank-pointer} Variable
10226 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10227 a variable. It points to something by being bound to the value of what
10228 it points to, like any other Lisp variable.
10231 Thus, if the value of the kill ring is:
10234 ("some text" "a different piece of text" "yet more text")
10239 and the @code{kill-ring-yank-pointer} points to the second clause, the
10240 value of @code{kill-ring-yank-pointer} is:
10243 ("a different piece of text" "yet more text")
10246 As explained in the previous chapter (@pxref{List Implementation}), the
10247 computer does not keep two different copies of the text being pointed to
10248 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10249 words ``a different piece of text'' and ``yet more text'' are not
10250 duplicated. Instead, the two Lisp variables point to the same pieces of
10251 text. Here is a diagram:
10253 @c cons-cell-diagram #5
10257 kill-ring kill-ring-yank-pointer
10259 | ___ ___ | ___ ___ ___ ___
10260 ---> | | | --> | | | | | |
10261 |___|___|----> |___|___|--> |___|___|--> nil
10264 | | --> "yet more text"
10266 | --> "a different piece of text"
10273 @ifset print-postscript-figures
10276 @center @image{cons-5}
10277 %%%% old method of including an image
10278 % \input /usr/local/lib/tex/inputs/psfig.tex
10279 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10284 @ifclear print-postscript-figures
10288 kill-ring kill-ring-yank-pointer
10290 | ___ ___ | ___ ___ ___ ___
10291 ---> | | | --> | | | | | |
10292 |___|___|----> |___|___|--> |___|___|--> nil
10295 | | --> "yet more text"
10297 | --> "a different piece of text
10306 Both the variable @code{kill-ring} and the variable
10307 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10308 usually described as if it were actually what it is composed of. The
10309 @code{kill-ring} is spoken of as if it were the list rather than that it
10310 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10311 spoken of as pointing to a list.
10313 These two ways of talking about the same thing sound confusing at first but
10314 make sense on reflection. The kill ring is generally thought of as the
10315 complete structure of data that holds the information of what has recently
10316 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10317 on the other hand, serves to indicate---that is, to `point to'---that part
10318 of the kill ring of which the first element (the @sc{car}) will be
10322 In GNU Emacs 22, the @code{kill-new} function calls
10324 @code{(setq kill-ring-yank-pointer kill-ring)}
10326 (defun rotate-yank-pointer (arg)
10327 "Rotate the yanking point in the kill ring.
10328 With argument, rotate that many kills forward (or backward, if negative)."
10330 (current-kill arg))
10332 (defun current-kill (n &optional do-not-move)
10333 "Rotate the yanking point by N places, and then return that kill.
10334 If N is zero, `interprogram-paste-function' is set, and calling it
10335 returns a string, then that string is added to the front of the
10336 kill ring and returned as the latest kill.
10337 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10338 yanking point; just return the Nth kill forward."
10339 (let ((interprogram-paste (and (= n 0)
10340 interprogram-paste-function
10341 (funcall interprogram-paste-function))))
10342 (if interprogram-paste
10344 ;; Disable the interprogram cut function when we add the new
10345 ;; text to the kill ring, so Emacs doesn't try to own the
10346 ;; selection, with identical text.
10347 (let ((interprogram-cut-function nil))
10348 (kill-new interprogram-paste))
10349 interprogram-paste)
10350 (or kill-ring (error "Kill ring is empty"))
10351 (let ((ARGth-kill-element
10352 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10353 (length kill-ring))
10356 (setq kill-ring-yank-pointer ARGth-kill-element))
10357 (car ARGth-kill-element)))))
10362 @node yank nthcdr Exercises, , kill-ring-yank-pointer, Yanking
10363 @section Exercises with @code{yank} and @code{nthcdr}
10367 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10368 your kill ring. Add several items to your kill ring; look at its
10369 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10370 around the kill ring. How many items were in your kill ring? Find
10371 the value of @code{kill-ring-max}. Was your kill ring full, or could
10372 you have kept more blocks of text within it?
10375 Using @code{nthcdr} and @code{car}, construct a series of expressions
10376 to return the first, second, third, and fourth elements of a list.
10379 @node Loops & Recursion, Regexp Search, Yanking, Top
10380 @comment node-name, next, previous, up
10381 @chapter Loops and Recursion
10382 @cindex Loops and recursion
10383 @cindex Recursion and loops
10384 @cindex Repetition (loops)
10386 Emacs Lisp has two primary ways to cause an expression, or a series of
10387 expressions, to be evaluated repeatedly: one uses a @code{while}
10388 loop, and the other uses @dfn{recursion}.
10390 Repetition can be very valuable. For example, to move forward four
10391 sentences, you need only write a program that will move forward one
10392 sentence and then repeat the process four times. Since a computer does
10393 not get bored or tired, such repetitive action does not have the
10394 deleterious effects that excessive or the wrong kinds of repetition can
10397 People mostly write Emacs Lisp functions using @code{while} loops and
10398 their kin; but you can use recursion, which provides a very powerful
10399 way to think about and then to solve problems@footnote{You can write
10400 recursive functions to be frugal or wasteful of mental or computer
10401 resources; as it happens, methods that people find easy---that are
10402 frugal of `mental resources'---sometimes use considerable computer
10403 resources. Emacs was designed to run on machines that we now consider
10404 limited and its default settings are conservative. You may want to
10405 increase the values of @code{max-specpdl-size} and
10406 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10407 15 and 30 times their default value.}.
10410 * while:: Causing a stretch of code to repeat.
10412 * Recursion:: Causing a function to call itself.
10413 * Looping exercise::
10416 @node while, dolist dotimes, Loops & Recursion, Loops & Recursion
10417 @comment node-name, next, previous, up
10418 @section @code{while}
10422 The @code{while} special form tests whether the value returned by
10423 evaluating its first argument is true or false. This is similar to what
10424 the Lisp interpreter does with an @code{if}; what the interpreter does
10425 next, however, is different.
10427 In a @code{while} expression, if the value returned by evaluating the
10428 first argument is false, the Lisp interpreter skips the rest of the
10429 expression (the @dfn{body} of the expression) and does not evaluate it.
10430 However, if the value is true, the Lisp interpreter evaluates the body
10431 of the expression and then again tests whether the first argument to
10432 @code{while} is true or false. If the value returned by evaluating the
10433 first argument is again true, the Lisp interpreter again evaluates the
10434 body of the expression.
10437 The template for a @code{while} expression looks like this:
10441 (while @var{true-or-false-test}
10447 * Looping with while:: Repeat so long as test returns true.
10448 * Loop Example:: A @code{while} loop that uses a list.
10449 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10450 * Incrementing Loop:: A loop with an incrementing counter.
10451 * Incrementing Loop Details::
10452 * Decrementing Loop:: A loop with a decrementing counter.
10455 @node Looping with while, Loop Example, while, while
10457 @unnumberedsubsec Looping with @code{while}
10460 So long as the true-or-false-test of the @code{while} expression
10461 returns a true value when it is evaluated, the body is repeatedly
10462 evaluated. This process is called a loop since the Lisp interpreter
10463 repeats the same thing again and again, like an airplane doing a loop.
10464 When the result of evaluating the true-or-false-test is false, the
10465 Lisp interpreter does not evaluate the rest of the @code{while}
10466 expression and `exits the loop'.
10468 Clearly, if the value returned by evaluating the first argument to
10469 @code{while} is always true, the body following will be evaluated
10470 again and again @dots{} and again @dots{} forever. Conversely, if the
10471 value returned is never true, the expressions in the body will never
10472 be evaluated. The craft of writing a @code{while} loop consists of
10473 choosing a mechanism such that the true-or-false-test returns true
10474 just the number of times that you want the subsequent expressions to
10475 be evaluated, and then have the test return false.
10477 The value returned by evaluating a @code{while} is the value of the
10478 true-or-false-test. An interesting consequence of this is that a
10479 @code{while} loop that evaluates without error will return @code{nil}
10480 or false regardless of whether it has looped 1 or 100 times or none at
10481 all. A @code{while} expression that evaluates successfully never
10482 returns a true value! What this means is that @code{while} is always
10483 evaluated for its side effects, which is to say, the consequences of
10484 evaluating the expressions within the body of the @code{while} loop.
10485 This makes sense. It is not the mere act of looping that is desired,
10486 but the consequences of what happens when the expressions in the loop
10487 are repeatedly evaluated.
10489 @node Loop Example, print-elements-of-list, Looping with while, while
10490 @comment node-name, next, previous, up
10491 @subsection A @code{while} Loop and a List
10493 A common way to control a @code{while} loop is to test whether a list
10494 has any elements. If it does, the loop is repeated; but if it does not,
10495 the repetition is ended. Since this is an important technique, we will
10496 create a short example to illustrate it.
10498 A simple way to test whether a list has elements is to evaluate the
10499 list: if it has no elements, it is an empty list and will return the
10500 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10501 the other hand, a list with elements will return those elements when it
10502 is evaluated. Since Emacs Lisp considers as true any value that is not
10503 @code{nil}, a list that returns elements will test true in a
10507 For example, you can set the variable @code{empty-list} to @code{nil} by
10508 evaluating the following @code{setq} expression:
10511 (setq empty-list ())
10515 After evaluating the @code{setq} expression, you can evaluate the
10516 variable @code{empty-list} in the usual way, by placing the cursor after
10517 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10524 On the other hand, if you set a variable to be a list with elements, the
10525 list will appear when you evaluate the variable, as you can see by
10526 evaluating the following two expressions:
10530 (setq animals '(gazelle giraffe lion tiger))
10536 Thus, to create a @code{while} loop that tests whether there are any
10537 items in the list @code{animals}, the first part of the loop will be
10548 When the @code{while} tests its first argument, the variable
10549 @code{animals} is evaluated. It returns a list. So long as the list
10550 has elements, the @code{while} considers the results of the test to be
10551 true; but when the list is empty, it considers the results of the test
10554 To prevent the @code{while} loop from running forever, some mechanism
10555 needs to be provided to empty the list eventually. An oft-used
10556 technique is to have one of the subsequent forms in the @code{while}
10557 expression set the value of the list to be the @sc{cdr} of the list.
10558 Each time the @code{cdr} function is evaluated, the list will be made
10559 shorter, until eventually only the empty list will be left. At this
10560 point, the test of the @code{while} loop will return false, and the
10561 arguments to the @code{while} will no longer be evaluated.
10563 For example, the list of animals bound to the variable @code{animals}
10564 can be set to be the @sc{cdr} of the original list with the
10565 following expression:
10568 (setq animals (cdr animals))
10572 If you have evaluated the previous expressions and then evaluate this
10573 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10574 area. If you evaluate the expression again, @code{(lion tiger)} will
10575 appear in the echo area. If you evaluate it again and yet again,
10576 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10578 A template for a @code{while} loop that uses the @code{cdr} function
10579 repeatedly to cause the true-or-false-test eventually to test false
10584 (while @var{test-whether-list-is-empty}
10586 @var{set-list-to-cdr-of-list})
10590 This test and use of @code{cdr} can be put together in a function that
10591 goes through a list and prints each element of the list on a line of its
10594 @node print-elements-of-list, Incrementing Loop, Loop Example, while
10595 @subsection An Example: @code{print-elements-of-list}
10596 @findex print-elements-of-list
10598 The @code{print-elements-of-list} function illustrates a @code{while}
10601 @cindex @file{*scratch*} buffer
10602 The function requires several lines for its output. If you are
10603 reading this in a recent instance of GNU Emacs,
10604 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10605 you can evaluate the following expression inside of Info, as usual.
10607 If you are using an earlier version of Emacs, you need to copy the
10608 necessary expressions to your @file{*scratch*} buffer and evaluate
10609 them there. This is because the echo area had only one line in the
10612 You can copy the expressions by marking the beginning of the region
10613 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10614 the end of the region and then copying the region using @kbd{M-w}
10615 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10616 then provides visual feedback). In the @file{*scratch*}
10617 buffer, you can yank the expressions back by typing @kbd{C-y}
10620 After you have copied the expressions to the @file{*scratch*} buffer,
10621 evaluate each expression in turn. Be sure to evaluate the last
10622 expression, @code{(print-elements-of-list animals)}, by typing
10623 @kbd{C-u C-x C-e}, that is, by giving an argument to
10624 @code{eval-last-sexp}. This will cause the result of the evaluation
10625 to be printed in the @file{*scratch*} buffer instead of being printed
10626 in the echo area. (Otherwise you will see something like this in your
10627 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10628 each @samp{^J} stands for a `newline'.)
10631 In a recent instance of GNU Emacs, you can evaluate these expressions
10632 directly in the Info buffer, and the echo area will grow to show the
10637 (setq animals '(gazelle giraffe lion tiger))
10639 (defun print-elements-of-list (list)
10640 "Print each element of LIST on a line of its own."
10643 (setq list (cdr list))))
10645 (print-elements-of-list animals)
10651 When you evaluate the three expressions in sequence, you will see
10667 Each element of the list is printed on a line of its own (that is what
10668 the function @code{print} does) and then the value returned by the
10669 function is printed. Since the last expression in the function is the
10670 @code{while} loop, and since @code{while} loops always return
10671 @code{nil}, a @code{nil} is printed after the last element of the list.
10673 @node Incrementing Loop, Incrementing Loop Details, print-elements-of-list, while
10674 @comment node-name, next, previous, up
10675 @subsection A Loop with an Incrementing Counter
10677 A loop is not useful unless it stops when it ought. Besides
10678 controlling a loop with a list, a common way of stopping a loop is to
10679 write the first argument as a test that returns false when the correct
10680 number of repetitions are complete. This means that the loop must
10681 have a counter---an expression that counts how many times the loop
10684 @node Incrementing Loop Details, Decrementing Loop, Incrementing Loop, while
10686 @unnumberedsubsec Details of an Incrementing Loop
10689 The test for a loop with an incrementing counter can be an expression
10690 such as @code{(< count desired-number)} which returns @code{t} for
10691 true if the value of @code{count} is less than the
10692 @code{desired-number} of repetitions and @code{nil} for false if the
10693 value of @code{count} is equal to or is greater than the
10694 @code{desired-number}. The expression that increments the count can
10695 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10696 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10697 argument. (The expression @w{@code{(1+ count)}} has the same result
10698 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10701 The template for a @code{while} loop controlled by an incrementing
10702 counter looks like this:
10706 @var{set-count-to-initial-value}
10707 (while (< count desired-number) ; @r{true-or-false-test}
10709 (setq count (1+ count))) ; @r{incrementer}
10714 Note that you need to set the initial value of @code{count}; usually it
10718 * Incrementing Example:: Counting pebbles in a triangle.
10719 * Inc Example parts:: The parts of the function definition.
10720 * Inc Example altogether:: Putting the function definition together.
10723 @node Incrementing Example, Inc Example parts, Incrementing Loop Details, Incrementing Loop Details
10724 @unnumberedsubsubsec Example with incrementing counter
10726 Suppose you are playing on the beach and decide to make a triangle of
10727 pebbles, putting one pebble in the first row, two in the second row,
10728 three in the third row and so on, like this:
10746 @bullet{} @bullet{}
10747 @bullet{} @bullet{} @bullet{}
10748 @bullet{} @bullet{} @bullet{} @bullet{}
10755 (About 2500 years ago, Pythagoras and others developed the beginnings of
10756 number theory by considering questions such as this.)
10758 Suppose you want to know how many pebbles you will need to make a
10759 triangle with 7 rows?
10761 Clearly, what you need to do is add up the numbers from 1 to 7. There
10762 are two ways to do this; start with the smallest number, one, and add up
10763 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10764 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10765 mechanisms illustrate common ways of writing @code{while} loops, we will
10766 create two examples, one counting up and the other counting down. In
10767 this first example, we will start with 1 and add 2, 3, 4 and so on.
10769 If you are just adding up a short list of numbers, the easiest way to do
10770 it is to add up all the numbers at once. However, if you do not know
10771 ahead of time how many numbers your list will have, or if you want to be
10772 prepared for a very long list, then you need to design your addition so
10773 that what you do is repeat a simple process many times instead of doing
10774 a more complex process once.
10776 For example, instead of adding up all the pebbles all at once, what you
10777 can do is add the number of pebbles in the first row, 1, to the number
10778 in the second row, 2, and then add the total of those two rows to the
10779 third row, 3. Then you can add the number in the fourth row, 4, to the
10780 total of the first three rows; and so on.
10782 The critical characteristic of the process is that each repetitive
10783 action is simple. In this case, at each step we add only two numbers,
10784 the number of pebbles in the row and the total already found. This
10785 process of adding two numbers is repeated again and again until the last
10786 row has been added to the total of all the preceding rows. In a more
10787 complex loop the repetitive action might not be so simple, but it will
10788 be simpler than doing everything all at once.
10790 @node Inc Example parts, Inc Example altogether, Incrementing Example, Incrementing Loop Details
10791 @unnumberedsubsubsec The parts of the function definition
10793 The preceding analysis gives us the bones of our function definition:
10794 first, we will need a variable that we can call @code{total} that will
10795 be the total number of pebbles. This will be the value returned by
10798 Second, we know that the function will require an argument: this
10799 argument will be the total number of rows in the triangle. It can be
10800 called @code{number-of-rows}.
10802 Finally, we need a variable to use as a counter. We could call this
10803 variable @code{counter}, but a better name is @code{row-number}. That
10804 is because what the counter does in this function is count rows, and a
10805 program should be written to be as understandable as possible.
10807 When the Lisp interpreter first starts evaluating the expressions in the
10808 function, the value of @code{total} should be set to zero, since we have
10809 not added anything to it. Then the function should add the number of
10810 pebbles in the first row to the total, and then add the number of
10811 pebbles in the second to the total, and then add the number of
10812 pebbles in the third row to the total, and so on, until there are no
10813 more rows left to add.
10815 Both @code{total} and @code{row-number} are used only inside the
10816 function, so they can be declared as local variables with @code{let}
10817 and given initial values. Clearly, the initial value for @code{total}
10818 should be 0. The initial value of @code{row-number} should be 1,
10819 since we start with the first row. This means that the @code{let}
10820 statement will look like this:
10830 After the internal variables are declared and bound to their initial
10831 values, we can begin the @code{while} loop. The expression that serves
10832 as the test should return a value of @code{t} for true so long as the
10833 @code{row-number} is less than or equal to the @code{number-of-rows}.
10834 (If the expression tests true only so long as the row number is less
10835 than the number of rows in the triangle, the last row will never be
10836 added to the total; hence the row number has to be either less than or
10837 equal to the number of rows.)
10840 @findex <= @r{(less than or equal)}
10841 Lisp provides the @code{<=} function that returns true if the value of
10842 its first argument is less than or equal to the value of its second
10843 argument and false otherwise. So the expression that the @code{while}
10844 will evaluate as its test should look like this:
10847 (<= row-number number-of-rows)
10850 The total number of pebbles can be found by repeatedly adding the number
10851 of pebbles in a row to the total already found. Since the number of
10852 pebbles in the row is equal to the row number, the total can be found by
10853 adding the row number to the total. (Clearly, in a more complex
10854 situation, the number of pebbles in the row might be related to the row
10855 number in a more complicated way; if this were the case, the row number
10856 would be replaced by the appropriate expression.)
10859 (setq total (+ total row-number))
10863 What this does is set the new value of @code{total} to be equal to the
10864 sum of adding the number of pebbles in the row to the previous total.
10866 After setting the value of @code{total}, the conditions need to be
10867 established for the next repetition of the loop, if there is one. This
10868 is done by incrementing the value of the @code{row-number} variable,
10869 which serves as a counter. After the @code{row-number} variable has
10870 been incremented, the true-or-false-test at the beginning of the
10871 @code{while} loop tests whether its value is still less than or equal to
10872 the value of the @code{number-of-rows} and if it is, adds the new value
10873 of the @code{row-number} variable to the @code{total} of the previous
10874 repetition of the loop.
10877 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10878 @code{row-number} variable can be incremented with this expression:
10881 (setq row-number (1+ row-number))
10884 @node Inc Example altogether, , Inc Example parts, Incrementing Loop Details
10885 @unnumberedsubsubsec Putting the function definition together
10887 We have created the parts for the function definition; now we need to
10891 First, the contents of the @code{while} expression:
10895 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10896 (setq total (+ total row-number))
10897 (setq row-number (1+ row-number))) ; @r{incrementer}
10901 Along with the @code{let} expression varlist, this very nearly
10902 completes the body of the function definition. However, it requires
10903 one final element, the need for which is somewhat subtle.
10905 The final touch is to place the variable @code{total} on a line by
10906 itself after the @code{while} expression. Otherwise, the value returned
10907 by the whole function is the value of the last expression that is
10908 evaluated in the body of the @code{let}, and this is the value
10909 returned by the @code{while}, which is always @code{nil}.
10911 This may not be evident at first sight. It almost looks as if the
10912 incrementing expression is the last expression of the whole function.
10913 But that expression is part of the body of the @code{while}; it is the
10914 last element of the list that starts with the symbol @code{while}.
10915 Moreover, the whole of the @code{while} loop is a list within the body
10919 In outline, the function will look like this:
10923 (defun @var{name-of-function} (@var{argument-list})
10924 "@var{documentation}@dots{}"
10925 (let (@var{varlist})
10926 (while (@var{true-or-false-test})
10927 @var{body-of-while}@dots{} )
10928 @dots{} )) ; @r{Need final expression here.}
10932 The result of evaluating the @code{let} is what is going to be returned
10933 by the @code{defun} since the @code{let} is not embedded within any
10934 containing list, except for the @code{defun} as a whole. However, if
10935 the @code{while} is the last element of the @code{let} expression, the
10936 function will always return @code{nil}. This is not what we want!
10937 Instead, what we want is the value of the variable @code{total}. This
10938 is returned by simply placing the symbol as the last element of the list
10939 starting with @code{let}. It gets evaluated after the preceding
10940 elements of the list are evaluated, which means it gets evaluated after
10941 it has been assigned the correct value for the total.
10943 It may be easier to see this by printing the list starting with
10944 @code{let} all on one line. This format makes it evident that the
10945 @var{varlist} and @code{while} expressions are the second and third
10946 elements of the list starting with @code{let}, and the @code{total} is
10951 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10956 Putting everything together, the @code{triangle} function definition
10961 (defun triangle (number-of-rows) ; @r{Version with}
10962 ; @r{ incrementing counter.}
10963 "Add up the number of pebbles in a triangle.
10964 The first row has one pebble, the second row two pebbles,
10965 the third row three pebbles, and so on.
10966 The argument is NUMBER-OF-ROWS."
10971 (while (<= row-number number-of-rows)
10972 (setq total (+ total row-number))
10973 (setq row-number (1+ row-number)))
10979 After you have installed @code{triangle} by evaluating the function, you
10980 can try it out. Here are two examples:
10991 The sum of the first four numbers is 10 and the sum of the first seven
10994 @node Decrementing Loop, , Incrementing Loop Details, while
10995 @comment node-name, next, previous, up
10996 @subsection Loop with a Decrementing Counter
10998 Another common way to write a @code{while} loop is to write the test
10999 so that it determines whether a counter is greater than zero. So long
11000 as the counter is greater than zero, the loop is repeated. But when
11001 the counter is equal to or less than zero, the loop is stopped. For
11002 this to work, the counter has to start out greater than zero and then
11003 be made smaller and smaller by a form that is evaluated
11006 The test will be an expression such as @code{(> counter 0)} which
11007 returns @code{t} for true if the value of @code{counter} is greater
11008 than zero, and @code{nil} for false if the value of @code{counter} is
11009 equal to or less than zero. The expression that makes the number
11010 smaller and smaller can be a simple @code{setq} such as @code{(setq
11011 counter (1- counter))}, where @code{1-} is a built-in function in
11012 Emacs Lisp that subtracts 1 from its argument.
11015 The template for a decrementing @code{while} loop looks like this:
11019 (while (> counter 0) ; @r{true-or-false-test}
11021 (setq counter (1- counter))) ; @r{decrementer}
11026 * Decrementing Example:: More pebbles on the beach.
11027 * Dec Example parts:: The parts of the function definition.
11028 * Dec Example altogether:: Putting the function definition together.
11031 @node Decrementing Example, Dec Example parts, Decrementing Loop, Decrementing Loop
11032 @unnumberedsubsubsec Example with decrementing counter
11034 To illustrate a loop with a decrementing counter, we will rewrite the
11035 @code{triangle} function so the counter decreases to zero.
11037 This is the reverse of the earlier version of the function. In this
11038 case, to find out how many pebbles are needed to make a triangle with
11039 3 rows, add the number of pebbles in the third row, 3, to the number
11040 in the preceding row, 2, and then add the total of those two rows to
11041 the row that precedes them, which is 1.
11043 Likewise, to find the number of pebbles in a triangle with 7 rows, add
11044 the number of pebbles in the seventh row, 7, to the number in the
11045 preceding row, which is 6, and then add the total of those two rows to
11046 the row that precedes them, which is 5, and so on. As in the previous
11047 example, each addition only involves adding two numbers, the total of
11048 the rows already added up and the number of pebbles in the row that is
11049 being added to the total. This process of adding two numbers is
11050 repeated again and again until there are no more pebbles to add.
11052 We know how many pebbles to start with: the number of pebbles in the
11053 last row is equal to the number of rows. If the triangle has seven
11054 rows, the number of pebbles in the last row is 7. Likewise, we know how
11055 many pebbles are in the preceding row: it is one less than the number in
11058 @node Dec Example parts, Dec Example altogether, Decrementing Example, Decrementing Loop
11059 @unnumberedsubsubsec The parts of the function definition
11061 We start with three variables: the total number of rows in the
11062 triangle; the number of pebbles in a row; and the total number of
11063 pebbles, which is what we want to calculate. These variables can be
11064 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
11065 @code{total}, respectively.
11067 Both @code{total} and @code{number-of-pebbles-in-row} are used only
11068 inside the function and are declared with @code{let}. The initial
11069 value of @code{total} should, of course, be zero. However, the
11070 initial value of @code{number-of-pebbles-in-row} should be equal to
11071 the number of rows in the triangle, since the addition will start with
11075 This means that the beginning of the @code{let} expression will look
11081 (number-of-pebbles-in-row number-of-rows))
11086 The total number of pebbles can be found by repeatedly adding the number
11087 of pebbles in a row to the total already found, that is, by repeatedly
11088 evaluating the following expression:
11091 (setq total (+ total number-of-pebbles-in-row))
11095 After the @code{number-of-pebbles-in-row} is added to the @code{total},
11096 the @code{number-of-pebbles-in-row} should be decremented by one, since
11097 the next time the loop repeats, the preceding row will be
11098 added to the total.
11100 The number of pebbles in a preceding row is one less than the number of
11101 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
11102 used to compute the number of pebbles in the preceding row. This can be
11103 done with the following expression:
11107 (setq number-of-pebbles-in-row
11108 (1- number-of-pebbles-in-row))
11112 Finally, we know that the @code{while} loop should stop making repeated
11113 additions when there are no pebbles in a row. So the test for
11114 the @code{while} loop is simply:
11117 (while (> number-of-pebbles-in-row 0)
11120 @node Dec Example altogether, , Dec Example parts, Decrementing Loop
11121 @unnumberedsubsubsec Putting the function definition together
11123 We can put these expressions together to create a function definition
11124 that works. However, on examination, we find that one of the local
11125 variables is unneeded!
11128 The function definition looks like this:
11132 ;;; @r{First subtractive version.}
11133 (defun triangle (number-of-rows)
11134 "Add up the number of pebbles in a triangle."
11136 (number-of-pebbles-in-row number-of-rows))
11137 (while (> number-of-pebbles-in-row 0)
11138 (setq total (+ total number-of-pebbles-in-row))
11139 (setq number-of-pebbles-in-row
11140 (1- number-of-pebbles-in-row)))
11145 As written, this function works.
11147 However, we do not need @code{number-of-pebbles-in-row}.
11149 @cindex Argument as local variable
11150 When the @code{triangle} function is evaluated, the symbol
11151 @code{number-of-rows} will be bound to a number, giving it an initial
11152 value. That number can be changed in the body of the function as if
11153 it were a local variable, without any fear that such a change will
11154 effect the value of the variable outside of the function. This is a
11155 very useful characteristic of Lisp; it means that the variable
11156 @code{number-of-rows} can be used anywhere in the function where
11157 @code{number-of-pebbles-in-row} is used.
11160 Here is a second version of the function written a bit more cleanly:
11164 (defun triangle (number) ; @r{Second version.}
11165 "Return sum of numbers 1 through NUMBER inclusive."
11167 (while (> number 0)
11168 (setq total (+ total number))
11169 (setq number (1- number)))
11174 In brief, a properly written @code{while} loop will consist of three parts:
11178 A test that will return false after the loop has repeated itself the
11179 correct number of times.
11182 An expression the evaluation of which will return the value desired
11183 after being repeatedly evaluated.
11186 An expression to change the value passed to the true-or-false-test so
11187 that the test returns false after the loop has repeated itself the right
11191 @node dolist dotimes, Recursion, while, Loops & Recursion
11192 @comment node-name, next, previous, up
11193 @section Save your time: @code{dolist} and @code{dotimes}
11195 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11196 provide for looping. Sometimes these are quicker to write than the
11197 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11198 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11200 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11201 list': @code{dolist} automatically shortens the list each time it
11202 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11203 each shorter version of the list to the first of its arguments.
11205 @code{dotimes} loops a specific number of times: you specify the number.
11212 @node dolist, dotimes, dolist dotimes, dolist dotimes
11213 @unnumberedsubsubsec The @code{dolist} Macro
11216 Suppose, for example, you want to reverse a list, so that
11217 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11220 In practice, you would use the @code{reverse} function, like this:
11224 (setq animals '(gazelle giraffe lion tiger))
11232 Here is how you could reverse the list using a @code{while} loop:
11236 (setq animals '(gazelle giraffe lion tiger))
11238 (defun reverse-list-with-while (list)
11239 "Using while, reverse the order of LIST."
11240 (let (value) ; make sure list starts empty
11242 (setq value (cons (car list) value))
11243 (setq list (cdr list)))
11246 (reverse-list-with-while animals)
11252 And here is how you could use the @code{dolist} macro:
11256 (setq animals '(gazelle giraffe lion tiger))
11258 (defun reverse-list-with-dolist (list)
11259 "Using dolist, reverse the order of LIST."
11260 (let (value) ; make sure list starts empty
11261 (dolist (element list value)
11262 (setq value (cons element value)))))
11264 (reverse-list-with-dolist animals)
11270 In Info, you can place your cursor after the closing parenthesis of
11271 each expression and type @kbd{C-x C-e}; in each case, you should see
11274 (tiger lion giraffe gazelle)
11280 For this example, the existing @code{reverse} function is obviously best.
11281 The @code{while} loop is just like our first example (@pxref{Loop
11282 Example, , A @code{while} Loop and a List}). The @code{while} first
11283 checks whether the list has elements; if so, it constructs a new list
11284 by adding the first element of the list to the existing list (which in
11285 the first iteration of the loop is @code{nil}). Since the second
11286 element is prepended in front of the first element, and the third
11287 element is prepended in front of the second element, the list is reversed.
11289 In the expression using a @code{while} loop,
11290 the @w{@code{(setq list (cdr list))}}
11291 expression shortens the list, so the @code{while} loop eventually
11292 stops. In addition, it provides the @code{cons} expression with a new
11293 first element by creating a new and shorter list at each repetition of
11296 The @code{dolist} expression does very much the same as the
11297 @code{while} expression, except that the @code{dolist} macro does some
11298 of the work you have to do when writing a @code{while} expression.
11300 Like a @code{while} loop, a @code{dolist} loops. What is different is
11301 that it automatically shortens the list each time it loops --- it
11302 `@sc{cdr}s down the list' on its own --- and it automatically binds
11303 the @sc{car} of each shorter version of the list to the first of its
11306 In the example, the @sc{car} of each shorter version of the list is
11307 referred to using the symbol @samp{element}, the list itself is called
11308 @samp{list}, and the value returned is called @samp{value}. The
11309 remainder of the @code{dolist} expression is the body.
11311 The @code{dolist} expression binds the @sc{car} of each shorter
11312 version of the list to @code{element} and then evaluates the body of
11313 the expression; and repeats the loop. The result is returned in
11316 @node dotimes, , dolist, dolist dotimes
11317 @unnumberedsubsubsec The @code{dotimes} Macro
11320 The @code{dotimes} macro is similar to @code{dolist}, except that it
11321 loops a specific number of times.
11323 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11324 and so forth each time around the loop, and the value of the third
11325 argument is returned. You need to provide the value of the second
11326 argument, which is how many times the macro loops.
11329 For example, the following binds the numbers from 0 up to, but not
11330 including, the number 3 to the first argument, @var{number}, and then
11331 constructs a list of the three numbers. (The first number is 0, the
11332 second number is 1, and the third number is 2; this makes a total of
11333 three numbers in all, starting with zero as the first number.)
11337 (let (value) ; otherwise a value is a void variable
11338 (dotimes (number 3 value)
11339 (setq value (cons number value))))
11346 @code{dotimes} returns @code{value}, so the way to use
11347 @code{dotimes} is to operate on some expression @var{number} number of
11348 times and then return the result, either as a list or an atom.
11351 Here is an example of a @code{defun} that uses @code{dotimes} to add
11352 up the number of pebbles in a triangle.
11356 (defun triangle-using-dotimes (number-of-rows)
11357 "Using dotimes, add up the number of pebbles in a triangle."
11358 (let ((total 0)) ; otherwise a total is a void variable
11359 (dotimes (number number-of-rows total)
11360 (setq total (+ total (1+ number))))))
11362 (triangle-using-dotimes 4)
11366 @node Recursion, Looping exercise, dolist dotimes, Loops & Recursion
11367 @comment node-name, next, previous, up
11371 A recursive function contains code that tells the Lisp interpreter to
11372 call a program that runs exactly like itself, but with slightly
11373 different arguments. The code runs exactly the same because it has
11374 the same name. However, even though the program has the same name, it
11375 is not the same entity. It is different. In the jargon, it is a
11376 different `instance'.
11378 Eventually, if the program is written correctly, the `slightly
11379 different arguments' will become sufficiently different from the first
11380 arguments that the final instance will stop.
11383 * Building Robots:: Same model, different serial number ...
11384 * Recursive Definition Parts:: Walk until you stop ...
11385 * Recursion with list:: Using a list as the test whether to recurse.
11386 * Recursive triangle function::
11387 * Recursion with cond::
11388 * Recursive Patterns:: Often used templates.
11389 * No Deferment:: Don't store up work ...
11390 * No deferment solution::
11393 @node Building Robots, Recursive Definition Parts, Recursion, Recursion
11394 @comment node-name, next, previous, up
11395 @subsection Building Robots: Extending the Metaphor
11396 @cindex Building robots
11397 @cindex Robots, building
11399 It is sometimes helpful to think of a running program as a robot that
11400 does a job. In doing its job, a recursive function calls on a second
11401 robot to help it. The second robot is identical to the first in every
11402 way, except that the second robot helps the first and has been
11403 passed different arguments than the first.
11405 In a recursive function, the second robot may call a third; and the
11406 third may call a fourth, and so on. Each of these is a different
11407 entity; but all are clones.
11409 Since each robot has slightly different instructions---the arguments
11410 will differ from one robot to the next---the last robot should know
11413 Let's expand on the metaphor in which a computer program is a robot.
11415 A function definition provides the blueprints for a robot. When you
11416 install a function definition, that is, when you evaluate a
11417 @code{defun} special form, you install the necessary equipment to
11418 build robots. It is as if you were in a factory, setting up an
11419 assembly line. Robots with the same name are built according to the
11420 same blueprints. So they have, as it were, the same `model number',
11421 but a different `serial number'.
11423 We often say that a recursive function `calls itself'. What we mean
11424 is that the instructions in a recursive function cause the Lisp
11425 interpreter to run a different function that has the same name and
11426 does the same job as the first, but with different arguments.
11428 It is important that the arguments differ from one instance to the
11429 next; otherwise, the process will never stop.
11431 @node Recursive Definition Parts, Recursion with list, Building Robots, Recursion
11432 @comment node-name, next, previous, up
11433 @subsection The Parts of a Recursive Definition
11434 @cindex Parts of a Recursive Definition
11435 @cindex Recursive Definition Parts
11437 A recursive function typically contains a conditional expression which
11442 A true-or-false-test that determines whether the function is called
11443 again, here called the @dfn{do-again-test}.
11446 The name of the function. When this name is called, a new instance of
11447 the function---a new robot, as it were---is created and told what to do.
11450 An expression that returns a different value each time the function is
11451 called, here called the @dfn{next-step-expression}. Consequently, the
11452 argument (or arguments) passed to the new instance of the function
11453 will be different from that passed to the previous instance. This
11454 causes the conditional expression, the @dfn{do-again-test}, to test
11455 false after the correct number of repetitions.
11458 Recursive functions can be much simpler than any other kind of
11459 function. Indeed, when people first start to use them, they often look
11460 so mysteriously simple as to be incomprehensible. Like riding a
11461 bicycle, reading a recursive function definition takes a certain knack
11462 which is hard at first but then seems simple.
11465 There are several different common recursive patterns. A very simple
11466 pattern looks like this:
11470 (defun @var{name-of-recursive-function} (@var{argument-list})
11471 "@var{documentation}@dots{}"
11472 (if @var{do-again-test}
11474 (@var{name-of-recursive-function}
11475 @var{next-step-expression})))
11479 Each time a recursive function is evaluated, a new instance of it is
11480 created and told what to do. The arguments tell the instance what to do.
11482 An argument is bound to the value of the next-step-expression. Each
11483 instance runs with a different value of the next-step-expression.
11485 The value in the next-step-expression is used in the do-again-test.
11487 The value returned by the next-step-expression is passed to the new
11488 instance of the function, which evaluates it (or some
11489 transmogrification of it) to determine whether to continue or stop.
11490 The next-step-expression is designed so that the do-again-test returns
11491 false when the function should no longer be repeated.
11493 The do-again-test is sometimes called the @dfn{stop condition},
11494 since it stops the repetitions when it tests false.
11496 @node Recursion with list, Recursive triangle function, Recursive Definition Parts, Recursion
11497 @comment node-name, next, previous, up
11498 @subsection Recursion with a List
11500 The example of a @code{while} loop that printed the elements of a list
11501 of numbers can be written recursively. Here is the code, including
11502 an expression to set the value of the variable @code{animals} to a list.
11504 If you are using GNU Emacs 20 or before, this example must be copied
11505 to the @file{*scratch*} buffer and each expression must be evaluated
11506 there. Use @kbd{C-u C-x C-e} to evaluate the
11507 @code{(print-elements-recursively animals)} expression so that the
11508 results are printed in the buffer; otherwise the Lisp interpreter will
11509 try to squeeze the results into the one line of the echo area.
11511 Also, place your cursor immediately after the last closing parenthesis
11512 of the @code{print-elements-recursively} function, before the comment.
11513 Otherwise, the Lisp interpreter will try to evaluate the comment.
11515 If you are using a more recent version of Emacs, you can evaluate this
11516 expression directly in Info.
11518 @findex print-elements-recursively
11521 (setq animals '(gazelle giraffe lion tiger))
11523 (defun print-elements-recursively (list)
11524 "Print each element of LIST on a line of its own.
11526 (when list ; @r{do-again-test}
11527 (print (car list)) ; @r{body}
11528 (print-elements-recursively ; @r{recursive call}
11529 (cdr list)))) ; @r{next-step-expression}
11531 (print-elements-recursively animals)
11535 The @code{print-elements-recursively} function first tests whether
11536 there is any content in the list; if there is, the function prints the
11537 first element of the list, the @sc{car} of the list. Then the
11538 function `invokes itself', but gives itself as its argument, not the
11539 whole list, but the second and subsequent elements of the list, the
11540 @sc{cdr} of the list.
11542 Put another way, if the list is not empty, the function invokes
11543 another instance of code that is similar to the initial code, but is a
11544 different thread of execution, with different arguments than the first
11547 Put in yet another way, if the list is not empty, the first robot
11548 assembles a second robot and tells it what to do; the second robot is
11549 a different individual from the first, but is the same model.
11551 When the second evaluation occurs, the @code{when} expression is
11552 evaluated and if true, prints the first element of the list it
11553 receives as its argument (which is the second element of the original
11554 list). Then the function `calls itself' with the @sc{cdr} of the list
11555 it is invoked with, which (the second time around) is the @sc{cdr} of
11556 the @sc{cdr} of the original list.
11558 Note that although we say that the function `calls itself', what we
11559 mean is that the Lisp interpreter assembles and instructs a new
11560 instance of the program. The new instance is a clone of the first,
11561 but is a separate individual.
11563 Each time the function `invokes itself', it invokes itself on a
11564 shorter version of the original list. It creates a new instance that
11565 works on a shorter list.
11567 Eventually, the function invokes itself on an empty list. It creates
11568 a new instance whose argument is @code{nil}. The conditional expression
11569 tests the value of @code{list}. Since the value of @code{list} is
11570 @code{nil}, the @code{when} expression tests false so the then-part is
11571 not evaluated. The function as a whole then returns @code{nil}.
11574 When you evaluate the expression @code{(print-elements-recursively
11575 animals)} in the @file{*scratch*} buffer, you see this result:
11591 @node Recursive triangle function, Recursion with cond, Recursion with list, Recursion
11592 @comment node-name, next, previous, up
11593 @subsection Recursion in Place of a Counter
11594 @findex triangle-recursively
11597 The @code{triangle} function described in a previous section can also
11598 be written recursively. It looks like this:
11602 (defun triangle-recursively (number)
11603 "Return the sum of the numbers 1 through NUMBER inclusive.
11605 (if (= number 1) ; @r{do-again-test}
11607 (+ number ; @r{else-part}
11608 (triangle-recursively ; @r{recursive call}
11609 (1- number))))) ; @r{next-step-expression}
11611 (triangle-recursively 7)
11616 You can install this function by evaluating it and then try it by
11617 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11618 cursor immediately after the last parenthesis of the function
11619 definition, before the comment.) The function evaluates to 28.
11621 To understand how this function works, let's consider what happens in the
11622 various cases when the function is passed 1, 2, 3, or 4 as the value of
11626 * Recursive Example arg of 1 or 2::
11627 * Recursive Example arg of 3 or 4::
11630 @node Recursive Example arg of 1 or 2, Recursive Example arg of 3 or 4, Recursive triangle function, Recursive triangle function
11632 @unnumberedsubsubsec An argument of 1 or 2
11635 First, what happens if the value of the argument is 1?
11637 The function has an @code{if} expression after the documentation
11638 string. It tests whether the value of @code{number} is equal to 1; if
11639 so, Emacs evaluates the then-part of the @code{if} expression, which
11640 returns the number 1 as the value of the function. (A triangle with
11641 one row has one pebble in it.)
11643 Suppose, however, that the value of the argument is 2. In this case,
11644 Emacs evaluates the else-part of the @code{if} expression.
11647 The else-part consists of an addition, the recursive call to
11648 @code{triangle-recursively} and a decrementing action; and it looks like
11652 (+ number (triangle-recursively (1- number)))
11655 When Emacs evaluates this expression, the innermost expression is
11656 evaluated first; then the other parts in sequence. Here are the steps
11660 @item Step 1 @w{ } Evaluate the innermost expression.
11662 The innermost expression is @code{(1- number)} so Emacs decrements the
11663 value of @code{number} from 2 to 1.
11665 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11667 The Lisp interpreter creates an individual instance of
11668 @code{triangle-recursively}. It does not matter that this function is
11669 contained within itself. Emacs passes the result Step 1 as the
11670 argument used by this instance of the @code{triangle-recursively}
11673 In this case, Emacs evaluates @code{triangle-recursively} with an
11674 argument of 1. This means that this evaluation of
11675 @code{triangle-recursively} returns 1.
11677 @item Step 3 @w{ } Evaluate the value of @code{number}.
11679 The variable @code{number} is the second element of the list that
11680 starts with @code{+}; its value is 2.
11682 @item Step 4 @w{ } Evaluate the @code{+} expression.
11684 The @code{+} expression receives two arguments, the first
11685 from the evaluation of @code{number} (Step 3) and the second from the
11686 evaluation of @code{triangle-recursively} (Step 2).
11688 The result of the addition is the sum of 2 plus 1, and the number 3 is
11689 returned, which is correct. A triangle with two rows has three
11693 @node Recursive Example arg of 3 or 4, , Recursive Example arg of 1 or 2, Recursive triangle function
11694 @unnumberedsubsubsec An argument of 3 or 4
11696 Suppose that @code{triangle-recursively} is called with an argument of
11700 @item Step 1 @w{ } Evaluate the do-again-test.
11702 The @code{if} expression is evaluated first. This is the do-again
11703 test and returns false, so the else-part of the @code{if} expression
11704 is evaluated. (Note that in this example, the do-again-test causes
11705 the function to call itself when it tests false, not when it tests
11708 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11710 The innermost expression of the else-part is evaluated, which decrements
11711 3 to 2. This is the next-step-expression.
11713 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11715 The number 2 is passed to the @code{triangle-recursively} function.
11717 We already know what happens when Emacs evaluates @code{triangle-recursively} with
11718 an argument of 2. After going through the sequence of actions described
11719 earlier, it returns a value of 3. So that is what will happen here.
11721 @item Step 4 @w{ } Evaluate the addition.
11723 3 will be passed as an argument to the addition and will be added to the
11724 number with which the function was called, which is 3.
11728 The value returned by the function as a whole will be 6.
11730 Now that we know what will happen when @code{triangle-recursively} is
11731 called with an argument of 3, it is evident what will happen if it is
11732 called with an argument of 4:
11736 In the recursive call, the evaluation of
11739 (triangle-recursively (1- 4))
11744 will return the value of evaluating
11747 (triangle-recursively 3)
11751 which is 6 and this value will be added to 4 by the addition in the
11756 The value returned by the function as a whole will be 10.
11758 Each time @code{triangle-recursively} is evaluated, it evaluates a
11759 version of itself---a different instance of itself---with a smaller
11760 argument, until the argument is small enough so that it does not
11763 Note that this particular design for a recursive function
11764 requires that operations be deferred.
11766 Before @code{(triangle-recursively 7)} can calculate its answer, it
11767 must call @code{(triangle-recursively 6)}; and before
11768 @code{(triangle-recursively 6)} can calculate its answer, it must call
11769 @code{(triangle-recursively 5)}; and so on. That is to say, the
11770 calculation that @code{(triangle-recursively 7)} makes must be
11771 deferred until @code{(triangle-recursively 6)} makes its calculation;
11772 and @code{(triangle-recursively 6)} must defer until
11773 @code{(triangle-recursively 5)} completes; and so on.
11775 If each of these instances of @code{triangle-recursively} are thought
11776 of as different robots, the first robot must wait for the second to
11777 complete its job, which must wait until the third completes, and so
11780 There is a way around this kind of waiting, which we will discuss in
11781 @ref{No Deferment, , Recursion without Deferments}.
11783 @node Recursion with cond, Recursive Patterns, Recursive triangle function, Recursion
11784 @comment node-name, next, previous, up
11785 @subsection Recursion Example Using @code{cond}
11788 The version of @code{triangle-recursively} described earlier is written
11789 with the @code{if} special form. It can also be written using another
11790 special form called @code{cond}. The name of the special form
11791 @code{cond} is an abbreviation of the word @samp{conditional}.
11793 Although the @code{cond} special form is not used as often in the
11794 Emacs Lisp sources as @code{if}, it is used often enough to justify
11798 The template for a @code{cond} expression looks like this:
11808 where the @var{body} is a series of lists.
11811 Written out more fully, the template looks like this:
11816 (@var{first-true-or-false-test} @var{first-consequent})
11817 (@var{second-true-or-false-test} @var{second-consequent})
11818 (@var{third-true-or-false-test} @var{third-consequent})
11823 When the Lisp interpreter evaluates the @code{cond} expression, it
11824 evaluates the first element (the @sc{car} or true-or-false-test) of
11825 the first expression in a series of expressions within the body of the
11828 If the true-or-false-test returns @code{nil} the rest of that
11829 expression, the consequent, is skipped and the true-or-false-test of the
11830 next expression is evaluated. When an expression is found whose
11831 true-or-false-test returns a value that is not @code{nil}, the
11832 consequent of that expression is evaluated. The consequent can be one
11833 or more expressions. If the consequent consists of more than one
11834 expression, the expressions are evaluated in sequence and the value of
11835 the last one is returned. If the expression does not have a consequent,
11836 the value of the true-or-false-test is returned.
11838 If none of the true-or-false-tests test true, the @code{cond} expression
11839 returns @code{nil}.
11842 Written using @code{cond}, the @code{triangle} function looks like this:
11846 (defun triangle-using-cond (number)
11847 (cond ((<= number 0) 0)
11850 (+ number (triangle-using-cond (1- number))))))
11855 In this example, the @code{cond} returns 0 if the number is less than or
11856 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11857 number (triangle-using-cond (1- number)))} if the number is greater than
11860 @node Recursive Patterns, No Deferment, Recursion with cond, Recursion
11861 @comment node-name, next, previous, up
11862 @subsection Recursive Patterns
11863 @cindex Recursive Patterns
11865 Here are three common recursive patterns. Each involves a list.
11866 Recursion does not need to involve lists, but Lisp is designed for lists
11867 and this provides a sense of its primal capabilities.
11875 @node Every, Accumulate, Recursive Patterns, Recursive Patterns
11876 @comment node-name, next, previous, up
11877 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11878 @cindex Every, type of recursive pattern
11879 @cindex Recursive pattern: every
11881 In the @code{every} recursive pattern, an action is performed on every
11885 The basic pattern is:
11889 If a list be empty, return @code{nil}.
11891 Else, act on the beginning of the list (the @sc{car} of the list)
11894 through a recursive call by the function on the rest (the
11895 @sc{cdr}) of the list,
11897 and, optionally, combine the acted-on element, using @code{cons},
11898 with the results of acting on the rest.
11907 (defun square-each (numbers-list)
11908 "Square each of a NUMBERS LIST, recursively."
11909 (if (not numbers-list) ; do-again-test
11912 (* (car numbers-list) (car numbers-list))
11913 (square-each (cdr numbers-list))))) ; next-step-expression
11917 (square-each '(1 2 3))
11924 If @code{numbers-list} is empty, do nothing. But if it has content,
11925 construct a list combining the square of the first number in the list
11926 with the result of the recursive call.
11928 (The example follows the pattern exactly: @code{nil} is returned if
11929 the numbers' list is empty. In practice, you would write the
11930 conditional so it carries out the action when the numbers' list is not
11933 The @code{print-elements-recursively} function (@pxref{Recursion with
11934 list, , Recursion with a List}) is another example of an @code{every}
11935 pattern, except in this case, rather than bring the results together
11936 using @code{cons}, we print each element of output.
11939 The @code{print-elements-recursively} function looks like this:
11943 (setq animals '(gazelle giraffe lion tiger))
11947 (defun print-elements-recursively (list)
11948 "Print each element of LIST on a line of its own.
11950 (when list ; @r{do-again-test}
11951 (print (car list)) ; @r{body}
11952 (print-elements-recursively ; @r{recursive call}
11953 (cdr list)))) ; @r{next-step-expression}
11955 (print-elements-recursively animals)
11960 The pattern for @code{print-elements-recursively} is:
11964 When the list is empty, do nothing.
11966 But when the list has at least one element,
11969 act on the beginning of the list (the @sc{car} of the list),
11971 and make a recursive call on the rest (the @sc{cdr}) of the list.
11975 @node Accumulate, Keep, Every, Recursive Patterns
11976 @comment node-name, next, previous, up
11977 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11978 @cindex Accumulate, type of recursive pattern
11979 @cindex Recursive pattern: accumulate
11981 Another recursive pattern is called the @code{accumulate} pattern. In
11982 the @code{accumulate} recursive pattern, an action is performed on
11983 every element of a list and the result of that action is accumulated
11984 with the results of performing the action on the other elements.
11986 This is very like the `every' pattern using @code{cons}, except that
11987 @code{cons} is not used, but some other combiner.
11994 If a list be empty, return zero or some other constant.
11996 Else, act on the beginning of the list (the @sc{car} of the list),
11999 and combine that acted-on element, using @code{+} or
12000 some other combining function, with
12002 a recursive call by the function on the rest (the @sc{cdr}) of the list.
12007 Here is an example:
12011 (defun add-elements (numbers-list)
12012 "Add the elements of NUMBERS-LIST together."
12013 (if (not numbers-list)
12015 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
12019 (add-elements '(1 2 3 4))
12024 @xref{Files List, , Making a List of Files}, for an example of the
12025 accumulate pattern.
12027 @node Keep, , Accumulate, Recursive Patterns
12028 @comment node-name, next, previous, up
12029 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
12030 @cindex Keep, type of recursive pattern
12031 @cindex Recursive pattern: keep
12033 A third recursive pattern is called the @code{keep} pattern.
12034 In the @code{keep} recursive pattern, each element of a list is tested;
12035 the element is acted on and the results are kept only if the element
12038 Again, this is very like the `every' pattern, except the element is
12039 skipped unless it meets a criterion.
12042 The pattern has three parts:
12046 If a list be empty, return @code{nil}.
12048 Else, if the beginning of the list (the @sc{car} of the list) passes
12052 act on that element and combine it, using @code{cons} with
12054 a recursive call by the function on the rest (the @sc{cdr}) of the list.
12057 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
12061 skip on that element,
12063 and, recursively call the function on the rest (the @sc{cdr}) of the list.
12068 Here is an example that uses @code{cond}:
12072 (defun keep-three-letter-words (word-list)
12073 "Keep three letter words in WORD-LIST."
12075 ;; First do-again-test: stop-condition
12076 ((not word-list) nil)
12078 ;; Second do-again-test: when to act
12079 ((eq 3 (length (symbol-name (car word-list))))
12080 ;; combine acted-on element with recursive call on shorter list
12081 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
12083 ;; Third do-again-test: when to skip element;
12084 ;; recursively call shorter list with next-step expression
12085 (t (keep-three-letter-words (cdr word-list)))))
12089 (keep-three-letter-words '(one two three four five six))
12090 @result{} (one two six)
12094 It goes without saying that you need not use @code{nil} as the test for
12095 when to stop; and you can, of course, combine these patterns.
12097 @node No Deferment, No deferment solution, Recursive Patterns, Recursion
12098 @subsection Recursion without Deferments
12099 @cindex Deferment in recursion
12100 @cindex Recursion without Deferments
12102 Let's consider again what happens with the @code{triangle-recursively}
12103 function. We will find that the intermediate calculations are
12104 deferred until all can be done.
12107 Here is the function definition:
12111 (defun triangle-recursively (number)
12112 "Return the sum of the numbers 1 through NUMBER inclusive.
12114 (if (= number 1) ; @r{do-again-test}
12116 (+ number ; @r{else-part}
12117 (triangle-recursively ; @r{recursive call}
12118 (1- number))))) ; @r{next-step-expression}
12122 What happens when we call this function with a argument of 7?
12124 The first instance of the @code{triangle-recursively} function adds
12125 the number 7 to the value returned by a second instance of
12126 @code{triangle-recursively}, an instance that has been passed an
12127 argument of 6. That is to say, the first calculation is:
12130 (+ 7 (triangle-recursively 6))
12134 The first instance of @code{triangle-recursively}---you may want to
12135 think of it as a little robot---cannot complete its job. It must hand
12136 off the calculation for @code{(triangle-recursively 6)} to a second
12137 instance of the program, to a second robot. This second individual is
12138 completely different from the first one; it is, in the jargon, a
12139 `different instantiation'. Or, put another way, it is a different
12140 robot. It is the same model as the first; it calculates triangle
12141 numbers recursively; but it has a different serial number.
12143 And what does @code{(triangle-recursively 6)} return? It returns the
12144 number 6 added to the value returned by evaluating
12145 @code{triangle-recursively} with an argument of 5. Using the robot
12146 metaphor, it asks yet another robot to help it.
12152 (+ 7 6 (triangle-recursively 5))
12156 And what happens next?
12159 (+ 7 6 5 (triangle-recursively 4))
12162 Each time @code{triangle-recursively} is called, except for the last
12163 time, it creates another instance of the program---another robot---and
12164 asks it to make a calculation.
12167 Eventually, the full addition is set up and performed:
12173 This design for the function defers the calculation of the first step
12174 until the second can be done, and defers that until the third can be
12175 done, and so on. Each deferment means the computer must remember what
12176 is being waited on. This is not a problem when there are only a few
12177 steps, as in this example. But it can be a problem when there are
12180 @node No deferment solution, , No Deferment, Recursion
12181 @subsection No Deferment Solution
12182 @cindex No deferment solution
12183 @cindex Defermentless solution
12184 @cindex Solution without deferment
12186 The solution to the problem of deferred operations is to write in a
12187 manner that does not defer operations@footnote{The phrase @dfn{tail
12188 recursive} is used to describe such a process, one that uses
12189 `constant space'.}. This requires
12190 writing to a different pattern, often one that involves writing two
12191 function definitions, an `initialization' function and a `helper'
12194 The `initialization' function sets up the job; the `helper' function
12198 Here are the two function definitions for adding up numbers. They are
12199 so simple, I find them hard to understand.
12203 (defun triangle-initialization (number)
12204 "Return the sum of the numbers 1 through NUMBER inclusive.
12205 This is the `initialization' component of a two function
12206 duo that uses recursion."
12207 (triangle-recursive-helper 0 0 number))
12213 (defun triangle-recursive-helper (sum counter number)
12214 "Return SUM, using COUNTER, through NUMBER inclusive.
12215 This is the `helper' component of a two function duo
12216 that uses recursion."
12217 (if (> counter number)
12219 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12220 (1+ counter) ; @r{counter}
12221 number))) ; @r{number}
12226 Install both function definitions by evaluating them, then call
12227 @code{triangle-initialization} with 2 rows:
12231 (triangle-initialization 2)
12236 The `initialization' function calls the first instance of the `helper'
12237 function with three arguments: zero, zero, and a number which is the
12238 number of rows in the triangle.
12240 The first two arguments passed to the `helper' function are
12241 initialization values. These values are changed when
12242 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12243 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12244 process that is iterative in a procedure that is recursive. The
12245 process is called iterative because the computer need only record the
12246 three values, @code{sum}, @code{counter}, and @code{number}; the
12247 procedure is recursive because the function `calls itself'. On the
12248 other hand, both the process and the procedure used by
12249 @code{triangle-recursively} are called recursive. The word
12250 `recursive' has different meanings in the two contexts.}
12252 Let's see what happens when we have a triangle that has one row. (This
12253 triangle will have one pebble in it!)
12256 @code{triangle-initialization} will call its helper with
12257 the arguments @w{@code{0 0 1}}. That function will run the conditional
12258 test whether @code{(> counter number)}:
12266 and find that the result is false, so it will invoke
12267 the else-part of the @code{if} clause:
12271 (triangle-recursive-helper
12272 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12273 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12274 number) ; @r{number stays the same}
12280 which will first compute:
12284 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12285 (1+ 0) ; @r{counter}
12289 (triangle-recursive-helper 0 1 1)
12293 Again, @code{(> counter number)} will be false, so again, the Lisp
12294 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12295 new instance with new arguments.
12298 This new instance will be;
12302 (triangle-recursive-helper
12303 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12304 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12305 number) ; @r{number stays the same}
12309 (triangle-recursive-helper 1 2 1)
12313 In this case, the @code{(> counter number)} test will be true! So the
12314 instance will return the value of the sum, which will be 1, as
12317 Now, let's pass @code{triangle-initialization} an argument
12318 of 2, to find out how many pebbles there are in a triangle with two rows.
12320 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12323 In stages, the instances called will be:
12327 @r{sum counter number}
12328 (triangle-recursive-helper 0 1 2)
12330 (triangle-recursive-helper 1 2 2)
12332 (triangle-recursive-helper 3 3 2)
12336 When the last instance is called, the @code{(> counter number)} test
12337 will be true, so the instance will return the value of @code{sum},
12340 This kind of pattern helps when you are writing functions that can use
12341 many resources in a computer.
12344 @node Looping exercise, , Recursion, Loops & Recursion
12345 @section Looping Exercise
12349 Write a function similar to @code{triangle} in which each row has a
12350 value which is the square of the row number. Use a @code{while} loop.
12353 Write a function similar to @code{triangle} that multiplies instead of
12357 Rewrite these two functions recursively. Rewrite these functions
12360 @c comma in printed title causes problem in Info cross reference
12362 Write a function for Texinfo mode that creates an index entry at the
12363 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12364 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12365 written in Texinfo.)
12367 Many of the functions you will need are described in two of the
12368 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12369 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12370 @code{forward-paragraph} to put the index entry at the beginning of
12371 the paragraph, you will have to use @w{@kbd{C-h f}}
12372 (@code{describe-function}) to find out how to make the command go
12375 For more information, see
12377 @ref{Indicating, , Indicating Definitions, texinfo}.
12380 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12381 a Texinfo manual in the current directory. Or, if you are on the
12383 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12386 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12387 Documentation Format}.
12391 @node Regexp Search, Counting Words, Loops & Recursion, Top
12392 @comment node-name, next, previous, up
12393 @chapter Regular Expression Searches
12394 @cindex Searches, illustrating
12395 @cindex Regular expression searches
12396 @cindex Patterns, searching for
12397 @cindex Motion by sentence and paragraph
12398 @cindex Sentences, movement by
12399 @cindex Paragraphs, movement by
12401 Regular expression searches are used extensively in GNU Emacs. The
12402 two functions, @code{forward-sentence} and @code{forward-paragraph},
12403 illustrate these searches well. They use regular expressions to find
12404 where to move point. The phrase `regular expression' is often written
12407 Regular expression searches are described in @ref{Regexp Search, ,
12408 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12409 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12410 Manual}. In writing this chapter, I am presuming that you have at
12411 least a mild acquaintance with them. The major point to remember is
12412 that regular expressions permit you to search for patterns as well as
12413 for literal strings of characters. For example, the code in
12414 @code{forward-sentence} searches for the pattern of possible
12415 characters that could mark the end of a sentence, and moves point to
12418 Before looking at the code for the @code{forward-sentence} function, it
12419 is worth considering what the pattern that marks the end of a sentence
12420 must be. The pattern is discussed in the next section; following that
12421 is a description of the regular expression search function,
12422 @code{re-search-forward}. The @code{forward-sentence} function
12423 is described in the section following. Finally, the
12424 @code{forward-paragraph} function is described in the last section of
12425 this chapter. @code{forward-paragraph} is a complex function that
12426 introduces several new features.
12429 * sentence-end:: The regular expression for @code{sentence-end}.
12430 * re-search-forward:: Very similar to @code{search-forward}.
12431 * forward-sentence:: A straightforward example of regexp search.
12432 * forward-paragraph:: A somewhat complex example.
12433 * etags:: How to create your own @file{TAGS} table.
12435 * re-search Exercises::
12438 @node sentence-end, re-search-forward, Regexp Search, Regexp Search
12439 @comment node-name, next, previous, up
12440 @section The Regular Expression for @code{sentence-end}
12441 @findex sentence-end
12443 The symbol @code{sentence-end} is bound to the pattern that marks the
12444 end of a sentence. What should this regular expression be?
12446 Clearly, a sentence may be ended by a period, a question mark, or an
12447 exclamation mark. Indeed, in English, only clauses that end with one
12448 of those three characters should be considered the end of a sentence.
12449 This means that the pattern should include the character set:
12455 However, we do not want @code{forward-sentence} merely to jump to a
12456 period, a question mark, or an exclamation mark, because such a character
12457 might be used in the middle of a sentence. A period, for example, is
12458 used after abbreviations. So other information is needed.
12460 According to convention, you type two spaces after every sentence, but
12461 only one space after a period, a question mark, or an exclamation mark in
12462 the body of a sentence. So a period, a question mark, or an exclamation
12463 mark followed by two spaces is a good indicator of an end of sentence.
12464 However, in a file, the two spaces may instead be a tab or the end of a
12465 line. This means that the regular expression should include these three
12466 items as alternatives.
12469 This group of alternatives will look like this:
12480 Here, @samp{$} indicates the end of the line, and I have pointed out
12481 where the tab and two spaces are inserted in the expression. Both are
12482 inserted by putting the actual characters into the expression.
12484 Two backslashes, @samp{\\}, are required before the parentheses and
12485 vertical bars: the first backslash quotes the following backslash in
12486 Emacs; and the second indicates that the following character, the
12487 parenthesis or the vertical bar, is special.
12490 Also, a sentence may be followed by one or more carriage returns, like
12501 Like tabs and spaces, a carriage return is inserted into a regular
12502 expression by inserting it literally. The asterisk indicates that the
12503 @key{RET} is repeated zero or more times.
12505 But a sentence end does not consist only of a period, a question mark or
12506 an exclamation mark followed by appropriate space: a closing quotation
12507 mark or a closing brace of some kind may precede the space. Indeed more
12508 than one such mark or brace may precede the space. These require a
12509 expression that looks like this:
12515 In this expression, the first @samp{]} is the first character in the
12516 expression; the second character is @samp{"}, which is preceded by a
12517 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12518 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12520 All this suggests what the regular expression pattern for matching the
12521 end of a sentence should be; and, indeed, if we evaluate
12522 @code{sentence-end} we find that it returns the following value:
12527 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12533 (Well, not in GNU Emacs 22; that is because of an effort to make the
12534 process simpler and to handle more glyphs and languages. When the
12535 value of @code{sentence-end} is @code{nil}, then use the value defined
12536 by the function @code{sentence-end}. (Here is a use of the difference
12537 between a value and a function in Emacs Lisp.) The function returns a
12538 value constructed from the variables @code{sentence-end-base},
12539 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12540 and @code{sentence-end-without-space}. The critical variable is
12541 @code{sentence-end-base}; its global value is similar to the one
12542 described above but it also contains two additional quotation marks.
12543 These have differing degrees of curliness. The
12544 @code{sentence-end-without-period} variable, when true, tells Emacs
12545 that a sentence may end without a period, such as text in Thai.)
12549 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12550 literally in the pattern.)
12552 This regular expression can be deciphered as follows:
12556 The first part of the pattern is the three characters, a period, a question
12557 mark and an exclamation mark, within square brackets. The pattern must
12558 begin with one or other of these characters.
12561 The second part of the pattern is the group of closing braces and
12562 quotation marks, which can appear zero or more times. These may follow
12563 the period, question mark or exclamation mark. In a regular expression,
12564 the backslash, @samp{\}, followed by the double quotation mark,
12565 @samp{"}, indicates the class of string-quote characters. Usually, the
12566 double quotation mark is the only character in this class. The
12567 asterisk, @samp{*}, indicates that the items in the previous group (the
12568 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12571 @item \\($\\| \\| \\)
12572 The third part of the pattern is one or other of: either the end of a
12573 line, or two blank spaces, or a tab. The double back-slashes are used
12574 to prevent Emacs from reading the parentheses and vertical bars as part
12575 of the search pattern; the parentheses are used to mark the group and
12576 the vertical bars are used to indicated that the patterns to either side
12577 of them are alternatives. The dollar sign is used to indicate the end
12578 of a line and both the two spaces and the tab are each inserted as is to
12579 indicate what they are.
12582 Finally, the last part of the pattern indicates that the end of the line
12583 or the whitespace following the period, question mark or exclamation
12584 mark may, but need not, be followed by one or more carriage returns. In
12585 the pattern, the carriage return is inserted as an actual carriage
12586 return between square brackets but here it is shown as @key{RET}.
12590 @node re-search-forward, forward-sentence, sentence-end, Regexp Search
12591 @comment node-name, next, previous, up
12592 @section The @code{re-search-forward} Function
12593 @findex re-search-forward
12595 The @code{re-search-forward} function is very like the
12596 @code{search-forward} function. (@xref{search-forward, , The
12597 @code{search-forward} Function}.)
12599 @code{re-search-forward} searches for a regular expression. If the
12600 search is successful, it leaves point immediately after the last
12601 character in the target. If the search is backwards, it leaves point
12602 just before the first character in the target. You may tell
12603 @code{re-search-forward} to return @code{t} for true. (Moving point
12604 is therefore a `side effect'.)
12606 Like @code{search-forward}, the @code{re-search-forward} function takes
12611 The first argument is the regular expression that the function searches
12612 for. The regular expression will be a string between quotations marks.
12615 The optional second argument limits how far the function will search; it is a
12616 bound, which is specified as a position in the buffer.
12619 The optional third argument specifies how the function responds to
12620 failure: @code{nil} as the third argument causes the function to
12621 signal an error (and print a message) when the search fails; any other
12622 value causes it to return @code{nil} if the search fails and @code{t}
12623 if the search succeeds.
12626 The optional fourth argument is the repeat count. A negative repeat
12627 count causes @code{re-search-forward} to search backwards.
12631 The template for @code{re-search-forward} looks like this:
12635 (re-search-forward "@var{regular-expression}"
12636 @var{limit-of-search}
12637 @var{what-to-do-if-search-fails}
12638 @var{repeat-count})
12642 The second, third, and fourth arguments are optional. However, if you
12643 want to pass a value to either or both of the last two arguments, you
12644 must also pass a value to all the preceding arguments. Otherwise, the
12645 Lisp interpreter will mistake which argument you are passing the value
12649 In the @code{forward-sentence} function, the regular expression will be
12650 the value of the variable @code{sentence-end}. In simple form, that is:
12654 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12660 The limit of the search will be the end of the paragraph (since a
12661 sentence cannot go beyond a paragraph). If the search fails, the
12662 function will return @code{nil}; and the repeat count will be provided
12663 by the argument to the @code{forward-sentence} function.
12665 @node forward-sentence, forward-paragraph, re-search-forward, Regexp Search
12666 @comment node-name, next, previous, up
12667 @section @code{forward-sentence}
12668 @findex forward-sentence
12670 The command to move the cursor forward a sentence is a straightforward
12671 illustration of how to use regular expression searches in Emacs Lisp.
12672 Indeed, the function looks longer and more complicated than it is; this
12673 is because the function is designed to go backwards as well as forwards;
12674 and, optionally, over more than one sentence. The function is usually
12675 bound to the key command @kbd{M-e}.
12678 * Complete forward-sentence::
12679 * fwd-sentence while loops:: Two @code{while} loops.
12680 * fwd-sentence re-search:: A regular expression search.
12683 @node Complete forward-sentence, fwd-sentence while loops, forward-sentence, forward-sentence
12685 @unnumberedsubsec Complete @code{forward-sentence} function definition
12689 Here is the code for @code{forward-sentence}:
12694 (defun forward-sentence (&optional arg)
12695 "Move forward to next `sentence-end'. With argument, repeat.
12696 With negative argument, move backward repeatedly to `sentence-beginning'.
12698 The variable `sentence-end' is a regular expression that matches ends of
12699 sentences. Also, every paragraph boundary terminates sentences as well."
12703 (or arg (setq arg 1))
12704 (let ((opoint (point))
12705 (sentence-end (sentence-end)))
12707 (let ((pos (point))
12708 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12709 (if (and (re-search-backward sentence-end par-beg t)
12710 (or (< (match-end 0) pos)
12711 (re-search-backward sentence-end par-beg t)))
12712 (goto-char (match-end 0))
12713 (goto-char par-beg)))
12714 (setq arg (1+ arg)))
12718 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12719 (if (re-search-forward sentence-end par-end t)
12720 (skip-chars-backward " \t\n")
12721 (goto-char par-end)))
12722 (setq arg (1- arg)))
12723 (constrain-to-field nil opoint t)))
12731 (defun forward-sentence (&optional arg)
12732 "Move forward to next sentence-end. With argument, repeat.
12733 With negative argument, move backward repeatedly to sentence-beginning.
12734 Sentence ends are identified by the value of sentence-end
12735 treated as a regular expression. Also, every paragraph boundary
12736 terminates sentences as well."
12740 (or arg (setq arg 1))
12743 (save-excursion (start-of-paragraph-text) (point))))
12744 (if (re-search-backward
12745 (concat sentence-end "[^ \t\n]") par-beg t)
12746 (goto-char (1- (match-end 0)))
12747 (goto-char par-beg)))
12748 (setq arg (1+ arg)))
12751 (save-excursion (end-of-paragraph-text) (point))))
12752 (if (re-search-forward sentence-end par-end t)
12753 (skip-chars-backward " \t\n")
12754 (goto-char par-end)))
12755 (setq arg (1- arg))))
12760 The function looks long at first sight and it is best to look at its
12761 skeleton first, and then its muscle. The way to see the skeleton is to
12762 look at the expressions that start in the left-most columns:
12766 (defun forward-sentence (&optional arg)
12767 "@var{documentation}@dots{}"
12769 (or arg (setq arg 1))
12770 (let ((opoint (point)) (sentence-end (sentence-end)))
12772 (let ((pos (point))
12773 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12774 @var{rest-of-body-of-while-loop-when-going-backwards}
12776 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12777 @var{rest-of-body-of-while-loop-when-going-forwards}
12778 @var{handle-forms-and-equivalent}
12782 This looks much simpler! The function definition consists of
12783 documentation, an @code{interactive} expression, an @code{or}
12784 expression, a @code{let} expression, and @code{while} loops.
12786 Let's look at each of these parts in turn.
12788 We note that the documentation is thorough and understandable.
12790 The function has an @code{interactive "p"} declaration. This means
12791 that the processed prefix argument, if any, is passed to the
12792 function as its argument. (This will be a number.) If the function
12793 is not passed an argument (it is optional) then the argument
12794 @code{arg} will be bound to 1.
12796 When @code{forward-sentence} is called non-interactively without an
12797 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12798 handles this. What it does is either leave the value of @code{arg} as
12799 it is, but only if @code{arg} is bound to a value; or it sets the
12800 value of @code{arg} to 1, in the case when @code{arg} is bound to
12803 Next is a @code{let}. That specifies the values of two local
12804 variables, @code{point} and @code{sentence-end}. The local value of
12805 point, from before the search, is used in the
12806 @code{constrain-to-field} function which handles forms and
12807 equivalents. The @code{sentence-end} variable is set by the
12808 @code{sentence-end} function.
12810 @node fwd-sentence while loops, fwd-sentence re-search, Complete forward-sentence, forward-sentence
12811 @unnumberedsubsec The @code{while} loops
12813 Two @code{while} loops follow. The first @code{while} has a
12814 true-or-false-test that tests true if the prefix argument for
12815 @code{forward-sentence} is a negative number. This is for going
12816 backwards. The body of this loop is similar to the body of the second
12817 @code{while} clause, but it is not exactly the same. We will skip
12818 this @code{while} loop and concentrate on the second @code{while}
12822 The second @code{while} loop is for moving point forward. Its skeleton
12827 (while (> arg 0) ; @r{true-or-false-test}
12829 (if (@var{true-or-false-test})
12832 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12836 The @code{while} loop is of the decrementing kind.
12837 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12838 has a true-or-false-test that tests true so long as the counter (in
12839 this case, the variable @code{arg}) is greater than zero; and it has a
12840 decrementer that subtracts 1 from the value of the counter every time
12843 If no prefix argument is given to @code{forward-sentence}, which is
12844 the most common way the command is used, this @code{while} loop will
12845 run once, since the value of @code{arg} will be 1.
12847 The body of the @code{while} loop consists of a @code{let} expression,
12848 which creates and binds a local variable, and has, as its body, an
12849 @code{if} expression.
12852 The body of the @code{while} loop looks like this:
12857 (save-excursion (end-of-paragraph-text) (point))))
12858 (if (re-search-forward sentence-end par-end t)
12859 (skip-chars-backward " \t\n")
12860 (goto-char par-end)))
12864 The @code{let} expression creates and binds the local variable
12865 @code{par-end}. As we shall see, this local variable is designed to
12866 provide a bound or limit to the regular expression search. If the
12867 search fails to find a proper sentence ending in the paragraph, it will
12868 stop on reaching the end of the paragraph.
12870 But first, let us examine how @code{par-end} is bound to the value of
12871 the end of the paragraph. What happens is that the @code{let} sets the
12872 value of @code{par-end} to the value returned when the Lisp interpreter
12873 evaluates the expression
12877 (save-excursion (end-of-paragraph-text) (point))
12882 In this expression, @code{(end-of-paragraph-text)} moves point to the
12883 end of the paragraph, @code{(point)} returns the value of point, and then
12884 @code{save-excursion} restores point to its original position. Thus,
12885 the @code{let} binds @code{par-end} to the value returned by the
12886 @code{save-excursion} expression, which is the position of the end of
12887 the paragraph. (The @code{end-of-paragraph-text} function uses
12888 @code{forward-paragraph}, which we will discuss shortly.)
12891 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12892 expression that looks like this:
12896 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12897 (skip-chars-backward " \t\n") ; @r{then-part}
12898 (goto-char par-end))) ; @r{else-part}
12902 The @code{if} tests whether its first argument is true and if so,
12903 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12904 evaluates the else-part. The true-or-false-test of the @code{if}
12905 expression is the regular expression search.
12907 It may seem odd to have what looks like the `real work' of
12908 the @code{forward-sentence} function buried here, but this is a common
12909 way this kind of operation is carried out in Lisp.
12911 @node fwd-sentence re-search, , fwd-sentence while loops, forward-sentence
12912 @unnumberedsubsec The regular expression search
12914 The @code{re-search-forward} function searches for the end of the
12915 sentence, that is, for the pattern defined by the @code{sentence-end}
12916 regular expression. If the pattern is found---if the end of the sentence is
12917 found---then the @code{re-search-forward} function does two things:
12921 The @code{re-search-forward} function carries out a side effect, which
12922 is to move point to the end of the occurrence found.
12925 The @code{re-search-forward} function returns a value of true. This is
12926 the value received by the @code{if}, and means that the search was
12931 The side effect, the movement of point, is completed before the
12932 @code{if} function is handed the value returned by the successful
12933 conclusion of the search.
12935 When the @code{if} function receives the value of true from a successful
12936 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12937 which is the expression @code{(skip-chars-backward " \t\n")}. This
12938 expression moves backwards over any blank spaces, tabs or carriage
12939 returns until a printed character is found and then leaves point after
12940 the character. Since point has already been moved to the end of the
12941 pattern that marks the end of the sentence, this action leaves point
12942 right after the closing printed character of the sentence, which is
12945 On the other hand, if the @code{re-search-forward} function fails to
12946 find a pattern marking the end of the sentence, the function returns
12947 false. The false then causes the @code{if} to evaluate its third
12948 argument, which is @code{(goto-char par-end)}: it moves point to the
12949 end of the paragraph.
12951 (And if the text is in a form or equivalent, and point may not move
12952 fully, then the @code{constrain-to-field} function comes into play.)
12954 Regular expression searches are exceptionally useful and the pattern
12955 illustrated by @code{re-search-forward}, in which the search is the
12956 test of an @code{if} expression, is handy. You will see or write code
12957 incorporating this pattern often.
12959 @node forward-paragraph, etags, forward-sentence, Regexp Search
12960 @comment node-name, next, previous, up
12961 @section @code{forward-paragraph}: a Goldmine of Functions
12962 @findex forward-paragraph
12966 (defun forward-paragraph (&optional arg)
12967 "Move forward to end of paragraph.
12968 With argument ARG, do it ARG times;
12969 a negative argument ARG = -N means move backward N paragraphs.
12971 A line which `paragraph-start' matches either separates paragraphs
12972 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12973 A paragraph end is the beginning of a line which is not part of the paragraph
12974 to which the end of the previous line belongs, or the end of the buffer.
12975 Returns the count of paragraphs left to move."
12977 (or arg (setq arg 1))
12978 (let* ((opoint (point))
12979 (fill-prefix-regexp
12980 (and fill-prefix (not (equal fill-prefix ""))
12981 (not paragraph-ignore-fill-prefix)
12982 (regexp-quote fill-prefix)))
12983 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12984 ;; These regexps shouldn't be anchored, because we look for them
12985 ;; starting at the left-margin. This allows paragraph commands to
12986 ;; work normally with indented text.
12987 ;; This hack will not find problem cases like "whatever\\|^something".
12988 (parstart (if (and (not (equal "" paragraph-start))
12989 (equal ?^ (aref paragraph-start 0)))
12990 (substring paragraph-start 1)
12992 (parsep (if (and (not (equal "" paragraph-separate))
12993 (equal ?^ (aref paragraph-separate 0)))
12994 (substring paragraph-separate 1)
12995 paragraph-separate))
12997 (if fill-prefix-regexp
12998 (concat parsep "\\|"
12999 fill-prefix-regexp "[ \t]*$")
13001 ;; This is used for searching.
13002 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
13004 (while (and (< arg 0) (not (bobp)))
13005 (if (and (not (looking-at parsep))
13006 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
13007 (looking-at parsep))
13008 (setq arg (1+ arg))
13009 (setq start (point))
13010 ;; Move back over paragraph-separating lines.
13011 (forward-char -1) (beginning-of-line)
13012 (while (and (not (bobp))
13013 (progn (move-to-left-margin)
13014 (looking-at parsep)))
13018 (setq arg (1+ arg))
13019 ;; Go to end of the previous (non-separating) line.
13021 ;; Search back for line that starts or separates paragraphs.
13022 (if (if fill-prefix-regexp
13023 ;; There is a fill prefix; it overrides parstart.
13024 (let (multiple-lines)
13025 (while (and (progn (beginning-of-line) (not (bobp)))
13026 (progn (move-to-left-margin)
13027 (not (looking-at parsep)))
13028 (looking-at fill-prefix-regexp))
13029 (unless (= (point) start)
13030 (setq multiple-lines t))
13032 (move-to-left-margin)
13033 ;; This deleted code caused a long hanging-indent line
13034 ;; not to be filled together with the following lines.
13035 ;; ;; Don't move back over a line before the paragraph
13036 ;; ;; which doesn't start with fill-prefix
13037 ;; ;; unless that is the only line we've moved over.
13038 ;; (and (not (looking-at fill-prefix-regexp))
13040 ;; (forward-line 1))
13042 (while (and (re-search-backward sp-parstart nil 1)
13043 (setq found-start t)
13044 ;; Found a candidate, but need to check if it is a
13046 (progn (setq start (point))
13047 (move-to-left-margin)
13048 (not (looking-at parsep)))
13049 (not (and (looking-at parstart)
13050 (or (not use-hard-newlines)
13053 (1- start) 'hard)))))
13054 (setq found-start nil)
13059 ;; Move forward over paragraph separators.
13060 ;; We know this cannot reach the place we started
13061 ;; because we know we moved back over a non-separator.
13062 (while (and (not (eobp))
13063 (progn (move-to-left-margin)
13064 (looking-at parsep)))
13066 ;; If line before paragraph is just margin, back up to there.
13068 (if (> (current-column) (current-left-margin))
13070 (skip-chars-backward " \t")
13072 (forward-line 1))))
13073 ;; No starter or separator line => use buffer beg.
13074 (goto-char (point-min))))))
13076 (while (and (> arg 0) (not (eobp)))
13077 ;; Move forward over separator lines...
13078 (while (and (not (eobp))
13079 (progn (move-to-left-margin) (not (eobp)))
13080 (looking-at parsep))
13082 (unless (eobp) (setq arg (1- arg)))
13083 ;; ... and one more line.
13085 (if fill-prefix-regexp
13086 ;; There is a fill prefix; it overrides parstart.
13087 (while (and (not (eobp))
13088 (progn (move-to-left-margin) (not (eobp)))
13089 (not (looking-at parsep))
13090 (looking-at fill-prefix-regexp))
13092 (while (and (re-search-forward sp-parstart nil 1)
13093 (progn (setq start (match-beginning 0))
13096 (progn (move-to-left-margin)
13097 (not (looking-at parsep)))
13098 (or (not (looking-at parstart))
13099 (and use-hard-newlines
13100 (not (get-text-property (1- start) 'hard)))))
13102 (if (< (point) (point-max))
13103 (goto-char start))))
13104 (constrain-to-field nil opoint t)
13105 ;; Return the number of steps that could not be done.
13109 The @code{forward-paragraph} function moves point forward to the end
13110 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
13111 number of functions that are important in themselves, including
13112 @code{let*}, @code{match-beginning}, and @code{looking-at}.
13114 The function definition for @code{forward-paragraph} is considerably
13115 longer than the function definition for @code{forward-sentence}
13116 because it works with a paragraph, each line of which may begin with a
13119 A fill prefix consists of a string of characters that are repeated at
13120 the beginning of each line. For example, in Lisp code, it is a
13121 convention to start each line of a paragraph-long comment with
13122 @samp{;;; }. In Text mode, four blank spaces make up another common
13123 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
13124 emacs, The GNU Emacs Manual}, for more information about fill
13127 The existence of a fill prefix means that in addition to being able to
13128 find the end of a paragraph whose lines begin on the left-most
13129 column, the @code{forward-paragraph} function must be able to find the
13130 end of a paragraph when all or many of the lines in the buffer begin
13131 with the fill prefix.
13133 Moreover, it is sometimes practical to ignore a fill prefix that
13134 exists, especially when blank lines separate paragraphs.
13135 This is an added complication.
13138 * forward-paragraph in brief:: Key parts of the function definition.
13139 * fwd-para let:: The @code{let*} expression.
13140 * fwd-para while:: The forward motion @code{while} loop.
13143 @node forward-paragraph in brief, fwd-para let, forward-paragraph, forward-paragraph
13145 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13148 Rather than print all of the @code{forward-paragraph} function, we
13149 will only print parts of it. Read without preparation, the function
13153 In outline, the function looks like this:
13157 (defun forward-paragraph (&optional arg)
13158 "@var{documentation}@dots{}"
13160 (or arg (setq arg 1))
13163 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13165 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13170 The first parts of the function are routine: the function's argument
13171 list consists of one optional argument. Documentation follows.
13173 The lower case @samp{p} in the @code{interactive} declaration means
13174 that the processed prefix argument, if any, is passed to the function.
13175 This will be a number, and is the repeat count of how many paragraphs
13176 point will move. The @code{or} expression in the next line handles
13177 the common case when no argument is passed to the function, which occurs
13178 if the function is called from other code rather than interactively.
13179 This case was described earlier. (@xref{forward-sentence, The
13180 @code{forward-sentence} function}.) Now we reach the end of the
13181 familiar part of this function.
13183 @node fwd-para let, fwd-para while, forward-paragraph in brief, forward-paragraph
13184 @unnumberedsubsec The @code{let*} expression
13186 The next line of the @code{forward-paragraph} function begins a
13187 @code{let*} expression. This is a different than @code{let}. The
13188 symbol is @code{let*} not @code{let}.
13190 The @code{let*} special form is like @code{let} except that Emacs sets
13191 each variable in sequence, one after another, and variables in the
13192 latter part of the varlist can make use of the values to which Emacs
13193 set variables in the earlier part of the varlist.
13196 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13199 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13201 In the @code{let*} expression in this function, Emacs binds a total of
13202 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13203 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13204 @code{found-start}.
13206 The variable @code{parsep} appears twice, first, to remove instances
13207 of @samp{^}, and second, to handle fill prefixes.
13209 The variable @code{opoint} is just the value of @code{point}. As you
13210 can guess, it is used in a @code{constrain-to-field} expression, just
13211 as in @code{forward-sentence}.
13213 The variable @code{fill-prefix-regexp} is set to the value returned by
13214 evaluating the following list:
13219 (not (equal fill-prefix ""))
13220 (not paragraph-ignore-fill-prefix)
13221 (regexp-quote fill-prefix))
13226 This is an expression whose first element is the @code{and} special form.
13228 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13229 function}), the @code{and} special form evaluates each of its
13230 arguments until one of the arguments returns a value of @code{nil}, in
13231 which case the @code{and} expression returns @code{nil}; however, if
13232 none of the arguments returns a value of @code{nil}, the value
13233 resulting from evaluating the last argument is returned. (Since such
13234 a value is not @code{nil}, it is considered true in Lisp.) In other
13235 words, an @code{and} expression returns a true value only if all its
13236 arguments are true.
13239 In this case, the variable @code{fill-prefix-regexp} is bound to a
13240 non-@code{nil} value only if the following four expressions produce a
13241 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13242 @code{fill-prefix-regexp} is bound to @code{nil}.
13246 When this variable is evaluated, the value of the fill prefix, if any,
13247 is returned. If there is no fill prefix, this variable returns
13250 @item (not (equal fill-prefix "")
13251 This expression checks whether an existing fill prefix is an empty
13252 string, that is, a string with no characters in it. An empty string is
13253 not a useful fill prefix.
13255 @item (not paragraph-ignore-fill-prefix)
13256 This expression returns @code{nil} if the variable
13257 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13258 true value such as @code{t}.
13260 @item (regexp-quote fill-prefix)
13261 This is the last argument to the @code{and} special form. If all the
13262 arguments to the @code{and} are true, the value resulting from
13263 evaluating this expression will be returned by the @code{and} expression
13264 and bound to the variable @code{fill-prefix-regexp},
13267 @findex regexp-quote
13269 The result of evaluating this @code{and} expression successfully is that
13270 @code{fill-prefix-regexp} will be bound to the value of
13271 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13272 What @code{regexp-quote} does is read a string and return a regular
13273 expression that will exactly match the string and match nothing else.
13274 This means that @code{fill-prefix-regexp} will be set to a value that
13275 will exactly match the fill prefix if the fill prefix exists.
13276 Otherwise, the variable will be set to @code{nil}.
13278 The next two local variables in the @code{let*} expression are
13279 designed to remove instances of @samp{^} from @code{parstart} and
13280 @code{parsep}, the local variables which indicate the paragraph start
13281 and the paragraph separator. The next expression sets @code{parsep}
13282 again. That is to handle fill prefixes.
13284 This is the setting that requires the definition call @code{let*}
13285 rather than @code{let}. The true-or-false-test for the @code{if}
13286 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13287 @code{nil} or some other value.
13289 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13290 the else-part of the @code{if} expression and binds @code{parsep} to
13291 its local value. (@code{parsep} is a regular expression that matches
13292 what separates paragraphs.)
13294 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13295 the then-part of the @code{if} expression and binds @code{parsep} to a
13296 regular expression that includes the @code{fill-prefix-regexp} as part
13299 Specifically, @code{parsep} is set to the original value of the
13300 paragraph separate regular expression concatenated with an alternative
13301 expression that consists of the @code{fill-prefix-regexp} followed by
13302 optional whitespace to the end of the line. The whitespace is defined
13303 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13304 regexp as an alternative to @code{parsep}.
13306 According to a comment in the code, the next local variable,
13307 @code{sp-parstart}, is used for searching, and then the final two,
13308 @code{start} and @code{found-start}, are set to @code{nil}.
13310 Now we get into the body of the @code{let*}. The first part of the body
13311 of the @code{let*} deals with the case when the function is given a
13312 negative argument and is therefore moving backwards. We will skip this
13315 @node fwd-para while, , fwd-para let, forward-paragraph
13316 @unnumberedsubsec The forward motion @code{while} loop
13318 The second part of the body of the @code{let*} deals with forward
13319 motion. It is a @code{while} loop that repeats itself so long as the
13320 value of @code{arg} is greater than zero. In the most common use of
13321 the function, the value of the argument is 1, so the body of the
13322 @code{while} loop is evaluated exactly once, and the cursor moves
13323 forward one paragraph.
13326 (while (and (> arg 0) (not (eobp)))
13328 ;; Move forward over separator lines...
13329 (while (and (not (eobp))
13330 (progn (move-to-left-margin) (not (eobp)))
13331 (looking-at parsep))
13333 (unless (eobp) (setq arg (1- arg)))
13334 ;; ... and one more line.
13337 (if fill-prefix-regexp
13338 ;; There is a fill prefix; it overrides parstart.
13339 (while (and (not (eobp))
13340 (progn (move-to-left-margin) (not (eobp)))
13341 (not (looking-at parsep))
13342 (looking-at fill-prefix-regexp))
13345 (while (and (re-search-forward sp-parstart nil 1)
13346 (progn (setq start (match-beginning 0))
13349 (progn (move-to-left-margin)
13350 (not (looking-at parsep)))
13351 (or (not (looking-at parstart))
13352 (and use-hard-newlines
13353 (not (get-text-property (1- start) 'hard)))))
13356 (if (< (point) (point-max))
13357 (goto-char start))))
13360 This part handles three situations: when point is between paragraphs,
13361 when there is a fill prefix and when there is no fill prefix.
13364 The @code{while} loop looks like this:
13368 ;; @r{going forwards and not at the end of the buffer}
13369 (while (and (> arg 0) (not (eobp)))
13371 ;; @r{between paragraphs}
13372 ;; Move forward over separator lines...
13373 (while (and (not (eobp))
13374 (progn (move-to-left-margin) (not (eobp)))
13375 (looking-at parsep))
13377 ;; @r{This decrements the loop}
13378 (unless (eobp) (setq arg (1- arg)))
13379 ;; ... and one more line.
13384 (if fill-prefix-regexp
13385 ;; There is a fill prefix; it overrides parstart;
13386 ;; we go forward line by line
13387 (while (and (not (eobp))
13388 (progn (move-to-left-margin) (not (eobp)))
13389 (not (looking-at parsep))
13390 (looking-at fill-prefix-regexp))
13395 ;; There is no fill prefix;
13396 ;; we go forward character by character
13397 (while (and (re-search-forward sp-parstart nil 1)
13398 (progn (setq start (match-beginning 0))
13401 (progn (move-to-left-margin)
13402 (not (looking-at parsep)))
13403 (or (not (looking-at parstart))
13404 (and use-hard-newlines
13405 (not (get-text-property (1- start) 'hard)))))
13410 ;; and if there is no fill prefix and if we are not at the end,
13411 ;; go to whatever was found in the regular expression search
13413 (if (< (point) (point-max))
13414 (goto-char start))))
13419 We can see that this is a decrementing counter @code{while} loop,
13420 using the expression @code{(setq arg (1- arg))} as the decrementer.
13421 That expression is not far from the @code{while}, but is hidden in
13422 another Lisp macro, an @code{unless} macro. Unless we are at the end
13423 of the buffer --- that is what the @code{eobp} function determines; it
13424 is an abbreviation of @samp{End Of Buffer P} --- we decrease the value
13425 of @code{arg} by one.
13427 (If we are at the end of the buffer, we cannot go forward any more and
13428 the next loop of the @code{while} expression will test false since the
13429 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13430 function means exactly as you expect; it is another name for
13431 @code{null}, a function that returns true when its argument is false.)
13433 Interestingly, the loop count is not decremented until we leave the
13434 space between paragraphs, unless we come to the end of buffer or stop
13435 seeing the local value of the paragraph separator.
13437 That second @code{while} also has a @code{(move-to-left-margin)}
13438 expression. The function is self-explanatory. It is inside a
13439 @code{progn} expression and not the last element of its body, so it is
13440 only invoked for its side effect, which is to move point to the left
13441 margin of the current line.
13444 The @code{looking-at} function is also self-explanatory; it returns
13445 true if the text after point matches the regular expression given as
13448 The rest of the body of the loop looks difficult at first, but makes
13449 sense as you come to understand it.
13452 First consider what happens if there is a fill prefix:
13456 (if fill-prefix-regexp
13457 ;; There is a fill prefix; it overrides parstart;
13458 ;; we go forward line by line
13459 (while (and (not (eobp))
13460 (progn (move-to-left-margin) (not (eobp)))
13461 (not (looking-at parsep))
13462 (looking-at fill-prefix-regexp))
13468 This expression moves point forward line by line so long
13469 as four conditions are true:
13473 Point is not at the end of the buffer.
13476 We can move to the left margin of the text and are
13477 not at the end of the buffer.
13480 The text following point does not separate paragraphs.
13483 The pattern following point is the fill prefix regular expression.
13486 The last condition may be puzzling, until you remember that point was
13487 moved to the beginning of the line early in the @code{forward-paragraph}
13488 function. This means that if the text has a fill prefix, the
13489 @code{looking-at} function will see it.
13492 Consider what happens when there is no fill prefix.
13496 (while (and (re-search-forward sp-parstart nil 1)
13497 (progn (setq start (match-beginning 0))
13500 (progn (move-to-left-margin)
13501 (not (looking-at parsep)))
13502 (or (not (looking-at parstart))
13503 (and use-hard-newlines
13504 (not (get-text-property (1- start) 'hard)))))
13510 This @code{while} loop has us searching forward for
13511 @code{sp-parstart}, which is the combination of possible whitespace
13512 with a the local value of the start of a paragraph or of a paragraph
13513 separator. (The latter two are within an expression starting
13514 @code{\(?:} so that they are not referenced by the
13515 @code{match-beginning} function.)
13518 The two expressions,
13522 (setq start (match-beginning 0))
13528 mean go to the start of the text matched by the regular expression
13531 The @code{(match-beginning 0)} expression is new. It returns a number
13532 specifying the location of the start of the text that was matched by
13535 The @code{match-beginning} function is used here because of a
13536 characteristic of a forward search: a successful forward search,
13537 regardless of whether it is a plain search or a regular expression
13538 search, moves point to the end of the text that is found. In this
13539 case, a successful search moves point to the end of the pattern for
13540 @code{sp-parstart}.
13542 However, we want to put point at the end of the current paragraph, not
13543 somewhere else. Indeed, since the search possibly includes the
13544 paragraph separator, point may end up at the beginning of the next one
13545 unless we use an expression that includes @code{match-beginning}.
13547 @findex match-beginning
13548 When given an argument of 0, @code{match-beginning} returns the
13549 position that is the start of the text matched by the most recent
13550 search. In this case, the most recent search looks for
13551 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13552 the beginning position of that pattern, rather than the end position
13555 (Incidentally, when passed a positive number as an argument, the
13556 @code{match-beginning} function returns the location of point at that
13557 parenthesized expression in the last search unless that parenthesized
13558 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13559 appears here since the argument is 0.)
13562 The last expression when there is no fill prefix is
13566 (if (< (point) (point-max))
13567 (goto-char start))))
13572 This says that if there is no fill prefix and if we are not at the
13573 end, point should move to the beginning of whatever was found by the
13574 regular expression search for @code{sp-parstart}.
13576 The full definition for the @code{forward-paragraph} function not only
13577 includes code for going forwards, but also code for going backwards.
13579 If you are reading this inside of GNU Emacs and you want to see the
13580 whole function, you can type @kbd{C-h f} (@code{describe-function})
13581 and the name of the function. This gives you the function
13582 documentation and the name of the library containing the function's
13583 source. Place point over the name of the library and press the RET
13584 key; you will be taken directly to the source. (Be sure to install
13585 your sources! Without them, you are like a person who tries to drive
13586 a car with his eyes shut!)
13588 @node etags, Regexp Review, forward-paragraph, Regexp Search
13589 @section Create Your Own @file{TAGS} File
13591 @cindex @file{TAGS} file, create own
13593 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13594 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13595 name of the function when prompted for it. This is a good habit to
13596 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13597 to the source for a function, variable, or node. The function depends
13598 on tags tables to tell it where to go.
13600 If the @code{find-tag} function first asks you for the name of a
13601 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13602 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13603 @file{TAGS} file depends on how your copy of Emacs was installed. I
13604 just told you the location that provides both my C and my Emacs Lisp
13607 You can also create your own @file{TAGS} file for directories that
13610 You often need to build and install tags tables yourself. They are
13611 not built automatically. A tags table is called a @file{TAGS} file;
13612 the name is in upper case letters.
13614 You can create a @file{TAGS} file by calling the @code{etags} program
13615 that comes as a part of the Emacs distribution. Usually, @code{etags}
13616 is compiled and installed when Emacs is built. (@code{etags} is not
13617 an Emacs Lisp function or a part of Emacs; it is a C program.)
13620 To create a @file{TAGS} file, first switch to the directory in which
13621 you want to create the file. In Emacs you can do this with the
13622 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13623 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13624 compile command, with @w{@code{etags *.el}} as the command to execute
13627 M-x compile RET etags *.el RET
13631 to create a @file{TAGS} file for Emacs Lisp.
13633 For example, if you have a large number of files in your
13634 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13635 of which I load 12---you can create a @file{TAGS} file for the Emacs
13636 Lisp files in that directory.
13639 The @code{etags} program takes all the usual shell `wildcards'. For
13640 example, if you have two directories for which you want a single
13641 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13642 @file{../elisp/} is the second directory:
13645 M-x compile RET etags *.el ../elisp/*.el RET
13652 M-x compile RET etags --help RET
13656 to see a list of the options accepted by @code{etags} as well as a
13657 list of supported languages.
13659 The @code{etags} program handles more than 20 languages, including
13660 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13661 LaTeX, Pascal, Perl, Postscript, Python, TeX, Texinfo, makefiles, and
13662 most assemblers. The program has no switches for specifying the
13663 language; it recognizes the language in an input file according to its
13664 file name and contents.
13666 @file{etags} is very helpful when you are writing code yourself and
13667 want to refer back to functions you have already written. Just run
13668 @code{etags} again at intervals as you write new functions, so they
13669 become part of the @file{TAGS} file.
13671 If you think an appropriate @file{TAGS} file already exists for what
13672 you want, but do not know where it is, you can use the @code{locate}
13673 program to attempt to find it.
13675 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13676 for you the full path names of all your @file{TAGS} files. On my
13677 system, this command lists 34 @file{TAGS} files. On the other hand, a
13678 `plain vanilla' system I recently installed did not contain any
13681 If the tags table you want has been created, you can use the @code{M-x
13682 visit-tags-table} command to specify it. Otherwise, you will need to
13683 create the tag table yourself and then use @code{M-x
13686 @subsubheading Building Tags in the Emacs sources
13687 @cindex Building Tags in the Emacs sources
13688 @cindex Tags in the Emacs sources
13691 The GNU Emacs sources come with a @file{Makefile} that contains a
13692 sophisticated @code{etags} command that creates, collects, and merges
13693 tags tables from all over the Emacs sources and puts the information
13694 into one @file{TAGS} file in the @file{src/} directory. (The
13695 @file{src/} directory is below the top level of your Emacs directory.)
13698 To build this @file{TAGS} file, go to the top level of your Emacs
13699 source directory and run the compile command @code{make tags}:
13702 M-x compile RET make tags RET
13706 (The @code{make tags} command works well with the GNU Emacs sources,
13707 as well as with some other source packages.)
13709 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13712 @node Regexp Review, re-search Exercises, etags, Regexp Search
13713 @comment node-name, next, previous, up
13716 Here is a brief summary of some recently introduced functions.
13720 Repeatedly evaluate the body of the expression so long as the first
13721 element of the body tests true. Then return @code{nil}. (The
13722 expression is evaluated only for its side effects.)
13731 (insert (format "foo is %d.\n" foo))
13732 (setq foo (1- foo))))
13734 @result{} foo is 2.
13741 (The @code{insert} function inserts its arguments at point; the
13742 @code{format} function returns a string formatted from its arguments
13743 the way @code{message} formats its arguments; @code{\n} produces a new
13746 @item re-search-forward
13747 Search for a pattern, and if the pattern is found, move point to rest
13751 Takes four arguments, like @code{search-forward}:
13755 A regular expression that specifies the pattern to search for.
13756 (Remember to put quotation marks around this argument!)
13759 Optionally, the limit of the search.
13762 Optionally, what to do if the search fails, return @code{nil} or an
13766 Optionally, how many times to repeat the search; if negative, the
13767 search goes backwards.
13771 Bind some variables locally to particular values,
13772 and then evaluate the remaining arguments, returning the value of the
13773 last one. While binding the local variables, use the local values of
13774 variables bound earlier, if any.
13783 (message "`bar' is %d." bar))
13784 @result{} `bar' is 21.
13788 @item match-beginning
13789 Return the position of the start of the text found by the last regular
13793 Return @code{t} for true if the text after point matches the argument,
13794 which should be a regular expression.
13797 Return @code{t} for true if point is at the end of the accessible part
13798 of a buffer. The end of the accessible part is the end of the buffer
13799 if the buffer is not narrowed; it is the end of the narrowed part if
13800 the buffer is narrowed.
13804 @node re-search Exercises, , Regexp Review, Regexp Search
13805 @section Exercises with @code{re-search-forward}
13809 Write a function to search for a regular expression that matches two
13810 or more blank lines in sequence.
13813 Write a function to search for duplicated words, such as `the the'.
13814 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13815 Manual}, for information on how to write a regexp (a regular
13816 expression) to match a string that is composed of two identical
13817 halves. You can devise several regexps; some are better than others.
13818 The function I use is described in an appendix, along with several
13819 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13822 @node Counting Words, Words in a defun, Regexp Search, Top
13823 @chapter Counting: Repetition and Regexps
13824 @cindex Repetition for word counting
13825 @cindex Regular expressions for word counting
13827 Repetition and regular expression searches are powerful tools that you
13828 often use when you write code in Emacs Lisp. This chapter illustrates
13829 the use of regular expression searches through the construction of
13830 word count commands using @code{while} loops and recursion.
13833 * Why Count Words::
13834 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
13835 * recursive-count-words:: Start with case of no words in region.
13836 * Counting Exercise::
13839 @node Why Count Words, @value{COUNT-WORDS}, Counting Words, Counting Words
13841 @unnumberedsec Counting words
13844 The standard Emacs distribution contains functions for counting the
13845 number of lines and words within a region.
13847 Certain types of writing ask you to count words. Thus, if you write
13848 an essay, you may be limited to 800 words; if you write a novel, you
13849 may discipline yourself to write 1000 words a day. It seems odd, but
13850 for a long time, Emacs lacked a word count command. Perhaps people used
13851 Emacs mostly for code or types of documentation that did not require
13852 word counts; or perhaps they restricted themselves to the operating
13853 system word count command, @code{wc}. Alternatively, people may have
13854 followed the publishers' convention and computed a word count by
13855 dividing the number of characters in a document by five.
13857 There are many ways to implement a command to count words. Here are
13858 some examples, which you may wish to compare with the standard Emacs
13859 command, @code{count-words-region}.
13861 @node @value{COUNT-WORDS}, recursive-count-words, Why Count Words, Counting Words
13862 @comment node-name, next, previous, up
13863 @section The @code{@value{COUNT-WORDS}} Function
13864 @findex @value{COUNT-WORDS}
13866 A word count command could count words in a line, paragraph, region,
13867 or buffer. What should the command cover? You could design the
13868 command to count the number of words in a complete buffer. However,
13869 the Emacs tradition encourages flexibility---you may want to count
13870 words in just a section, rather than all of a buffer. So it makes
13871 more sense to design the command to count the number of words in a
13872 region. Once you have a command to count words in a region, you can,
13873 if you wish, count words in a whole buffer by marking it with
13874 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13876 Clearly, counting words is a repetitive act: starting from the
13877 beginning of the region, you count the first word, then the second
13878 word, then the third word, and so on, until you reach the end of the
13879 region. This means that word counting is ideally suited to recursion
13880 or to a @code{while} loop.
13883 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
13884 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
13887 @node Design @value{COUNT-WORDS}, Whitespace Bug, @value{COUNT-WORDS}, @value{COUNT-WORDS}
13889 @unnumberedsubsec Designing @code{@value{COUNT-WORDS}}
13892 First, we will implement the word count command with a @code{while}
13893 loop, then with recursion. The command will, of course, be
13897 The template for an interactive function definition is, as always:
13901 (defun @var{name-of-function} (@var{argument-list})
13902 "@var{documentation}@dots{}"
13903 (@var{interactive-expression}@dots{})
13908 What we need to do is fill in the slots.
13910 The name of the function should be self-explanatory and similar to the
13911 existing @code{count-lines-region} name. This makes the name easier
13912 to remember. @code{count-words-region} is the obvious choice. Since
13913 that name is now used for the standard Emacs command to count words, we
13914 will name our implementation @code{@value{COUNT-WORDS}}.
13916 The function counts words within a region. This means that the
13917 argument list must contain symbols that are bound to the two
13918 positions, the beginning and end of the region. These two positions
13919 can be called @samp{beginning} and @samp{end} respectively. The first
13920 line of the documentation should be a single sentence, since that is
13921 all that is printed as documentation by a command such as
13922 @code{apropos}. The interactive expression will be of the form
13923 @samp{(interactive "r")}, since that will cause Emacs to pass the
13924 beginning and end of the region to the function's argument list. All
13927 The body of the function needs to be written to do three tasks:
13928 first, to set up conditions under which the @code{while} loop can
13929 count words, second, to run the @code{while} loop, and third, to send
13930 a message to the user.
13932 When a user calls @code{@value{COUNT-WORDS}}, point may be at the
13933 beginning or the end of the region. However, the counting process
13934 must start at the beginning of the region. This means we will want
13935 to put point there if it is not already there. Executing
13936 @code{(goto-char beginning)} ensures this. Of course, we will want to
13937 return point to its expected position when the function finishes its
13938 work. For this reason, the body must be enclosed in a
13939 @code{save-excursion} expression.
13941 The central part of the body of the function consists of a
13942 @code{while} loop in which one expression jumps point forward word by
13943 word, and another expression counts those jumps. The true-or-false-test
13944 of the @code{while} loop should test true so long as point should jump
13945 forward, and false when point is at the end of the region.
13947 We could use @code{(forward-word 1)} as the expression for moving point
13948 forward word by word, but it is easier to see what Emacs identifies as a
13949 `word' if we use a regular expression search.
13951 A regular expression search that finds the pattern for which it is
13952 searching leaves point after the last character matched. This means
13953 that a succession of successful word searches will move point forward
13956 As a practical matter, we want the regular expression search to jump
13957 over whitespace and punctuation between words as well as over the
13958 words themselves. A regexp that refuses to jump over interword
13959 whitespace would never jump more than one word! This means that
13960 the regexp should include the whitespace and punctuation that follows
13961 a word, if any, as well as the word itself. (A word may end a buffer
13962 and not have any following whitespace or punctuation, so that part of
13963 the regexp must be optional.)
13965 Thus, what we want for the regexp is a pattern defining one or more
13966 word constituent characters followed, optionally, by one or more
13967 characters that are not word constituents. The regular expression for
13975 The buffer's syntax table determines which characters are and are not
13976 word constituents. (@xref{Syntax, , What Constitutes a Word or
13977 Symbol?}, for more about syntax. Also, see @ref{Syntax, Syntax, The
13978 Syntax Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, ,
13979 Syntax Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
13982 The search expression looks like this:
13985 (re-search-forward "\\w+\\W*")
13989 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13990 single backslash has special meaning to the Emacs Lisp interpreter.
13991 It indicates that the following character is interpreted differently
13992 than usual. For example, the two characters, @samp{\n}, stand for
13993 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13994 backslashes in a row stand for an ordinary, `unspecial' backslash, so
13995 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13996 letter. So it discovers the letter is special.)
13998 We need a counter to count how many words there are; this variable
13999 must first be set to 0 and then incremented each time Emacs goes
14000 around the @code{while} loop. The incrementing expression is simply:
14003 (setq count (1+ count))
14006 Finally, we want to tell the user how many words there are in the
14007 region. The @code{message} function is intended for presenting this
14008 kind of information to the user. The message has to be phrased so
14009 that it reads properly regardless of how many words there are in the
14010 region: we don't want to say that ``there are 1 words in the region''.
14011 The conflict between singular and plural is ungrammatical. We can
14012 solve this problem by using a conditional expression that evaluates
14013 different messages depending on the number of words in the region.
14014 There are three possibilities: no words in the region, one word in the
14015 region, and more than one word. This means that the @code{cond}
14016 special form is appropriate.
14019 All this leads to the following function definition:
14023 ;;; @r{First version; has bugs!}
14024 (defun @value{COUNT-WORDS} (beginning end)
14025 "Print number of words in the region.
14026 Words are defined as at least one word-constituent
14027 character followed by at least one character that
14028 is not a word-constituent. The buffer's syntax
14029 table determines which characters these are."
14031 (message "Counting words in region ... ")
14035 ;;; @r{1. Set up appropriate conditions.}
14037 (goto-char beginning)
14042 ;;; @r{2. Run the} while @r{loop.}
14043 (while (< (point) end)
14044 (re-search-forward "\\w+\\W*")
14045 (setq count (1+ count)))
14049 ;;; @r{3. Send a message to the user.}
14050 (cond ((zerop count)
14052 "The region does NOT have any words."))
14055 "The region has 1 word."))
14058 "The region has %d words." count))))))
14063 As written, the function works, but not in all circumstances.
14065 @node Whitespace Bug, , Design @value{COUNT-WORDS}, @value{COUNT-WORDS}
14066 @comment node-name, next, previous, up
14067 @subsection The Whitespace Bug in @code{@value{COUNT-WORDS}}
14069 The @code{@value{COUNT-WORDS}} command described in the preceding
14070 section has two bugs, or rather, one bug with two manifestations.
14071 First, if you mark a region containing only whitespace in the middle
14072 of some text, the @code{@value{COUNT-WORDS}} command tells you that the
14073 region contains one word! Second, if you mark a region containing
14074 only whitespace at the end of the buffer or the accessible portion of
14075 a narrowed buffer, the command displays an error message that looks
14079 Search failed: "\\w+\\W*"
14082 If you are reading this in Info in GNU Emacs, you can test for these
14085 First, evaluate the function in the usual manner to install it.
14087 Here is a copy of the definition. Place your cursor after the closing
14088 parenthesis and type @kbd{C-x C-e} to install it.
14092 ;; @r{First version; has bugs!}
14093 (defun @value{COUNT-WORDS} (beginning end)
14094 "Print number of words in the region.
14095 Words are defined as at least one word-constituent character followed
14096 by at least one character that is not a word-constituent. The buffer's
14097 syntax table determines which characters these are."
14101 (message "Counting words in region ... ")
14105 ;;; @r{1. Set up appropriate conditions.}
14107 (goto-char beginning)
14112 ;;; @r{2. Run the} while @r{loop.}
14113 (while (< (point) end)
14114 (re-search-forward "\\w+\\W*")
14115 (setq count (1+ count)))
14119 ;;; @r{3. Send a message to the user.}
14120 (cond ((zerop count)
14121 (message "The region does NOT have any words."))
14122 ((= 1 count) (message "The region has 1 word."))
14123 (t (message "The region has %d words." count))))))
14129 If you wish, you can also install this keybinding by evaluating it:
14132 (global-set-key "\C-c=" '@value{COUNT-WORDS})
14135 To conduct the first test, set mark and point to the beginning and end
14136 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14137 @value{COUNT-WORDS}} if you have not bound @kbd{C-c =}):
14144 Emacs will tell you, correctly, that the region has three words.
14146 Repeat the test, but place mark at the beginning of the line and place
14147 point just @emph{before} the word @samp{one}. Again type the command
14148 @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}). Emacs should tell you
14149 that the region has no words, since it is composed only of the
14150 whitespace at the beginning of the line. But instead Emacs tells you
14151 that the region has one word!
14153 For the third test, copy the sample line to the end of the
14154 @file{*scratch*} buffer and then type several spaces at the end of the
14155 line. Place mark right after the word @samp{three} and point at the
14156 end of line. (The end of the line will be the end of the buffer.)
14157 Type @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}) as you did before.
14158 Again, Emacs should tell you that the region has no words, since it is
14159 composed only of the whitespace at the end of the line. Instead,
14160 Emacs displays an error message saying @samp{Search failed}.
14162 The two bugs stem from the same problem.
14164 Consider the first manifestation of the bug, in which the command
14165 tells you that the whitespace at the beginning of the line contains
14166 one word. What happens is this: The @code{M-x @value{COUNT-WORDS}}
14167 command moves point to the beginning of the region. The @code{while}
14168 tests whether the value of point is smaller than the value of
14169 @code{end}, which it is. Consequently, the regular expression search
14170 looks for and finds the first word. It leaves point after the word.
14171 @code{count} is set to one. The @code{while} loop repeats; but this
14172 time the value of point is larger than the value of @code{end}, the
14173 loop is exited; and the function displays a message saying the number
14174 of words in the region is one. In brief, the regular expression
14175 search looks for and finds the word even though it is outside
14178 In the second manifestation of the bug, the region is whitespace at
14179 the end of the buffer. Emacs says @samp{Search failed}. What happens
14180 is that the true-or-false-test in the @code{while} loop tests true, so
14181 the search expression is executed. But since there are no more words
14182 in the buffer, the search fails.
14184 In both manifestations of the bug, the search extends or attempts to
14185 extend outside of the region.
14187 The solution is to limit the search to the region---this is a fairly
14188 simple action, but as you may have come to expect, it is not quite as
14189 simple as you might think.
14191 As we have seen, the @code{re-search-forward} function takes a search
14192 pattern as its first argument. But in addition to this first,
14193 mandatory argument, it accepts three optional arguments. The optional
14194 second argument bounds the search. The optional third argument, if
14195 @code{t}, causes the function to return @code{nil} rather than signal
14196 an error if the search fails. The optional fourth argument is a
14197 repeat count. (In Emacs, you can see a function's documentation by
14198 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14200 In the @code{@value{COUNT-WORDS}} definition, the value of the end of
14201 the region is held by the variable @code{end} which is passed as an
14202 argument to the function. Thus, we can add @code{end} as an argument
14203 to the regular expression search expression:
14206 (re-search-forward "\\w+\\W*" end)
14209 However, if you make only this change to the @code{@value{COUNT-WORDS}}
14210 definition and then test the new version of the definition on a
14211 stretch of whitespace, you will receive an error message saying
14212 @samp{Search failed}.
14214 What happens is this: the search is limited to the region, and fails
14215 as you expect because there are no word-constituent characters in the
14216 region. Since it fails, we receive an error message. But we do not
14217 want to receive an error message in this case; we want to receive the
14218 message that "The region does NOT have any words."
14220 The solution to this problem is to provide @code{re-search-forward}
14221 with a third argument of @code{t}, which causes the function to return
14222 @code{nil} rather than signal an error if the search fails.
14224 However, if you make this change and try it, you will see the message
14225 ``Counting words in region ... '' and @dots{} you will keep on seeing
14226 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14228 Here is what happens: the search is limited to the region, as before,
14229 and it fails because there are no word-constituent characters in the
14230 region, as expected. Consequently, the @code{re-search-forward}
14231 expression returns @code{nil}. It does nothing else. In particular,
14232 it does not move point, which it does as a side effect if it finds the
14233 search target. After the @code{re-search-forward} expression returns
14234 @code{nil}, the next expression in the @code{while} loop is evaluated.
14235 This expression increments the count. Then the loop repeats. The
14236 true-or-false-test tests true because the value of point is still less
14237 than the value of end, since the @code{re-search-forward} expression
14238 did not move point. @dots{} and the cycle repeats @dots{}
14240 The @code{@value{COUNT-WORDS}} definition requires yet another
14241 modification, to cause the true-or-false-test of the @code{while} loop
14242 to test false if the search fails. Put another way, there are two
14243 conditions that must be satisfied in the true-or-false-test before the
14244 word count variable is incremented: point must still be within the
14245 region and the search expression must have found a word to count.
14247 Since both the first condition and the second condition must be true
14248 together, the two expressions, the region test and the search
14249 expression, can be joined with an @code{and} special form and embedded in
14250 the @code{while} loop as the true-or-false-test, like this:
14253 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14256 @c colon in printed section title causes problem in Info cross reference
14257 @c also trouble with an overfull hbox
14260 (For information about @code{and}, see
14261 @ref{kill-new function, , The @code{kill-new} function}.)
14265 (@xref{kill-new function, , The @code{kill-new} function}, for
14266 information about @code{and}.)
14269 The @code{re-search-forward} expression returns @code{t} if the search
14270 succeeds and as a side effect moves point. Consequently, as words are
14271 found, point is moved through the region. When the search expression
14272 fails to find another word, or when point reaches the end of the
14273 region, the true-or-false-test tests false, the @code{while} loop
14274 exits, and the @code{@value{COUNT-WORDS}} function displays one or
14275 other of its messages.
14277 After incorporating these final changes, the @code{@value{COUNT-WORDS}}
14278 works without bugs (or at least, without bugs that I have found!).
14279 Here is what it looks like:
14283 ;;; @r{Final version:} @code{while}
14284 (defun @value{COUNT-WORDS} (beginning end)
14285 "Print number of words in the region."
14287 (message "Counting words in region ... ")
14291 ;;; @r{1. Set up appropriate conditions.}
14294 (goto-char beginning)
14298 ;;; @r{2. Run the} while @r{loop.}
14299 (while (and (< (point) end)
14300 (re-search-forward "\\w+\\W*" end t))
14301 (setq count (1+ count)))
14305 ;;; @r{3. Send a message to the user.}
14306 (cond ((zerop count)
14308 "The region does NOT have any words."))
14311 "The region has 1 word."))
14314 "The region has %d words." count))))))
14318 @node recursive-count-words, Counting Exercise, @value{COUNT-WORDS}, Counting Words
14319 @comment node-name, next, previous, up
14320 @section Count Words Recursively
14321 @cindex Count words recursively
14322 @cindex Recursively counting words
14323 @cindex Words, counted recursively
14325 You can write the function for counting words recursively as well as
14326 with a @code{while} loop. Let's see how this is done.
14328 First, we need to recognize that the @code{@value{COUNT-WORDS}}
14329 function has three jobs: it sets up the appropriate conditions for
14330 counting to occur; it counts the words in the region; and it sends a
14331 message to the user telling how many words there are.
14333 If we write a single recursive function to do everything, we will
14334 receive a message for every recursive call. If the region contains 13
14335 words, we will receive thirteen messages, one right after the other.
14336 We don't want this! Instead, we must write two functions to do the
14337 job, one of which (the recursive function) will be used inside of the
14338 other. One function will set up the conditions and display the
14339 message; the other will return the word count.
14341 Let us start with the function that causes the message to be displayed.
14342 We can continue to call this @code{@value{COUNT-WORDS}}.
14344 This is the function that the user will call. It will be interactive.
14345 Indeed, it will be similar to our previous versions of this
14346 function, except that it will call @code{recursive-count-words} to
14347 determine how many words are in the region.
14350 We can readily construct a template for this function, based on our
14355 ;; @r{Recursive version; uses regular expression search}
14356 (defun @value{COUNT-WORDS} (beginning end)
14357 "@var{documentation}@dots{}"
14358 (@var{interactive-expression}@dots{})
14362 ;;; @r{1. Set up appropriate conditions.}
14363 (@var{explanatory message})
14364 (@var{set-up functions}@dots{}
14368 ;;; @r{2. Count the words.}
14369 @var{recursive call}
14373 ;;; @r{3. Send a message to the user.}
14374 @var{message providing word count}))
14378 The definition looks straightforward, except that somehow the count
14379 returned by the recursive call must be passed to the message
14380 displaying the word count. A little thought suggests that this can be
14381 done by making use of a @code{let} expression: we can bind a variable
14382 in the varlist of a @code{let} expression to the number of words in
14383 the region, as returned by the recursive call; and then the
14384 @code{cond} expression, using binding, can display the value to the
14387 Often, one thinks of the binding within a @code{let} expression as
14388 somehow secondary to the `primary' work of a function. But in this
14389 case, what you might consider the `primary' job of the function,
14390 counting words, is done within the @code{let} expression.
14393 Using @code{let}, the function definition looks like this:
14397 (defun @value{COUNT-WORDS} (beginning end)
14398 "Print number of words in the region."
14403 ;;; @r{1. Set up appropriate conditions.}
14404 (message "Counting words in region ... ")
14406 (goto-char beginning)
14410 ;;; @r{2. Count the words.}
14411 (let ((count (recursive-count-words end)))
14415 ;;; @r{3. Send a message to the user.}
14416 (cond ((zerop count)
14418 "The region does NOT have any words."))
14421 "The region has 1 word."))
14424 "The region has %d words." count))))))
14428 Next, we need to write the recursive counting function.
14430 A recursive function has at least three parts: the `do-again-test', the
14431 `next-step-expression', and the recursive call.
14433 The do-again-test determines whether the function will or will not be
14434 called again. Since we are counting words in a region and can use a
14435 function that moves point forward for every word, the do-again-test
14436 can check whether point is still within the region. The do-again-test
14437 should find the value of point and determine whether point is before,
14438 at, or after the value of the end of the region. We can use the
14439 @code{point} function to locate point. Clearly, we must pass the
14440 value of the end of the region to the recursive counting function as an
14443 In addition, the do-again-test should also test whether the search finds a
14444 word. If it does not, the function should not call itself again.
14446 The next-step-expression changes a value so that when the recursive
14447 function is supposed to stop calling itself, it stops. More
14448 precisely, the next-step-expression changes a value so that at the
14449 right time, the do-again-test stops the recursive function from
14450 calling itself again. In this case, the next-step-expression can be
14451 the expression that moves point forward, word by word.
14453 The third part of a recursive function is the recursive call.
14455 Somewhere, also, we also need a part that does the `work' of the
14456 function, a part that does the counting. A vital part!
14459 But already, we have an outline of the recursive counting function:
14463 (defun recursive-count-words (region-end)
14464 "@var{documentation}@dots{}"
14465 @var{do-again-test}
14466 @var{next-step-expression}
14467 @var{recursive call})
14471 Now we need to fill in the slots. Let's start with the simplest cases
14472 first: if point is at or beyond the end of the region, there cannot
14473 be any words in the region, so the function should return zero.
14474 Likewise, if the search fails, there are no words to count, so the
14475 function should return zero.
14477 On the other hand, if point is within the region and the search
14478 succeeds, the function should call itself again.
14481 Thus, the do-again-test should look like this:
14485 (and (< (point) region-end)
14486 (re-search-forward "\\w+\\W*" region-end t))
14490 Note that the search expression is part of the do-again-test---the
14491 function returns @code{t} if its search succeeds and @code{nil} if it
14492 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14493 @code{@value{COUNT-WORDS}}}, for an explanation of how
14494 @code{re-search-forward} works.)
14496 The do-again-test is the true-or-false test of an @code{if} clause.
14497 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14498 clause should call the function again; but if it fails, the else-part
14499 should return zero since either point is outside the region or the
14500 search failed because there were no words to find.
14502 But before considering the recursive call, we need to consider the
14503 next-step-expression. What is it? Interestingly, it is the search
14504 part of the do-again-test.
14506 In addition to returning @code{t} or @code{nil} for the
14507 do-again-test, @code{re-search-forward} moves point forward as a side
14508 effect of a successful search. This is the action that changes the
14509 value of point so that the recursive function stops calling itself
14510 when point completes its movement through the region. Consequently,
14511 the @code{re-search-forward} expression is the next-step-expression.
14514 In outline, then, the body of the @code{recursive-count-words}
14515 function looks like this:
14519 (if @var{do-again-test-and-next-step-combined}
14521 @var{recursive-call-returning-count}
14527 How to incorporate the mechanism that counts?
14529 If you are not used to writing recursive functions, a question like
14530 this can be troublesome. But it can and should be approached
14533 We know that the counting mechanism should be associated in some way
14534 with the recursive call. Indeed, since the next-step-expression moves
14535 point forward by one word, and since a recursive call is made for
14536 each word, the counting mechanism must be an expression that adds one
14537 to the value returned by a call to @code{recursive-count-words}.
14540 Consider several cases:
14544 If there are two words in the region, the function should return
14545 a value resulting from adding one to the value returned when it counts
14546 the first word, plus the number returned when it counts the remaining
14547 words in the region, which in this case is one.
14550 If there is one word in the region, the function should return
14551 a value resulting from adding one to the value returned when it counts
14552 that word, plus the number returned when it counts the remaining
14553 words in the region, which in this case is zero.
14556 If there are no words in the region, the function should return zero.
14559 From the sketch we can see that the else-part of the @code{if} returns
14560 zero for the case of no words. This means that the then-part of the
14561 @code{if} must return a value resulting from adding one to the value
14562 returned from a count of the remaining words.
14565 The expression will look like this, where @code{1+} is a function that
14566 adds one to its argument.
14569 (1+ (recursive-count-words region-end))
14573 The whole @code{recursive-count-words} function will then look like
14578 (defun recursive-count-words (region-end)
14579 "@var{documentation}@dots{}"
14581 ;;; @r{1. do-again-test}
14582 (if (and (< (point) region-end)
14583 (re-search-forward "\\w+\\W*" region-end t))
14587 ;;; @r{2. then-part: the recursive call}
14588 (1+ (recursive-count-words region-end))
14590 ;;; @r{3. else-part}
14596 Let's examine how this works:
14598 If there are no words in the region, the else part of the @code{if}
14599 expression is evaluated and consequently the function returns zero.
14601 If there is one word in the region, the value of point is less than
14602 the value of @code{region-end} and the search succeeds. In this case,
14603 the true-or-false-test of the @code{if} expression tests true, and the
14604 then-part of the @code{if} expression is evaluated. The counting
14605 expression is evaluated. This expression returns a value (which will
14606 be the value returned by the whole function) that is the sum of one
14607 added to the value returned by a recursive call.
14609 Meanwhile, the next-step-expression has caused point to jump over the
14610 first (and in this case only) word in the region. This means that
14611 when @code{(recursive-count-words region-end)} is evaluated a second
14612 time, as a result of the recursive call, the value of point will be
14613 equal to or greater than the value of region end. So this time,
14614 @code{recursive-count-words} will return zero. The zero will be added
14615 to one, and the original evaluation of @code{recursive-count-words}
14616 will return one plus zero, which is one, which is the correct amount.
14618 Clearly, if there are two words in the region, the first call to
14619 @code{recursive-count-words} returns one added to the value returned
14620 by calling @code{recursive-count-words} on a region containing the
14621 remaining word---that is, it adds one to one, producing two, which is
14622 the correct amount.
14624 Similarly, if there are three 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 two words---and so on and so on.
14631 With full documentation the two functions look like this:
14635 The recursive function:
14637 @findex recursive-count-words
14640 (defun recursive-count-words (region-end)
14641 "Number of words between point and REGION-END."
14645 ;;; @r{1. do-again-test}
14646 (if (and (< (point) region-end)
14647 (re-search-forward "\\w+\\W*" region-end t))
14651 ;;; @r{2. then-part: the recursive call}
14652 (1+ (recursive-count-words region-end))
14654 ;;; @r{3. else-part}
14665 ;;; @r{Recursive version}
14666 (defun @value{COUNT-WORDS} (beginning end)
14667 "Print number of words in the region.
14671 Words are defined as at least one word-constituent
14672 character followed by at least one character that is
14673 not a word-constituent. The buffer's syntax table
14674 determines which characters these are."
14678 (message "Counting words in region ... ")
14680 (goto-char beginning)
14681 (let ((count (recursive-count-words end)))
14684 (cond ((zerop count)
14686 "The region does NOT have any words."))
14690 (message "The region has 1 word."))
14693 "The region has %d words." count))))))
14697 @node Counting Exercise, , recursive-count-words, Counting Words
14698 @section Exercise: Counting Punctuation
14700 Using a @code{while} loop, write a function to count the number of
14701 punctuation marks in a region---period, comma, semicolon, colon,
14702 exclamation mark, and question mark. Do the same using recursion.
14704 @node Words in a defun, Readying a Graph, Counting Words, Top
14705 @chapter Counting Words in a @code{defun}
14706 @cindex Counting words in a @code{defun}
14707 @cindex Word counting in a @code{defun}
14709 Our next project is to count the number of words in a function
14710 definition. Clearly, this can be done using some variant of
14711 @code{@value{COUNT-WORDS}}. @xref{Counting Words, , Counting Words:
14712 Repetition and Regexps}. If we are just going to count the words in
14713 one definition, it is easy enough to mark the definition with the
14714 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14715 @code{@value{COUNT-WORDS}}.
14717 However, I am more ambitious: I want to count the words and symbols in
14718 every definition in the Emacs sources and then print a graph that
14719 shows how many functions there are of each length: how many contain 40
14720 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14721 and so on. I have often been curious how long a typical function is,
14722 and this will tell.
14725 * Divide and Conquer::
14726 * Words and Symbols:: What to count?
14727 * Syntax:: What constitutes a word or symbol?
14728 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
14729 * Several defuns:: Counting several defuns in a file.
14730 * Find a File:: Do you want to look at a file?
14731 * lengths-list-file:: A list of the lengths of many definitions.
14732 * Several files:: Counting in definitions in different files.
14733 * Several files recursively:: Recursively counting in different files.
14734 * Prepare the data:: Prepare the data for display in a graph.
14737 @node Divide and Conquer, Words and Symbols, Words in a defun, Words in a defun
14739 @unnumberedsec Divide and Conquer
14742 Described in one phrase, the histogram project is daunting; but
14743 divided into numerous small steps, each of which we can take one at a
14744 time, the project becomes less fearsome. Let us consider what the
14749 First, write a function to count the words in one definition. This
14750 includes the problem of handling symbols as well as words.
14753 Second, write a function to list the numbers of words in each function
14754 in a file. This function can use the @code{count-words-in-defun}
14758 Third, write a function to list the numbers of words in each function
14759 in each of several files. This entails automatically finding the
14760 various files, switching to them, and counting the words in the
14761 definitions within them.
14764 Fourth, write a function to convert the list of numbers that we
14765 created in step three to a form that will be suitable for printing as
14769 Fifth, write a function to print the results as a graph.
14772 This is quite a project! But if we take each step slowly, it will not
14775 @node Words and Symbols, Syntax, Divide and Conquer, Words in a defun
14776 @section What to Count?
14777 @cindex Words and symbols in defun
14779 When we first start thinking about how to count the words in a
14780 function definition, the first question is (or ought to be) what are
14781 we going to count? When we speak of `words' with respect to a Lisp
14782 function definition, we are actually speaking, in large part, of
14783 `symbols'. For example, the following @code{multiply-by-seven}
14784 function contains the five symbols @code{defun},
14785 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14786 addition, in the documentation string, it contains the four words
14787 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14788 symbol @samp{number} is repeated, so the definition contains a total
14789 of ten words and symbols.
14793 (defun multiply-by-seven (number)
14794 "Multiply NUMBER by seven."
14800 However, if we mark the @code{multiply-by-seven} definition with
14801 @kbd{C-M-h} (@code{mark-defun}), and then call
14802 @code{@value{COUNT-WORDS}} on it, we will find that
14803 @code{@value{COUNT-WORDS}} claims the definition has eleven words, not
14804 ten! Something is wrong!
14806 The problem is twofold: @code{@value{COUNT-WORDS}} does not count the
14807 @samp{*} as a word, and it counts the single symbol,
14808 @code{multiply-by-seven}, as containing three words. The hyphens are
14809 treated as if they were interword spaces rather than intraword
14810 connectors: @samp{multiply-by-seven} is counted as if it were written
14811 @samp{multiply by seven}.
14813 The cause of this confusion is the regular expression search within
14814 the @code{@value{COUNT-WORDS}} definition that moves point forward word
14815 by word. In the canonical version of @code{@value{COUNT-WORDS}}, the
14823 This regular expression is a pattern defining one or more word
14824 constituent characters possibly followed by one or more characters
14825 that are not word constituents. What is meant by `word constituent
14826 characters' brings us to the issue of syntax, which is worth a section
14829 @node Syntax, count-words-in-defun, Words and Symbols, Words in a defun
14830 @section What Constitutes a Word or Symbol?
14831 @cindex Syntax categories and tables
14833 Emacs treats different characters as belonging to different
14834 @dfn{syntax categories}. For example, the regular expression,
14835 @samp{\\w+}, is a pattern specifying one or more @emph{word
14836 constituent} characters. Word constituent characters are members of
14837 one syntax category. Other syntax categories include the class of
14838 punctuation characters, such as the period and the comma, and the
14839 class of whitespace characters, such as the blank space and the tab
14840 character. (For more information, see @ref{Syntax, Syntax, The Syntax
14841 Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, , Syntax
14842 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14844 Syntax tables specify which characters belong to which categories.
14845 Usually, a hyphen is not specified as a `word constituent character'.
14846 Instead, it is specified as being in the `class of characters that are
14847 part of symbol names but not words.' This means that the
14848 @code{@value{COUNT-WORDS}} function treats it in the same way it treats
14849 an interword white space, which is why @code{@value{COUNT-WORDS}}
14850 counts @samp{multiply-by-seven} as three words.
14852 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14853 one symbol: modify the syntax table or modify the regular expression.
14855 We could redefine a hyphen as a word constituent character by
14856 modifying the syntax table that Emacs keeps for each mode. This
14857 action would serve our purpose, except that a hyphen is merely the
14858 most common character within symbols that is not typically a word
14859 constituent character; there are others, too.
14861 Alternatively, we can redefine the regular expression used in the
14862 @code{@value{COUNT-WORDS}} definition so as to include symbols. This
14863 procedure has the merit of clarity, but the task is a little tricky.
14866 The first part is simple enough: the pattern must match ``at least one
14867 character that is a word or symbol constituent''. Thus:
14870 "\\(\\w\\|\\s_\\)+"
14874 The @samp{\\(} is the first part of the grouping construct that
14875 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14876 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14877 character and the @samp{\\s_} matches any character that is part of a
14878 symbol name but not a word-constituent character. The @samp{+}
14879 following the group indicates that the word or symbol constituent
14880 characters must be matched at least once.
14882 However, the second part of the regexp is more difficult to design.
14883 What we want is to follow the first part with ``optionally one or more
14884 characters that are not constituents of a word or symbol''. At first,
14885 I thought I could define this with the following:
14888 "\\(\\W\\|\\S_\\)*"
14892 The upper case @samp{W} and @samp{S} match characters that are
14893 @emph{not} word or symbol constituents. Unfortunately, this
14894 expression matches any character that is either not a word constituent
14895 or not a symbol constituent. This matches any character!
14897 I then noticed that every word or symbol in my test region was
14898 followed by white space (blank space, tab, or newline). So I tried
14899 placing a pattern to match one or more blank spaces after the pattern
14900 for one or more word or symbol constituents. This failed, too. Words
14901 and symbols are often separated by whitespace, but in actual code
14902 parentheses may follow symbols and punctuation may follow words. So
14903 finally, I designed a pattern in which the word or symbol constituents
14904 are followed optionally by characters that are not white space and
14905 then followed optionally by white space.
14908 Here is the full regular expression:
14911 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14914 @node count-words-in-defun, Several defuns, Syntax, Words in a defun
14915 @section The @code{count-words-in-defun} Function
14916 @cindex Counting words in a @code{defun}
14918 We have seen that there are several ways to write a
14919 @code{count-words-region} function. To write a
14920 @code{count-words-in-defun}, we need merely adapt one of these
14923 The version that uses a @code{while} loop is easy to understand, so I
14924 am going to adapt that. Because @code{count-words-in-defun} will be
14925 part of a more complex program, it need not be interactive and it need
14926 not display a message but just return the count. These considerations
14927 simplify the definition a little.
14929 On the other hand, @code{count-words-in-defun} will be used within a
14930 buffer that contains function definitions. Consequently, it is
14931 reasonable to ask that the function determine whether it is called
14932 when point is within a function definition, and if it is, to return
14933 the count for that definition. This adds complexity to the
14934 definition, but saves us from needing to pass arguments to the
14938 These considerations lead us to prepare the following template:
14942 (defun count-words-in-defun ()
14943 "@var{documentation}@dots{}"
14944 (@var{set up}@dots{}
14945 (@var{while loop}@dots{})
14946 @var{return count})
14951 As usual, our job is to fill in the slots.
14955 We are presuming that this function will be called within a buffer
14956 containing function definitions. Point will either be within a
14957 function definition or not. For @code{count-words-in-defun} to work,
14958 point must move to the beginning of the definition, a counter must
14959 start at zero, and the counting loop must stop when point reaches the
14960 end of the definition.
14962 The @code{beginning-of-defun} function searches backwards for an
14963 opening delimiter such as a @samp{(} at the beginning of a line, and
14964 moves point to that position, or else to the limit of the search. In
14965 practice, this means that @code{beginning-of-defun} moves point to the
14966 beginning of an enclosing or preceding function definition, or else to
14967 the beginning of the buffer. We can use @code{beginning-of-defun} to
14968 place point where we wish to start.
14970 The @code{while} loop requires a counter to keep track of the words or
14971 symbols being counted. A @code{let} expression can be used to create
14972 a local variable for this purpose, and bind it to an initial value of zero.
14974 The @code{end-of-defun} function works like @code{beginning-of-defun}
14975 except that it moves point to the end of the definition.
14976 @code{end-of-defun} can be used as part of an expression that
14977 determines the position of the end of the definition.
14979 The set up for @code{count-words-in-defun} takes shape rapidly: first
14980 we move point to the beginning of the definition, then we create a
14981 local variable to hold the count, and finally, we record the position
14982 of the end of the definition so the @code{while} loop will know when to stop
14986 The code looks like this:
14990 (beginning-of-defun)
14992 (end (save-excursion (end-of-defun) (point))))
14997 The code is simple. The only slight complication is likely to concern
14998 @code{end}: it is bound to the position of the end of the definition
14999 by a @code{save-excursion} expression that returns the value of point
15000 after @code{end-of-defun} temporarily moves it to the end of the
15003 The second part of the @code{count-words-in-defun}, after the set up,
15004 is the @code{while} loop.
15006 The loop must contain an expression that jumps point forward word by
15007 word and symbol by symbol, and another expression that counts the
15008 jumps. The true-or-false-test for the @code{while} loop should test
15009 true so long as point should jump forward, and false when point is at
15010 the end of the definition. We have already redefined the regular
15011 expression for this (@pxref{Syntax}), so the loop is straightforward:
15015 (while (and (< (point) end)
15017 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t)
15018 (setq count (1+ count)))
15022 The third part of the function definition returns the count of words
15023 and symbols. This part is the last expression within the body of the
15024 @code{let} expression, and can be, very simply, the local variable
15025 @code{count}, which when evaluated returns the count.
15028 Put together, the @code{count-words-in-defun} definition looks like this:
15030 @findex count-words-in-defun
15033 (defun count-words-in-defun ()
15034 "Return the number of words and symbols in a defun."
15035 (beginning-of-defun)
15037 (end (save-excursion (end-of-defun) (point))))
15041 (and (< (point) end)
15043 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
15045 (setq count (1+ count)))
15050 How to test this? The function is not interactive, but it is easy to
15051 put a wrapper around the function to make it interactive; we can use
15052 almost the same code as for the recursive version of
15053 @code{@value{COUNT-WORDS}}:
15057 ;;; @r{Interactive version.}
15058 (defun count-words-defun ()
15059 "Number of words and symbols in a function definition."
15062 "Counting words and symbols in function definition ... ")
15065 (let ((count (count-words-in-defun)))
15069 "The definition does NOT have any words or symbols."))
15074 "The definition has 1 word or symbol."))
15077 "The definition has %d words or symbols." count)))))
15083 Let's re-use @kbd{C-c =} as a convenient keybinding:
15086 (global-set-key "\C-c=" 'count-words-defun)
15089 Now we can try out @code{count-words-defun}: install both
15090 @code{count-words-in-defun} and @code{count-words-defun}, and set the
15091 keybinding, and then place the cursor within the following definition:
15095 (defun multiply-by-seven (number)
15096 "Multiply NUMBER by seven."
15103 Success! The definition has 10 words and symbols.
15105 The next problem is to count the numbers of words and symbols in
15106 several definitions within a single file.
15108 @node Several defuns, Find a File, count-words-in-defun, Words in a defun
15109 @section Count Several @code{defuns} Within a File
15111 A file such as @file{simple.el} may have a hundred or more function
15112 definitions within it. Our long term goal is to collect statistics on
15113 many files, but as a first step, our immediate goal is to collect
15114 statistics on one file.
15116 The information will be a series of numbers, each number being the
15117 length of a function definition. We can store the numbers in a list.
15119 We know that we will want to incorporate the information regarding one
15120 file with information about many other files; this means that the
15121 function for counting definition lengths within one file need only
15122 return the list of lengths. It need not and should not display any
15125 The word count commands contain one expression to jump point forward
15126 word by word and another expression to count the jumps. The function
15127 to return the lengths of definitions can be designed to work the same
15128 way, with one expression to jump point forward definition by
15129 definition and another expression to construct the lengths' list.
15131 This statement of the problem makes it elementary to write the
15132 function definition. Clearly, we will start the count at the
15133 beginning of the file, so the first command will be @code{(goto-char
15134 (point-min))}. Next, we start the @code{while} loop; and the
15135 true-or-false test of the loop can be a regular expression search for
15136 the next function definition---so long as the search succeeds, point
15137 is moved forward and then the body of the loop is evaluated. The body
15138 needs an expression that constructs the lengths' list. @code{cons},
15139 the list construction command, can be used to create the list. That
15140 is almost all there is to it.
15143 Here is what this fragment of code looks like:
15147 (goto-char (point-min))
15148 (while (re-search-forward "^(defun" nil t)
15150 (cons (count-words-in-defun) lengths-list)))
15154 What we have left out is the mechanism for finding the file that
15155 contains the function definitions.
15157 In previous examples, we either used this, the Info file, or we
15158 switched back and forth to some other buffer, such as the
15159 @file{*scratch*} buffer.
15161 Finding a file is a new process that we have not yet discussed.
15163 @node Find a File, lengths-list-file, Several defuns, Words in a defun
15164 @comment node-name, next, previous, up
15165 @section Find a File
15166 @cindex Find a File
15168 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15169 command. This command is almost, but not quite right for the lengths
15173 Let's look at the source for @code{find-file}:
15177 (defun find-file (filename)
15178 "Edit file FILENAME.
15179 Switch to a buffer visiting file FILENAME,
15180 creating one if none already exists."
15181 (interactive "FFind file: ")
15182 (switch-to-buffer (find-file-noselect filename)))
15187 (The most recent version of the @code{find-file} function definition
15188 permits you to specify optional wildcards to visit multiple files; that
15189 makes the definition more complex and we will not discuss it here,
15190 since it is not relevant. You can see its source using either
15191 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15195 (defun find-file (filename &optional wildcards)
15196 "Edit file FILENAME.
15197 Switch to a buffer visiting file FILENAME,
15198 creating one if none already exists.
15199 Interactively, the default if you just type RET is the current directory,
15200 but the visited file name is available through the minibuffer history:
15201 type M-n to pull it into the minibuffer.
15203 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15204 expand wildcards (if any) and visit multiple files. You can
15205 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15207 To visit a file without any kind of conversion and without
15208 automatically choosing a major mode, use \\[find-file-literally]."
15209 (interactive (find-file-read-args "Find file: " nil))
15210 (let ((value (find-file-noselect filename nil nil wildcards)))
15212 (mapcar 'switch-to-buffer (nreverse value))
15213 (switch-to-buffer value))))
15216 The definition I am showing possesses short but complete documentation
15217 and an interactive specification that prompts you for a file name when
15218 you use the command interactively. The body of the definition
15219 contains two functions, @code{find-file-noselect} and
15220 @code{switch-to-buffer}.
15222 According to its documentation as shown by @kbd{C-h f} (the
15223 @code{describe-function} command), the @code{find-file-noselect}
15224 function reads the named file into a buffer and returns the buffer.
15225 (Its most recent version includes an optional wildcards argument,
15226 too, as well as another to read a file literally and an other you
15227 suppress warning messages. These optional arguments are irrelevant.)
15229 However, the @code{find-file-noselect} function does not select the
15230 buffer in which it puts the file. Emacs does not switch its attention
15231 (or yours if you are using @code{find-file-noselect}) to the selected
15232 buffer. That is what @code{switch-to-buffer} does: it switches the
15233 buffer to which Emacs attention is directed; and it switches the
15234 buffer displayed in the window to the new buffer. We have discussed
15235 buffer switching elsewhere. (@xref{Switching Buffers}.)
15237 In this histogram project, we do not need to display each file on the
15238 screen as the program determines the length of each definition within
15239 it. Instead of employing @code{switch-to-buffer}, we can work with
15240 @code{set-buffer}, which redirects the attention of the computer
15241 program to a different buffer but does not redisplay it on the screen.
15242 So instead of calling on @code{find-file} to do the job, we must write
15243 our own expression.
15245 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15247 @node lengths-list-file, Several files, Find a File, Words in a defun
15248 @section @code{lengths-list-file} in Detail
15250 The core of the @code{lengths-list-file} function is a @code{while}
15251 loop containing a function to move point forward `defun by defun' and
15252 a function to count the number of words and symbols in each defun.
15253 This core must be surrounded by functions that do various other tasks,
15254 including finding the file, and ensuring that point starts out at the
15255 beginning of the file. The function definition looks like this:
15256 @findex lengths-list-file
15260 (defun lengths-list-file (filename)
15261 "Return list of definitions' lengths within FILE.
15262 The returned list is a list of numbers.
15263 Each number is the number of words or
15264 symbols in one function definition."
15267 (message "Working on `%s' ... " filename)
15269 (let ((buffer (find-file-noselect filename))
15271 (set-buffer buffer)
15272 (setq buffer-read-only t)
15274 (goto-char (point-min))
15275 (while (re-search-forward "^(defun" nil t)
15277 (cons (count-words-in-defun) lengths-list)))
15278 (kill-buffer buffer)
15284 The function is passed one argument, the name of the file on which it
15285 will work. It has four lines of documentation, but no interactive
15286 specification. Since people worry that a computer is broken if they
15287 don't see anything going on, the first line of the body is a
15290 The next line contains a @code{save-excursion} that returns Emacs'
15291 attention to the current buffer when the function completes. This is
15292 useful in case you embed this function in another function that
15293 presumes point is restored to the original buffer.
15295 In the varlist of the @code{let} expression, Emacs finds the file and
15296 binds the local variable @code{buffer} to the buffer containing the
15297 file. At the same time, Emacs creates @code{lengths-list} as a local
15300 Next, Emacs switches its attention to the buffer.
15302 In the following line, Emacs makes the buffer read-only. Ideally,
15303 this line is not necessary. None of the functions for counting words
15304 and symbols in a function definition should change the buffer.
15305 Besides, the buffer is not going to be saved, even if it were changed.
15306 This line is entirely the consequence of great, perhaps excessive,
15307 caution. The reason for the caution is that this function and those
15308 it calls work on the sources for Emacs and it is inconvenient if they
15309 are inadvertently modified. It goes without saying that I did not
15310 realize a need for this line until an experiment went awry and started
15311 to modify my Emacs source files @dots{}
15313 Next comes a call to widen the buffer if it is narrowed. This
15314 function is usually not needed---Emacs creates a fresh buffer if none
15315 already exists; but if a buffer visiting the file already exists Emacs
15316 returns that one. In this case, the buffer may be narrowed and must
15317 be widened. If we wanted to be fully `user-friendly', we would
15318 arrange to save the restriction and the location of point, but we
15321 The @code{(goto-char (point-min))} expression moves point to the
15322 beginning of the buffer.
15324 Then comes a @code{while} loop in which the `work' of the function is
15325 carried out. In the loop, Emacs determines the length of each
15326 definition and constructs a lengths' list containing the information.
15328 Emacs kills the buffer after working through it. This is to save
15329 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15330 source files of interest; GNU Emacs 22 contains over a thousand source
15331 files. Another function will apply @code{lengths-list-file} to each
15334 Finally, the last expression within the @code{let} expression is the
15335 @code{lengths-list} variable; its value is returned as the value of
15336 the whole function.
15338 You can try this function by installing it in the usual fashion. Then
15339 place your cursor after the following expression and type @kbd{C-x
15340 C-e} (@code{eval-last-sexp}).
15342 @c !!! 22.1.1 lisp sources location here
15345 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15349 (You may need to change the pathname of the file; the one here is for
15350 GNU Emacs version 22.1.1. To change the expression, copy it to
15351 the @file{*scratch*} buffer and edit it.
15355 (Also, to see the full length of the list, rather than a truncated
15356 version, you may have to evaluate the following:
15359 (custom-set-variables '(eval-expression-print-length nil))
15363 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15364 Then evaluate the @code{lengths-list-file} expression.)
15367 The lengths' list for @file{debug.el} takes less than a second to
15368 produce and looks like this in GNU Emacs 22:
15371 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15375 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15376 took seven seconds to produce and looked like this:
15379 (75 41 80 62 20 45 44 68 45 12 34 235)
15382 (The newer version of @file{debug.el} contains more defuns than the
15383 earlier one; and my new machine is much faster than the old one.)
15385 Note that the length of the last definition in the file is first in
15388 @node Several files, Several files recursively, lengths-list-file, Words in a defun
15389 @section Count Words in @code{defuns} in Different Files
15391 In the previous section, we created a function that returns a list of
15392 the lengths of each definition in a file. Now, we want to define a
15393 function to return a master list of the lengths of the definitions in
15396 Working on each of a list of files is a repetitious act, so we can use
15397 either a @code{while} loop or recursion.
15400 * lengths-list-many-files:: Return a list of the lengths of defuns.
15401 * append:: Attach one list to another.
15404 @node lengths-list-many-files, append, Several files, Several files
15406 @unnumberedsubsec Determine the lengths of @code{defuns}
15409 The design using a @code{while} loop is routine. The argument passed
15410 the function is a list of files. As we saw earlier (@pxref{Loop
15411 Example}), you can write a @code{while} loop so that the body of the
15412 loop is evaluated if such a list contains elements, but to exit the
15413 loop if the list is empty. For this design to work, the body of the
15414 loop must contain an expression that shortens the list each time the
15415 body is evaluated, so that eventually the list is empty. The usual
15416 technique is to set the value of the list to the value of the @sc{cdr}
15417 of the list each time the body is evaluated.
15420 The template looks like this:
15424 (while @var{test-whether-list-is-empty}
15426 @var{set-list-to-cdr-of-list})
15430 Also, we remember that a @code{while} loop returns @code{nil} (the
15431 result of evaluating the true-or-false-test), not the result of any
15432 evaluation within its body. (The evaluations within the body of the
15433 loop are done for their side effects.) However, the expression that
15434 sets the lengths' list is part of the body---and that is the value
15435 that we want returned by the function as a whole. To do this, we
15436 enclose the @code{while} loop within a @code{let} expression, and
15437 arrange that the last element of the @code{let} expression contains
15438 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15439 Example with an Incrementing Counter}.)
15441 @findex lengths-list-many-files
15443 These considerations lead us directly to the function itself:
15447 ;;; @r{Use @code{while} loop.}
15448 (defun lengths-list-many-files (list-of-files)
15449 "Return list of lengths of defuns in LIST-OF-FILES."
15452 (let (lengths-list)
15454 ;;; @r{true-or-false-test}
15455 (while list-of-files
15460 ;;; @r{Generate a lengths' list.}
15462 (expand-file-name (car list-of-files)))))
15466 ;;; @r{Make files' list shorter.}
15467 (setq list-of-files (cdr list-of-files)))
15469 ;;; @r{Return final value of lengths' list.}
15474 @code{expand-file-name} is a built-in function that converts a file
15475 name to the absolute, long, path name form. The function employs the
15476 name of the directory in which the function is called.
15478 @c !!! 22.1.1 lisp sources location here
15480 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15481 Emacs is visiting the
15482 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15492 @c !!! 22.1.1 lisp sources location here
15494 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15497 The only other new element of this function definition is the as yet
15498 unstudied function @code{append}, which merits a short section for
15501 @node append, , lengths-list-many-files, Several files
15502 @subsection The @code{append} Function
15505 The @code{append} function attaches one list to another. Thus,
15508 (append '(1 2 3 4) '(5 6 7 8))
15519 This is exactly how we want to attach two lengths' lists produced by
15520 @code{lengths-list-file} to each other. The results contrast with
15524 (cons '(1 2 3 4) '(5 6 7 8))
15529 which constructs a new list in which the first argument to @code{cons}
15530 becomes the first element of the new list:
15533 ((1 2 3 4) 5 6 7 8)
15536 @node Several files recursively, Prepare the data, Several files, Words in a defun
15537 @section Recursively Count Words in Different Files
15539 Besides a @code{while} loop, you can work on each of a list of files
15540 with recursion. A recursive version of @code{lengths-list-many-files}
15541 is short and simple.
15543 The recursive function has the usual parts: the `do-again-test', the
15544 `next-step-expression', and the recursive call. The `do-again-test'
15545 determines whether the function should call itself again, which it
15546 will do if the @code{list-of-files} contains any remaining elements;
15547 the `next-step-expression' resets the @code{list-of-files} to the
15548 @sc{cdr} of itself, so eventually the list will be empty; and the
15549 recursive call calls itself on the shorter list. The complete
15550 function is shorter than this description!
15551 @findex recursive-lengths-list-many-files
15555 (defun recursive-lengths-list-many-files (list-of-files)
15556 "Return list of lengths of each defun in LIST-OF-FILES."
15557 (if list-of-files ; @r{do-again-test}
15560 (expand-file-name (car list-of-files)))
15561 (recursive-lengths-list-many-files
15562 (cdr list-of-files)))))
15567 In a sentence, the function returns the lengths' list for the first of
15568 the @code{list-of-files} appended to the result of calling itself on
15569 the rest of the @code{list-of-files}.
15571 Here is a test of @code{recursive-lengths-list-many-files}, along with
15572 the results of running @code{lengths-list-file} on each of the files
15575 Install @code{recursive-lengths-list-many-files} and
15576 @code{lengths-list-file}, if necessary, and then evaluate the
15577 following expressions. You may need to change the files' pathnames;
15578 those here work when this Info file and the Emacs sources are located
15579 in their customary places. To change the expressions, copy them to
15580 the @file{*scratch*} buffer, edit them, and then evaluate them.
15582 The results are shown after the @samp{@result{}}. (These results are
15583 for files from Emacs version 22.1.1; files from other versions of
15584 Emacs may produce different results.)
15586 @c !!! 22.1.1 lisp sources location here
15589 (cd "/usr/local/share/emacs/22.1.1/")
15591 (lengths-list-file "./lisp/macros.el")
15592 @result{} (283 263 480 90)
15596 (lengths-list-file "./lisp/mail/mailalias.el")
15597 @result{} (38 32 29 95 178 180 321 218 324)
15601 (lengths-list-file "./lisp/makesum.el")
15606 (recursive-lengths-list-many-files
15607 '("./lisp/macros.el"
15608 "./lisp/mail/mailalias.el"
15609 "./lisp/makesum.el"))
15610 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15614 The @code{recursive-lengths-list-many-files} function produces the
15617 The next step is to prepare the data in the list for display in a graph.
15619 @node Prepare the data, , Several files recursively, Words in a defun
15620 @section Prepare the Data for Display in a Graph
15622 The @code{recursive-lengths-list-many-files} function returns a list
15623 of numbers. Each number records the length of a function definition.
15624 What we need to do now is transform this data into a list of numbers
15625 suitable for generating a graph. The new list will tell how many
15626 functions definitions contain less than 10 words and
15627 symbols, how many contain between 10 and 19 words and symbols, how
15628 many contain between 20 and 29 words and symbols, and so on.
15630 In brief, we need to go through the lengths' list produced by the
15631 @code{recursive-lengths-list-many-files} function and count the number
15632 of defuns within each range of lengths, and produce a list of those
15636 * Data for Display in Detail::
15637 * Sorting:: Sorting lists.
15638 * Files List:: Making a list of files.
15639 * Counting function definitions::
15642 @node Data for Display in Detail, Sorting, Prepare the data, Prepare the data
15644 @unnumberedsubsec The Data for Display in Detail
15647 Based on what we have done before, we can readily foresee that it
15648 should not be too hard to write a function that `@sc{cdr}s' down the
15649 lengths' list, looks at each element, determines which length range it
15650 is in, and increments a counter for that range.
15652 However, before beginning to write such a function, we should consider
15653 the advantages of sorting the lengths' list first, so the numbers are
15654 ordered from smallest to largest. First, sorting will make it easier
15655 to count the numbers in each range, since two adjacent numbers will
15656 either be in the same length range or in adjacent ranges. Second, by
15657 inspecting a sorted list, we can discover the highest and lowest
15658 number, and thereby determine the largest and smallest length range
15661 @node Sorting, Files List, Data for Display in Detail, Prepare the data
15662 @subsection Sorting Lists
15665 Emacs contains a function to sort lists, called (as you might guess)
15666 @code{sort}. The @code{sort} function takes two arguments, the list
15667 to be sorted, and a predicate that determines whether the first of
15668 two list elements is ``less'' than the second.
15670 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15671 Type Object as an Argument}), a predicate is a function that
15672 determines whether some property is true or false. The @code{sort}
15673 function will reorder a list according to whatever property the
15674 predicate uses; this means that @code{sort} can be used to sort
15675 non-numeric lists by non-numeric criteria---it can, for example,
15676 alphabetize a list.
15679 The @code{<} function is used when sorting a numeric list. For example,
15682 (sort '(4 8 21 17 33 7 21 7) '<)
15690 (4 7 7 8 17 21 21 33)
15694 (Note that in this example, both the arguments are quoted so that the
15695 symbols are not evaluated before being passed to @code{sort} as
15698 Sorting the list returned by the
15699 @code{recursive-lengths-list-many-files} function is straightforward;
15700 it uses the @code{<} function:
15704 In GNU Emacs 22, eval
15706 (cd "/usr/local/share/emacs/22.0.50/")
15708 (recursive-lengths-list-many-files
15709 '("./lisp/macros.el"
15710 "./lisp/mail/mailalias.el"
15711 "./lisp/makesum.el"))
15719 (recursive-lengths-list-many-files
15720 '("./lisp/macros.el"
15721 "./lisp/mailalias.el"
15722 "./lisp/makesum.el"))
15732 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15736 (Note that in this example, the first argument to @code{sort} is not
15737 quoted, since the expression must be evaluated so as to produce the
15738 list that is passed to @code{sort}.)
15740 @node Files List, Counting function definitions, Sorting, Prepare the data
15741 @subsection Making a List of Files
15743 The @code{recursive-lengths-list-many-files} function requires a list
15744 of files as its argument. For our test examples, we constructed such
15745 a list by hand; but the Emacs Lisp source directory is too large for
15746 us to do for that. Instead, we will write a function to do the job
15747 for us. In this function, we will use both a @code{while} loop and a
15750 @findex directory-files
15751 We did not have to write a function like this for older versions of
15752 GNU Emacs, since they placed all the @samp{.el} files in one
15753 directory. Instead, we were able to use the @code{directory-files}
15754 function, which lists the names of files that match a specified
15755 pattern within a single directory.
15757 However, recent versions of Emacs place Emacs Lisp files in
15758 sub-directories of the top level @file{lisp} directory. This
15759 re-arrangement eases navigation. For example, all the mail related
15760 files are in a @file{lisp} sub-directory called @file{mail}. But at
15761 the same time, this arrangement forces us to create a file listing
15762 function that descends into the sub-directories.
15764 @findex files-in-below-directory
15765 We can create this function, called @code{files-in-below-directory},
15766 using familiar functions such as @code{car}, @code{nthcdr}, and
15767 @code{substring} in conjunction with an existing function called
15768 @code{directory-files-and-attributes}. This latter function not only
15769 lists all the filenames in a directory, including the names
15770 of sub-directories, but also their attributes.
15772 To restate our goal: to create a function that will enable us
15773 to feed filenames to @code{recursive-lengths-list-many-files}
15774 as a list that looks like this (but with more elements):
15778 ("./lisp/macros.el"
15779 "./lisp/mail/rmail.el"
15780 "./lisp/makesum.el")
15784 The @code{directory-files-and-attributes} function returns a list of
15785 lists. Each of the lists within the main list consists of 13
15786 elements. The first element is a string that contains the name of the
15787 file -- which, in GNU/Linux, may be a `directory file', that is to
15788 say, a file with the special attributes of a directory. The second
15789 element of the list is @code{t} for a directory, a string
15790 for symbolic link (the string is the name linked to), or @code{nil}.
15792 For example, the first @samp{.el} file in the @file{lisp/} directory
15793 is @file{abbrev.el}. Its name is
15794 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15795 directory or a symbolic link.
15798 This is how @code{directory-files-and-attributes} lists that file and
15824 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15825 directory. The beginning of its listing looks like this:
15836 (To learn about the different attributes, look at the documentation of
15837 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15838 function does not list the filename, so its first element is
15839 @code{directory-files-and-attributes}'s second element.)
15841 We will want our new function, @code{files-in-below-directory}, to
15842 list the @samp{.el} files in the directory it is told to check, and in
15843 any directories below that directory.
15845 This gives us a hint on how to construct
15846 @code{files-in-below-directory}: within a directory, the function
15847 should add @samp{.el} filenames to a list; and if, within a directory,
15848 the function comes upon a sub-directory, it should go into that
15849 sub-directory and repeat its actions.
15851 However, we should note that every directory contains a name that
15852 refers to itself, called @file{.}, (``dot'') and a name that refers to
15853 its parent directory, called @file{..} (``double dot''). (In
15854 @file{/}, the root directory, @file{..} refers to itself, since
15855 @file{/} has no parent.) Clearly, we do not want our
15856 @code{files-in-below-directory} function to enter those directories,
15857 since they always lead us, directly or indirectly, to the current
15860 Consequently, our @code{files-in-below-directory} function must do
15865 Check to see whether it is looking at a filename that ends in
15866 @samp{.el}; and if so, add its name to a list.
15869 Check to see whether it is looking at a filename that is the name of a
15870 directory; and if so,
15874 Check to see whether it is looking at @file{.} or @file{..}; and if
15878 Or else, go into that directory and repeat the process.
15882 Let's write a function definition to do these tasks. We will use a
15883 @code{while} loop to move from one filename to another within a
15884 directory, checking what needs to be done; and we will use a recursive
15885 call to repeat the actions on each sub-directory. The recursive
15886 pattern is `accumulate'
15887 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15888 using @code{append} as the combiner.
15891 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15892 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15894 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15895 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15898 @c /usr/local/share/emacs/22.1.1/lisp/
15901 Here is the function:
15905 (defun files-in-below-directory (directory)
15906 "List the .el files in DIRECTORY and in its sub-directories."
15907 ;; Although the function will be used non-interactively,
15908 ;; it will be easier to test if we make it interactive.
15909 ;; The directory will have a name such as
15910 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15911 (interactive "DDirectory name: ")
15914 (let (el-files-list
15915 (current-directory-list
15916 (directory-files-and-attributes directory t)))
15917 ;; while we are in the current directory
15918 (while current-directory-list
15922 ;; check to see whether filename ends in `.el'
15923 ;; and if so, append its name to a list.
15924 ((equal ".el" (substring (car (car current-directory-list)) -3))
15925 (setq el-files-list
15926 (cons (car (car current-directory-list)) el-files-list)))
15929 ;; check whether filename is that of a directory
15930 ((eq t (car (cdr (car current-directory-list))))
15931 ;; decide whether to skip or recurse
15934 (substring (car (car current-directory-list)) -1))
15935 ;; then do nothing since filename is that of
15936 ;; current directory or parent, "." or ".."
15940 ;; else descend into the directory and repeat the process
15941 (setq el-files-list
15943 (files-in-below-directory
15944 (car (car current-directory-list)))
15946 ;; move to the next filename in the list; this also
15947 ;; shortens the list so the while loop eventually comes to an end
15948 (setq current-directory-list (cdr current-directory-list)))
15949 ;; return the filenames
15954 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15955 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15957 The @code{files-in-below-directory} @code{directory-files} function
15958 takes one argument, the name of a directory.
15961 Thus, on my system,
15963 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15965 @c !!! 22.1.1 lisp sources location here
15969 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15974 tells me that in and below my Lisp sources directory are 1031
15977 @code{files-in-below-directory} returns a list in reverse alphabetical
15978 order. An expression to sort the list in alphabetical order looks
15984 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15991 "Test how long it takes to find lengths of all sorted elisp defuns."
15992 (insert "\n" (current-time-string) "\n")
15995 (recursive-lengths-list-many-files
15996 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15998 (insert (format "%s" (current-time-string))))
16001 @node Counting function definitions, , Files List, Prepare the data
16002 @subsection Counting function definitions
16004 Our immediate goal is to generate a list that tells us how many
16005 function definitions contain fewer than 10 words and symbols, how many
16006 contain between 10 and 19 words and symbols, how many contain between
16007 20 and 29 words and symbols, and so on.
16009 With a sorted list of numbers, this is easy: count how many elements
16010 of the list are smaller than 10, then, after moving past the numbers
16011 just counted, count how many are smaller than 20, then, after moving
16012 past the numbers just counted, count how many are smaller than 30, and
16013 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
16014 larger than the top of that range. We can call the list of such
16015 numbers the @code{top-of-ranges} list.
16018 If we wished, we could generate this list automatically, but it is
16019 simpler to write a list manually. Here it is:
16020 @vindex top-of-ranges
16024 (defvar top-of-ranges
16027 110 120 130 140 150
16028 160 170 180 190 200
16029 210 220 230 240 250
16030 260 270 280 290 300)
16031 "List specifying ranges for `defuns-per-range'.")
16035 To change the ranges, we edit this list.
16037 Next, we need to write the function that creates the list of the
16038 number of definitions within each range. Clearly, this function must
16039 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
16042 The @code{defuns-per-range} function must do two things again and
16043 again: it must count the number of definitions within a range
16044 specified by the current top-of-range value; and it must shift to the
16045 next higher value in the @code{top-of-ranges} list after counting the
16046 number of definitions in the current range. Since each of these
16047 actions is repetitive, we can use @code{while} loops for the job.
16048 One loop counts the number of definitions in the range defined by the
16049 current top-of-range value, and the other loop selects each of the
16050 top-of-range values in turn.
16052 Several entries of the @code{sorted-lengths} list are counted for each
16053 range; this means that the loop for the @code{sorted-lengths} list
16054 will be inside the loop for the @code{top-of-ranges} list, like a
16055 small gear inside a big gear.
16057 The inner loop counts the number of definitions within the range. It
16058 is a simple counting loop of the type we have seen before.
16059 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
16060 The true-or-false test of the loop tests whether the value from the
16061 @code{sorted-lengths} list is smaller than the current value of the
16062 top of the range. If it is, the function increments the counter and
16063 tests the next value from the @code{sorted-lengths} list.
16066 The inner loop looks like this:
16070 (while @var{length-element-smaller-than-top-of-range}
16071 (setq number-within-range (1+ number-within-range))
16072 (setq sorted-lengths (cdr sorted-lengths)))
16076 The outer loop must start with the lowest value of the
16077 @code{top-of-ranges} list, and then be set to each of the succeeding
16078 higher values in turn. This can be done with a loop like this:
16082 (while top-of-ranges
16083 @var{body-of-loop}@dots{}
16084 (setq top-of-ranges (cdr top-of-ranges)))
16089 Put together, the two loops look like this:
16093 (while top-of-ranges
16095 ;; @r{Count the number of elements within the current range.}
16096 (while @var{length-element-smaller-than-top-of-range}
16097 (setq number-within-range (1+ number-within-range))
16098 (setq sorted-lengths (cdr sorted-lengths)))
16100 ;; @r{Move to next range.}
16101 (setq top-of-ranges (cdr top-of-ranges)))
16105 In addition, in each circuit of the outer loop, Emacs should record
16106 the number of definitions within that range (the value of
16107 @code{number-within-range}) in a list. We can use @code{cons} for
16108 this purpose. (@xref{cons, , @code{cons}}.)
16110 The @code{cons} function works fine, except that the list it
16111 constructs will contain the number of definitions for the highest
16112 range at its beginning and the number of definitions for the lowest
16113 range at its end. This is because @code{cons} attaches new elements
16114 of the list to the beginning of the list, and since the two loops are
16115 working their way through the lengths' list from the lower end first,
16116 the @code{defuns-per-range-list} will end up largest number first.
16117 But we will want to print our graph with smallest values first and the
16118 larger later. The solution is to reverse the order of the
16119 @code{defuns-per-range-list}. We can do this using the
16120 @code{nreverse} function, which reverses the order of a list.
16127 (nreverse '(1 2 3 4))
16138 Note that the @code{nreverse} function is ``destructive''---that is,
16139 it changes the list to which it is applied; this contrasts with the
16140 @code{car} and @code{cdr} functions, which are non-destructive. In
16141 this case, we do not want the original @code{defuns-per-range-list},
16142 so it does not matter that it is destroyed. (The @code{reverse}
16143 function provides a reversed copy of a list, leaving the original list
16148 Put all together, the @code{defuns-per-range} looks like this:
16152 (defun defuns-per-range (sorted-lengths top-of-ranges)
16153 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16154 (let ((top-of-range (car top-of-ranges))
16155 (number-within-range 0)
16156 defuns-per-range-list)
16161 (while top-of-ranges
16167 ;; @r{Need number for numeric test.}
16168 (car sorted-lengths)
16169 (< (car sorted-lengths) top-of-range))
16173 ;; @r{Count number of definitions within current range.}
16174 (setq number-within-range (1+ number-within-range))
16175 (setq sorted-lengths (cdr sorted-lengths)))
16177 ;; @r{Exit inner loop but remain within outer loop.}
16181 (setq defuns-per-range-list
16182 (cons number-within-range defuns-per-range-list))
16183 (setq number-within-range 0) ; @r{Reset count to zero.}
16187 ;; @r{Move to next range.}
16188 (setq top-of-ranges (cdr top-of-ranges))
16189 ;; @r{Specify next top of range value.}
16190 (setq top-of-range (car top-of-ranges)))
16194 ;; @r{Exit outer loop and count the number of defuns larger than}
16195 ;; @r{ the largest top-of-range value.}
16196 (setq defuns-per-range-list
16198 (length sorted-lengths)
16199 defuns-per-range-list))
16203 ;; @r{Return a list of the number of definitions within each range,}
16204 ;; @r{ smallest to largest.}
16205 (nreverse defuns-per-range-list)))
16211 The function is straightforward except for one subtle feature. The
16212 true-or-false test of the inner loop looks like this:
16216 (and (car sorted-lengths)
16217 (< (car sorted-lengths) top-of-range))
16223 instead of like this:
16226 (< (car sorted-lengths) top-of-range)
16229 The purpose of the test is to determine whether the first item in the
16230 @code{sorted-lengths} list is less than the value of the top of the
16233 The simple version of the test works fine unless the
16234 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16235 @code{(car sorted-lengths)} expression function returns
16236 @code{nil}. The @code{<} function cannot compare a number to
16237 @code{nil}, which is an empty list, so Emacs signals an error and
16238 stops the function from attempting to continue to execute.
16240 The @code{sorted-lengths} list always becomes @code{nil} when the
16241 counter reaches the end of the list. This means that any attempt to
16242 use the @code{defuns-per-range} function with the simple version of
16243 the test will fail.
16245 We solve the problem by using the @code{(car sorted-lengths)}
16246 expression in conjunction with the @code{and} expression. The
16247 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16248 value so long as the list has at least one number within it, but
16249 returns @code{nil} if the list is empty. The @code{and} expression
16250 first evaluates the @code{(car sorted-lengths)} expression, and
16251 if it is @code{nil}, returns false @emph{without} evaluating the
16252 @code{<} expression. But if the @code{(car sorted-lengths)}
16253 expression returns a non-@code{nil} value, the @code{and} expression
16254 evaluates the @code{<} expression, and returns that value as the value
16255 of the @code{and} expression.
16257 @c colon in printed section title causes problem in Info cross reference
16258 This way, we avoid an error.
16261 (For information about @code{and}, see
16262 @ref{kill-new function, , The @code{kill-new} function}.)
16266 (@xref{kill-new function, , The @code{kill-new} function}, for
16267 information about @code{and}.)
16270 Here is a short test of the @code{defuns-per-range} function. First,
16271 evaluate the expression that binds (a shortened)
16272 @code{top-of-ranges} list to the list of values, then evaluate the
16273 expression for binding the @code{sorted-lengths} list, and then
16274 evaluate the @code{defuns-per-range} function.
16278 ;; @r{(Shorter list than we will use later.)}
16279 (setq top-of-ranges
16280 '(110 120 130 140 150
16281 160 170 180 190 200))
16283 (setq sorted-lengths
16284 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16286 (defuns-per-range sorted-lengths top-of-ranges)
16292 The list returned looks like this:
16295 (2 2 2 0 0 1 0 2 0 0 4)
16299 Indeed, there are two elements of the @code{sorted-lengths} list
16300 smaller than 110, two elements between 110 and 119, two elements
16301 between 120 and 129, and so on. There are four elements with a value
16304 @c The next step is to turn this numbers' list into a graph.
16305 @node Readying a Graph, Emacs Initialization, Words in a defun, Top
16306 @chapter Readying a Graph
16307 @cindex Readying a graph
16308 @cindex Graph prototype
16309 @cindex Prototype graph
16310 @cindex Body of graph
16312 Our goal is to construct a graph showing the numbers of function
16313 definitions of various lengths in the Emacs lisp sources.
16315 As a practical matter, if you were creating a graph, you would
16316 probably use a program such as @code{gnuplot} to do the job.
16317 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16318 however, we create one from scratch, and in the process we will
16319 re-acquaint ourselves with some of what we learned before and learn
16322 In this chapter, we will first write a simple graph printing function.
16323 This first definition will be a @dfn{prototype}, a rapidly written
16324 function that enables us to reconnoiter this unknown graph-making
16325 territory. We will discover dragons, or find that they are myth.
16326 After scouting the terrain, we will feel more confident and enhance
16327 the function to label the axes automatically.
16330 * Columns of a graph::
16331 * graph-body-print:: How to print the body of a graph.
16332 * recursive-graph-body-print::
16334 * Line Graph Exercise::
16337 @node Columns of a graph, graph-body-print, Readying a Graph, Readying a Graph
16339 @unnumberedsec Printing the Columns of a Graph
16342 Since Emacs is designed to be flexible and work with all kinds of
16343 terminals, including character-only terminals, the graph will need to
16344 be made from one of the `typewriter' symbols. An asterisk will do; as
16345 we enhance the graph-printing function, we can make the choice of
16346 symbol a user option.
16348 We can call this function @code{graph-body-print}; it will take a
16349 @code{numbers-list} as its only argument. At this stage, we will not
16350 label the graph, but only print its body.
16352 The @code{graph-body-print} function inserts a vertical column of
16353 asterisks for each element in the @code{numbers-list}. The height of
16354 each line is determined by the value of that element of the
16355 @code{numbers-list}.
16357 Inserting columns is a repetitive act; that means that this function can
16358 be written either with a @code{while} loop or recursively.
16360 Our first challenge is to discover how to print a column of asterisks.
16361 Usually, in Emacs, we print characters onto a screen horizontally,
16362 line by line, by typing. We have two routes we can follow: write our
16363 own column-insertion function or discover whether one exists in Emacs.
16365 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16366 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16367 command, except that the latter finds only those functions that are
16368 commands. The @kbd{M-x apropos} command lists all symbols that match
16369 a regular expression, including functions that are not interactive.
16372 What we want to look for is some command that prints or inserts
16373 columns. Very likely, the name of the function will contain either
16374 the word `print' or the word `insert' or the word `column'.
16375 Therefore, we can simply type @kbd{M-x apropos RET
16376 print\|insert\|column RET} and look at the result. On my system, this
16377 command once too takes quite some time, and then produced a list of 79
16378 functions and variables. Now it does not take much time at all and
16379 produces a list of 211 functions and variables. Scanning down the
16380 list, the only function that looks as if it might do the job is
16381 @code{insert-rectangle}.
16384 Indeed, this is the function we want; its documentation says:
16389 Insert text of RECTANGLE with upper left corner at point.
16390 RECTANGLE's first line is inserted at point,
16391 its second line is inserted at a point vertically under point, etc.
16392 RECTANGLE should be a list of strings.
16393 After this command, the mark is at the upper left corner
16394 and point is at the lower right corner.
16398 We can run a quick test, to make sure it does what we expect of it.
16400 Here is the result of placing the cursor after the
16401 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16402 (@code{eval-last-sexp}). The function inserts the strings
16403 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16404 point. Also the function returns @code{nil}.
16408 (insert-rectangle '("first" "second" "third"))first
16415 Of course, we won't be inserting the text of the
16416 @code{insert-rectangle} expression itself into the buffer in which we
16417 are making the graph, but will call the function from our program. We
16418 shall, however, have to make sure that point is in the buffer at the
16419 place where the @code{insert-rectangle} function will insert its
16422 If you are reading this in Info, you can see how this works by
16423 switching to another buffer, such as the @file{*scratch*} buffer,
16424 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16425 @code{insert-rectangle} expression into the minibuffer at the prompt,
16426 and then typing @key{RET}. This causes Emacs to evaluate the
16427 expression in the minibuffer, but to use as the value of point the
16428 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16429 keybinding for @code{eval-expression}. Also, @code{nil} does not
16430 appear in the @file{*scratch*} buffer since the expression is
16431 evaluated in the minibuffer.)
16433 We find when we do this that point ends up at the end of the last
16434 inserted line---that is to say, this function moves point as a
16435 side-effect. If we were to repeat the command, with point at this
16436 position, the next insertion would be below and to the right of the
16437 previous insertion. We don't want this! If we are going to make a
16438 bar graph, the columns need to be beside each other.
16440 So we discover that each cycle of the column-inserting @code{while}
16441 loop must reposition point to the place we want it, and that place
16442 will be at the top, not the bottom, of the column. Moreover, we
16443 remember that when we print a graph, we do not expect all the columns
16444 to be the same height. This means that the top of each column may be
16445 at a different height from the previous one. We cannot simply
16446 reposition point to the same line each time, but moved over to the
16447 right---or perhaps we can@dots{}
16449 We are planning to make the columns of the bar graph out of asterisks.
16450 The number of asterisks in the column is the number specified by the
16451 current element of the @code{numbers-list}. We need to construct a
16452 list of asterisks of the right length for each call to
16453 @code{insert-rectangle}. If this list consists solely of the requisite
16454 number of asterisks, then we will have position point the right number
16455 of lines above the base for the graph to print correctly. This could
16458 Alternatively, if we can figure out some way to pass
16459 @code{insert-rectangle} a list of the same length each time, then we
16460 can place point on the same line each time, but move it over one
16461 column to the right for each new column. If we do this, however, some
16462 of the entries in the list passed to @code{insert-rectangle} must be
16463 blanks rather than asterisks. For example, if the maximum height of
16464 the graph is 5, but the height of the column is 3, then
16465 @code{insert-rectangle} requires an argument that looks like this:
16468 (" " " " "*" "*" "*")
16471 This last proposal is not so difficult, so long as we can determine
16472 the column height. There are two ways for us to specify the column
16473 height: we can arbitrarily state what it will be, which would work
16474 fine for graphs of that height; or we can search through the list of
16475 numbers and use the maximum height of the list as the maximum height
16476 of the graph. If the latter operation were difficult, then the former
16477 procedure would be easiest, but there is a function built into Emacs
16478 that determines the maximum of its arguments. We can use that
16479 function. The function is called @code{max} and it returns the
16480 largest of all its arguments, which must be numbers. Thus, for
16488 returns 7. (A corresponding function called @code{min} returns the
16489 smallest of all its arguments.)
16493 However, we cannot simply call @code{max} on the @code{numbers-list};
16494 the @code{max} function expects numbers as its argument, not a list of
16495 numbers. Thus, the following expression,
16498 (max '(3 4 6 5 7 3))
16503 produces the following error message;
16506 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16510 We need a function that passes a list of arguments to a function.
16511 This function is @code{apply}. This function `applies' its first
16512 argument (a function) to its remaining arguments, the last of which
16519 (apply 'max 3 4 7 3 '(4 8 5))
16525 (Incidentally, I don't know how you would learn of this function
16526 without a book such as this. It is possible to discover other
16527 functions, like @code{search-forward} or @code{insert-rectangle}, by
16528 guessing at a part of their names and then using @code{apropos}. Even
16529 though its base in metaphor is clear---`apply' its first argument to
16530 the rest---I doubt a novice would come up with that particular word
16531 when using @code{apropos} or other aid. Of course, I could be wrong;
16532 after all, the function was first named by someone who had to invent
16535 The second and subsequent arguments to @code{apply} are optional, so
16536 we can use @code{apply} to call a function and pass the elements of a
16537 list to it, like this, which also returns 8:
16540 (apply 'max '(4 8 5))
16543 This latter way is how we will use @code{apply}. The
16544 @code{recursive-lengths-list-many-files} function returns a numbers'
16545 list to which we can apply @code{max} (we could also apply @code{max} to
16546 the sorted numbers' list; it does not matter whether the list is
16550 Hence, the operation for finding the maximum height of the graph is this:
16553 (setq max-graph-height (apply 'max numbers-list))
16556 Now we can return to the question of how to create a list of strings
16557 for a column of the graph. Told the maximum height of the graph
16558 and the number of asterisks that should appear in the column, the
16559 function should return a list of strings for the
16560 @code{insert-rectangle} command to insert.
16562 Each column is made up of asterisks or blanks. Since the function is
16563 passed the value of the height of the column and the number of
16564 asterisks in the column, the number of blanks can be found by
16565 subtracting the number of asterisks from the height of the column.
16566 Given the number of blanks and the number of asterisks, two
16567 @code{while} loops can be used to construct the list:
16571 ;;; @r{First version.}
16572 (defun column-of-graph (max-graph-height actual-height)
16573 "Return list of strings that is one column of a graph."
16574 (let ((insert-list nil)
16575 (number-of-top-blanks
16576 (- max-graph-height actual-height)))
16580 ;; @r{Fill in asterisks.}
16581 (while (> actual-height 0)
16582 (setq insert-list (cons "*" insert-list))
16583 (setq actual-height (1- actual-height)))
16587 ;; @r{Fill in blanks.}
16588 (while (> number-of-top-blanks 0)
16589 (setq insert-list (cons " " insert-list))
16590 (setq number-of-top-blanks
16591 (1- number-of-top-blanks)))
16595 ;; @r{Return whole list.}
16600 If you install this function and then evaluate the following
16601 expression you will see that it returns the list as desired:
16604 (column-of-graph 5 3)
16612 (" " " " "*" "*" "*")
16615 As written, @code{column-of-graph} contains a major flaw: the symbols
16616 used for the blank and for the marked entries in the column are
16617 `hard-coded' as a space and asterisk. This is fine for a prototype,
16618 but you, or another user, may wish to use other symbols. For example,
16619 in testing the graph function, you many want to use a period in place
16620 of the space, to make sure the point is being repositioned properly
16621 each time the @code{insert-rectangle} function is called; or you might
16622 want to substitute a @samp{+} sign or other symbol for the asterisk.
16623 You might even want to make a graph-column that is more than one
16624 display column wide. The program should be more flexible. The way to
16625 do that is to replace the blank and the asterisk with two variables
16626 that we can call @code{graph-blank} and @code{graph-symbol} and define
16627 those variables separately.
16629 Also, the documentation is not well written. These considerations
16630 lead us to the second version of the function:
16634 (defvar graph-symbol "*"
16635 "String used as symbol in graph, usually an asterisk.")
16639 (defvar graph-blank " "
16640 "String used as blank in graph, usually a blank space.
16641 graph-blank must be the same number of columns wide
16647 (For an explanation of @code{defvar}, see
16648 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16652 ;;; @r{Second version.}
16653 (defun column-of-graph (max-graph-height actual-height)
16654 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16658 The graph-symbols are contiguous entries at the end
16660 The list will be inserted as one column of a graph.
16661 The strings are either graph-blank or graph-symbol."
16665 (let ((insert-list nil)
16666 (number-of-top-blanks
16667 (- max-graph-height actual-height)))
16671 ;; @r{Fill in @code{graph-symbols}.}
16672 (while (> actual-height 0)
16673 (setq insert-list (cons graph-symbol insert-list))
16674 (setq actual-height (1- actual-height)))
16678 ;; @r{Fill in @code{graph-blanks}.}
16679 (while (> number-of-top-blanks 0)
16680 (setq insert-list (cons graph-blank insert-list))
16681 (setq number-of-top-blanks
16682 (1- number-of-top-blanks)))
16684 ;; @r{Return whole list.}
16689 If we wished, we could rewrite @code{column-of-graph} a third time to
16690 provide optionally for a line graph as well as for a bar graph. This
16691 would not be hard to do. One way to think of a line graph is that it
16692 is no more than a bar graph in which the part of each bar that is
16693 below the top is blank. To construct a column for a line graph, the
16694 function first constructs a list of blanks that is one shorter than
16695 the value, then it uses @code{cons} to attach a graph symbol to the
16696 list; then it uses @code{cons} again to attach the `top blanks' to
16699 It is easy to see how to write such a function, but since we don't
16700 need it, we will not do it. But the job could be done, and if it were
16701 done, it would be done with @code{column-of-graph}. Even more
16702 important, it is worth noting that few changes would have to be made
16703 anywhere else. The enhancement, if we ever wish to make it, is
16706 Now, finally, we come to our first actual graph printing function.
16707 This prints the body of a graph, not the labels for the vertical and
16708 horizontal axes, so we can call this @code{graph-body-print}.
16710 @node graph-body-print, recursive-graph-body-print, Columns of a graph, Readying a Graph
16711 @section The @code{graph-body-print} Function
16712 @findex graph-body-print
16714 After our preparation in the preceding section, the
16715 @code{graph-body-print} function is straightforward. The function
16716 will print column after column of asterisks and blanks, using the
16717 elements of a numbers' list to specify the number of asterisks in each
16718 column. This is a repetitive act, which means we can use a
16719 decrementing @code{while} loop or recursive function for the job. In
16720 this section, we will write the definition using a @code{while} loop.
16722 The @code{column-of-graph} function requires the height of the graph
16723 as an argument, so we should determine and record that as a local variable.
16725 This leads us to the following template for the @code{while} loop
16726 version of this function:
16730 (defun graph-body-print (numbers-list)
16731 "@var{documentation}@dots{}"
16732 (let ((height @dots{}
16737 (while numbers-list
16738 @var{insert-columns-and-reposition-point}
16739 (setq numbers-list (cdr numbers-list)))))
16744 We need to fill in the slots of the template.
16746 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16747 determine the height of the graph.
16749 The @code{while} loop will cycle through the @code{numbers-list} one
16750 element at a time. As it is shortened by the @code{(setq numbers-list
16751 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16752 list is the value of the argument for @code{column-of-graph}.
16754 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16755 function inserts the list returned by @code{column-of-graph}. Since
16756 the @code{insert-rectangle} function moves point to the lower right of
16757 the inserted rectangle, we need to save the location of point at the
16758 time the rectangle is inserted, move back to that position after the
16759 rectangle is inserted, and then move horizontally to the next place
16760 from which @code{insert-rectangle} is called.
16762 If the inserted columns are one character wide, as they will be if
16763 single blanks and asterisks are used, the repositioning command is
16764 simply @code{(forward-char 1)}; however, the width of a column may be
16765 greater than one. This means that the repositioning command should be
16766 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16767 itself is the length of a @code{graph-blank} and can be found using
16768 the expression @code{(length graph-blank)}. The best place to bind
16769 the @code{symbol-width} variable to the value of the width of graph
16770 column is in the varlist of the @code{let} expression.
16773 These considerations lead to the following function definition:
16777 (defun graph-body-print (numbers-list)
16778 "Print a bar graph of the NUMBERS-LIST.
16779 The numbers-list consists of the Y-axis values."
16781 (let ((height (apply 'max numbers-list))
16782 (symbol-width (length graph-blank))
16787 (while numbers-list
16788 (setq from-position (point))
16790 (column-of-graph height (car numbers-list)))
16791 (goto-char from-position)
16792 (forward-char symbol-width)
16795 ;; @r{Draw graph column by column.}
16797 (setq numbers-list (cdr numbers-list)))
16800 ;; @r{Place point for X axis labels.}
16801 (forward-line height)
16808 The one unexpected expression in this function is the
16809 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16810 expression makes the graph printing operation more interesting to
16811 watch than it would be otherwise. The expression causes Emacs to
16812 `sit' or do nothing for a zero length of time and then redraw the
16813 screen. Placed here, it causes Emacs to redraw the screen column by
16814 column. Without it, Emacs would not redraw the screen until the
16817 We can test @code{graph-body-print} with a short list of numbers.
16821 Install @code{graph-symbol}, @code{graph-blank},
16822 @code{column-of-graph}, which are in
16824 @ref{Readying a Graph, , Readying a Graph},
16827 @ref{Columns of a graph},
16829 and @code{graph-body-print}.
16833 Copy the following expression:
16836 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16840 Switch to the @file{*scratch*} buffer and place the cursor where you
16841 want the graph to start.
16844 Type @kbd{M-:} (@code{eval-expression}).
16847 Yank the @code{graph-body-print} expression into the minibuffer
16848 with @kbd{C-y} (@code{yank)}.
16851 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16855 Emacs will print a graph like this:
16869 @node recursive-graph-body-print, Printed Axes, graph-body-print, Readying a Graph
16870 @section The @code{recursive-graph-body-print} Function
16871 @findex recursive-graph-body-print
16873 The @code{graph-body-print} function may also be written recursively.
16874 The recursive solution is divided into two parts: an outside `wrapper'
16875 that uses a @code{let} expression to determine the values of several
16876 variables that need only be found once, such as the maximum height of
16877 the graph, and an inside function that is called recursively to print
16881 The `wrapper' is uncomplicated:
16885 (defun recursive-graph-body-print (numbers-list)
16886 "Print a bar graph of the NUMBERS-LIST.
16887 The numbers-list consists of the Y-axis values."
16888 (let ((height (apply 'max numbers-list))
16889 (symbol-width (length graph-blank))
16891 (recursive-graph-body-print-internal
16898 The recursive function is a little more difficult. It has four parts:
16899 the `do-again-test', the printing code, the recursive call, and the
16900 `next-step-expression'. The `do-again-test' is a @code{when}
16901 expression that determines whether the @code{numbers-list} contains
16902 any remaining elements; if it does, the function prints one column of
16903 the graph using the printing code and calls itself again. The
16904 function calls itself again according to the value produced by the
16905 `next-step-expression' which causes the call to act on a shorter
16906 version of the @code{numbers-list}.
16910 (defun recursive-graph-body-print-internal
16911 (numbers-list height symbol-width)
16912 "Print a bar graph.
16913 Used within recursive-graph-body-print function."
16918 (setq from-position (point))
16920 (column-of-graph height (car numbers-list)))
16923 (goto-char from-position)
16924 (forward-char symbol-width)
16925 (sit-for 0) ; @r{Draw graph column by column.}
16926 (recursive-graph-body-print-internal
16927 (cdr numbers-list) height symbol-width)))
16932 After installation, this expression can be tested; here is a sample:
16935 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16939 Here is what @code{recursive-graph-body-print} produces:
16953 Either of these two functions, @code{graph-body-print} or
16954 @code{recursive-graph-body-print}, create the body of a graph.
16956 @node Printed Axes, Line Graph Exercise, recursive-graph-body-print, Readying a Graph
16957 @section Need for Printed Axes
16959 A graph needs printed axes, so you can orient yourself. For a do-once
16960 project, it may be reasonable to draw the axes by hand using Emacs'
16961 Picture mode; but a graph drawing function may be used more than once.
16963 For this reason, I have written enhancements to the basic
16964 @code{print-graph-body} function that automatically print labels for
16965 the horizontal and vertical axes. Since the label printing functions
16966 do not contain much new material, I have placed their description in
16967 an appendix. @xref{Full Graph, , A Graph with Labelled Axes}.
16969 @node Line Graph Exercise, , Printed Axes, Readying a Graph
16972 Write a line graph version of the graph printing functions.
16974 @node Emacs Initialization, Debugging, Readying a Graph, Top
16975 @chapter Your @file{.emacs} File
16976 @cindex @file{.emacs} file
16977 @cindex Customizing your @file{.emacs} file
16978 @cindex Initialization file
16980 ``You don't have to like Emacs to like it'' -- this seemingly
16981 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16982 the box' Emacs is a generic tool. Most people who use it, customize
16983 it to suit themselves.
16985 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16986 expressions in Emacs Lisp you can change or extend Emacs.
16989 * Default Configuration::
16990 * Site-wide Init:: You can write site-wide init files.
16991 * defcustom:: Emacs will write code for you.
16992 * Beginning a .emacs File:: How to write a @code{.emacs file}.
16993 * Text and Auto-fill:: Automatically wrap lines.
16994 * Mail Aliases:: Use abbreviations for email addresses.
16995 * Indent Tabs Mode:: Don't use tabs with @TeX{}
16996 * Keybindings:: Create some personal keybindings.
16997 * Keymaps:: More about key binding.
16998 * Loading Files:: Load (i.e., evaluate) files automatically.
16999 * Autoload:: Make functions available.
17000 * Simple Extension:: Define a function; bind it to a key.
17001 * X11 Colors:: Colors in X.
17003 * Mode Line:: How to customize your mode line.
17006 @node Default Configuration, Site-wide Init, Emacs Initialization, Emacs Initialization
17008 @unnumberedsec Emacs' Default Configuration
17011 There are those who appreciate Emacs' default configuration. After
17012 all, Emacs starts you in C mode when you edit a C file, starts you in
17013 Fortran mode when you edit a Fortran file, and starts you in
17014 Fundamental mode when you edit an unadorned file. This all makes
17015 sense, if you do not know who is going to use Emacs. Who knows what a
17016 person hopes to do with an unadorned file? Fundamental mode is the
17017 right default for such a file, just as C mode is the right default for
17018 editing C code. (Enough programming languages have syntaxes
17019 that enable them to share or nearly share features, so C mode is
17020 now provided by CC mode, the `C Collection'.)
17022 But when you do know who is going to use Emacs---you,
17023 yourself---then it makes sense to customize Emacs.
17025 For example, I seldom want Fundamental mode when I edit an
17026 otherwise undistinguished file; I want Text mode. This is why I
17027 customize Emacs: so it suits me.
17029 You can customize and extend Emacs by writing or adapting a
17030 @file{~/.emacs} file. This is your personal initialization file; its
17031 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
17032 may also add @file{.el} to @file{~/.emacs} and call it a
17033 @file{~/.emacs.el} file. In the past, you were forbidden to type the
17034 extra keystrokes that the name @file{~/.emacs.el} requires, but now
17035 you may. The new format is consistent with the Emacs Lisp file
17036 naming conventions; the old format saves typing.}
17038 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
17039 code yourself; or you can use Emacs' @code{customize} feature to write
17040 the code for you. You can combine your own expressions and
17041 auto-written Customize expressions in your @file{.emacs} file.
17043 (I myself prefer to write my own expressions, except for those,
17044 particularly fonts, that I find easier to manipulate using the
17045 @code{customize} command. I combine the two methods.)
17047 Most of this chapter is about writing expressions yourself. It
17048 describes a simple @file{.emacs} file; for more information, see
17049 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
17050 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
17053 @node Site-wide Init, defcustom, Default Configuration, Emacs Initialization
17054 @section Site-wide Initialization Files
17056 @cindex @file{default.el} init file
17057 @cindex @file{site-init.el} init file
17058 @cindex @file{site-load.el} init file
17059 In addition to your personal initialization file, Emacs automatically
17060 loads various site-wide initialization files, if they exist. These
17061 have the same form as your @file{.emacs} file, but are loaded by
17064 Two site-wide initialization files, @file{site-load.el} and
17065 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
17066 `dumped' version of Emacs is created, as is most common. (Dumped
17067 copies of Emacs load more quickly. However, once a file is loaded and
17068 dumped, a change to it does not lead to a change in Emacs unless you
17069 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
17070 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
17071 @file{INSTALL} file.)
17073 Three other site-wide initialization files are loaded automatically
17074 each time you start Emacs, if they exist. These are
17075 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
17076 file, and @file{default.el}, and the terminal type file, which are both
17077 loaded @emph{after} your @file{.emacs} file.
17079 Settings and definitions in your @file{.emacs} file will overwrite
17080 conflicting settings and definitions in a @file{site-start.el} file,
17081 if it exists; but the settings and definitions in a @file{default.el}
17082 or terminal type file will overwrite those in your @file{.emacs} file.
17083 (You can prevent interference from a terminal type file by setting
17084 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
17085 Simple Extension}.)
17087 @c Rewritten to avoid overfull hbox.
17088 The @file{INSTALL} file that comes in the distribution contains
17089 descriptions of the @file{site-init.el} and @file{site-load.el} files.
17091 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
17092 control loading. These files are in the @file{lisp} directory of the
17093 Emacs distribution and are worth perusing.
17095 The @file{loaddefs.el} file contains a good many suggestions as to
17096 what to put into your own @file{.emacs} file, or into a site-wide
17097 initialization file.
17099 @node defcustom, Beginning a .emacs File, Site-wide Init, Emacs Initialization
17100 @section Specifying Variables using @code{defcustom}
17103 You can specify variables using @code{defcustom} so that you and
17104 others can then use Emacs' @code{customize} feature to set their
17105 values. (You cannot use @code{customize} to write function
17106 definitions; but you can write @code{defuns} in your @file{.emacs}
17107 file. Indeed, you can write any Lisp expression in your @file{.emacs}
17110 The @code{customize} feature depends on the @code{defcustom} special
17111 form. Although you can use @code{defvar} or @code{setq} for variables
17112 that users set, the @code{defcustom} special form is designed for the
17115 You can use your knowledge of @code{defvar} for writing the
17116 first three arguments for @code{defcustom}. The first argument to
17117 @code{defcustom} is the name of the variable. The second argument is
17118 the variable's initial value, if any; and this value is set only if
17119 the value has not already been set. The third argument is the
17122 The fourth and subsequent arguments to @code{defcustom} specify types
17123 and options; these are not featured in @code{defvar}. (These
17124 arguments are optional.)
17126 Each of these arguments consists of a keyword followed by a value.
17127 Each keyword starts with the colon character @samp{:}.
17130 For example, the customizable user option variable
17131 @code{text-mode-hook} looks like this:
17135 (defcustom text-mode-hook nil
17136 "Normal hook run when entering Text mode and many related modes."
17138 :options '(turn-on-auto-fill flyspell-mode)
17144 The name of the variable is @code{text-mode-hook}; it has no default
17145 value; and its documentation string tells you what it does.
17147 The @code{:type} keyword tells Emacs the kind of data to which
17148 @code{text-mode-hook} should be set and how to display the value in a
17149 Customization buffer.
17151 The @code{:options} keyword specifies a suggested list of values for
17152 the variable. Usually, @code{:options} applies to a hook.
17153 The list is only a suggestion; it is not exclusive; a person who sets
17154 the variable may set it to other values; the list shown following the
17155 @code{:options} keyword is intended to offer convenient choices to a
17158 Finally, the @code{:group} keyword tells the Emacs Customization
17159 command in which group the variable is located. This tells where to
17162 The @code{defcustom} function recognizes more than a dozen keywords.
17163 For more information, see @ref{Customization, , Writing Customization
17164 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17166 Consider @code{text-mode-hook} as an example.
17168 There are two ways to customize this variable. You can use the
17169 customization command or write the appropriate expressions yourself.
17172 Using the customization command, you can type:
17179 and find that the group for editing files of data is called `data'.
17180 Enter that group. Text Mode Hook is the first member. You can click
17181 on its various options, such as @code{turn-on-auto-fill}, to set the
17182 values. After you click on the button to
17185 Save for Future Sessions
17189 Emacs will write an expression into your @file{.emacs} file.
17190 It will look like this:
17194 (custom-set-variables
17195 ;; custom-set-variables was added by Custom.
17196 ;; If you edit it by hand, you could mess it up, so be careful.
17197 ;; Your init file should contain only one such instance.
17198 ;; If there is more than one, they won't work right.
17199 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17204 (The @code{text-mode-hook-identify} function tells
17205 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17206 It comes on automatically.)
17208 The @code{custom-set-variables} function works somewhat differently
17209 than a @code{setq}. While I have never learned the differences, I
17210 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17211 file by hand: I make the changes in what appears to me to be a
17212 reasonable manner and have not had any problems. Others prefer to use
17213 the Customization command and let Emacs do the work for them.
17215 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17216 This function sets the various font faces. Over time, I have set a
17217 considerable number of faces. Some of the time, I re-set them using
17218 @code{customize}; other times, I simply edit the
17219 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17221 The second way to customize your @code{text-mode-hook} is to set it
17222 yourself in your @file{.emacs} file using code that has nothing to do
17223 with the @code{custom-set-@dots{}} functions.
17226 When you do this, and later use @code{customize}, you will see a
17230 CHANGED outside Customize; operating on it here may be unreliable.
17234 This message is only a warning. If you click on the button to
17237 Save for Future Sessions
17241 Emacs will write a @code{custom-set-@dots{}} expression near the end
17242 of your @file{.emacs} file that will be evaluated after your
17243 hand-written expression. It will, therefore, overrule your
17244 hand-written expression. No harm will be done. When you do this,
17245 however, be careful to remember which expression is active; if you
17246 forget, you may confuse yourself.
17248 So long as you remember where the values are set, you will have no
17249 trouble. In any event, the values are always set in your
17250 initialization file, which is usually called @file{.emacs}.
17252 I myself use @code{customize} for hardly anything. Mostly, I write
17253 expressions myself.
17257 Incidentally, to be more complete concerning defines: @code{defsubst}
17258 defines an inline function. The syntax is just like that of
17259 @code{defun}. @code{defconst} defines a symbol as a constant. The
17260 intent is that neither programs nor users should ever change a value
17261 set by @code{defconst}. (You can change it; the value set is a
17262 variable; but please do not.)
17264 @node Beginning a .emacs File, Text and Auto-fill, defcustom, Emacs Initialization
17265 @section Beginning a @file{.emacs} File
17266 @cindex @file{.emacs} file, beginning of
17268 When you start Emacs, it loads your @file{.emacs} file unless you tell
17269 it not to by specifying @samp{-q} on the command line. (The
17270 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17272 A @file{.emacs} file contains Lisp expressions. Often, these are no
17273 more than expressions to set values; sometimes they are function
17276 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17277 Manual}, for a short description of initialization files.
17279 This chapter goes over some of the same ground, but is a walk among
17280 extracts from a complete, long-used @file{.emacs} file---my own.
17282 The first part of the file consists of comments: reminders to myself.
17283 By now, of course, I remember these things, but when I started, I did
17289 ;;;; Bob's .emacs file
17290 ; Robert J. Chassell
17291 ; 26 September 1985
17296 Look at that date! I started this file a long time ago. I have been
17297 adding to it ever since.
17301 ; Each section in this file is introduced by a
17302 ; line beginning with four semicolons; and each
17303 ; entry is introduced by a line beginning with
17304 ; three semicolons.
17309 This describes the usual conventions for comments in Emacs Lisp.
17310 Everything on a line that follows a semicolon is a comment. Two,
17311 three, and four semicolons are used as subsection and section markers.
17312 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17313 more about comments.)
17318 ; Control-h is the help key;
17319 ; after typing control-h, type a letter to
17320 ; indicate the subject about which you want help.
17321 ; For an explanation of the help facility,
17322 ; type control-h two times in a row.
17327 Just remember: type @kbd{C-h} two times for help.
17331 ; To find out about any mode, type control-h m
17332 ; while in that mode. For example, to find out
17333 ; about mail mode, enter mail mode and then type
17339 `Mode help', as I call this, is very helpful. Usually, it tells you
17340 all you need to know.
17342 Of course, you don't need to include comments like these in your
17343 @file{.emacs} file. I included them in mine because I kept forgetting
17344 about Mode help or the conventions for comments---but I was able to
17345 remember to look here to remind myself.
17347 @node Text and Auto-fill, Mail Aliases, Beginning a .emacs File, Emacs Initialization
17348 @section Text and Auto Fill Mode
17350 Now we come to the part that `turns on' Text mode and
17355 ;;; Text mode and Auto Fill mode
17356 ;; The next two lines put Emacs into Text mode
17357 ;; and Auto Fill mode, and are for writers who
17358 ;; want to start writing prose rather than code.
17359 (setq-default major-mode 'text-mode)
17360 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17364 Here is the first part of this @file{.emacs} file that does something
17365 besides remind a forgetful human!
17367 The first of the two lines in parentheses tells Emacs to turn on Text
17368 mode when you find a file, @emph{unless} that file should go into some
17369 other mode, such as C mode.
17371 @cindex Per-buffer, local variables list
17372 @cindex Local variables list, per-buffer,
17373 @cindex Automatic mode selection
17374 @cindex Mode selection, automatic
17375 When Emacs reads a file, it looks at the extension to the file name,
17376 if any. (The extension is the part that comes after a @samp{.}.) If
17377 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17378 on C mode. Also, Emacs looks at first nonblank line of the file; if
17379 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17380 possesses a list of extensions and specifications that it uses
17381 automatically. In addition, Emacs looks near the last page for a
17382 per-buffer, ``local variables list'', if any.
17385 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17388 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17392 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17393 Files'' in @cite{The GNU Emacs Manual}.
17396 Now, back to the @file{.emacs} file.
17399 Here is the line again; how does it work?
17401 @cindex Text Mode turned on
17403 (setq major-mode 'text-mode)
17407 This line is a short, but complete Emacs Lisp expression.
17409 We are already familiar with @code{setq}. It sets the following variable,
17410 @code{major-mode}, to the subsequent value, which is @code{text-mode}.
17411 The single quote mark before @code{text-mode} tells Emacs to deal directly
17412 with the @code{text-mode} symbol, not with whatever it might stand for.
17413 @xref{set & setq, , Setting the Value of a Variable},
17414 for a reminder of how @code{setq} works.
17415 The main point is that there is no difference between the procedure you
17416 use to set a value in your @file{.emacs} file and the procedure you use
17417 anywhere else in Emacs.
17420 Here is the next line:
17422 @cindex Auto Fill mode turned on
17425 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17429 In this line, the @code{add-hook} command adds
17430 @code{turn-on-auto-fill} to the variable.
17432 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17433 it!, turns on Auto Fill mode.
17435 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17436 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17437 turns on Auto Fill mode.
17439 In brief, the first line causes Emacs to enter Text mode when you edit a
17440 file, unless the file name extension, a first non-blank line, or local
17441 variables to tell Emacs otherwise.
17443 Text mode among other actions, sets the syntax table to work
17444 conveniently for writers. In Text mode, Emacs considers an apostrophe
17445 as part of a word like a letter; but Emacs does not consider a period
17446 or a space as part of a word. Thus, @kbd{M-f} moves you over
17447 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17448 the @samp{t} of @samp{it's}.
17450 The second line causes Emacs to turn on Auto Fill mode when it turns
17451 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17452 that is too wide and brings the excessively wide part of the line down
17453 to the next line. Emacs breaks lines between words, not within them.
17455 When Auto Fill mode is turned off, lines continue to the right as you
17456 type them. Depending on how you set the value of
17457 @code{truncate-lines}, the words you type either disappear off the
17458 right side of the screen, or else are shown, in a rather ugly and
17459 unreadable manner, as a continuation line on the screen.
17462 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17463 fill commands to insert two spaces after a colon:
17466 (setq colon-double-space t)
17469 @node Mail Aliases, Indent Tabs Mode, Text and Auto-fill, Emacs Initialization
17470 @section Mail Aliases
17472 Here is a @code{setq} that `turns on' mail aliases, along with more
17478 ; To enter mail mode, type `C-x m'
17479 ; To enter RMAIL (for reading mail),
17481 (setq mail-aliases t)
17485 @cindex Mail aliases
17487 This @code{setq} command sets the value of the variable
17488 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17489 says, in effect, ``Yes, use mail aliases.''
17491 Mail aliases are convenient short names for long email addresses or
17492 for lists of email addresses. The file where you keep your `aliases'
17493 is @file{~/.mailrc}. You write an alias like this:
17496 alias geo george@@foobar.wiz.edu
17500 When you write a message to George, address it to @samp{geo}; the
17501 mailer will automatically expand @samp{geo} to the full address.
17503 @node Indent Tabs Mode, Keybindings, Mail Aliases, Emacs Initialization
17504 @section Indent Tabs Mode
17505 @cindex Tabs, preventing
17506 @findex indent-tabs-mode
17508 By default, Emacs inserts tabs in place of multiple spaces when it
17509 formats a region. (For example, you might indent many lines of text
17510 all at once with the @code{indent-region} command.) Tabs look fine on
17511 a terminal or with ordinary printing, but they produce badly indented
17512 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17515 The following turns off Indent Tabs mode:
17519 ;;; Prevent Extraneous Tabs
17520 (setq-default indent-tabs-mode nil)
17524 Note that this line uses @code{setq-default} rather than the
17525 @code{setq} command that we have seen before. The @code{setq-default}
17526 command sets values only in buffers that do not have their own local
17527 values for the variable.
17530 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17532 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17536 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17537 Files'' in @cite{The GNU Emacs Manual}.
17541 @node Keybindings, Keymaps, Indent Tabs Mode, Emacs Initialization
17542 @section Some Keybindings
17544 Now for some personal keybindings:
17548 ;;; Compare windows
17549 (global-set-key "\C-cw" 'compare-windows)
17553 @findex compare-windows
17554 @code{compare-windows} is a nifty command that compares the text in
17555 your current window with text in the next window. It makes the
17556 comparison by starting at point in each window, moving over text in
17557 each window as far as they match. I use this command all the time.
17559 This also shows how to set a key globally, for all modes.
17561 @cindex Setting a key globally
17562 @cindex Global set key
17563 @cindex Key setting globally
17564 @findex global-set-key
17565 The command is @code{global-set-key}. It is followed by the
17566 keybinding. In a @file{.emacs} file, the keybinding is written as
17567 shown: @code{\C-c} stands for `control-c', which means `press the
17568 control key and the @key{c} key at the same time'. The @code{w} means
17569 `press the @key{w} key'. The keybinding is surrounded by double
17570 quotation marks. In documentation, you would write this as
17571 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17572 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17573 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17574 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17577 The command invoked by the keys is @code{compare-windows}. Note that
17578 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17579 would first try to evaluate the symbol to determine its value.
17581 These three things, the double quotation marks, the backslash before
17582 the @samp{C}, and the single quote mark are necessary parts of
17583 keybinding that I tend to forget. Fortunately, I have come to
17584 remember that I should look at my existing @file{.emacs} file, and
17585 adapt what is there.
17587 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17588 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17589 set of keys, @kbd{C-c} followed by a single character, is strictly
17590 reserved for individuals' own use. (I call these `own' keys, since
17591 these are for my own use.) You should always be able to create such a
17592 keybinding for your own use without stomping on someone else's
17593 keybinding. If you ever write an extension to Emacs, please avoid
17594 taking any of these keys for public use. Create a key like @kbd{C-c
17595 C-w} instead. Otherwise, we will run out of `own' keys.
17598 Here is another keybinding, with a comment:
17602 ;;; Keybinding for `occur'
17603 ; I use occur a lot, so let's bind it to a key:
17604 (global-set-key "\C-co" 'occur)
17609 The @code{occur} command shows all the lines in the current buffer
17610 that contain a match for a regular expression. Matching lines are
17611 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17612 to jump to occurrences.
17614 @findex global-unset-key
17615 @cindex Unbinding key
17616 @cindex Key unbinding
17618 Here is how to unbind a key, so it does not
17624 (global-unset-key "\C-xf")
17628 There is a reason for this unbinding: I found I inadvertently typed
17629 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17630 file, as I intended, I accidentally set the width for filled text,
17631 almost always to a width I did not want. Since I hardly ever reset my
17632 default width, I simply unbound the key.
17634 @findex list-buffers, @r{rebound}
17635 @findex buffer-menu, @r{bound to key}
17637 The following rebinds an existing key:
17641 ;;; Rebind `C-x C-b' for `buffer-menu'
17642 (global-set-key "\C-x\C-b" 'buffer-menu)
17646 By default, @kbd{C-x C-b} runs the
17647 @code{list-buffers} command. This command lists
17648 your buffers in @emph{another} window. Since I
17649 almost always want to do something in that
17650 window, I prefer the @code{buffer-menu}
17651 command, which not only lists the buffers,
17652 but moves point into that window.
17654 @node Keymaps, Loading Files, Keybindings, Emacs Initialization
17657 @cindex Rebinding keys
17659 Emacs uses @dfn{keymaps} to record which keys call which commands.
17660 When you use @code{global-set-key} to set the keybinding for a single
17661 command in all parts of Emacs, you are specifying the keybinding in
17662 @code{current-global-map}.
17664 Specific modes, such as C mode or Text mode, have their own keymaps;
17665 the mode-specific keymaps override the global map that is shared by
17668 The @code{global-set-key} function binds, or rebinds, the global
17669 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17670 function @code{buffer-menu}:
17673 (global-set-key "\C-x\C-b" 'buffer-menu)
17676 Mode-specific keymaps are bound using the @code{define-key} function,
17677 which takes a specific keymap as an argument, as well as the key and
17678 the command. For example, my @file{.emacs} file contains the
17679 following expression to bind the @code{texinfo-insert-@@group} command
17680 to @kbd{C-c C-c g}:
17684 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17689 The @code{texinfo-insert-@@group} function itself is a little extension
17690 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17691 use this command all the time and prefer to type the three strokes
17692 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17693 (@samp{@@group} and its matching @samp{@@end group} are commands that
17694 keep all enclosed text together on one page; many multi-line examples
17695 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17698 Here is the @code{texinfo-insert-@@group} function definition:
17702 (defun texinfo-insert-@@group ()
17703 "Insert the string @@group in a Texinfo buffer."
17705 (beginning-of-line)
17706 (insert "@@group\n"))
17710 (Of course, I could have used Abbrev mode to save typing, rather than
17711 write a function to insert a word; but I prefer key strokes consistent
17712 with other Texinfo mode key bindings.)
17714 You will see numerous @code{define-key} expressions in
17715 @file{loaddefs.el} as well as in the various mode libraries, such as
17716 @file{cc-mode.el} and @file{lisp-mode.el}.
17718 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17719 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17720 Reference Manual}, for more information about keymaps.
17722 @node Loading Files, Autoload, Keymaps, Emacs Initialization
17723 @section Loading Files
17724 @cindex Loading files
17727 Many people in the GNU Emacs community have written extensions to
17728 Emacs. As time goes by, these extensions are often included in new
17729 releases. For example, the Calendar and Diary packages are now part
17730 of the standard GNU Emacs, as is Calc.
17732 You can use a @code{load} command to evaluate a complete file and
17733 thereby install all the functions and variables in the file into Emacs.
17736 @c (auto-compression-mode t)
17739 (load "~/emacs/slowsplit")
17742 This evaluates, i.e.@: loads, the @file{slowsplit.el} file or if it
17743 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17744 @file{emacs} sub-directory of your home directory. The file contains
17745 the function @code{split-window-quietly}, which John Robinson wrote in
17748 The @code{split-window-quietly} function splits a window with the
17749 minimum of redisplay. I installed it in 1989 because it worked well
17750 with the slow 1200 baud terminals I was then using. Nowadays, I only
17751 occasionally come across such a slow connection, but I continue to use
17752 the function because I like the way it leaves the bottom half of a
17753 buffer in the lower of the new windows and the top half in the upper
17757 To replace the key binding for the default
17758 @code{split-window-vertically}, you must also unset that key and bind
17759 the keys to @code{split-window-quietly}, like this:
17763 (global-unset-key "\C-x2")
17764 (global-set-key "\C-x2" 'split-window-quietly)
17769 If you load many extensions, as I do, then instead of specifying the
17770 exact location of the extension file, as shown above, you can specify
17771 that directory as part of Emacs' @code{load-path}. Then, when Emacs
17772 loads a file, it will search that directory as well as its default
17773 list of directories. (The default list is specified in @file{paths.h}
17774 when Emacs is built.)
17777 The following command adds your @file{~/emacs} directory to the
17778 existing load path:
17782 ;;; Emacs Load Path
17783 (setq load-path (cons "~/emacs" load-path))
17787 Incidentally, @code{load-library} is an interactive interface to the
17788 @code{load} function. The complete function looks like this:
17790 @findex load-library
17793 (defun load-library (library)
17794 "Load the library named LIBRARY.
17795 This is an interface to the function `load'."
17797 (list (completing-read "Load library: "
17798 (apply-partially 'locate-file-completion-table
17800 (get-load-suffixes)))))
17805 The name of the function, @code{load-library}, comes from the use of
17806 `library' as a conventional synonym for `file'. The source for the
17807 @code{load-library} command is in the @file{files.el} library.
17809 Another interactive command that does a slightly different job is
17810 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17811 Emacs, emacs, The GNU Emacs Manual}, for information on the
17812 distinction between @code{load-library} and this command.
17814 @node Autoload, Simple Extension, Loading Files, Emacs Initialization
17815 @section Autoloading
17818 Instead of installing a function by loading the file that contains it,
17819 or by evaluating the function definition, you can make the function
17820 available but not actually install it until it is first called. This
17821 is called @dfn{autoloading}.
17823 When you execute an autoloaded function, Emacs automatically evaluates
17824 the file that contains the definition, and then calls the function.
17826 Emacs starts quicker with autoloaded functions, since their libraries
17827 are not loaded right away; but you need to wait a moment when you
17828 first use such a function, while its containing file is evaluated.
17830 Rarely used functions are frequently autoloaded. The
17831 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17832 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17833 come to use a `rare' function frequently. When you do, you should
17834 load that function's file with a @code{load} expression in your
17835 @file{.emacs} file.
17837 In my @file{.emacs} file, I load 14 libraries that contain functions
17838 that would otherwise be autoloaded. (Actually, it would have been
17839 better to include these files in my `dumped' Emacs, but I forgot.
17840 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17841 Reference Manual}, and the @file{INSTALL} file for more about
17844 You may also want to include autoloaded expressions in your @file{.emacs}
17845 file. @code{autoload} is a built-in function that takes up to five
17846 arguments, the final three of which are optional. The first argument
17847 is the name of the function to be autoloaded; the second is the name
17848 of the file to be loaded. The third argument is documentation for the
17849 function, and the fourth tells whether the function can be called
17850 interactively. The fifth argument tells what type of
17851 object---@code{autoload} can handle a keymap or macro as well as a
17852 function (the default is a function).
17855 Here is a typical example:
17859 (autoload 'html-helper-mode
17860 "html-helper-mode" "Edit HTML documents" t)
17865 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17866 which is a standard part of the distribution.)
17869 This expression autoloads the @code{html-helper-mode} function. It
17870 takes it from the @file{html-helper-mode.el} file (or from the byte
17871 compiled version @file{html-helper-mode.elc}, if that exists.) The
17872 file must be located in a directory specified by @code{load-path}.
17873 The documentation says that this is a mode to help you edit documents
17874 written in the HyperText Markup Language. You can call this mode
17875 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17876 duplicate the function's regular documentation in the autoload
17877 expression because the regular function is not yet loaded, so its
17878 documentation is not available.)
17880 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17881 Manual}, for more information.
17883 @node Simple Extension, X11 Colors, Autoload, Emacs Initialization
17884 @section A Simple Extension: @code{line-to-top-of-window}
17885 @findex line-to-top-of-window
17886 @cindex Simple extension in @file{.emacs} file
17888 Here is a simple extension to Emacs that moves the line point is on to
17889 the top of the window. I use this all the time, to make text easier
17892 You can put the following code into a separate file and then load it
17893 from your @file{.emacs} file, or you can include it within your
17894 @file{.emacs} file.
17897 Here is the definition:
17901 ;;; Line to top of window;
17902 ;;; replace three keystroke sequence C-u 0 C-l
17903 (defun line-to-top-of-window ()
17904 "Move the line point is on to top of window."
17911 Now for the keybinding.
17913 Nowadays, function keys as well as mouse button events and
17914 non-@sc{ascii} characters are written within square brackets, without
17915 quotation marks. (In Emacs version 18 and before, you had to write
17916 different function key bindings for each different make of terminal.)
17918 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17922 (global-set-key [f6] 'line-to-top-of-window)
17925 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17926 Your Init File, emacs, The GNU Emacs Manual}.
17928 @cindex Conditional 'twixt two versions of Emacs
17929 @cindex Version of Emacs, choosing
17930 @cindex Emacs version, choosing
17931 If you run two versions of GNU Emacs, such as versions 22 and 23, and
17932 use one @file{.emacs} file, you can select which code to evaluate with
17933 the following conditional:
17938 ((= 22 emacs-major-version)
17939 ;; evaluate version 22 code
17941 ((= 23 emacs-major-version)
17942 ;; evaluate version 23 code
17947 For example, in contrast to version 20, more recent versions blink
17948 their cursors by default. I hate such blinking, as well as other
17949 features, so I placed the following in my @file{.emacs}
17950 file@footnote{When I start instances of Emacs that do not load my
17951 @file{.emacs} file or any site file, I also turn off blinking:
17954 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17956 @exdent Or nowadays, using an even more sophisticated set of options,
17964 (when (>= emacs-major-version 21)
17965 (blink-cursor-mode 0)
17966 ;; Insert newline when you press `C-n' (next-line)
17967 ;; at the end of the buffer
17968 (setq next-line-add-newlines t)
17971 ;; Turn on image viewing
17972 (auto-image-file-mode t)
17975 ;; Turn on menu bar (this bar has text)
17976 ;; (Use numeric argument to turn on)
17980 ;; Turn off tool bar (this bar has icons)
17981 ;; (Use numeric argument to turn on)
17982 (tool-bar-mode nil)
17985 ;; Turn off tooltip mode for tool bar
17986 ;; (This mode causes icon explanations to pop up)
17987 ;; (Use numeric argument to turn on)
17989 ;; If tooltips turned on, make tips appear promptly
17990 (setq tooltip-delay 0.1) ; default is 0.7 second
17995 @node X11 Colors, Miscellaneous, Simple Extension, Emacs Initialization
17996 @section X11 Colors
17998 You can specify colors when you use Emacs with the MIT X Windowing
18001 I dislike the default colors and specify my own.
18004 Here are the expressions in my @file{.emacs}
18005 file that set values:
18009 ;; Set cursor color
18010 (set-cursor-color "white")
18013 (set-mouse-color "white")
18015 ;; Set foreground and background
18016 (set-foreground-color "white")
18017 (set-background-color "darkblue")
18021 ;;; Set highlighting colors for isearch and drag
18022 (set-face-foreground 'highlight "white")
18023 (set-face-background 'highlight "blue")
18027 (set-face-foreground 'region "cyan")
18028 (set-face-background 'region "blue")
18032 (set-face-foreground 'secondary-selection "skyblue")
18033 (set-face-background 'secondary-selection "darkblue")
18037 ;; Set calendar highlighting colors
18038 (setq calendar-load-hook
18040 (set-face-foreground 'diary-face "skyblue")
18041 (set-face-background 'holiday-face "slate blue")
18042 (set-face-foreground 'holiday-face "white")))
18046 The various shades of blue soothe my eye and prevent me from seeing
18047 the screen flicker.
18049 Alternatively, I could have set my specifications in various X
18050 initialization files. For example, I could set the foreground,
18051 background, cursor, and pointer (i.e., mouse) colors in my
18052 @file{~/.Xresources} file like this:
18056 Emacs*foreground: white
18057 Emacs*background: darkblue
18058 Emacs*cursorColor: white
18059 Emacs*pointerColor: white
18063 In any event, since it is not part of Emacs, I set the root color of
18064 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
18065 run more modern window managers, such as Enlightenment, Gnome, or KDE;
18066 in those cases, I often specify an image rather than a plain color.}:
18069 xsetroot -solid Navy -fg white &
18073 @node Miscellaneous, Mode Line, X11 Colors, Emacs Initialization
18074 @section Miscellaneous Settings for a @file{.emacs} File
18077 Here are a few miscellaneous settings:
18082 Set the shape and color of the mouse cursor:
18086 ; Cursor shapes are defined in
18087 ; `/usr/include/X11/cursorfont.h';
18088 ; for example, the `target' cursor is number 128;
18089 ; the `top_left_arrow' cursor is number 132.
18093 (let ((mpointer (x-get-resource "*mpointer"
18094 "*emacs*mpointer")))
18095 ;; If you have not set your mouse pointer
18096 ;; then set it, otherwise leave as is:
18097 (if (eq mpointer nil)
18098 (setq mpointer "132")) ; top_left_arrow
18101 (setq x-pointer-shape (string-to-int mpointer))
18102 (set-mouse-color "white"))
18107 Or you can set the values of a variety of features in an alist, like
18113 default-frame-alist
18114 '((cursor-color . "white")
18115 (mouse-color . "white")
18116 (foreground-color . "white")
18117 (background-color . "DodgerBlue4")
18118 ;; (cursor-type . bar)
18119 (cursor-type . box)
18122 (tool-bar-lines . 0)
18123 (menu-bar-lines . 1)
18127 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
18133 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
18134 into @kbd{@key{CTRL}-h}.@*
18135 (Some older keyboards needed this, although I have not seen the
18140 ;; Translate `C-h' to <DEL>.
18141 ; (keyboard-translate ?\C-h ?\C-?)
18143 ;; Translate <DEL> to `C-h'.
18144 (keyboard-translate ?\C-? ?\C-h)
18148 @item Turn off a blinking cursor!
18152 (if (fboundp 'blink-cursor-mode)
18153 (blink-cursor-mode -1))
18158 or start GNU Emacs with the command @code{emacs -nbc}.
18161 @item When using `grep'@*
18162 @samp{-i}@w{ } Ignore case distinctions@*
18163 @samp{-n}@w{ } Prefix each line of output with line number@*
18164 @samp{-H}@w{ } Print the filename for each match.@*
18165 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18168 (setq grep-command "grep -i -nH -e ")
18172 @c Evidently, no longer needed in GNU Emacs 22
18174 item Automatically uncompress compressed files when visiting them
18177 (load "uncompress")
18182 @item Find an existing buffer, even if it has a different name@*
18183 This avoids problems with symbolic links.
18186 (setq find-file-existing-other-name t)
18189 @item Set your language environment and default input method
18193 (set-language-environment "latin-1")
18194 ;; Remember you can enable or disable multilingual text input
18195 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18196 (setq default-input-method "latin-1-prefix")
18200 If you want to write with Chinese `GB' characters, set this instead:
18204 (set-language-environment "Chinese-GB")
18205 (setq default-input-method "chinese-tonepy")
18210 @subsubheading Fixing Unpleasant Key Bindings
18211 @cindex Key bindings, fixing
18212 @cindex Bindings, key, fixing unpleasant
18214 Some systems bind keys unpleasantly. Sometimes, for example, the
18215 @key{CTRL} key appears in an awkward spot rather than at the far left
18218 Usually, when people fix these sorts of keybindings, they do not
18219 change their @file{~/.emacs} file. Instead, they bind the proper keys
18220 on their consoles with the @code{loadkeys} or @code{install-keymap}
18221 commands in their boot script and then include @code{xmodmap} commands
18222 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18230 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18232 install-keymap emacs2
18238 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18239 Lock} key is at the far left of the home row:
18243 # Bind the key labeled `Caps Lock' to `Control'
18244 # (Such a broken user interface suggests that keyboard manufacturers
18245 # think that computers are typewriters from 1885.)
18247 xmodmap -e "clear Lock"
18248 xmodmap -e "add Control = Caps_Lock"
18254 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18255 key to a @key{META} key:
18259 # Some ill designed keyboards have a key labeled ALT and no Meta
18260 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18265 @node Mode Line, , Miscellaneous, Emacs Initialization
18266 @section A Modified Mode Line
18267 @vindex mode-line-format
18268 @cindex Mode line format
18270 Finally, a feature I really like: a modified mode line.
18272 When I work over a network, I forget which machine I am using. Also,
18273 I tend to I lose track of where I am, and which line point is on.
18275 So I reset my mode line to look like this:
18278 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18281 I am visiting a file called @file{foo.texi}, on my machine
18282 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18283 Texinfo mode, and am at the top of the buffer.
18286 My @file{.emacs} file has a section that looks like this:
18290 ;; Set a Mode Line that tells me which machine, which directory,
18291 ;; and which line I am on, plus the other customary information.
18292 (setq-default mode-line-format
18296 "mouse-1: select window, mouse-2: delete others ..."))
18297 mode-line-mule-info
18299 mode-line-frame-identification
18303 mode-line-buffer-identification
18306 (system-name) 0 (string-match "\\..+" (system-name))))
18311 "mouse-1: select window, mouse-2: delete others ..."))
18312 (line-number-mode " Line %l ")
18318 "mouse-1: select window, mouse-2: delete others ..."))
18319 (:eval (mode-line-mode-name))
18322 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18331 Here, I redefine the default mode line. Most of the parts are from
18332 the original; but I make a few changes. I set the @emph{default} mode
18333 line format so as to permit various modes, such as Info, to override
18336 Many elements in the list are self-explanatory:
18337 @code{mode-line-modified} is a variable that tells whether the buffer
18338 has been modified, @code{mode-name} tells the name of the mode, and so
18339 on. However, the format looks complicated because of two features we
18340 have not discussed.
18342 @cindex Properties, in mode line example
18343 The first string in the mode line is a dash or hyphen, @samp{-}. In
18344 the old days, it would have been specified simply as @code{"-"}. But
18345 nowadays, Emacs can add properties to a string, such as highlighting
18346 or, as in this case, a help feature. If you place your mouse cursor
18347 over the hyphen, some help information appears (By default, you must
18348 wait seven-tenths of a second before the information appears. You can
18349 change that timing by changing the value of @code{tooltip-delay}.)
18352 The new string format has a special syntax:
18355 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18359 The @code{#(} begins a list. The first element of the list is the
18360 string itself, just one @samp{-}. The second and third
18361 elements specify the range over which the fourth element applies. A
18362 range starts @emph{after} a character, so a zero means the range
18363 starts just before the first character; a 1 means that the range ends
18364 just after the first character. The third element is the property for
18365 the range. It consists of a property list, a
18366 property name, in this case, @samp{help-echo}, followed by a value, in this
18367 case, a string. The second, third, and fourth elements of this new
18368 string format can be repeated.
18370 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18371 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18372 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18374 @code{mode-line-buffer-identification}
18375 displays the current buffer name. It is a list
18376 beginning @code{(#("%12b" 0 4 @dots{}}.
18377 The @code{#(} begins the list.
18379 The @samp{"%12b"} displays the current buffer name, using the
18380 @code{buffer-name} function with which we are familiar; the `12'
18381 specifies the maximum number of characters that will be displayed.
18382 When a name has fewer characters, whitespace is added to fill out to
18383 this number. (Buffer names can and often should be longer than 12
18384 characters; this length works well in a typical 80 column wide
18387 @code{:eval} says to evaluate the following form and use the result as
18388 a string to display. In this case, the expression displays the first
18389 component of the full system name. The end of the first component is
18390 a @samp{.} (`period'), so I use the @code{string-match} function to
18391 tell me the length of the first component. The substring from the
18392 zeroth character to that length is the name of the machine.
18395 This is the expression:
18400 (system-name) 0 (string-match "\\..+" (system-name))))
18404 @samp{%[} and @samp{%]} cause a pair of square brackets
18405 to appear for each recursive editing level. @samp{%n} says `Narrow'
18406 when narrowing is in effect. @samp{%P} tells you the percentage of
18407 the buffer that is above the bottom of the window, or `Top', `Bottom',
18408 or `All'. (A lower case @samp{p} tell you the percentage above the
18409 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18412 Remember, ``You don't have to like Emacs to like it'' --- your own
18413 Emacs can have different colors, different commands, and different
18414 keys than a default Emacs.
18416 On the other hand, if you want to bring up a plain `out of the box'
18417 Emacs, with no customization, type:
18424 This will start an Emacs that does @emph{not} load your
18425 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18428 @node Debugging, Conclusion, Emacs Initialization, Top
18432 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18433 first is built into the internals of Emacs and is always with you;
18434 the second requires that you instrument a function before you can use it.
18436 Both debuggers are described extensively in @ref{Debugging, ,
18437 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18438 In this chapter, I will walk through a short example of each.
18441 * debug:: How to use the built-in debugger.
18442 * debug-on-entry:: Start debugging when you call a function.
18443 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18444 * edebug:: How to use Edebug, a source level debugger.
18445 * Debugging Exercises::
18448 @node debug, debug-on-entry, Debugging, Debugging
18449 @section @code{debug}
18452 Suppose you have written a function definition that is intended to
18453 return the sum of the numbers 1 through a given number. (This is the
18454 @code{triangle} function discussed earlier. @xref{Decrementing
18455 Example, , Example with Decrementing Counter}, for a discussion.)
18456 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18458 However, your function definition has a bug. You have mistyped
18459 @samp{1=} for @samp{1-}. Here is the broken definition:
18461 @findex triangle-bugged
18464 (defun triangle-bugged (number)
18465 "Return sum of numbers 1 through NUMBER inclusive."
18467 (while (> number 0)
18468 (setq total (+ total number))
18469 (setq number (1= number))) ; @r{Error here.}
18474 If you are reading this in Info, you can evaluate this definition in
18475 the normal fashion. You will see @code{triangle-bugged} appear in the
18479 Now evaluate the @code{triangle-bugged} function with an
18483 (triangle-bugged 4)
18487 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18493 ---------- Buffer: *Backtrace* ----------
18494 Debugger entered--Lisp error: (void-function 1=)
18496 (setq number (1= number))
18497 (while (> number 0) (setq total (+ total number))
18498 (setq number (1= number)))
18499 (let ((total 0)) (while (> number 0) (setq total ...)
18500 (setq number ...)) total)
18504 eval((triangle-bugged 4))
18505 eval-last-sexp-1(nil)
18506 eval-last-sexp(nil)
18507 call-interactively(eval-last-sexp)
18508 ---------- Buffer: *Backtrace* ----------
18513 (I have reformatted this example slightly; the debugger does not fold
18514 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18515 the @file{*Backtrace*} buffer.)
18517 In practice, for a bug as simple as this, the `Lisp error' line will
18518 tell you what you need to know to correct the definition. The
18519 function @code{1=} is `void'.
18523 In GNU Emacs 20 and before, you will see:
18526 Symbol's function definition is void:@: 1=
18530 which has the same meaning as the @file{*Backtrace*} buffer line in
18534 However, suppose you are not quite certain what is going on?
18535 You can read the complete backtrace.
18537 In this case, you need to run a recent GNU Emacs, which automatically
18538 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18539 else, you need to start the debugger manually as described below.
18541 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18542 what Emacs did that led to the error. Emacs made an interactive call
18543 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18544 of the @code{triangle-bugged} expression. Each line above tells you
18545 what the Lisp interpreter evaluated next.
18548 The third line from the top of the buffer is
18551 (setq number (1= number))
18555 Emacs tried to evaluate this expression; in order to do so, it tried
18556 to evaluate the inner expression shown on the second line from the
18565 This is where the error occurred; as the top line says:
18568 Debugger entered--Lisp error: (void-function 1=)
18572 You can correct the mistake, re-evaluate the function definition, and
18573 then run your test again.
18575 @node debug-on-entry, debug-on-quit, debug, Debugging
18576 @section @code{debug-on-entry}
18577 @findex debug-on-entry
18579 A recent GNU Emacs starts the debugger automatically when your
18580 function has an error.
18583 GNU Emacs version 20 and before did not; it simply
18584 presented you with an error message. You had to start the debugger
18588 Incidentally, you can start the debugger manually for all versions of
18589 Emacs; the advantage is that the debugger runs even if you do not have
18590 a bug in your code. Sometimes your code will be free of bugs!
18592 You can enter the debugger when you call the function by calling
18593 @code{debug-on-entry}.
18600 M-x debug-on-entry RET triangle-bugged RET
18605 Now, evaluate the following:
18608 (triangle-bugged 5)
18612 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18613 you that it is beginning to evaluate the @code{triangle-bugged}
18618 ---------- Buffer: *Backtrace* ----------
18619 Debugger entered--entering a function:
18620 * triangle-bugged(5)
18621 eval((triangle-bugged 5))
18624 eval-last-sexp-1(nil)
18625 eval-last-sexp(nil)
18626 call-interactively(eval-last-sexp)
18627 ---------- Buffer: *Backtrace* ----------
18631 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18632 the first expression in @code{triangle-bugged}; the buffer will look
18637 ---------- Buffer: *Backtrace* ----------
18638 Debugger entered--beginning evaluation of function call form:
18639 * (let ((total 0)) (while (> number 0) (setq total ...)
18640 (setq number ...)) total)
18641 * triangle-bugged(5)
18642 eval((triangle-bugged 5))
18645 eval-last-sexp-1(nil)
18646 eval-last-sexp(nil)
18647 call-interactively(eval-last-sexp)
18648 ---------- Buffer: *Backtrace* ----------
18653 Now, type @kbd{d} again, eight times, slowly. Each time you type
18654 @kbd{d}, Emacs will evaluate another expression in the function
18658 Eventually, the buffer will look like this:
18662 ---------- Buffer: *Backtrace* ----------
18663 Debugger entered--beginning evaluation of function call form:
18664 * (setq number (1= number))
18665 * (while (> number 0) (setq total (+ total number))
18666 (setq number (1= number)))
18669 * (let ((total 0)) (while (> number 0) (setq total ...)
18670 (setq number ...)) total)
18671 * triangle-bugged(5)
18672 eval((triangle-bugged 5))
18675 eval-last-sexp-1(nil)
18676 eval-last-sexp(nil)
18677 call-interactively(eval-last-sexp)
18678 ---------- Buffer: *Backtrace* ----------
18684 Finally, after you type @kbd{d} two more times, Emacs will reach the
18685 error, and the top two lines of the @file{*Backtrace*} buffer will look
18690 ---------- Buffer: *Backtrace* ----------
18691 Debugger entered--Lisp error: (void-function 1=)
18694 ---------- Buffer: *Backtrace* ----------
18698 By typing @kbd{d}, you were able to step through the function.
18700 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18701 quits the trace, but does not cancel @code{debug-on-entry}.
18703 @findex cancel-debug-on-entry
18704 To cancel the effect of @code{debug-on-entry}, call
18705 @code{cancel-debug-on-entry} and the name of the function, like this:
18708 M-x cancel-debug-on-entry RET triangle-bugged RET
18712 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18714 @node debug-on-quit, edebug, debug-on-entry, Debugging
18715 @section @code{debug-on-quit} and @code{(debug)}
18717 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18718 there are two other ways to start @code{debug}.
18720 @findex debug-on-quit
18721 You can start @code{debug} whenever you type @kbd{C-g}
18722 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18723 @code{t}. This is useful for debugging infinite loops.
18726 @cindex @code{(debug)} in code
18727 Or, you can insert a line that says @code{(debug)} into your code
18728 where you want the debugger to start, like this:
18732 (defun triangle-bugged (number)
18733 "Return sum of numbers 1 through NUMBER inclusive."
18735 (while (> number 0)
18736 (setq total (+ total number))
18737 (debug) ; @r{Start debugger.}
18738 (setq number (1= number))) ; @r{Error here.}
18743 The @code{debug} function is described in detail in @ref{Debugger, ,
18744 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18746 @node edebug, Debugging Exercises, debug-on-quit, Debugging
18747 @section The @code{edebug} Source Level Debugger
18748 @cindex Source level debugger
18751 Edebug is a source level debugger. Edebug normally displays the
18752 source of the code you are debugging, with an arrow at the left that
18753 shows which line you are currently executing.
18755 You can walk through the execution of a function, line by line, or run
18756 quickly until reaching a @dfn{breakpoint} where execution stops.
18758 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18759 Lisp Reference Manual}.
18762 Here is a bugged function definition for @code{triangle-recursively}.
18763 @xref{Recursive triangle function, , Recursion in place of a counter},
18764 for a review of it.
18768 (defun triangle-recursively-bugged (number)
18769 "Return sum of numbers 1 through NUMBER inclusive.
18774 (triangle-recursively-bugged
18775 (1= number))))) ; @r{Error here.}
18780 Normally, you would install this definition by positioning your cursor
18781 after the function's closing parenthesis and typing @kbd{C-x C-e}
18782 (@code{eval-last-sexp}) or else by positioning your cursor within the
18783 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18784 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18788 However, to prepare this function definition for Edebug, you must
18789 first @dfn{instrument} the code using a different command. You can do
18790 this by positioning your cursor within or just after the definition
18794 M-x edebug-defun RET
18798 This will cause Emacs to load Edebug automatically if it is not
18799 already loaded, and properly instrument the function.
18801 After instrumenting the function, place your cursor after the
18802 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18805 (triangle-recursively-bugged 3)
18809 You will be jumped back to the source for
18810 @code{triangle-recursively-bugged} and the cursor positioned at the
18811 beginning of the @code{if} line of the function. Also, you will see
18812 an arrowhead at the left hand side of that line. The arrowhead marks
18813 the line where the function is executing. (In the following examples,
18814 we show the arrowhead with @samp{=>}; in a windowing system, you may
18815 see the arrowhead as a solid triangle in the window `fringe'.)
18818 =>@point{}(if (= number 1)
18823 In the example, the location of point is displayed with a star,
18824 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18827 In the example, the location of point is displayed as @samp{@point{}}
18828 (in a printed book, it is displayed with a five pointed star).
18831 If you now press @key{SPC}, point will move to the next expression to
18832 be executed; the line will look like this:
18835 =>(if @point{}(= number 1)
18839 As you continue to press @key{SPC}, point will move from expression to
18840 expression. At the same time, whenever an expression returns a value,
18841 that value will be displayed in the echo area. For example, after you
18842 move point past @code{number}, you will see the following:
18845 Result: 3 (#o3, #x3, ?\C-c)
18849 This means the value of @code{number} is 3, which is octal three,
18850 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18851 alphabet, in case you need to know this information).
18853 You can continue moving through the code until you reach the line with
18854 the error. Before evaluation, that line looks like this:
18857 => @point{}(1= number))))) ; @r{Error here.}
18862 When you press @key{SPC} once again, you will produce an error message
18866 Symbol's function definition is void:@: 1=
18872 Press @kbd{q} to quit Edebug.
18874 To remove instrumentation from a function definition, simply
18875 re-evaluate it with a command that does not instrument it.
18876 For example, you could place your cursor after the definition's
18877 closing parenthesis and type @kbd{C-x C-e}.
18879 Edebug does a great deal more than walk with you through a function.
18880 You can set it so it races through on its own, stopping only at an
18881 error or at specified stopping points; you can cause it to display the
18882 changing values of various expressions; you can find out how many
18883 times a function is called, and more.
18885 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18886 Lisp Reference Manual}.
18889 @node Debugging Exercises, , edebug, Debugging
18890 @section Debugging Exercises
18894 Install the @code{@value{COUNT-WORDS}} function and then cause it to
18895 enter the built-in debugger when you call it. Run the command on a
18896 region containing two words. You will need to press @kbd{d} a
18897 remarkable number of times. On your system, is a `hook' called after
18898 the command finishes? (For information on hooks, see @ref{Command
18899 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18903 Copy @code{@value{COUNT-WORDS}} into the @file{*scratch*} buffer,
18904 instrument the function for Edebug, and walk through its execution.
18905 The function does not need to have a bug, although you can introduce
18906 one if you wish. If the function lacks a bug, the walk-through
18907 completes without problems.
18910 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18911 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.@:
18912 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18913 for commands made outside of the Edebug debugging buffer.)
18916 In the Edebug debugging buffer, use the @kbd{p}
18917 (@code{edebug-bounce-point}) command to see where in the region the
18918 @code{@value{COUNT-WORDS}} is working.
18921 Move point to some spot further down the function and then type the
18922 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18925 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18926 walk through the function on its own; use an upper case @kbd{T} for
18927 @code{edebug-Trace-fast-mode}.
18930 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18934 @node Conclusion, the-the, Debugging, Top
18935 @chapter Conclusion
18937 We have now reached the end of this Introduction. You have now
18938 learned enough about programming in Emacs Lisp to set values, to write
18939 simple @file{.emacs} files for yourself and your friends, and write
18940 simple customizations and extensions to Emacs.
18942 This is a place to stop. Or, if you wish, you can now go onward, and
18945 You have learned some of the basic nuts and bolts of programming. But
18946 only some. There are a great many more brackets and hinges that are
18947 easy to use that we have not touched.
18949 A path you can follow right now lies among the sources to GNU Emacs
18952 @cite{The GNU Emacs Lisp Reference Manual}.
18955 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18956 Emacs Lisp Reference Manual}.
18959 The Emacs Lisp sources are an adventure. When you read the sources and
18960 come across a function or expression that is unfamiliar, you need to
18961 figure out or find out what it does.
18963 Go to the Reference Manual. It is a thorough, complete, and fairly
18964 easy-to-read description of Emacs Lisp. It is written not only for
18965 experts, but for people who know what you know. (The @cite{Reference
18966 Manual} comes with the standard GNU Emacs distribution. Like this
18967 introduction, it comes as a Texinfo source file, so you can read it
18968 on-line and as a typeset, printed book.)
18970 Go to the other on-line help that is part of GNU Emacs: the on-line
18971 documentation for all functions and variables, and @code{find-tag},
18972 the program that takes you to sources.
18974 Here is an example of how I explore the sources. Because of its name,
18975 @file{simple.el} is the file I looked at first, a long time ago. As
18976 it happens some of the functions in @file{simple.el} are complicated,
18977 or at least look complicated at first sight. The @code{open-line}
18978 function, for example, looks complicated.
18980 You may want to walk through this function slowly, as we did with the
18981 @code{forward-sentence} function. (@xref{forward-sentence, The
18982 @code{forward-sentence} function}.) Or you may want to skip that
18983 function and look at another, such as @code{split-line}. You don't
18984 need to read all the functions. According to
18985 @code{count-words-in-defun}, the @code{split-line} function contains
18986 102 words and symbols.
18988 Even though it is short, @code{split-line} contains expressions
18989 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18990 @code{current-column} and @code{insert-and-inherit}.
18992 Consider the @code{skip-chars-forward} function. (It is part of the
18993 function definition for @code{back-to-indentation}, which is shown in
18994 @ref{Review, , Review}.)
18996 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18997 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18998 function. This gives you the function documentation.
19000 You may be able to guess what is done by a well named function such as
19001 @code{indent-to}; or you can look it up, too. Incidentally, the
19002 @code{describe-function} function itself is in @file{help.el}; it is
19003 one of those long, but decipherable functions. You can look up
19004 @code{describe-function} using the @kbd{C-h f} command!
19006 In this instance, since the code is Lisp, the @file{*Help*} buffer
19007 contains the name of the library containing the function's source.
19008 You can put point over the name of the library and press the RET key,
19009 which in this situation is bound to @code{help-follow}, and be taken
19010 directly to the source, in the same way as @kbd{M-.}
19013 The definition for @code{describe-function} illustrates how to
19014 customize the @code{interactive} expression without using the standard
19015 character codes; and it shows how to create a temporary buffer.
19017 (The @code{indent-to} function is written in C rather than Emacs Lisp;
19018 it is a `built-in' function. @code{help-follow} takes you to its
19019 source as does @code{find-tag}, when properly set up.)
19021 You can look at a function's source using @code{find-tag}, which is
19022 bound to @kbd{M-.} Finally, you can find out what the Reference
19023 Manual has to say by visiting the manual in Info, and typing @kbd{i}
19024 (@code{Info-index}) and the name of the function, or by looking up the
19025 function in the index to a printed copy of the manual.
19027 Similarly, you can find out what is meant by
19028 @code{insert-and-inherit}.
19030 Other interesting source files include @file{paragraphs.el},
19031 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
19032 file includes short, easily understood functions as well as longer
19033 ones. The @file{loaddefs.el} file contains the many standard
19034 autoloads and many keymaps. I have never looked at it all; only at
19035 parts. @file{loadup.el} is the file that loads the standard parts of
19036 Emacs; it tells you a great deal about how Emacs is built.
19037 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
19038 Reference Manual}, for more about building.)
19040 As I said, you have learned some nuts and bolts; however, and very
19041 importantly, we have hardly touched major aspects of programming; I
19042 have said nothing about how to sort information, except to use the
19043 predefined @code{sort} function; I have said nothing about how to store
19044 information, except to use variables and lists; I have said nothing
19045 about how to write programs that write programs. These are topics for
19046 another, and different kind of book, a different kind of learning.
19048 What you have done is learn enough for much practical work with GNU
19049 Emacs. What you have done is get started. This is the end of a
19052 @c ================ Appendix ================
19054 @node the-the, Kill Ring, Conclusion, Top
19055 @appendix The @code{the-the} Function
19057 @cindex Duplicated words function
19058 @cindex Words, duplicated
19060 Sometimes when you you write text, you duplicate words---as with ``you
19061 you'' near the beginning of this sentence. I find that most
19062 frequently, I duplicate ``the''; hence, I call the function for
19063 detecting duplicated words, @code{the-the}.
19066 As a first step, you could use the following regular expression to
19067 search for duplicates:
19070 \\(\\w+[ \t\n]+\\)\\1
19074 This regexp matches one or more word-constituent characters followed
19075 by one or more spaces, tabs, or newlines. However, it does not detect
19076 duplicated words on different lines, since the ending of the first
19077 word, the end of the line, is different from the ending of the second
19078 word, a space. (For more information about regular expressions, see
19079 @ref{Regexp Search, , Regular Expression Searches}, as well as
19080 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
19081 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
19082 The GNU Emacs Lisp Reference Manual}.)
19084 You might try searching just for duplicated word-constituent
19085 characters but that does not work since the pattern detects doubles
19086 such as the two occurrences of `th' in `with the'.
19088 Another possible regexp searches for word-constituent characters
19089 followed by non-word-constituent characters, reduplicated. Here,
19090 @w{@samp{\\w+}} matches one or more word-constituent characters and
19091 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
19094 \\(\\(\\w+\\)\\W*\\)\\1
19100 Here is the pattern that I use. It is not perfect, but good enough.
19101 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
19102 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
19103 any characters that are @emph{not} an @@-sign, space, newline, or tab.
19106 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
19109 One can write more complicated expressions, but I found that this
19110 expression is good enough, so I use it.
19112 Here is the @code{the-the} function, as I include it in my
19113 @file{.emacs} file, along with a handy global key binding:
19118 "Search forward for for a duplicated word."
19120 (message "Searching for for duplicated words ...")
19124 ;; This regexp is not perfect
19125 ;; but is fairly good over all:
19126 (if (re-search-forward
19127 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
19128 (message "Found duplicated word.")
19129 (message "End of buffer")))
19133 ;; Bind `the-the' to C-c \
19134 (global-set-key "\C-c\\" 'the-the)
19143 one two two three four five
19148 You can substitute the other regular expressions shown above in the
19149 function definition and try each of them on this list.
19151 @node Kill Ring, Full Graph, the-the, Top
19152 @appendix Handling the Kill Ring
19153 @cindex Kill ring handling
19154 @cindex Handling the kill ring
19155 @cindex Ring, making a list like a
19157 The kill ring is a list that is transformed into a ring by the
19158 workings of the @code{current-kill} function. The @code{yank} and
19159 @code{yank-pop} commands use the @code{current-kill} function.
19161 This appendix describes the @code{current-kill} function as well as
19162 both the @code{yank} and the @code{yank-pop} commands, but first,
19163 consider the workings of the kill ring.
19166 * What the Kill Ring Does::
19168 * yank:: Paste a copy of a clipped element.
19169 * yank-pop:: Insert element pointed to.
19173 @node What the Kill Ring Does, current-kill, Kill Ring, Kill Ring
19175 @unnumberedsec What the Kill Ring Does
19179 The kill ring has a default maximum length of sixty items; this number
19180 is too large for an explanation. Instead, set it to four. Please
19181 evaluate the following:
19185 (setq old-kill-ring-max kill-ring-max)
19186 (setq kill-ring-max 4)
19191 Then, please copy each line of the following indented example into the
19192 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19196 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19197 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19198 merely copy it to the kill ring. However, your machine may beep at
19199 you. Alternatively, for silence, you may copy the region of each line
19200 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19201 each line for this command to succeed, but it does not matter at which
19202 end you put point or mark.)
19206 Please invoke the calls in order, so that five elements attempt to
19207 fill the kill ring:
19212 second piece of text
19214 fourth line of text
19221 Then find the value of @code{kill-ring} by evaluating
19233 ("fifth bit of text" "fourth line of text"
19234 "third line" "second piece of text")
19239 The first element, @samp{first some text}, was dropped.
19242 To return to the old value for the length of the kill ring, evaluate:
19245 (setq kill-ring-max old-kill-ring-max)
19248 @node current-kill, yank, What the Kill Ring Does, Kill Ring
19249 @comment node-name, next, previous, up
19250 @appendixsec The @code{current-kill} Function
19251 @findex current-kill
19253 The @code{current-kill} function changes the element in the kill ring
19254 to which @code{kill-ring-yank-pointer} points. (Also, the
19255 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19256 to the latest element of the kill ring. The @code{kill-new}
19257 function is used directly or indirectly by @code{kill-append},
19258 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19259 and @code{kill-region}.)
19262 * Code for current-kill::
19263 * Understanding current-kill::
19266 @node Code for current-kill, Understanding current-kill, current-kill, current-kill
19268 @unnumberedsubsec The code for @code{current-kill}
19273 The @code{current-kill} function is used by @code{yank} and by
19274 @code{yank-pop}. Here is the code for @code{current-kill}:
19278 (defun current-kill (n &optional do-not-move)
19279 "Rotate the yanking point by N places, and then return that kill.
19280 If N is zero, `interprogram-paste-function' is set, and calling it
19281 returns a string, then that string is added to the front of the
19282 kill ring and returned as the latest kill.
19285 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19286 yanking point; just return the Nth kill forward."
19287 (let ((interprogram-paste (and (= n 0)
19288 interprogram-paste-function
19289 (funcall interprogram-paste-function))))
19292 (if interprogram-paste
19294 ;; Disable the interprogram cut function when we add the new
19295 ;; text to the kill ring, so Emacs doesn't try to own the
19296 ;; selection, with identical text.
19297 (let ((interprogram-cut-function nil))
19298 (kill-new interprogram-paste))
19299 interprogram-paste)
19302 (or kill-ring (error "Kill ring is empty"))
19303 (let ((ARGth-kill-element
19304 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19305 (length kill-ring))
19308 (setq kill-ring-yank-pointer ARGth-kill-element))
19309 (car ARGth-kill-element)))))
19313 Remember also that the @code{kill-new} function sets
19314 @code{kill-ring-yank-pointer} to the latest element of the kill
19315 ring, which means that all the functions that call it set the value
19316 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19317 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19320 Here is the line in @code{kill-new}, which is explained in
19321 @ref{kill-new function, , The @code{kill-new} function}.
19324 (setq kill-ring-yank-pointer kill-ring)
19327 @node Understanding current-kill, , Code for current-kill, current-kill
19329 @unnumberedsubsec @code{current-kill} in Outline
19332 The @code{current-kill} function looks complex, but as usual, it can
19333 be understood by taking it apart piece by piece. First look at it in
19338 (defun current-kill (n &optional do-not-move)
19339 "Rotate the yanking point by N places, and then return that kill."
19345 This function takes two arguments, one of which is optional. It has a
19346 documentation string. It is @emph{not} interactive.
19349 * Body of current-kill::
19350 * Digression concerning error:: How to mislead humans, but not computers.
19351 * Determining the Element::
19354 @node Body of current-kill, Digression concerning error, Understanding current-kill, Understanding current-kill
19356 @unnumberedsubsubsec The Body of @code{current-kill}
19359 The body of the function definition is a @code{let} expression, which
19360 itself has a body as well as a @var{varlist}.
19362 The @code{let} expression declares a variable that will be only usable
19363 within the bounds of this function. This variable is called
19364 @code{interprogram-paste} and is for copying to another program. It
19365 is not for copying within this instance of GNU Emacs. Most window
19366 systems provide a facility for interprogram pasting. Sadly, that
19367 facility usually provides only for the last element. Most windowing
19368 systems have not adopted a ring of many possibilities, even though
19369 Emacs has provided it for decades.
19371 The @code{if} expression has two parts, one if there exists
19372 @code{interprogram-paste} and one if not.
19375 Let us consider the `if not' or else-part of the @code{current-kill}
19376 function. (The then-part uses the @code{kill-new} function, which
19377 we have already described. @xref{kill-new function, , The
19378 @code{kill-new} function}.)
19382 (or kill-ring (error "Kill ring is empty"))
19383 (let ((ARGth-kill-element
19384 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19385 (length kill-ring))
19388 (setq kill-ring-yank-pointer ARGth-kill-element))
19389 (car ARGth-kill-element))
19394 The code first checks whether the kill ring has content; otherwise it
19398 Note that the @code{or} expression is very similar to testing length
19405 (if (zerop (length kill-ring)) ; @r{if-part}
19406 (error "Kill ring is empty")) ; @r{then-part}
19412 If there is not anything in the kill ring, its length must be zero and
19413 an error message sent to the user: @samp{Kill ring is empty}. The
19414 @code{current-kill} function uses an @code{or} expression which is
19415 simpler. But an @code{if} expression reminds us what goes on.
19417 This @code{if} expression uses the function @code{zerop} which returns
19418 true if the value it is testing is zero. When @code{zerop} tests
19419 true, the then-part of the @code{if} is evaluated. The then-part is a
19420 list starting with the function @code{error}, which is a function that
19421 is similar to the @code{message} function
19422 (@pxref{message, , The @code{message} Function}) in that
19423 it prints a one-line message in the echo area. However, in addition
19424 to printing a message, @code{error} also stops evaluation of the
19425 function within which it is embedded. This means that the rest of the
19426 function will not be evaluated if the length of the kill ring is zero.
19428 Then the @code{current-kill} function selects the element to return.
19429 The selection depends on the number of places that @code{current-kill}
19430 rotates and on where @code{kill-ring-yank-pointer} points.
19432 Next, either the optional @code{do-not-move} argument is true or the
19433 current value of @code{kill-ring-yank-pointer} is set to point to the
19434 list. Finally, another expression returns the first element of the
19435 list even if the @code{do-not-move} argument is true.
19437 @node Digression concerning error, Determining the Element, Body of current-kill, Understanding current-kill
19439 @unnumberedsubsubsec Digression about the word `error'
19442 In my opinion, it is slightly misleading, at least to humans, to use
19443 the term `error' as the name of the @code{error} function. A better
19444 term would be `cancel'. Strictly speaking, of course, you cannot
19445 point to, much less rotate a pointer to a list that has no length, so
19446 from the point of view of the computer, the word `error' is correct.
19447 But a human expects to attempt this sort of thing, if only to find out
19448 whether the kill ring is full or empty. This is an act of
19451 From the human point of view, the act of exploration and discovery is
19452 not necessarily an error, and therefore should not be labelled as one,
19453 even in the bowels of a computer. As it is, the code in Emacs implies
19454 that a human who is acting virtuously, by exploring his or her
19455 environment, is making an error. This is bad. Even though the computer
19456 takes the same steps as it does when there is an `error', a term such as
19457 `cancel' would have a clearer connotation.
19459 @node Determining the Element, , Digression concerning error, Understanding current-kill
19461 @unnumberedsubsubsec Determining the Element
19464 Among other actions, the else-part of the @code{if} expression sets
19465 the value of @code{kill-ring-yank-pointer} to
19466 @code{ARGth-kill-element} when the kill ring has something in it and
19467 the value of @code{do-not-move} is @code{nil}.
19470 The code looks like this:
19474 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19475 (length kill-ring))
19480 This needs some examination. Unless it is not supposed to move the
19481 pointer, the @code{current-kill} function changes where
19482 @code{kill-ring-yank-pointer} points.
19484 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19485 expression does. Also, clearly, @code{ARGth-kill-element} is being
19486 set to be equal to some @sc{cdr} of the kill ring, using the
19487 @code{nthcdr} function that is described in an earlier section.
19488 (@xref{copy-region-as-kill}.) How does it do this?
19490 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19491 works by repeatedly taking the @sc{cdr} of a list---it takes the
19492 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19495 The two following expressions produce the same result:
19499 (setq kill-ring-yank-pointer (cdr kill-ring))
19501 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19505 However, the @code{nthcdr} expression is more complicated. It uses
19506 the @code{mod} function to determine which @sc{cdr} to select.
19508 (You will remember to look at inner functions first; indeed, we will
19509 have to go inside the @code{mod}.)
19511 The @code{mod} function returns the value of its first argument modulo
19512 the second; that is to say, it returns the remainder after dividing
19513 the first argument by the second. The value returned has the same
19514 sign as the second argument.
19522 @result{} 0 ;; @r{because there is no remainder}
19529 In this case, the first argument is often smaller than the second.
19541 We can guess what the @code{-} function does. It is like @code{+} but
19542 subtracts instead of adds; the @code{-} function subtracts its second
19543 argument from its first. Also, we already know what the @code{length}
19544 function does (@pxref{length}). It returns the length of a list.
19546 And @code{n} is the name of the required argument to the
19547 @code{current-kill} function.
19550 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19551 expression returns the whole list, as you can see by evaluating the
19556 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19557 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19558 (nthcdr (mod (- 0 4) 4)
19559 '("fourth line of text"
19561 "second piece of text"
19562 "first some text"))
19567 When the first argument to the @code{current-kill} function is one,
19568 the @code{nthcdr} expression returns the list without its first
19573 (nthcdr (mod (- 1 4) 4)
19574 '("fourth line of text"
19576 "second piece of text"
19577 "first some text"))
19581 @cindex @samp{global variable} defined
19582 @cindex @samp{variable, global}, defined
19583 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19584 are @dfn{global variables}. That means that any expression in Emacs
19585 Lisp can access them. They are not like the local variables set by
19586 @code{let} or like the symbols in an argument list.
19587 Local variables can only be accessed
19588 within the @code{let} that defines them or the function that specifies
19589 them in an argument list (and within expressions called by them).
19592 @c texi2dvi fails when the name of the section is within ifnottex ...
19593 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19594 @ref{defun, , The @code{defun} Special Form}.)
19597 @node yank, yank-pop, current-kill, Kill Ring
19598 @comment node-name, next, previous, up
19599 @appendixsec @code{yank}
19602 After learning about @code{current-kill}, the code for the
19603 @code{yank} function is almost easy.
19605 The @code{yank} function does not use the
19606 @code{kill-ring-yank-pointer} variable directly. It calls
19607 @code{insert-for-yank} which calls @code{current-kill} which sets the
19608 @code{kill-ring-yank-pointer} variable.
19611 The code looks like this:
19616 (defun yank (&optional arg)
19617 "Reinsert (\"paste\") the last stretch of killed text.
19618 More precisely, reinsert the stretch of killed text most recently
19619 killed OR yanked. Put point at end, and set mark at beginning.
19620 With just \\[universal-argument] as argument, same but put point at
19621 beginning (and mark at end). With argument N, reinsert the Nth most
19622 recently killed stretch of killed text.
19624 When this command inserts killed text into the buffer, it honors
19625 `yank-excluded-properties' and `yank-handler' as described in the
19626 doc string for `insert-for-yank-1', which see.
19628 See also the command \\[yank-pop]."
19632 (setq yank-window-start (window-start))
19633 ;; If we don't get all the way thru, make last-command indicate that
19634 ;; for the following command.
19635 (setq this-command t)
19636 (push-mark (point))
19639 (insert-for-yank (current-kill (cond
19644 ;; This is like exchange-point-and-mark,
19645 ;; but doesn't activate the mark.
19646 ;; It is cleaner to avoid activation, even though the command
19647 ;; loop would deactivate the mark because we inserted text.
19648 (goto-char (prog1 (mark t)
19649 (set-marker (mark-marker) (point) (current-buffer)))))
19652 ;; If we do get all the way thru, make this-command indicate that.
19653 (if (eq this-command t)
19654 (setq this-command 'yank))
19659 The key expression is @code{insert-for-yank}, which inserts the string
19660 returned by @code{current-kill}, but removes some text properties from
19663 However, before getting to that expression, the function sets the value
19664 of @code{yank-window-start} to the position returned by the
19665 @code{(window-start)} expression, the position at which the display
19666 currently starts. The @code{yank} function also sets
19667 @code{this-command} and pushes the mark.
19669 After it yanks the appropriate element, if the optional argument is a
19670 @sc{cons} rather than a number or nothing, it puts point at beginning
19671 of the yanked text and mark at its end.
19673 (The @code{prog1} function is like @code{progn} but returns the value
19674 of its first argument rather than the value of its last argument. Its
19675 first argument is forced to return the buffer's mark as an integer.
19676 You can see the documentation for these functions by placing point
19677 over them in this buffer and then typing @kbd{C-h f}
19678 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19681 The last part of the function tells what to do when it succeeds.
19683 @node yank-pop, ring file, yank, Kill Ring
19684 @comment node-name, next, previous, up
19685 @appendixsec @code{yank-pop}
19688 After understanding @code{yank} and @code{current-kill}, you know how
19689 to approach the @code{yank-pop} function. Leaving out the
19690 documentation to save space, it looks like this:
19695 (defun yank-pop (&optional arg)
19698 (if (not (eq last-command 'yank))
19699 (error "Previous command was not a yank"))
19702 (setq this-command 'yank)
19703 (unless arg (setq arg 1))
19704 (let ((inhibit-read-only t)
19705 (before (< (point) (mark t))))
19709 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19710 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19711 (setq yank-undo-function nil)
19714 (set-marker (mark-marker) (point) (current-buffer))
19715 (insert-for-yank (current-kill arg))
19716 ;; Set the window start back where it was in the yank command,
19718 (set-window-start (selected-window) yank-window-start t)
19722 ;; This is like exchange-point-and-mark,
19723 ;; but doesn't activate the mark.
19724 ;; It is cleaner to avoid activation, even though the command
19725 ;; loop would deactivate the mark because we inserted text.
19726 (goto-char (prog1 (mark t)
19727 (set-marker (mark-marker)
19729 (current-buffer))))))
19734 The function is interactive with a small @samp{p} so the prefix
19735 argument is processed and passed to the function. The command can
19736 only be used after a previous yank; otherwise an error message is
19737 sent. This check uses the variable @code{last-command} which is set
19738 by @code{yank} and is discussed elsewhere.
19739 (@xref{copy-region-as-kill}.)
19741 The @code{let} clause sets the variable @code{before} to true or false
19742 depending whether point is before or after mark and then the region
19743 between point and mark is deleted. This is the region that was just
19744 inserted by the previous yank and it is this text that will be
19747 @code{funcall} calls its first argument as a function, passing
19748 remaining arguments to it. The first argument is whatever the
19749 @code{or} expression returns. The two remaining arguments are the
19750 positions of point and mark set by the preceding @code{yank} command.
19752 There is more, but that is the hardest part.
19754 @node ring file, , yank-pop, Kill Ring
19755 @comment node-name, next, previous, up
19756 @appendixsec The @file{ring.el} File
19757 @cindex @file{ring.el} file
19759 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19760 provides many of the features we just discussed. But functions such
19761 as @code{kill-ring-yank-pointer} do not use this library, possibly
19762 because they were written earlier.
19764 @node Full Graph, Free Software and Free Manuals, Kill Ring, Top
19765 @appendix A Graph with Labelled Axes
19767 Printed axes help you understand a graph. They convey scale. In an
19768 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19769 wrote the code to print the body of a graph. Here we write the code
19770 for printing and labelling vertical and horizontal axes, along with the
19774 * Labelled Example::
19775 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19776 * print-Y-axis:: Print a label for the vertical axis.
19777 * print-X-axis:: Print a horizontal label.
19778 * Print Whole Graph:: The function to print a complete graph.
19781 @node Labelled Example, print-graph Varlist, Full Graph, Full Graph
19783 @unnumberedsec Labelled Example Graph
19786 Since insertions fill a buffer to the right and below point, the new
19787 graph printing function should first print the Y or vertical axis,
19788 then the body of the graph, and finally the X or horizontal axis.
19789 This sequence lays out for us the contents of the function:
19799 Print body of graph.
19806 Here is an example of how a finished graph should look:
19819 1 - ****************
19826 In this graph, both the vertical and the horizontal axes are labelled
19827 with numbers. However, in some graphs, the horizontal axis is time
19828 and would be better labelled with months, like this:
19842 Indeed, with a little thought, we can easily come up with a variety of
19843 vertical and horizontal labelling schemes. Our task could become
19844 complicated. But complications breed confusion. Rather than permit
19845 this, it is better choose a simple labelling scheme for our first
19846 effort, and to modify or replace it later.
19849 These considerations suggest the following outline for the
19850 @code{print-graph} function:
19854 (defun print-graph (numbers-list)
19855 "@var{documentation}@dots{}"
19856 (let ((height @dots{}
19860 (print-Y-axis height @dots{} )
19861 (graph-body-print numbers-list)
19862 (print-X-axis @dots{} )))
19866 We can work on each part of the @code{print-graph} function definition
19869 @node print-graph Varlist, print-Y-axis, Labelled Example, Full Graph
19870 @comment node-name, next, previous, up
19871 @appendixsec The @code{print-graph} Varlist
19872 @cindex @code{print-graph} varlist
19874 In writing the @code{print-graph} function, the first task is to write
19875 the varlist in the @code{let} expression. (We will leave aside for the
19876 moment any thoughts about making the function interactive or about the
19877 contents of its documentation string.)
19879 The varlist should set several values. Clearly, the top of the label
19880 for the vertical axis must be at least the height of the graph, which
19881 means that we must obtain this information here. Note that the
19882 @code{print-graph-body} function also requires this information. There
19883 is no reason to calculate the height of the graph in two different
19884 places, so we should change @code{print-graph-body} from the way we
19885 defined it earlier to take advantage of the calculation.
19887 Similarly, both the function for printing the X axis labels and the
19888 @code{print-graph-body} function need to learn the value of the width of
19889 each symbol. We can perform the calculation here and change the
19890 definition for @code{print-graph-body} from the way we defined it in the
19893 The length of the label for the horizontal axis must be at least as long
19894 as the graph. However, this information is used only in the function
19895 that prints the horizontal axis, so it does not need to be calculated here.
19897 These thoughts lead us directly to the following form for the varlist
19898 in the @code{let} for @code{print-graph}:
19902 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19903 (symbol-width (length graph-blank)))
19908 As we shall see, this expression is not quite right.
19911 @node print-Y-axis, print-X-axis, print-graph Varlist, Full Graph
19912 @comment node-name, next, previous, up
19913 @appendixsec The @code{print-Y-axis} Function
19914 @cindex Axis, print vertical
19915 @cindex Y axis printing
19916 @cindex Vertical axis printing
19917 @cindex Print vertical axis
19919 The job of the @code{print-Y-axis} function is to print a label for
19920 the vertical axis that looks like this:
19938 The function should be passed the height of the graph, and then should
19939 construct and insert the appropriate numbers and marks.
19942 * print-Y-axis in Detail::
19943 * Height of label:: What height for the Y axis?
19944 * Compute a Remainder:: How to compute the remainder of a division.
19945 * Y Axis Element:: Construct a line for the Y axis.
19946 * Y-axis-column:: Generate a list of Y axis labels.
19947 * print-Y-axis Penultimate:: A not quite final version.
19950 @node print-Y-axis in Detail, Height of label, print-Y-axis, print-Y-axis
19952 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19955 It is easy enough to see in the figure what the Y axis label should
19956 look like; but to say in words, and then to write a function
19957 definition to do the job is another matter. It is not quite true to
19958 say that we want a number and a tic every five lines: there are only
19959 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19960 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19961 and 9). It is better to say that we want a number and a tic mark on
19962 the base line (number 1) and then that we want a number and a tic on
19963 the fifth line from the bottom and on every line that is a multiple of
19966 @node Height of label, Compute a Remainder, print-Y-axis in Detail, print-Y-axis
19968 @unnumberedsubsec What height should the label be?
19971 The next issue is what height the label should be? Suppose the maximum
19972 height of tallest column of the graph is seven. Should the highest
19973 label on the Y axis be @samp{5 -}, and should the graph stick up above
19974 the label? Or should the highest label be @samp{7 -}, and mark the peak
19975 of the graph? Or should the highest label be @code{10 -}, which is a
19976 multiple of five, and be higher than the topmost value of the graph?
19978 The latter form is preferred. Most graphs are drawn within rectangles
19979 whose sides are an integral number of steps long---5, 10, 15, and so
19980 on for a step distance of five. But as soon as we decide to use a
19981 step height for the vertical axis, we discover that the simple
19982 expression in the varlist for computing the height is wrong. The
19983 expression is @code{(apply 'max numbers-list)}. This returns the
19984 precise height, not the maximum height plus whatever is necessary to
19985 round up to the nearest multiple of five. A more complex expression
19988 As usual in cases like this, a complex problem becomes simpler if it is
19989 divided into several smaller problems.
19991 First, consider the case when the highest value of the graph is an
19992 integral multiple of five---when it is 5, 10, 15, or some higher
19993 multiple of five. We can use this value as the Y axis height.
19995 A fairly simply way to determine whether a number is a multiple of
19996 five is to divide it by five and see if the division results in a
19997 remainder. If there is no remainder, the number is a multiple of
19998 five. Thus, seven divided by five has a remainder of two, and seven
19999 is not an integral multiple of five. Put in slightly different
20000 language, more reminiscent of the classroom, five goes into seven
20001 once, with a remainder of two. However, five goes into ten twice,
20002 with no remainder: ten is an integral multiple of five.
20004 @node Compute a Remainder, Y Axis Element, Height of label, print-Y-axis
20005 @appendixsubsec Side Trip: Compute a Remainder
20007 @findex % @r{(remainder function)}
20008 @cindex Remainder function, @code{%}
20009 In Lisp, the function for computing a remainder is @code{%}. The
20010 function returns the remainder of its first argument divided by its
20011 second argument. As it happens, @code{%} is a function in Emacs Lisp
20012 that you cannot discover using @code{apropos}: you find nothing if you
20013 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
20014 learn of the existence of @code{%} is to read about it in a book such
20015 as this or in the Emacs Lisp sources.
20017 You can try the @code{%} function by evaluating the following two
20029 The first expression returns 2 and the second expression returns 0.
20031 To test whether the returned value is zero or some other number, we
20032 can use the @code{zerop} function. This function returns @code{t} if
20033 its argument, which must be a number, is zero.
20045 Thus, the following expression will return @code{t} if the height
20046 of the graph is evenly divisible by five:
20049 (zerop (% height 5))
20053 (The value of @code{height}, of course, can be found from @code{(apply
20054 'max numbers-list)}.)
20056 On the other hand, if the value of @code{height} is not a multiple of
20057 five, we want to reset the value to the next higher multiple of five.
20058 This is straightforward arithmetic using functions with which we are
20059 already familiar. First, we divide the value of @code{height} by five
20060 to determine how many times five goes into the number. Thus, five
20061 goes into twelve twice. If we add one to this quotient and multiply by
20062 five, we will obtain the value of the next multiple of five that is
20063 larger than the height. Five goes into twelve twice. Add one to two,
20064 and multiply by five; the result is fifteen, which is the next multiple
20065 of five that is higher than twelve. The Lisp expression for this is:
20068 (* (1+ (/ height 5)) 5)
20072 For example, if you evaluate the following, the result is 15:
20075 (* (1+ (/ 12 5)) 5)
20078 All through this discussion, we have been using `five' as the value
20079 for spacing labels on the Y axis; but we may want to use some other
20080 value. For generality, we should replace `five' with a variable to
20081 which we can assign a value. The best name I can think of for this
20082 variable is @code{Y-axis-label-spacing}.
20085 Using this term, and an @code{if} expression, we produce the
20090 (if (zerop (% height Y-axis-label-spacing))
20093 (* (1+ (/ height Y-axis-label-spacing))
20094 Y-axis-label-spacing))
20099 This expression returns the value of @code{height} itself if the height
20100 is an even multiple of the value of the @code{Y-axis-label-spacing} or
20101 else it computes and returns a value of @code{height} that is equal to
20102 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
20104 We can now include this expression in the @code{let} expression of the
20105 @code{print-graph} function (after first setting the value of
20106 @code{Y-axis-label-spacing}):
20107 @vindex Y-axis-label-spacing
20111 (defvar Y-axis-label-spacing 5
20112 "Number of lines from one Y axis label to next.")
20117 (let* ((height (apply 'max numbers-list))
20118 (height-of-top-line
20119 (if (zerop (% height Y-axis-label-spacing))
20124 (* (1+ (/ height Y-axis-label-spacing))
20125 Y-axis-label-spacing)))
20126 (symbol-width (length graph-blank))))
20132 (Note use of the @code{let*} function: the initial value of height is
20133 computed once by the @code{(apply 'max numbers-list)} expression and
20134 then the resulting value of @code{height} is used to compute its
20135 final value. @xref{fwd-para let, , The @code{let*} expression}, for
20136 more about @code{let*}.)
20138 @node Y Axis Element, Y-axis-column, Compute a Remainder, print-Y-axis
20139 @appendixsubsec Construct a Y Axis Element
20141 When we print the vertical axis, we want to insert strings such as
20142 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20143 Moreover, we want the numbers and dashes to line up, so shorter
20144 numbers must be padded with leading spaces. If some of the strings
20145 use two digit numbers, the strings with single digit numbers must
20146 include a leading blank space before the number.
20148 @findex number-to-string
20149 To figure out the length of the number, the @code{length} function is
20150 used. But the @code{length} function works only with a string, not with
20151 a number. So the number has to be converted from being a number to
20152 being a string. This is done with the @code{number-to-string} function.
20157 (length (number-to-string 35))
20160 (length (number-to-string 100))
20166 (@code{number-to-string} is also called @code{int-to-string}; you will
20167 see this alternative name in various sources.)
20169 In addition, in each label, each number is followed by a string such
20170 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20171 This variable is defined with @code{defvar}:
20176 (defvar Y-axis-tic " - "
20177 "String that follows number in a Y axis label.")
20181 The length of the Y label is the sum of the length of the Y axis tic
20182 mark and the length of the number of the top of the graph.
20185 (length (concat (number-to-string height) Y-axis-tic)))
20188 This value will be calculated by the @code{print-graph} function in
20189 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20190 did not think to include this in the varlist when we first proposed it.)
20192 To make a complete vertical axis label, a tic mark is concatenated
20193 with a number; and the two together may be preceded by one or more
20194 spaces depending on how long the number is. The label consists of
20195 three parts: the (optional) leading spaces, the number, and the tic
20196 mark. The function is passed the value of the number for the specific
20197 row, and the value of the width of the top line, which is calculated
20198 (just once) by @code{print-graph}.
20202 (defun Y-axis-element (number full-Y-label-width)
20203 "Construct a NUMBERed label element.
20204 A numbered element looks like this ` 5 - ',
20205 and is padded as needed so all line up with
20206 the element for the largest number."
20209 (let* ((leading-spaces
20210 (- full-Y-label-width
20212 (concat (number-to-string number)
20217 (make-string leading-spaces ? )
20218 (number-to-string number)
20223 The @code{Y-axis-element} function concatenates together the leading
20224 spaces, if any; the number, as a string; and the tic mark.
20226 To figure out how many leading spaces the label will need, the
20227 function subtracts the actual length of the label---the length of the
20228 number plus the length of the tic mark---from the desired label width.
20230 @findex make-string
20231 Blank spaces are inserted using the @code{make-string} function. This
20232 function takes two arguments: the first tells it how long the string
20233 will be and the second is a symbol for the character to insert, in a
20234 special format. The format is a question mark followed by a blank
20235 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20236 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20237 syntax for characters. (Of course, you might want to replace the
20238 blank space by some other character @dots{} You know what to do.)
20240 The @code{number-to-string} function is used in the concatenation
20241 expression, to convert the number to a string that is concatenated
20242 with the leading spaces and the tic mark.
20244 @node Y-axis-column, print-Y-axis Penultimate, Y Axis Element, print-Y-axis
20245 @appendixsubsec Create a Y Axis Column
20247 The preceding functions provide all the tools needed to construct a
20248 function that generates a list of numbered and blank strings to insert
20249 as the label for the vertical axis:
20251 @findex Y-axis-column
20254 (defun Y-axis-column (height width-of-label)
20255 "Construct list of Y axis labels and blank strings.
20256 For HEIGHT of line above base and WIDTH-OF-LABEL."
20260 (while (> height 1)
20261 (if (zerop (% height Y-axis-label-spacing))
20262 ;; @r{Insert label.}
20265 (Y-axis-element height width-of-label)
20269 ;; @r{Else, insert blanks.}
20272 (make-string width-of-label ? )
20274 (setq height (1- height)))
20275 ;; @r{Insert base line.}
20277 (cons (Y-axis-element 1 width-of-label) Y-axis))
20278 (nreverse Y-axis)))
20282 In this function, we start with the value of @code{height} and
20283 repetitively subtract one from its value. After each subtraction, we
20284 test to see whether the value is an integral multiple of the
20285 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20286 using the @code{Y-axis-element} function; if not, we construct a
20287 blank label using the @code{make-string} function. The base line
20288 consists of the number one followed by a tic mark.
20291 @node print-Y-axis Penultimate, , Y-axis-column, print-Y-axis
20292 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20294 The list constructed by the @code{Y-axis-column} function is passed to
20295 the @code{print-Y-axis} function, which inserts the list as a column.
20297 @findex print-Y-axis
20300 (defun print-Y-axis (height full-Y-label-width)
20301 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20302 Height must be the maximum height of the graph.
20303 Full width is the width of the highest label element."
20304 ;; Value of height and full-Y-label-width
20305 ;; are passed by `print-graph'.
20308 (let ((start (point)))
20310 (Y-axis-column height full-Y-label-width))
20311 ;; @r{Place point ready for inserting graph.}
20313 ;; @r{Move point forward by value of} full-Y-label-width
20314 (forward-char full-Y-label-width)))
20318 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20319 insert the Y axis labels created by the @code{Y-axis-column} function.
20320 In addition, it places point at the correct position for printing the body of
20323 You can test @code{print-Y-axis}:
20331 Y-axis-label-spacing
20340 Copy the following expression:
20343 (print-Y-axis 12 5)
20347 Switch to the @file{*scratch*} buffer and place the cursor where you
20348 want the axis labels to start.
20351 Type @kbd{M-:} (@code{eval-expression}).
20354 Yank the @code{graph-body-print} expression into the minibuffer
20355 with @kbd{C-y} (@code{yank)}.
20358 Press @key{RET} to evaluate the expression.
20361 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20362 }}}. (The @code{print-graph} function will pass the value of
20363 @code{height-of-top-line}, which in this case will end up as 15,
20364 thereby getting rid of what might appear as a bug.)
20367 @node print-X-axis, Print Whole Graph, print-Y-axis, Full Graph
20368 @appendixsec The @code{print-X-axis} Function
20369 @cindex Axis, print horizontal
20370 @cindex X axis printing
20371 @cindex Print horizontal axis
20372 @cindex Horizontal axis printing
20374 X axis labels are much like Y axis labels, except that the ticks are on a
20375 line above the numbers. Labels should look like this:
20384 The first tic is under the first column of the graph and is preceded by
20385 several blank spaces. These spaces provide room in rows above for the Y
20386 axis labels. The second, third, fourth, and subsequent ticks are all
20387 spaced equally, according to the value of @code{X-axis-label-spacing}.
20389 The second row of the X axis consists of numbers, preceded by several
20390 blank spaces and also separated according to the value of the variable
20391 @code{X-axis-label-spacing}.
20393 The value of the variable @code{X-axis-label-spacing} should itself be
20394 measured in units of @code{symbol-width}, since you may want to change
20395 the width of the symbols that you are using to print the body of the
20396 graph without changing the ways the graph is labelled.
20399 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
20400 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20403 @node Similarities differences, X Axis Tic Marks, print-X-axis, print-X-axis
20405 @unnumberedsubsec Similarities and differences
20408 The @code{print-X-axis} function is constructed in more or less the
20409 same fashion as the @code{print-Y-axis} function except that it has
20410 two lines: the line of tic marks and the numbers. We will write a
20411 separate function to print each line and then combine them within the
20412 @code{print-X-axis} function.
20414 This is a three step process:
20418 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20421 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20424 Write a function to print both lines, the @code{print-X-axis} function,
20425 using @code{print-X-axis-tic-line} and
20426 @code{print-X-axis-numbered-line}.
20429 @node X Axis Tic Marks, , Similarities differences, print-X-axis
20430 @appendixsubsec X Axis Tic Marks
20432 The first function should print the X axis tic marks. We must specify
20433 the tic marks themselves and their spacing:
20437 (defvar X-axis-label-spacing
20438 (if (boundp 'graph-blank)
20439 (* 5 (length graph-blank)) 5)
20440 "Number of units from one X axis label to next.")
20445 (Note that the value of @code{graph-blank} is set by another
20446 @code{defvar}. The @code{boundp} predicate checks whether it has
20447 already been set; @code{boundp} returns @code{nil} if it has not. If
20448 @code{graph-blank} were unbound and we did not use this conditional
20449 construction, in a recent GNU Emacs, we would enter the debugger and
20450 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20451 @w{(void-variable graph-blank)}}.)
20454 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20458 (defvar X-axis-tic-symbol "|"
20459 "String to insert to point to a column in X axis.")
20464 The goal is to make a line that looks like this:
20470 The first tic is indented so that it is under the first column, which is
20471 indented to provide space for the Y axis labels.
20473 A tic element consists of the blank spaces that stretch from one tic to
20474 the next plus a tic symbol. The number of blanks is determined by the
20475 width of the tic symbol and the @code{X-axis-label-spacing}.
20478 The code looks like this:
20482 ;;; X-axis-tic-element
20486 ;; @r{Make a string of blanks.}
20487 (- (* symbol-width X-axis-label-spacing)
20488 (length X-axis-tic-symbol))
20490 ;; @r{Concatenate blanks with tic symbol.}
20496 Next, we determine how many blanks are needed to indent the first tic
20497 mark to the first column of the graph. This uses the value of
20498 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20501 The code to make @code{X-axis-leading-spaces}
20506 ;; X-axis-leading-spaces
20508 (make-string full-Y-label-width ? )
20513 We also need to determine the length of the horizontal axis, which is
20514 the length of the numbers list, and the number of ticks in the horizontal
20521 (length numbers-list)
20527 (* symbol-width X-axis-label-spacing)
20531 ;; number-of-X-ticks
20532 (if (zerop (% (X-length tic-width)))
20533 (/ (X-length tic-width))
20534 (1+ (/ (X-length tic-width))))
20539 All this leads us directly to the function for printing the X axis tic line:
20541 @findex print-X-axis-tic-line
20544 (defun print-X-axis-tic-line
20545 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20546 "Print ticks for X axis."
20547 (insert X-axis-leading-spaces)
20548 (insert X-axis-tic-symbol) ; @r{Under first column.}
20551 ;; @r{Insert second tic in the right spot.}
20554 (- (* symbol-width X-axis-label-spacing)
20555 ;; @r{Insert white space up to second tic symbol.}
20556 (* 2 (length X-axis-tic-symbol)))
20558 X-axis-tic-symbol))
20561 ;; @r{Insert remaining ticks.}
20562 (while (> number-of-X-tics 1)
20563 (insert X-axis-tic-element)
20564 (setq number-of-X-tics (1- number-of-X-tics))))
20568 The line of numbers is equally straightforward:
20571 First, we create a numbered element with blank spaces before each number:
20573 @findex X-axis-element
20576 (defun X-axis-element (number)
20577 "Construct a numbered X axis element."
20578 (let ((leading-spaces
20579 (- (* symbol-width X-axis-label-spacing)
20580 (length (number-to-string number)))))
20581 (concat (make-string leading-spaces ? )
20582 (number-to-string number))))
20586 Next, we create the function to print the numbered line, starting with
20587 the number ``1'' under the first column:
20589 @findex print-X-axis-numbered-line
20592 (defun print-X-axis-numbered-line
20593 (number-of-X-tics X-axis-leading-spaces)
20594 "Print line of X-axis numbers"
20595 (let ((number X-axis-label-spacing))
20596 (insert X-axis-leading-spaces)
20602 ;; @r{Insert white space up to next number.}
20603 (- (* symbol-width X-axis-label-spacing) 2)
20605 (number-to-string number)))
20608 ;; @r{Insert remaining numbers.}
20609 (setq number (+ number X-axis-label-spacing))
20610 (while (> number-of-X-tics 1)
20611 (insert (X-axis-element number))
20612 (setq number (+ number X-axis-label-spacing))
20613 (setq number-of-X-tics (1- number-of-X-tics)))))
20617 Finally, we need to write the @code{print-X-axis} that uses
20618 @code{print-X-axis-tic-line} and
20619 @code{print-X-axis-numbered-line}.
20621 The function must determine the local values of the variables used by both
20622 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20623 then it must call them. Also, it must print the carriage return that
20624 separates the two lines.
20626 The function consists of a varlist that specifies five local variables,
20627 and calls to each of the two line printing functions:
20629 @findex print-X-axis
20632 (defun print-X-axis (numbers-list)
20633 "Print X axis labels to length of NUMBERS-LIST."
20634 (let* ((leading-spaces
20635 (make-string full-Y-label-width ? ))
20638 ;; symbol-width @r{is provided by} graph-body-print
20639 (tic-width (* symbol-width X-axis-label-spacing))
20640 (X-length (length numbers-list))
20648 ;; @r{Make a string of blanks.}
20649 (- (* symbol-width X-axis-label-spacing)
20650 (length X-axis-tic-symbol))
20654 ;; @r{Concatenate blanks with tic symbol.}
20655 X-axis-tic-symbol))
20659 (if (zerop (% X-length tic-width))
20660 (/ X-length tic-width)
20661 (1+ (/ X-length tic-width)))))
20664 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20666 (print-X-axis-numbered-line tic-number leading-spaces)))
20671 You can test @code{print-X-axis}:
20675 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20676 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20677 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20680 Copy the following expression:
20685 (let ((full-Y-label-width 5)
20688 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20693 Switch to the @file{*scratch*} buffer and place the cursor where you
20694 want the axis labels to start.
20697 Type @kbd{M-:} (@code{eval-expression}).
20700 Yank the test expression into the minibuffer
20701 with @kbd{C-y} (@code{yank)}.
20704 Press @key{RET} to evaluate the expression.
20708 Emacs will print the horizontal axis like this:
20718 @node Print Whole Graph, , print-X-axis, Full Graph
20719 @appendixsec Printing the Whole Graph
20720 @cindex Printing the whole graph
20721 @cindex Whole graph printing
20722 @cindex Graph, printing all
20724 Now we are nearly ready to print the whole graph.
20726 The function to print the graph with the proper labels follows the
20727 outline we created earlier (@pxref{Full Graph, , A Graph with Labelled
20728 Axes}), but with additions.
20731 Here is the outline:
20735 (defun print-graph (numbers-list)
20736 "@var{documentation}@dots{}"
20737 (let ((height @dots{}
20741 (print-Y-axis height @dots{} )
20742 (graph-body-print numbers-list)
20743 (print-X-axis @dots{} )))
20748 * The final version:: A few changes.
20749 * Test print-graph:: Run a short test.
20750 * Graphing words in defuns:: Executing the final code.
20751 * lambda:: How to write an anonymous function.
20752 * mapcar:: Apply a function to elements of a list.
20753 * Another Bug:: Yet another bug @dots{} most insidious.
20754 * Final printed graph:: The graph itself!
20757 @node The final version, Test print-graph, Print Whole Graph, Print Whole Graph
20759 @unnumberedsubsec Changes for the Final Version
20762 The final version is different from what we planned in two ways:
20763 first, it contains additional values calculated once in the varlist;
20764 second, it carries an option to specify the labels' increment per row.
20765 This latter feature turns out to be essential; otherwise, a graph may
20766 have more rows than fit on a display or on a sheet of paper.
20769 This new feature requires a change to the @code{Y-axis-column}
20770 function, to add @code{vertical-step} to it. The function looks like
20773 @findex Y-axis-column @r{Final version.}
20776 ;;; @r{Final version.}
20777 (defun Y-axis-column
20778 (height width-of-label &optional vertical-step)
20779 "Construct list of labels for Y axis.
20780 HEIGHT is maximum height of graph.
20781 WIDTH-OF-LABEL is maximum width of label.
20782 VERTICAL-STEP, an option, is a positive integer
20783 that specifies how much a Y axis label increments
20784 for each line. For example, a step of 5 means
20785 that each line is five units of the graph."
20789 (number-per-line (or vertical-step 1)))
20790 (while (> height 1)
20791 (if (zerop (% height Y-axis-label-spacing))
20794 ;; @r{Insert label.}
20798 (* height number-per-line)
20803 ;; @r{Else, insert blanks.}
20806 (make-string width-of-label ? )
20808 (setq height (1- height)))
20811 ;; @r{Insert base line.}
20812 (setq Y-axis (cons (Y-axis-element
20813 (or vertical-step 1)
20816 (nreverse Y-axis)))
20820 The values for the maximum height of graph and the width of a symbol
20821 are computed by @code{print-graph} in its @code{let} expression; so
20822 @code{graph-body-print} must be changed to accept them.
20824 @findex graph-body-print @r{Final version.}
20827 ;;; @r{Final version.}
20828 (defun graph-body-print (numbers-list height symbol-width)
20829 "Print a bar graph of the NUMBERS-LIST.
20830 The numbers-list consists of the Y-axis values.
20831 HEIGHT is maximum height of graph.
20832 SYMBOL-WIDTH is number of each column."
20835 (let (from-position)
20836 (while numbers-list
20837 (setq from-position (point))
20839 (column-of-graph height (car numbers-list)))
20840 (goto-char from-position)
20841 (forward-char symbol-width)
20844 ;; @r{Draw graph column by column.}
20846 (setq numbers-list (cdr numbers-list)))
20847 ;; @r{Place point for X axis labels.}
20848 (forward-line height)
20854 Finally, the code for the @code{print-graph} function:
20856 @findex print-graph @r{Final version.}
20859 ;;; @r{Final version.}
20861 (numbers-list &optional vertical-step)
20862 "Print labelled bar graph of the NUMBERS-LIST.
20863 The numbers-list consists of the Y-axis values.
20867 Optionally, VERTICAL-STEP, a positive integer,
20868 specifies how much a Y axis label increments for
20869 each line. For example, a step of 5 means that
20870 each row is five units."
20873 (let* ((symbol-width (length graph-blank))
20874 ;; @code{height} @r{is both the largest number}
20875 ;; @r{and the number with the most digits.}
20876 (height (apply 'max numbers-list))
20879 (height-of-top-line
20880 (if (zerop (% height Y-axis-label-spacing))
20883 (* (1+ (/ height Y-axis-label-spacing))
20884 Y-axis-label-spacing)))
20887 (vertical-step (or vertical-step 1))
20888 (full-Y-label-width
20894 (* height-of-top-line vertical-step))
20900 height-of-top-line full-Y-label-width vertical-step)
20904 numbers-list height-of-top-line symbol-width)
20905 (print-X-axis numbers-list)))
20909 @node Test print-graph, Graphing words in defuns, The final version, Print Whole Graph
20910 @appendixsubsec Testing @code{print-graph}
20913 We can test the @code{print-graph} function with a short list of numbers:
20917 Install the final versions of @code{Y-axis-column},
20918 @code{graph-body-print}, and @code{print-graph} (in addition to the
20922 Copy the following expression:
20925 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20929 Switch to the @file{*scratch*} buffer and place the cursor where you
20930 want the axis labels to start.
20933 Type @kbd{M-:} (@code{eval-expression}).
20936 Yank the test expression into the minibuffer
20937 with @kbd{C-y} (@code{yank)}.
20940 Press @key{RET} to evaluate the expression.
20944 Emacs will print a graph that looks like this:
20965 On the other hand, if you pass @code{print-graph} a
20966 @code{vertical-step} value of 2, by evaluating this expression:
20969 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20974 The graph looks like this:
20995 (A question: is the `2' on the bottom of the vertical axis a bug or a
20996 feature? If you think it is a bug, and should be a `1' instead, (or
20997 even a `0'), you can modify the sources.)
20999 @node Graphing words in defuns, lambda, Test print-graph, Print Whole Graph
21000 @appendixsubsec Graphing Numbers of Words and Symbols
21002 Now for the graph for which all this code was written: a graph that
21003 shows how many function definitions contain fewer than 10 words and
21004 symbols, how many contain between 10 and 19 words and symbols, how
21005 many contain between 20 and 29 words and symbols, and so on.
21007 This is a multi-step process. First make sure you have loaded all the
21011 It is a good idea to reset the value of @code{top-of-ranges} in case
21012 you have set it to some different value. You can evaluate the
21017 (setq top-of-ranges
21020 110 120 130 140 150
21021 160 170 180 190 200
21022 210 220 230 240 250
21023 260 270 280 290 300)
21028 Next create a list of the number of words and symbols in each range.
21032 Evaluate the following:
21036 (setq list-for-graph
21039 (recursive-lengths-list-many-files
21040 (directory-files "/usr/local/emacs/lisp"
21048 On my old machine, this took about an hour. It looked though 303 Lisp
21049 files in my copy of Emacs version 19.23. After all that computing,
21050 the @code{list-for-graph} had this value:
21054 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
21055 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
21060 This means that my copy of Emacs had 537 function definitions with
21061 fewer than 10 words or symbols in them, 1,027 function definitions
21062 with 10 to 19 words or symbols in them, 955 function definitions with
21063 20 to 29 words or symbols in them, and so on.
21065 Clearly, just by looking at this list we can see that most function
21066 definitions contain ten to thirty words and symbols.
21068 Now for printing. We do @emph{not} want to print a graph that is
21069 1,030 lines high @dots{} Instead, we should print a graph that is
21070 fewer than twenty-five lines high. A graph that height can be
21071 displayed on almost any monitor, and easily printed on a sheet of paper.
21073 This means that each value in @code{list-for-graph} must be reduced to
21074 one-fiftieth its present value.
21076 Here is a short function to do just that, using two functions we have
21077 not yet seen, @code{mapcar} and @code{lambda}.
21081 (defun one-fiftieth (full-range)
21082 "Return list, each number one-fiftieth of previous."
21083 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21087 @node lambda, mapcar, Graphing words in defuns, Print Whole Graph
21088 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
21089 @cindex Anonymous function
21092 @code{lambda} is the symbol for an anonymous function, a function
21093 without a name. Every time you use an anonymous function, you need to
21094 include its whole body.
21101 (lambda (arg) (/ arg 50))
21105 is a function definition that says `return the value resulting from
21106 dividing whatever is passed to me as @code{arg} by 50'.
21109 Earlier, for example, we had a function @code{multiply-by-seven}; it
21110 multiplied its argument by 7. This function is similar, except it
21111 divides its argument by 50; and, it has no name. The anonymous
21112 equivalent of @code{multiply-by-seven} is:
21115 (lambda (number) (* 7 number))
21119 (@xref{defun, , The @code{defun} Special Form}.)
21123 If we want to multiply 3 by 7, we can write:
21125 @c !!! Clear print-postscript-figures if the computer formatting this
21126 @c document is too small and cannot handle all the diagrams and figures.
21127 @c clear print-postscript-figures
21128 @c set print-postscript-figures
21129 @c lambda example diagram #1
21133 (multiply-by-seven 3)
21134 \_______________/ ^
21140 @ifset print-postscript-figures
21143 @center @image{lambda-1}
21144 %%%% old method of including an image
21145 % \input /usr/local/lib/tex/inputs/psfig.tex
21146 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21151 @ifclear print-postscript-figures
21155 (multiply-by-seven 3)
21156 \_______________/ ^
21165 This expression returns 21.
21169 Similarly, we can write:
21171 @c lambda example diagram #2
21175 ((lambda (number) (* 7 number)) 3)
21176 \____________________________/ ^
21178 anonymous function argument
21182 @ifset print-postscript-figures
21185 @center @image{lambda-2}
21186 %%%% old method of including an image
21187 % \input /usr/local/lib/tex/inputs/psfig.tex
21188 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21193 @ifclear print-postscript-figures
21197 ((lambda (number) (* 7 number)) 3)
21198 \____________________________/ ^
21200 anonymous function argument
21208 If we want to divide 100 by 50, we can write:
21210 @c lambda example diagram #3
21214 ((lambda (arg) (/ arg 50)) 100)
21215 \______________________/ \_/
21217 anonymous function argument
21221 @ifset print-postscript-figures
21224 @center @image{lambda-3}
21225 %%%% old method of including an image
21226 % \input /usr/local/lib/tex/inputs/psfig.tex
21227 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21232 @ifclear print-postscript-figures
21236 ((lambda (arg) (/ arg 50)) 100)
21237 \______________________/ \_/
21239 anonymous function argument
21246 This expression returns 2. The 100 is passed to the function, which
21247 divides that number by 50.
21249 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21250 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21251 expressions derive from the Lambda Calculus.
21253 @node mapcar, Another Bug, lambda, Print Whole Graph
21254 @appendixsubsec The @code{mapcar} Function
21257 @code{mapcar} is a function that calls its first argument with each
21258 element of its second argument, in turn. The second argument must be
21261 The @samp{map} part of the name comes from the mathematical phrase,
21262 `mapping over a domain', meaning to apply a function to each of the
21263 elements in a domain. The mathematical phrase is based on the
21264 metaphor of a surveyor walking, one step at a time, over an area he is
21265 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21274 (mapcar '1+ '(2 4 6))
21280 The function @code{1+} which adds one to its argument, is executed on
21281 @emph{each} element of the list, and a new list is returned.
21283 Contrast this with @code{apply}, which applies its first argument to
21285 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21289 In the definition of @code{one-fiftieth}, the first argument is the
21290 anonymous function:
21293 (lambda (arg) (/ arg 50))
21297 and the second argument is @code{full-range}, which will be bound to
21298 @code{list-for-graph}.
21301 The whole expression looks like this:
21304 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21307 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21308 Lisp Reference Manual}, for more about @code{mapcar}.
21310 Using the @code{one-fiftieth} function, we can generate a list in
21311 which each element is one-fiftieth the size of the corresponding
21312 element in @code{list-for-graph}.
21316 (setq fiftieth-list-for-graph
21317 (one-fiftieth list-for-graph))
21322 The resulting list looks like this:
21326 (10 20 19 15 11 9 6 5 4 3 3 2 2
21327 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21332 This, we are almost ready to print! (We also notice the loss of
21333 information: many of the higher ranges are 0, meaning that fewer than
21334 50 defuns had that many words or symbols---but not necessarily meaning
21335 that none had that many words or symbols.)
21337 @node Another Bug, Final printed graph, mapcar, Print Whole Graph
21338 @appendixsubsec Another Bug @dots{} Most Insidious
21339 @cindex Bug, most insidious type
21340 @cindex Insidious type of bug
21342 I said `almost ready to print'! Of course, there is a bug in the
21343 @code{print-graph} function @dots{} It has a @code{vertical-step}
21344 option, but not a @code{horizontal-step} option. The
21345 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21346 @code{print-graph} function will print only by ones.
21348 This is a classic example of what some consider the most insidious
21349 type of bug, the bug of omission. This is not the kind of bug you can
21350 find by studying the code, for it is not in the code; it is an omitted
21351 feature. Your best actions are to try your program early and often;
21352 and try to arrange, as much as you can, to write code that is easy to
21353 understand and easy to change. Try to be aware, whenever you can,
21354 that whatever you have written, @emph{will} be rewritten, if not soon,
21355 eventually. A hard maxim to follow.
21357 It is the @code{print-X-axis-numbered-line} function that needs the
21358 work; and then the @code{print-X-axis} and the @code{print-graph}
21359 functions need to be adapted. Not much needs to be done; there is one
21360 nicety: the numbers ought to line up under the tic marks. This takes
21364 Here is the corrected @code{print-X-axis-numbered-line}:
21368 (defun print-X-axis-numbered-line
21369 (number-of-X-tics X-axis-leading-spaces
21370 &optional horizontal-step)
21371 "Print line of X-axis numbers"
21372 (let ((number X-axis-label-spacing)
21373 (horizontal-step (or horizontal-step 1)))
21376 (insert X-axis-leading-spaces)
21377 ;; @r{Delete extra leading spaces.}
21380 (length (number-to-string horizontal-step)))))
21385 ;; @r{Insert white space.}
21387 X-axis-label-spacing)
21390 (number-to-string horizontal-step)))
21394 (* number horizontal-step))))
21397 ;; @r{Insert remaining numbers.}
21398 (setq number (+ number X-axis-label-spacing))
21399 (while (> number-of-X-tics 1)
21400 (insert (X-axis-element
21401 (* number horizontal-step)))
21402 (setq number (+ number X-axis-label-spacing))
21403 (setq number-of-X-tics (1- number-of-X-tics)))))
21408 If you are reading this in Info, you can see the new versions of
21409 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21410 reading this in a printed book, you can see the changed lines here
21411 (the full text is too much to print).
21416 (defun print-X-axis (numbers-list horizontal-step)
21418 (print-X-axis-numbered-line
21419 tic-number leading-spaces horizontal-step))
21427 &optional vertical-step horizontal-step)
21429 (print-X-axis numbers-list horizontal-step))
21437 (defun print-X-axis (numbers-list horizontal-step)
21438 "Print X axis labels to length of NUMBERS-LIST.
21439 Optionally, HORIZONTAL-STEP, a positive integer,
21440 specifies how much an X axis label increments for
21444 ;; Value of symbol-width and full-Y-label-width
21445 ;; are passed by `print-graph'.
21446 (let* ((leading-spaces
21447 (make-string full-Y-label-width ? ))
21448 ;; symbol-width @r{is provided by} graph-body-print
21449 (tic-width (* symbol-width X-axis-label-spacing))
21450 (X-length (length numbers-list))
21456 ;; @r{Make a string of blanks.}
21457 (- (* symbol-width X-axis-label-spacing)
21458 (length X-axis-tic-symbol))
21462 ;; @r{Concatenate blanks with tic symbol.}
21463 X-axis-tic-symbol))
21465 (if (zerop (% X-length tic-width))
21466 (/ X-length tic-width)
21467 (1+ (/ X-length tic-width)))))
21471 (print-X-axis-tic-line
21472 tic-number leading-spaces X-tic)
21474 (print-X-axis-numbered-line
21475 tic-number leading-spaces horizontal-step)))
21482 (numbers-list &optional vertical-step horizontal-step)
21483 "Print labelled bar graph of the NUMBERS-LIST.
21484 The numbers-list consists of the Y-axis values.
21488 Optionally, VERTICAL-STEP, a positive integer,
21489 specifies how much a Y axis label increments for
21490 each line. For example, a step of 5 means that
21491 each row is five units.
21495 Optionally, HORIZONTAL-STEP, a positive integer,
21496 specifies how much an X axis label increments for
21498 (let* ((symbol-width (length graph-blank))
21499 ;; @code{height} @r{is both the largest number}
21500 ;; @r{and the number with the most digits.}
21501 (height (apply 'max numbers-list))
21504 (height-of-top-line
21505 (if (zerop (% height Y-axis-label-spacing))
21508 (* (1+ (/ height Y-axis-label-spacing))
21509 Y-axis-label-spacing)))
21512 (vertical-step (or vertical-step 1))
21513 (full-Y-label-width
21517 (* height-of-top-line vertical-step))
21522 height-of-top-line full-Y-label-width vertical-step)
21524 numbers-list height-of-top-line symbol-width)
21525 (print-X-axis numbers-list horizontal-step)))
21532 Graphing Definitions Re-listed
21535 Here are all the graphing definitions in their final form:
21539 (defvar top-of-ranges
21542 110 120 130 140 150
21543 160 170 180 190 200
21544 210 220 230 240 250)
21545 "List specifying ranges for `defuns-per-range'.")
21549 (defvar graph-symbol "*"
21550 "String used as symbol in graph, usually an asterisk.")
21554 (defvar graph-blank " "
21555 "String used as blank in graph, usually a blank space.
21556 graph-blank must be the same number of columns wide
21561 (defvar Y-axis-tic " - "
21562 "String that follows number in a Y axis label.")
21566 (defvar Y-axis-label-spacing 5
21567 "Number of lines from one Y axis label to next.")
21571 (defvar X-axis-tic-symbol "|"
21572 "String to insert to point to a column in X axis.")
21576 (defvar X-axis-label-spacing
21577 (if (boundp 'graph-blank)
21578 (* 5 (length graph-blank)) 5)
21579 "Number of units from one X axis label to next.")
21585 (defun count-words-in-defun ()
21586 "Return the number of words and symbols in a defun."
21587 (beginning-of-defun)
21589 (end (save-excursion (end-of-defun) (point))))
21594 (and (< (point) end)
21596 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21598 (setq count (1+ count)))
21605 (defun lengths-list-file (filename)
21606 "Return list of definitions' lengths within FILE.
21607 The returned list is a list of numbers.
21608 Each number is the number of words or
21609 symbols in one function definition."
21613 (message "Working on `%s' ... " filename)
21615 (let ((buffer (find-file-noselect filename))
21617 (set-buffer buffer)
21618 (setq buffer-read-only t)
21620 (goto-char (point-min))
21624 (while (re-search-forward "^(defun" nil t)
21626 (cons (count-words-in-defun) lengths-list)))
21627 (kill-buffer buffer)
21634 (defun lengths-list-many-files (list-of-files)
21635 "Return list of lengths of defuns in LIST-OF-FILES."
21636 (let (lengths-list)
21637 ;;; @r{true-or-false-test}
21638 (while list-of-files
21644 ;;; @r{Generate a lengths' list.}
21646 (expand-file-name (car list-of-files)))))
21647 ;;; @r{Make files' list shorter.}
21648 (setq list-of-files (cdr list-of-files)))
21649 ;;; @r{Return final value of lengths' list.}
21656 (defun defuns-per-range (sorted-lengths top-of-ranges)
21657 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21658 (let ((top-of-range (car top-of-ranges))
21659 (number-within-range 0)
21660 defuns-per-range-list)
21665 (while top-of-ranges
21669 ;; @r{Need number for numeric test.}
21670 (car sorted-lengths)
21671 (< (car sorted-lengths) top-of-range))
21673 ;; @r{Count number of definitions within current range.}
21674 (setq number-within-range (1+ number-within-range))
21675 (setq sorted-lengths (cdr sorted-lengths)))
21679 ;; @r{Exit inner loop but remain within outer loop.}
21681 (setq defuns-per-range-list
21682 (cons number-within-range defuns-per-range-list))
21683 (setq number-within-range 0) ; @r{Reset count to zero.}
21685 ;; @r{Move to next range.}
21686 (setq top-of-ranges (cdr top-of-ranges))
21687 ;; @r{Specify next top of range value.}
21688 (setq top-of-range (car top-of-ranges)))
21692 ;; @r{Exit outer loop and count the number of defuns larger than}
21693 ;; @r{ the largest top-of-range value.}
21694 (setq defuns-per-range-list
21696 (length sorted-lengths)
21697 defuns-per-range-list))
21699 ;; @r{Return a list of the number of definitions within each range,}
21700 ;; @r{ smallest to largest.}
21701 (nreverse defuns-per-range-list)))
21707 (defun column-of-graph (max-graph-height actual-height)
21708 "Return list of MAX-GRAPH-HEIGHT strings;
21709 ACTUAL-HEIGHT are graph-symbols.
21710 The graph-symbols are contiguous entries at the end
21712 The list will be inserted as one column of a graph.
21713 The strings are either graph-blank or graph-symbol."
21717 (let ((insert-list nil)
21718 (number-of-top-blanks
21719 (- max-graph-height actual-height)))
21721 ;; @r{Fill in @code{graph-symbols}.}
21722 (while (> actual-height 0)
21723 (setq insert-list (cons graph-symbol insert-list))
21724 (setq actual-height (1- actual-height)))
21728 ;; @r{Fill in @code{graph-blanks}.}
21729 (while (> number-of-top-blanks 0)
21730 (setq insert-list (cons graph-blank insert-list))
21731 (setq number-of-top-blanks
21732 (1- number-of-top-blanks)))
21734 ;; @r{Return whole list.}
21741 (defun Y-axis-element (number full-Y-label-width)
21742 "Construct a NUMBERed label element.
21743 A numbered element looks like this ` 5 - ',
21744 and is padded as needed so all line up with
21745 the element for the largest number."
21748 (let* ((leading-spaces
21749 (- full-Y-label-width
21751 (concat (number-to-string number)
21756 (make-string leading-spaces ? )
21757 (number-to-string number)
21764 (defun print-Y-axis
21765 (height full-Y-label-width &optional vertical-step)
21766 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21767 Height must be the maximum height of the graph.
21768 Full width is the width of the highest label element.
21769 Optionally, print according to VERTICAL-STEP."
21772 ;; Value of height and full-Y-label-width
21773 ;; are passed by `print-graph'.
21774 (let ((start (point)))
21776 (Y-axis-column height full-Y-label-width vertical-step))
21779 ;; @r{Place point ready for inserting graph.}
21781 ;; @r{Move point forward by value of} full-Y-label-width
21782 (forward-char full-Y-label-width)))
21788 (defun print-X-axis-tic-line
21789 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21790 "Print ticks for X axis."
21791 (insert X-axis-leading-spaces)
21792 (insert X-axis-tic-symbol) ; @r{Under first column.}
21795 ;; @r{Insert second tic in the right spot.}
21798 (- (* symbol-width X-axis-label-spacing)
21799 ;; @r{Insert white space up to second tic symbol.}
21800 (* 2 (length X-axis-tic-symbol)))
21802 X-axis-tic-symbol))
21805 ;; @r{Insert remaining ticks.}
21806 (while (> number-of-X-tics 1)
21807 (insert X-axis-tic-element)
21808 (setq number-of-X-tics (1- number-of-X-tics))))
21814 (defun X-axis-element (number)
21815 "Construct a numbered X axis element."
21816 (let ((leading-spaces
21817 (- (* symbol-width X-axis-label-spacing)
21818 (length (number-to-string number)))))
21819 (concat (make-string leading-spaces ? )
21820 (number-to-string number))))
21826 (defun graph-body-print (numbers-list height symbol-width)
21827 "Print a bar graph of the NUMBERS-LIST.
21828 The numbers-list consists of the Y-axis values.
21829 HEIGHT is maximum height of graph.
21830 SYMBOL-WIDTH is number of each column."
21833 (let (from-position)
21834 (while numbers-list
21835 (setq from-position (point))
21837 (column-of-graph height (car numbers-list)))
21838 (goto-char from-position)
21839 (forward-char symbol-width)
21842 ;; @r{Draw graph column by column.}
21844 (setq numbers-list (cdr numbers-list)))
21845 ;; @r{Place point for X axis labels.}
21846 (forward-line height)
21853 (defun Y-axis-column
21854 (height width-of-label &optional vertical-step)
21855 "Construct list of labels for Y axis.
21856 HEIGHT is maximum height of graph.
21857 WIDTH-OF-LABEL is maximum width of label.
21860 VERTICAL-STEP, an option, is a positive integer
21861 that specifies how much a Y axis label increments
21862 for each line. For example, a step of 5 means
21863 that each line is five units of the graph."
21865 (number-per-line (or vertical-step 1)))
21868 (while (> height 1)
21869 (if (zerop (% height Y-axis-label-spacing))
21870 ;; @r{Insert label.}
21874 (* height number-per-line)
21879 ;; @r{Else, insert blanks.}
21882 (make-string width-of-label ? )
21884 (setq height (1- height)))
21887 ;; @r{Insert base line.}
21888 (setq Y-axis (cons (Y-axis-element
21889 (or vertical-step 1)
21892 (nreverse Y-axis)))
21898 (defun print-X-axis-numbered-line
21899 (number-of-X-tics X-axis-leading-spaces
21900 &optional horizontal-step)
21901 "Print line of X-axis numbers"
21902 (let ((number X-axis-label-spacing)
21903 (horizontal-step (or horizontal-step 1)))
21906 (insert X-axis-leading-spaces)
21908 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21911 ;; @r{Insert white space up to next number.}
21912 (- (* symbol-width X-axis-label-spacing)
21913 (1- (length (number-to-string horizontal-step)))
21916 (number-to-string (* number horizontal-step))))
21919 ;; @r{Insert remaining numbers.}
21920 (setq number (+ number X-axis-label-spacing))
21921 (while (> number-of-X-tics 1)
21922 (insert (X-axis-element (* number horizontal-step)))
21923 (setq number (+ number X-axis-label-spacing))
21924 (setq number-of-X-tics (1- number-of-X-tics)))))
21930 (defun print-X-axis (numbers-list horizontal-step)
21931 "Print X axis labels to length of NUMBERS-LIST.
21932 Optionally, HORIZONTAL-STEP, a positive integer,
21933 specifies how much an X axis label increments for
21937 ;; Value of symbol-width and full-Y-label-width
21938 ;; are passed by `print-graph'.
21939 (let* ((leading-spaces
21940 (make-string full-Y-label-width ? ))
21941 ;; symbol-width @r{is provided by} graph-body-print
21942 (tic-width (* symbol-width X-axis-label-spacing))
21943 (X-length (length numbers-list))
21949 ;; @r{Make a string of blanks.}
21950 (- (* symbol-width X-axis-label-spacing)
21951 (length X-axis-tic-symbol))
21955 ;; @r{Concatenate blanks with tic symbol.}
21956 X-axis-tic-symbol))
21958 (if (zerop (% X-length tic-width))
21959 (/ X-length tic-width)
21960 (1+ (/ X-length tic-width)))))
21964 (print-X-axis-tic-line
21965 tic-number leading-spaces X-tic)
21967 (print-X-axis-numbered-line
21968 tic-number leading-spaces horizontal-step)))
21974 (defun one-fiftieth (full-range)
21975 "Return list, each number of which is 1/50th previous."
21976 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21983 (numbers-list &optional vertical-step horizontal-step)
21984 "Print labelled bar graph of the NUMBERS-LIST.
21985 The numbers-list consists of the Y-axis values.
21989 Optionally, VERTICAL-STEP, a positive integer,
21990 specifies how much a Y axis label increments for
21991 each line. For example, a step of 5 means that
21992 each row is five units.
21996 Optionally, HORIZONTAL-STEP, a positive integer,
21997 specifies how much an X axis label increments for
21999 (let* ((symbol-width (length graph-blank))
22000 ;; @code{height} @r{is both the largest number}
22001 ;; @r{and the number with the most digits.}
22002 (height (apply 'max numbers-list))
22005 (height-of-top-line
22006 (if (zerop (% height Y-axis-label-spacing))
22009 (* (1+ (/ height Y-axis-label-spacing))
22010 Y-axis-label-spacing)))
22013 (vertical-step (or vertical-step 1))
22014 (full-Y-label-width
22018 (* height-of-top-line vertical-step))
22024 height-of-top-line full-Y-label-width vertical-step)
22026 numbers-list height-of-top-line symbol-width)
22027 (print-X-axis numbers-list horizontal-step)))
22034 @node Final printed graph, , Another Bug, Print Whole Graph
22035 @appendixsubsec The Printed Graph
22037 When made and installed, you can call the @code{print-graph} command
22043 (print-graph fiftieth-list-for-graph 50 10)
22073 50 - ***************** * *
22075 10 50 100 150 200 250 300 350
22082 The largest group of functions contain 10 -- 19 words and symbols each.
22084 @node Free Software and Free Manuals, GNU Free Documentation License, Full Graph, Top
22085 @appendix Free Software and Free Manuals
22087 @strong{by Richard M. Stallman}
22090 The biggest deficiency in free operating systems is not in the
22091 software---it is the lack of good free manuals that we can include in
22092 these systems. Many of our most important programs do not come with
22093 full manuals. Documentation is an essential part of any software
22094 package; when an important free software package does not come with a
22095 free manual, that is a major gap. We have many such gaps today.
22097 Once upon a time, many years ago, I thought I would learn Perl. I got
22098 a copy of a free manual, but I found it hard to read. When I asked
22099 Perl users about alternatives, they told me that there were better
22100 introductory manuals---but those were not free.
22102 Why was this? The authors of the good manuals had written them for
22103 O'Reilly Associates, which published them with restrictive terms---no
22104 copying, no modification, source files not available---which exclude
22105 them from the free software community.
22107 That wasn't the first time this sort of thing has happened, and (to
22108 our community's great loss) it was far from the last. Proprietary
22109 manual publishers have enticed a great many authors to restrict their
22110 manuals since then. Many times I have heard a GNU user eagerly tell me
22111 about a manual that he is writing, with which he expects to help the
22112 GNU project---and then had my hopes dashed, as he proceeded to explain
22113 that he had signed a contract with a publisher that would restrict it
22114 so that we cannot use it.
22116 Given that writing good English is a rare skill among programmers, we
22117 can ill afford to lose manuals this way.
22119 Free documentation, like free software, is a matter of freedom, not
22120 price. The problem with these manuals was not that O'Reilly Associates
22121 charged a price for printed copies---that in itself is fine. The Free
22122 Software Foundation @uref{http://shop.fsf.org, sells printed copies} of
22123 free @uref{http://www.gnu.org/doc/doc.html, GNU manuals}, too.
22124 But GNU manuals are available in source code form, while these manuals
22125 are available only on paper. GNU manuals come with permission to copy
22126 and modify; the Perl manuals do not. These restrictions are the
22129 The criterion for a free manual is pretty much the same as for free
22130 software: it is a matter of giving all users certain
22131 freedoms. Redistribution (including commercial redistribution) must be
22132 permitted, so that the manual can accompany every copy of the program,
22133 on-line or on paper. Permission for modification is crucial too.
22135 As a general rule, I don't believe that it is essential for people to
22136 have permission to modify all sorts of articles and books. The issues
22137 for writings are not necessarily the same as those for software. For
22138 example, I don't think you or I are obliged to give permission to
22139 modify articles like this one, which describe our actions and our
22142 But there is a particular reason why the freedom to modify is crucial
22143 for documentation for free software. When people exercise their right
22144 to modify the software, and add or change its features, if they are
22145 conscientious they will change the manual too---so they can provide
22146 accurate and usable documentation with the modified program. A manual
22147 which forbids programmers to be conscientious and finish the job, or
22148 more precisely requires them to write a new manual from scratch if
22149 they change the program, does not fill our community's needs.
22151 While a blanket prohibition on modification is unacceptable, some
22152 kinds of limits on the method of modification pose no problem. For
22153 example, requirements to preserve the original author's copyright
22154 notice, the distribution terms, or the list of authors, are ok. It is
22155 also no problem to require modified versions to include notice that
22156 they were modified, even to have entire sections that may not be
22157 deleted or changed, as long as these sections deal with nontechnical
22158 topics. (Some GNU manuals have them.)
22160 These kinds of restrictions are not a problem because, as a practical
22161 matter, they don't stop the conscientious programmer from adapting the
22162 manual to fit the modified program. In other words, they don't block
22163 the free software community from making full use of the manual.
22165 However, it must be possible to modify all the technical content of
22166 the manual, and then distribute the result in all the usual media,
22167 through all the usual channels; otherwise, the restrictions do block
22168 the community, the manual is not free, and so we need another manual.
22170 Unfortunately, it is often hard to find someone to write another
22171 manual when a proprietary manual exists. The obstacle is that many
22172 users think that a proprietary manual is good enough---so they don't
22173 see the need to write a free manual. They do not see that the free
22174 operating system has a gap that needs filling.
22176 Why do users think that proprietary manuals are good enough? Some have
22177 not considered the issue. I hope this article will do something to
22180 Other users consider proprietary manuals acceptable for the same
22181 reason so many people consider proprietary software acceptable: they
22182 judge in purely practical terms, not using freedom as a
22183 criterion. These people are entitled to their opinions, but since
22184 those opinions spring from values which do not include freedom, they
22185 are no guide for those of us who do value freedom.
22187 Please spread the word about this issue. We continue to lose manuals
22188 to proprietary publishing. If we spread the word that proprietary
22189 manuals are not sufficient, perhaps the next person who wants to help
22190 GNU by writing documentation will realize, before it is too late, that
22191 he must above all make it free.
22193 We can also encourage commercial publishers to sell free, copylefted
22194 manuals instead of proprietary ones. One way you can help this is to
22195 check the distribution terms of a manual before you buy it, and prefer
22196 copylefted manuals to non-copylefted ones.
22200 Note: The Free Software Foundation maintains a page on its Web site
22201 that lists free books available from other publishers:@*
22202 @uref{http://www.gnu.org/doc/other-free-books.html}
22204 @node GNU Free Documentation License, Index, Free Software and Free Manuals, Top
22205 @appendix GNU Free Documentation License
22207 @cindex FDL, GNU Free Documentation License
22208 @include doclicense.texi
22210 @node Index, About the Author, GNU Free Documentation License, Top
22211 @comment node-name, next, previous, up
22215 MENU ENTRY: NODE NAME.
22221 @c Place biographical information on right-hand (verso) page
22224 \par\vfill\supereject
22226 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22227 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22230 % \par\vfill\supereject
22231 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22232 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22233 %\page\hbox{}%\page
22234 %\page\hbox{}%\page
22241 @c ================ Biographical information ================
22245 @center About the Author
22250 @node About the Author, , Index, Top
22251 @unnumbered About the Author
22255 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22256 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22257 world on software freedom. Chassell was a founding Director and
22258 Treasurer of the Free Software Foundation, Inc. He is co-author of
22259 the @cite{Texinfo} manual, and has edited more than a dozen other
22260 books. He graduated from Cambridge University, in England. He has an
22261 abiding interest in social and economic history and flies his own
22268 @c @c Prevent page number on blank verso, so eject it first.
22270 @c \par\vfill\supereject
22275 @c @evenheading @thispage @| @| @thistitle
22276 @c @oddheading @| @| @thispage