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
11 @include emacsver.texi
14 @c <<<< For hard copy printing, this file is now
15 @c set for smallbook, which works for all sizes
16 @c of paper, and with PostScript figures >>>>
22 @set print-postscript-figures
24 @c clear print-postscript-figures
27 @comment %**end of header
29 @c per rms and peterb, use 10pt fonts for the main text, mostly to
30 @c save on paper cost.
31 @c Do this inside @tex for now, so current makeinfo does not complain.
37 \global\hbadness=6666 % don't worry about not-too-underfull boxes
40 @c These refer to the printed book sold by the FSF.
41 @set edition-number 3.10
42 @set update-date 28 October 2009
45 ## Summary of shell commands to create various output formats:
47 pushd /usr/local/src/emacs/lispintro/
51 makeinfo --paragraph-indent=0 --verbose emacs-lisp-intro.texi
53 ## ;; (progn (when (bufferp (get-buffer "*info*")) (kill-buffer "*info*")) (info "/usr/local/src/emacs/info/eintr"))
56 texi2dvi emacs-lisp-intro.texi
58 ## xdvi -margins 24pt -topmargin 4pt -offsets 24pt -geometry 760x1140 -s 5 -useTeXpages -mousemode 1 emacs-lisp-intro.dvi &
61 makeinfo --html --no-split --verbose emacs-lisp-intro.texi
63 ## galeon emacs-lisp-intro.html
66 makeinfo --fill-column=70 --no-split --paragraph-indent=0 \
67 --verbose --no-headers --output=emacs-lisp-intro.txt emacs-lisp-intro.texi
73 mtusb # mount -v -t ext3 /dev/sda /mnt
74 cp -v /u/intro/emacs-lisp-intro.texi /mnt/backup/intro/emacs-lisp-intro.texi
75 umtusb # umount -v /mnt
78 ## Other shell commands
80 pushd /usr/local/src/emacs/lispintro/
84 texi2dvi --pdf emacs-lisp-intro.texi
85 # xpdf emacs-lisp-intro.pdf &
87 ## DocBook -- note file extension
88 makeinfo --docbook --no-split --paragraph-indent=0 \
89 --verbose --output=emacs-lisp-intro.docbook emacs-lisp-intro.texi
91 ## XML with a Texinfo DTD -- note file extension
92 makeinfo --xml --no-split --paragraph-indent=0 \
93 --verbose --output=emacs-lisp-intro.texinfoxml emacs-lisp-intro.texi
95 ## PostScript (needs DVI)
96 # gv emacs-lisp-intro.ps &
97 # Create DVI if we lack it
98 # texi2dvi emacs-lisp-intro.texi
99 dvips emacs-lisp-intro.dvi -o emacs-lisp-intro.ps
102 # Use OpenOffice to view RTF
103 # Create HTML if we lack it
104 # makeinfo --no-split --html emacs-lisp-intro.texi
105 /usr/local/src/html2rtf.pl emacs-lisp-intro.html
108 /usr/bin/rtf2latex emacs-lisp-intro.rtf
114 @c ================ Included Figures ================
116 @c Set print-postscript-figures if you print PostScript figures.
117 @c If you clear this, the ten figures will be printed as ASCII diagrams.
118 @c (This is not relevant to Info, since Info only handles ASCII.)
119 @c Your site may require editing changes to print PostScript; in this
120 @c case, search for `print-postscript-figures' and make appropriate changes.
122 @c ================ How to Create an Info file ================
124 @c If you have `makeinfo' installed, run the following command
126 @c makeinfo emacs-lisp-intro.texi
128 @c or, if you want a single, large Info file, and no paragraph indents:
129 @c makeinfo --no-split --paragraph-indent=0 --verbose emacs-lisp-intro.texi
131 @c After creating the Info file, edit your Info `dir' file, if the
132 @c `dircategory' section below does not enable your system to
133 @c install the manual automatically.
134 @c (The `dir' file is often in the `/usr/local/share/info/' directory.)
136 @c ================ How to Create an HTML file ================
138 @c To convert to HTML format
139 @c makeinfo --html --no-split --verbose emacs-lisp-intro.texi
141 @c ================ How to Print a Book in Various Sizes ================
143 @c This book can be printed in any of three different sizes.
144 @c In the above header, set @-commands appropriately.
154 @c European A4 size paper:
159 @c ================ How to Typeset and Print ================
161 @c If you do not include PostScript figures, run either of the
162 @c following command sequences, or similar commands suited to your
165 @c texi2dvi emacs-lisp-intro.texi
166 @c lpr -d emacs-lisp-intro.dvi
170 @c tex emacs-lisp-intro.texi
171 @c texindex emacs-lisp-intro.??
172 @c tex emacs-lisp-intro.texi
173 @c lpr -d emacs-lisp-intro.dvi
175 @c If you include the PostScript figures, and you have old software,
176 @c you may need to convert the .dvi file to a .ps file before
177 @c printing. Run either of the following command sequences, or one
180 @c dvips -f < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
184 @c postscript -p < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
187 @c (Note: if you edit the book so as to change the length of the
188 @c table of contents, you may have to change the value of `pageno' below.)
190 @c ================ End of Formatting Sections ================
192 @c For next or subsequent edition:
193 @c create function using with-output-to-temp-buffer
194 @c create a major mode, with keymaps
195 @c run an asynchronous process, like grep or diff
197 @c For 8.5 by 11 inch format: do not use such a small amount of
198 @c whitespace between paragraphs as smallbook format
201 \global\parskip 6pt plus 1pt
205 @c For all sized formats: print within-book cross
206 @c reference with ``...'' rather than [...]
208 @c This works with the texinfo.tex file, version 2003-05-04.08,
209 @c in the Texinfo version 4.6 of the 2003 Jun 13 distribution.
212 \if \xrefprintnodename
213 \global\def\xrefprintnodename#1{\unskip, ``#1''}
215 \global\def\xrefprintnodename#1{ ``#1''}
217 % \global\def\xrefprintnodename#1{, ``#1''}
220 @c ----------------------------------------------------
222 @dircategory GNU Emacs Lisp
224 * Emacs Lisp Intro: (eintr).
225 A simple introduction to Emacs Lisp programming.
229 This is an @cite{Introduction to Programming in Emacs Lisp}, for
230 people who are not programmers.
233 Edition @value{edition-number}, @value{update-date}
236 Distributed with Emacs version @value{EMACSVER}.
240 <p>The homepage for GNU Emacs is at
241 <a href="http://www.gnu.org/software/emacs/">http://www.gnu.org/software/emacs/</a>.
242 <br>To view this manual in other formats, click
243 <a href="/software/emacs/emacs-lisp-intro/emacs-lisp-intro.html">here</a>.
247 Copyright @copyright{} 1990--1995, 1997, 2001--2013 Free Software
254 GNU Press, @hfill @uref{http://www.fsf.org/licensing/gnu-press/}@*
255 a division of the @hfill email: @email{sales@@fsf.org}@*
256 Free Software Foundation, Inc. @hfill Tel: +1 (617) 542-5942@*
257 51 Franklin Street, Fifth Floor @hfill Fax: +1 (617) 542-2652@*
258 Boston, MA 02110-1301 USA
262 Printed copies available from @uref{http://shop.fsf.org/}. Published by:
265 GNU Press, http://www.fsf.org/licensing/gnu-press/
266 a division of the email: sales@@fsf.org
267 Free Software Foundation, Inc. Tel: +1 (617) 542-5942
268 51 Franklin Street, Fifth Floor Fax: +1 (617) 542-2652
269 Boston, MA 02110-1301 USA
276 Permission is granted to copy, distribute and/or modify this document
277 under the terms of the GNU Free Documentation License, Version 1.3 or
278 any later version published by the Free Software Foundation; there
279 being no Invariant Section, with the Front-Cover Texts being ``A GNU
280 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
281 the license is included in the section entitled ``GNU Free
282 Documentation License''.
284 (a) The FSF's Back-Cover Text is: ``You have the freedom to
285 copy and modify this GNU manual. Buying copies from the FSF
286 supports it in developing GNU and promoting software freedom.''
289 @c half title; two lines here, so do not use `shorttitlepage'
292 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
294 {\begingroup\hbox{}\vskip 0.25in \chaprm%
295 \centerline{Programming in Emacs Lisp}%
296 \endgroup\page\hbox{}\page}
301 @center @titlefont{An Introduction to}
303 @center @titlefont{Programming in Emacs Lisp}
305 @center Revised Third Edition
307 @center by Robert J. Chassell
310 @vskip 0pt plus 1filll
316 @evenheading @thispage @| @| @thischapter
317 @oddheading @thissection @| @| @thispage
321 @c Keep T.O.C. short by tightening up for largebook
324 \global\parskip 2pt plus 1pt
325 \global\advance\baselineskip by -1pt
335 @top An Introduction to Programming in Emacs Lisp
339 This master menu first lists each chapter and index; then it lists
340 every node in every chapter.
343 @c >>>> Set pageno appropriately <<<<
345 @c The first page of the Preface is a roman numeral; it is the first
346 @c right handed page after the Table of Contents; hence the following
347 @c setting must be for an odd negative number.
350 @c global@pageno = -11
353 @set COUNT-WORDS count-words-example
354 @c Length of variable name chosen so that things still line up when expanded.
357 * Preface:: What to look for.
358 * List Processing:: What is Lisp?
359 * Practicing Evaluation:: Running several programs.
360 * Writing Defuns:: How to write function definitions.
361 * Buffer Walk Through:: Exploring a few buffer-related functions.
362 * More Complex:: A few, even more complex functions.
363 * Narrowing & Widening:: Restricting your and Emacs attention to
365 * car cdr & cons:: Fundamental functions in Lisp.
366 * Cutting & Storing Text:: Removing text and saving it.
367 * List Implementation:: How lists are implemented in the computer.
368 * Yanking:: Pasting stored text.
369 * Loops & Recursion:: How to repeat a process.
370 * Regexp Search:: Regular expression searches.
371 * Counting Words:: A review of repetition and regexps.
372 * Words in a defun:: Counting words in a @code{defun}.
373 * Readying a Graph:: A prototype graph printing function.
374 * Emacs Initialization:: How to write a @file{.emacs} file.
375 * Debugging:: How to run the Emacs Lisp debuggers.
376 * Conclusion:: Now you have the basics.
377 * the-the:: An appendix: how to find reduplicated words.
378 * Kill Ring:: An appendix: how the kill ring works.
379 * Full Graph:: How to create a graph with labeled axes.
380 * Free Software and Free Manuals::
381 * GNU Free Documentation License::
386 --- The Detailed Node Listing ---
390 * Why:: Why learn Emacs Lisp?
391 * On Reading this Text:: Read, gain familiarity, pick up habits....
392 * Who You Are:: For whom this is written.
394 * Note for Novices:: You can read this as a novice.
399 * Lisp Lists:: What are lists?
400 * Run a Program:: Any list in Lisp is a program ready to run.
401 * Making Errors:: Generating an error message.
402 * Names & Definitions:: Names of symbols and function definitions.
403 * Lisp Interpreter:: What the Lisp interpreter does.
404 * Evaluation:: Running a program.
405 * Variables:: Returning a value from a variable.
406 * Arguments:: Passing information to a function.
407 * set & setq:: Setting the value of a variable.
408 * Summary:: The major points.
409 * Error Message Exercises::
413 * Numbers Lists:: List have numbers, other lists, in them.
414 * Lisp Atoms:: Elemental entities.
415 * Whitespace in Lists:: Formatting lists to be readable.
416 * Typing Lists:: How GNU Emacs helps you type lists.
420 * Complications:: Variables, Special forms, Lists within.
421 * Byte Compiling:: Specially processing code for speed.
425 * How the Interpreter Acts:: Returns and Side Effects...
426 * Evaluating Inner Lists:: Lists within lists...
430 * fill-column Example::
431 * Void Function:: The error message for a symbol
433 * Void Variable:: The error message for a symbol without a value.
437 * Data types:: Types of data passed to a function.
438 * Args as Variable or List:: An argument can be the value
439 of a variable or list.
440 * Variable Number of Arguments:: Some functions may take a
441 variable number of arguments.
442 * Wrong Type of Argument:: Passing an argument of the wrong type
444 * message:: A useful function for sending messages.
446 Setting the Value of a Variable
448 * Using set:: Setting values.
449 * Using setq:: Setting a quoted value.
450 * Counting:: Using @code{setq} to count.
452 Practicing Evaluation
454 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
456 * Buffer Names:: Buffers and files are different.
457 * Getting Buffers:: Getting a buffer itself, not merely its name.
458 * Switching Buffers:: How to change to another buffer.
459 * Buffer Size & Locations:: Where point is located and the size of
461 * Evaluation Exercise::
463 How To Write Function Definitions
465 * Primitive Functions::
466 * defun:: The @code{defun} macro.
467 * Install:: Install a function definition.
468 * Interactive:: Making a function interactive.
469 * Interactive Options:: Different options for @code{interactive}.
470 * Permanent Installation:: Installing code permanently.
471 * let:: Creating and initializing local variables.
473 * else:: If--then--else expressions.
474 * Truth & Falsehood:: What Lisp considers false and true.
475 * save-excursion:: Keeping track of point, mark, and buffer.
479 Install a Function Definition
481 * Effect of installation::
482 * Change a defun:: How to change a function definition.
484 Make a Function Interactive
486 * Interactive multiply-by-seven:: An overview.
487 * multiply-by-seven in detail:: The interactive version.
491 * Prevent confusion::
492 * Parts of let Expression::
493 * Sample let Expression::
494 * Uninitialized let Variables::
496 The @code{if} Special Form
498 * if in more detail::
499 * type-of-animal in detail:: An example of an @code{if} expression.
501 Truth and Falsehood in Emacs Lisp
503 * nil explained:: @code{nil} has two meanings.
505 @code{save-excursion}
507 * Point and mark:: A review of various locations.
508 * Template for save-excursion::
510 A Few Buffer--Related Functions
512 * Finding More:: How to find more information.
513 * simplified-beginning-of-buffer:: Shows @code{goto-char},
514 @code{point-min}, and @code{push-mark}.
515 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
516 * append-to-buffer:: Uses @code{save-excursion} and
517 @code{insert-buffer-substring}.
518 * Buffer Related Review:: Review.
521 The Definition of @code{mark-whole-buffer}
523 * mark-whole-buffer overview::
524 * Body of mark-whole-buffer:: Only three lines of code.
526 The Definition of @code{append-to-buffer}
528 * append-to-buffer overview::
529 * append interactive:: A two part interactive expression.
530 * append-to-buffer body:: Incorporates a @code{let} expression.
531 * append save-excursion:: How the @code{save-excursion} works.
533 A Few More Complex Functions
535 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
536 * insert-buffer:: Read-only, and with @code{or}.
537 * beginning-of-buffer:: Shows @code{goto-char},
538 @code{point-min}, and @code{push-mark}.
539 * Second Buffer Related Review::
540 * optional Exercise::
542 The Definition of @code{insert-buffer}
544 * insert-buffer code::
545 * insert-buffer interactive:: When you can read, but not write.
546 * insert-buffer body:: The body has an @code{or} and a @code{let}.
547 * if & or:: Using an @code{if} instead of an @code{or}.
548 * Insert or:: How the @code{or} expression works.
549 * Insert let:: Two @code{save-excursion} expressions.
550 * New insert-buffer::
552 The Interactive Expression in @code{insert-buffer}
554 * Read-only buffer:: When a buffer cannot be modified.
555 * b for interactive:: An existing buffer or else its name.
557 Complete Definition of @code{beginning-of-buffer}
559 * Optional Arguments::
560 * beginning-of-buffer opt arg:: Example with optional argument.
561 * beginning-of-buffer complete::
563 @code{beginning-of-buffer} with an Argument
565 * Disentangle beginning-of-buffer::
566 * Large buffer case::
567 * Small buffer case::
569 Narrowing and Widening
571 * Narrowing advantages:: The advantages of narrowing
572 * save-restriction:: The @code{save-restriction} special form.
573 * what-line:: The number of the line that point is on.
576 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
578 * Strange Names:: An historical aside: why the strange names?
579 * car & cdr:: Functions for extracting part of a list.
580 * cons:: Constructing a list.
581 * nthcdr:: Calling @code{cdr} repeatedly.
583 * setcar:: Changing the first element of a list.
584 * setcdr:: Changing the rest of a list.
590 * length:: How to find the length of a list.
592 Cutting and Storing Text
594 * Storing Text:: Text is stored in a list.
595 * zap-to-char:: Cutting out text up to a character.
596 * kill-region:: Cutting text out of a region.
597 * copy-region-as-kill:: A definition for copying text.
598 * Digression into C:: Minor note on C programming language macros.
599 * defvar:: How to give a variable an initial value.
600 * cons & search-fwd Review::
605 * Complete zap-to-char:: The complete implementation.
606 * zap-to-char interactive:: A three part interactive expression.
607 * zap-to-char body:: A short overview.
608 * search-forward:: How to search for a string.
609 * progn:: The @code{progn} special form.
610 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
614 * Complete kill-region:: The function definition.
615 * condition-case:: Dealing with a problem.
618 @code{copy-region-as-kill}
620 * Complete copy-region-as-kill:: The complete function definition.
621 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
623 The Body of @code{copy-region-as-kill}
625 * last-command & this-command::
626 * kill-append function::
627 * kill-new function::
629 Initializing a Variable with @code{defvar}
631 * See variable current value::
632 * defvar and asterisk::
634 How Lists are Implemented
637 * Symbols as Chest:: Exploring a powerful metaphor.
642 * Kill Ring Overview::
643 * kill-ring-yank-pointer:: The kill ring is a list.
644 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
648 * while:: Causing a stretch of code to repeat.
650 * Recursion:: Causing a function to call itself.
655 * Looping with while:: Repeat so long as test returns true.
656 * Loop Example:: A @code{while} loop that uses a list.
657 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
658 * Incrementing Loop:: A loop with an incrementing counter.
659 * Incrementing Loop Details::
660 * Decrementing Loop:: A loop with a decrementing counter.
662 Details of an Incrementing Loop
664 * Incrementing Example:: Counting pebbles in a triangle.
665 * Inc Example parts:: The parts of the function definition.
666 * Inc Example altogether:: Putting the function definition together.
668 Loop with a Decrementing Counter
670 * Decrementing Example:: More pebbles on the beach.
671 * Dec Example parts:: The parts of the function definition.
672 * Dec Example altogether:: Putting the function definition together.
674 Save your time: @code{dolist} and @code{dotimes}
681 * Building Robots:: Same model, different serial number ...
682 * Recursive Definition Parts:: Walk until you stop ...
683 * Recursion with list:: Using a list as the test whether to recurse.
684 * Recursive triangle function::
685 * Recursion with cond::
686 * Recursive Patterns:: Often used templates.
687 * No Deferment:: Don't store up work ...
688 * No deferment solution::
690 Recursion in Place of a Counter
692 * Recursive Example arg of 1 or 2::
693 * Recursive Example arg of 3 or 4::
701 Regular Expression Searches
703 * sentence-end:: The regular expression for @code{sentence-end}.
704 * re-search-forward:: Very similar to @code{search-forward}.
705 * forward-sentence:: A straightforward example of regexp search.
706 * forward-paragraph:: A somewhat complex example.
707 * etags:: How to create your own @file{TAGS} table.
709 * re-search Exercises::
711 @code{forward-sentence}
713 * Complete forward-sentence::
714 * fwd-sentence while loops:: Two @code{while} loops.
715 * fwd-sentence re-search:: A regular expression search.
717 @code{forward-paragraph}: a Goldmine of Functions
719 * forward-paragraph in brief:: Key parts of the function definition.
720 * fwd-para let:: The @code{let*} expression.
721 * fwd-para while:: The forward motion @code{while} loop.
723 Counting: Repetition and Regexps
726 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
727 * recursive-count-words:: Start with case of no words in region.
728 * Counting Exercise::
730 The @code{@value{COUNT-WORDS}} Function
732 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
733 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
735 Counting Words in a @code{defun}
737 * Divide and Conquer::
738 * Words and Symbols:: What to count?
739 * Syntax:: What constitutes a word or symbol?
740 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
741 * Several defuns:: Counting several defuns in a file.
742 * Find a File:: Do you want to look at a file?
743 * lengths-list-file:: A list of the lengths of many definitions.
744 * Several files:: Counting in definitions in different files.
745 * Several files recursively:: Recursively counting in different files.
746 * Prepare the data:: Prepare the data for display in a graph.
748 Count Words in @code{defuns} in Different Files
750 * lengths-list-many-files:: Return a list of the lengths of defuns.
751 * append:: Attach one list to another.
753 Prepare the Data for Display in a Graph
755 * Data for Display in Detail::
756 * Sorting:: Sorting lists.
757 * Files List:: Making a list of files.
758 * Counting function definitions::
762 * Columns of a graph::
763 * graph-body-print:: How to print the body of a graph.
764 * recursive-graph-body-print::
766 * Line Graph Exercise::
768 Your @file{.emacs} File
770 * Default Configuration::
771 * Site-wide Init:: You can write site-wide init files.
772 * defcustom:: Emacs will write code for you.
773 * Beginning a .emacs File:: How to write a @code{.emacs file}.
774 * Text and Auto-fill:: Automatically wrap lines.
775 * Mail Aliases:: Use abbreviations for email addresses.
776 * Indent Tabs Mode:: Don't use tabs with @TeX{}
777 * Keybindings:: Create some personal keybindings.
778 * Keymaps:: More about key binding.
779 * Loading Files:: Load (i.e., evaluate) files automatically.
780 * Autoload:: Make functions available.
781 * Simple Extension:: Define a function; bind it to a key.
782 * X11 Colors:: Colors in X.
784 * Mode Line:: How to customize your mode line.
788 * debug:: How to use the built-in debugger.
789 * debug-on-entry:: Start debugging when you call a function.
790 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
791 * edebug:: How to use Edebug, a source level debugger.
792 * Debugging Exercises::
794 Handling the Kill Ring
796 * What the Kill Ring Does::
798 * yank:: Paste a copy of a clipped element.
799 * yank-pop:: Insert element pointed to.
802 The @code{current-kill} Function
804 * Code for current-kill::
805 * Understanding current-kill::
807 @code{current-kill} in Outline
809 * Body of current-kill::
810 * Digression concerning error:: How to mislead humans, but not computers.
811 * Determining the Element::
813 A Graph with Labeled Axes
816 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
817 * print-Y-axis:: Print a label for the vertical axis.
818 * print-X-axis:: Print a horizontal label.
819 * Print Whole Graph:: The function to print a complete graph.
821 The @code{print-Y-axis} Function
823 * print-Y-axis in Detail::
824 * Height of label:: What height for the Y axis?
825 * Compute a Remainder:: How to compute the remainder of a division.
826 * Y Axis Element:: Construct a line for the Y axis.
827 * Y-axis-column:: Generate a list of Y axis labels.
828 * print-Y-axis Penultimate:: A not quite final version.
830 The @code{print-X-axis} Function
832 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
833 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
835 Printing the Whole Graph
837 * The final version:: A few changes.
838 * Test print-graph:: Run a short test.
839 * Graphing words in defuns:: Executing the final code.
840 * lambda:: How to write an anonymous function.
841 * mapcar:: Apply a function to elements of a list.
842 * Another Bug:: Yet another bug @dots{} most insidious.
843 * Final printed graph:: The graph itself!
851 Most of the GNU Emacs integrated environment is written in the programming
852 language called Emacs Lisp. The code written in this programming
853 language is the software---the sets of instructions---that tell the
854 computer what to do when you give it commands. Emacs is designed so
855 that you can write new code in Emacs Lisp and easily install it as an
856 extension to the editor.
858 (GNU Emacs is sometimes called an ``extensible editor'', but it does
859 much more than provide editing capabilities. It is better to refer to
860 Emacs as an ``extensible computing environment''. However, that
861 phrase is quite a mouthful. It is easier to refer to Emacs simply as
862 an editor. Moreover, everything you do in Emacs---find the Mayan date
863 and phases of the moon, simplify polynomials, debug code, manage
864 files, read letters, write books---all these activities are kinds of
865 editing in the most general sense of the word.)
868 * Why:: Why learn Emacs Lisp?
869 * On Reading this Text:: Read, gain familiarity, pick up habits....
870 * Who You Are:: For whom this is written.
872 * Note for Novices:: You can read this as a novice.
878 @unnumberedsec Why Study Emacs Lisp?
881 Although Emacs Lisp is usually thought of in association only with Emacs,
882 it is a full computer programming language. You can use Emacs Lisp as
883 you would any other programming language.
885 Perhaps you want to understand programming; perhaps you want to extend
886 Emacs; or perhaps you want to become a programmer. This introduction to
887 Emacs Lisp is designed to get you started: to guide you in learning the
888 fundamentals of programming, and more importantly, to show you how you
889 can teach yourself to go further.
891 @node On Reading this Text
892 @unnumberedsec On Reading this Text
894 All through this document, you will see little sample programs you can
895 run inside of Emacs. If you read this document in Info inside of GNU
896 Emacs, you can run the programs as they appear. (This is easy to do and
897 is explained when the examples are presented.) Alternatively, you can
898 read this introduction as a printed book while sitting beside a computer
899 running Emacs. (This is what I like to do; I like printed books.) If
900 you don't have a running Emacs beside you, you can still read this book,
901 but in this case, it is best to treat it as a novel or as a travel guide
902 to a country not yet visited: interesting, but not the same as being
905 Much of this introduction is dedicated to walkthroughs or guided tours
906 of code used in GNU Emacs. These tours are designed for two purposes:
907 first, to give you familiarity with real, working code (code you use
908 every day); and, second, to give you familiarity with the way Emacs
909 works. It is interesting to see how a working environment is
912 hope that you will pick up the habit of browsing through source code.
913 You can learn from it and mine it for ideas. Having GNU Emacs is like
914 having a dragon's cave of treasures.
916 In addition to learning about Emacs as an editor and Emacs Lisp as a
917 programming language, the examples and guided tours will give you an
918 opportunity to get acquainted with Emacs as a Lisp programming
919 environment. GNU Emacs supports programming and provides tools that
920 you will want to become comfortable using, such as @kbd{M-.} (the key
921 which invokes the @code{find-tag} command). You will also learn about
922 buffers and other objects that are part of the environment.
923 Learning about these features of Emacs is like learning new routes
924 around your home town.
927 In addition, I have written several programs as extended examples.
928 Although these are examples, the programs are real. I use them.
929 Other people use them. You may use them. Beyond the fragments of
930 programs used for illustrations, there is very little in here that is
931 `just for teaching purposes'; what you see is used. This is a great
932 advantage of Emacs Lisp: it is easy to learn to use it for work.
935 Finally, I hope to convey some of the skills for using Emacs to
936 learn aspects of programming that you don't know. You can often use
937 Emacs to help you understand what puzzles you or to find out how to do
938 something new. This self-reliance is not only a pleasure, but an
942 @unnumberedsec For Whom This is Written
944 This text is written as an elementary introduction for people who are
945 not programmers. If you are a programmer, you may not be satisfied with
946 this primer. The reason is that you may have become expert at reading
947 reference manuals and be put off by the way this text is organized.
949 An expert programmer who reviewed this text said to me:
952 @i{I prefer to learn from reference manuals. I ``dive into'' each
953 paragraph, and ``come up for air'' between paragraphs.}
955 @i{When I get to the end of a paragraph, I assume that that subject is
956 done, finished, that I know everything I need (with the
957 possible exception of the case when the next paragraph starts talking
958 about it in more detail). I expect that a well written reference manual
959 will not have a lot of redundancy, and that it will have excellent
960 pointers to the (one) place where the information I want is.}
963 This introduction is not written for this person!
965 Firstly, I try to say everything at least three times: first, to
966 introduce it; second, to show it in context; and third, to show it in a
967 different context, or to review it.
969 Secondly, I hardly ever put all the information about a subject in one
970 place, much less in one paragraph. To my way of thinking, that imposes
971 too heavy a burden on the reader. Instead I try to explain only what
972 you need to know at the time. (Sometimes I include a little extra
973 information so you won't be surprised later when the additional
974 information is formally introduced.)
976 When you read this text, you are not expected to learn everything the
977 first time. Frequently, you need only make, as it were, a `nodding
978 acquaintance' with some of the items mentioned. My hope is that I have
979 structured the text and given you enough hints that you will be alert to
980 what is important, and concentrate on it.
982 You will need to ``dive into'' some paragraphs; there is no other way
983 to read them. But I have tried to keep down the number of such
984 paragraphs. This book is intended as an approachable hill, rather than
985 as a daunting mountain.
987 This introduction to @cite{Programming in Emacs Lisp} has a companion
990 @cite{The GNU Emacs Lisp Reference Manual}.
993 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
994 Emacs Lisp Reference Manual}.
996 The reference manual has more detail than this introduction. In the
997 reference manual, all the information about one topic is concentrated
998 in one place. You should turn to it if you are like the programmer
999 quoted above. And, of course, after you have read this
1000 @cite{Introduction}, you will find the @cite{Reference Manual} useful
1001 when you are writing your own programs.
1004 @unnumberedsec Lisp History
1005 @cindex Lisp history
1007 Lisp was first developed in the late 1950s at the Massachusetts
1008 Institute of Technology for research in artificial intelligence. The
1009 great power of the Lisp language makes it superior for other purposes as
1010 well, such as writing editor commands and integrated environments.
1014 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
1015 in the 1960s. It is somewhat inspired by Common Lisp, which became a
1016 standard in the 1980s. However, Emacs Lisp is much simpler than Common
1017 Lisp. (The standard Emacs distribution contains an optional extensions
1018 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
1020 @node Note for Novices
1021 @unnumberedsec A Note for Novices
1023 If you don't know GNU Emacs, you can still read this document
1024 profitably. However, I recommend you learn Emacs, if only to learn to
1025 move around your computer screen. You can teach yourself how to use
1026 Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This
1027 means you press and release the @key{CTRL} key and the @kbd{h} at the
1028 same time, and then press and release @kbd{t}.)
1030 Also, I often refer to one of Emacs's standard commands by listing the
1031 keys which you press to invoke the command and then giving the name of
1032 the command in parentheses, like this: @kbd{M-C-\}
1033 (@code{indent-region}). What this means is that the
1034 @code{indent-region} command is customarily invoked by typing
1035 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
1036 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
1037 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
1038 @key{META} key, @key{CTRL} key and @key{\} key all at the same time.
1039 (On many modern keyboards the @key{META} key is labeled
1041 Sometimes a combination like this is called a keychord, since it is
1042 similar to the way you play a chord on a piano. If your keyboard does
1043 not have a @key{META} key, the @key{ESC} key prefix is used in place
1044 of it. In this case, @kbd{M-C-\} means that you press and release your
1045 @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at
1046 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
1047 along with the key that is labeled @key{ALT} and, at the same time,
1048 press the @key{\} key.
1050 In addition to typing a lone keychord, you can prefix what you type
1051 with @kbd{C-u}, which is called the `universal argument'. The
1052 @kbd{C-u} keychord passes an argument to the subsequent command.
1053 Thus, to indent a region of plain text by 6 spaces, mark the region,
1054 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
1055 Emacs either passes the number 4 to the command or otherwise runs the
1056 command differently than it would otherwise.) @xref{Arguments, ,
1057 Numeric Arguments, emacs, The GNU Emacs Manual}.
1059 If you are reading this in Info using GNU Emacs, you can read through
1060 this whole document just by pressing the space bar, @key{SPC}.
1061 (To learn about Info, type @kbd{C-h i} and then select Info.)
1063 A note on terminology: when I use the word Lisp alone, I often am
1064 referring to the various dialects of Lisp in general, but when I speak
1065 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
1068 @unnumberedsec Thank You
1070 My thanks to all who helped me with this book. My especial thanks to
1071 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
1072 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.
1073 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
1074 @w{Philip Johnson} and @w{David Stampe} for their patient
1075 encouragement. My mistakes are my own.
1079 @email{bob@@gnu.org}
1082 @c ================ Beginning of main text ================
1084 @c Start main text on right-hand (verso) page
1087 \par\vfill\supereject
1090 \par\vfill\supereject
1092 \par\vfill\supereject
1094 \par\vfill\supereject
1098 @c Note: this resetting of the page number back to 1 causes TeX to gripe
1099 @c about already having seen page numbers 1-4 before (in the preface):
1100 @c pdfTeX warning (ext4): destination with the same identifier (name{1})
1101 @c has been already used, duplicate ignored
1102 @c I guess that is harmless (what happens if a later part of the text
1103 @c makes a link to something in the first 4 pages though?).
1104 @c E.g., note that the Emacs manual has a preface, but does not bother
1105 @c resetting the page numbers back to 1 after that.
1108 @evenheading @thispage @| @| @thischapter
1109 @oddheading @thissection @| @| @thispage
1113 @node List Processing
1114 @chapter List Processing
1116 To the untutored eye, Lisp is a strange programming language. In Lisp
1117 code there are parentheses everywhere. Some people even claim that
1118 the name stands for `Lots of Isolated Silly Parentheses'. But the
1119 claim is unwarranted. Lisp stands for LISt Processing, and the
1120 programming language handles @emph{lists} (and lists of lists) by
1121 putting them between parentheses. The parentheses mark the boundaries
1122 of the list. Sometimes a list is preceded by a single apostrophe or
1123 quotation mark, @samp{'}@footnote{The single apostrophe or quotation
1124 mark is an abbreviation for the function @code{quote}; you need not
1125 think about functions now; functions are defined in @ref{Making
1126 Errors, , Generate an Error Message}.} Lists are the basis of Lisp.
1129 * Lisp Lists:: What are lists?
1130 * Run a Program:: Any list in Lisp is a program ready to run.
1131 * Making Errors:: Generating an error message.
1132 * Names & Definitions:: Names of symbols and function definitions.
1133 * Lisp Interpreter:: What the Lisp interpreter does.
1134 * Evaluation:: Running a program.
1135 * Variables:: Returning a value from a variable.
1136 * Arguments:: Passing information to a function.
1137 * set & setq:: Setting the value of a variable.
1138 * Summary:: The major points.
1139 * Error Message Exercises::
1146 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1147 This list is preceded by a single apostrophe. It could just as well be
1148 written as follows, which looks more like the kind of list you are likely
1149 to be familiar with:
1161 The elements of this list are the names of the four different flowers,
1162 separated from each other by whitespace and surrounded by parentheses,
1163 like flowers in a field with a stone wall around them.
1164 @cindex Flowers in a field
1167 * Numbers Lists:: List have numbers, other lists, in them.
1168 * Lisp Atoms:: Elemental entities.
1169 * Whitespace in Lists:: Formatting lists to be readable.
1170 * Typing Lists:: How GNU Emacs helps you type lists.
1175 @unnumberedsubsec Numbers, Lists inside of Lists
1178 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1179 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1180 separated by whitespace.
1182 In Lisp, both data and programs are represented the same way; that is,
1183 they are both lists of words, numbers, or other lists, separated by
1184 whitespace and surrounded by parentheses. (Since a program looks like
1185 data, one program may easily serve as data for another; this is a very
1186 powerful feature of Lisp.) (Incidentally, these two parenthetical
1187 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1188 @samp{.} as punctuation marks.)
1191 Here is another list, this time with a list inside of it:
1194 '(this list has (a list inside of it))
1197 The components of this list are the words @samp{this}, @samp{list},
1198 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1199 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1200 @samp{of}, @samp{it}.
1203 @subsection Lisp Atoms
1206 In Lisp, what we have been calling words are called @dfn{atoms}. This
1207 term comes from the historical meaning of the word atom, which means
1208 `indivisible'. As far as Lisp is concerned, the words we have been
1209 using in the lists cannot be divided into any smaller parts and still
1210 mean the same thing as part of a program; likewise with numbers and
1211 single character symbols like @samp{+}. On the other hand, unlike an
1212 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1213 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1215 In a list, atoms are separated from each other by whitespace. They can be
1216 right next to a parenthesis.
1218 @cindex @samp{empty list} defined
1219 Technically speaking, a list in Lisp consists of parentheses surrounding
1220 atoms separated by whitespace or surrounding other lists or surrounding
1221 both atoms and other lists. A list can have just one atom in it or
1222 have nothing in it at all. A list with nothing in it looks like this:
1223 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1224 empty list is considered both an atom and a list at the same time.
1226 @cindex Symbolic expressions, introduced
1227 @cindex @samp{expression} defined
1228 @cindex @samp{form} defined
1229 The printed representation of both atoms and lists are called
1230 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1231 The word @dfn{expression} by itself can refer to either the printed
1232 representation, or to the atom or list as it is held internally in the
1233 computer. Often, people use the term @dfn{expression}
1234 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1235 as a synonym for expression.)
1237 Incidentally, the atoms that make up our universe were named such when
1238 they were thought to be indivisible; but it has been found that physical
1239 atoms are not indivisible. Parts can split off an atom or it can
1240 fission into two parts of roughly equal size. Physical atoms were named
1241 prematurely, before their truer nature was found. In Lisp, certain
1242 kinds of atom, such as an array, can be separated into parts; but the
1243 mechanism for doing this is different from the mechanism for splitting a
1244 list. As far as list operations are concerned, the atoms of a list are
1247 As in English, the meanings of the component letters of a Lisp atom
1248 are different from the meaning the letters make as a word. For
1249 example, the word for the South American sloth, the @samp{ai}, is
1250 completely different from the two words, @samp{a}, and @samp{i}.
1252 There are many kinds of atom in nature but only a few in Lisp: for
1253 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1254 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1255 listed in the examples above are all symbols. In everyday Lisp
1256 conversation, the word ``atom'' is not often used, because programmers
1257 usually try to be more specific about what kind of atom they are dealing
1258 with. Lisp programming is mostly about symbols (and sometimes numbers)
1259 within lists. (Incidentally, the preceding three word parenthetical
1260 remark is a proper list in Lisp, since it consists of atoms, which in
1261 this case are symbols, separated by whitespace and enclosed by
1262 parentheses, without any non-Lisp punctuation.)
1265 Text between double quotation marks---even sentences or
1266 paragraphs---is also an atom. Here is an example:
1267 @cindex Text between double quotation marks
1270 '(this list includes "text between quotation marks.")
1273 @cindex @samp{string} defined
1275 In Lisp, all of the quoted text including the punctuation mark and the
1276 blank spaces is a single atom. This kind of atom is called a
1277 @dfn{string} (for `string of characters') and is the sort of thing that
1278 is used for messages that a computer can print for a human to read.
1279 Strings are a different kind of atom than numbers or symbols and are
1282 @node Whitespace in Lists
1283 @subsection Whitespace in Lists
1284 @cindex Whitespace in lists
1287 The amount of whitespace in a list does not matter. From the point of view
1288 of the Lisp language,
1299 is exactly the same as this:
1302 '(this list looks like this)
1305 Both examples show what to Lisp is the same list, the list made up of
1306 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1307 @samp{this} in that order.
1309 Extra whitespace and newlines are designed to make a list more readable
1310 by humans. When Lisp reads the expression, it gets rid of all the extra
1311 whitespace (but it needs to have at least one space between atoms in
1312 order to tell them apart.)
1314 Odd as it seems, the examples we have seen cover almost all of what Lisp
1315 lists look like! Every other list in Lisp looks more or less like one
1316 of these examples, except that the list may be longer and more complex.
1317 In brief, a list is between parentheses, a string is between quotation
1318 marks, a symbol looks like a word, and a number looks like a number.
1319 (For certain situations, square brackets, dots and a few other special
1320 characters may be used; however, we will go quite far without them.)
1323 @subsection GNU Emacs Helps You Type Lists
1324 @cindex Help typing lists
1325 @cindex Formatting help
1327 When you type a Lisp expression in GNU Emacs using either Lisp
1328 Interaction mode or Emacs Lisp mode, you have available to you several
1329 commands to format the Lisp expression so it is easy to read. For
1330 example, pressing the @key{TAB} key automatically indents the line the
1331 cursor is on by the right amount. A command to properly indent the
1332 code in a region is customarily bound to @kbd{M-C-\}. Indentation is
1333 designed so that you can see which elements of a list belong to which
1334 list---elements of a sub-list are indented more than the elements of
1337 In addition, when you type a closing parenthesis, Emacs momentarily
1338 jumps the cursor back to the matching opening parenthesis, so you can
1339 see which one it is. This is very useful, since every list you type
1340 in Lisp must have its closing parenthesis match its opening
1341 parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs
1342 Manual}, for more information about Emacs's modes.)
1345 @section Run a Program
1346 @cindex Run a program
1347 @cindex Program, running one
1349 @cindex @samp{evaluate} defined
1350 A list in Lisp---any list---is a program ready to run. If you run it
1351 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1352 of three things: do nothing except return to you the list itself; send
1353 you an error message; or, treat the first symbol in the list as a
1354 command to do something. (Usually, of course, it is the last of these
1355 three things that you really want!)
1357 @c use code for the single apostrophe, not samp.
1358 The single apostrophe, @code{'}, that I put in front of some of the
1359 example lists in preceding sections is called a @dfn{quote}; when it
1360 precedes a list, it tells Lisp to do nothing with the list, other than
1361 take it as it is written. But if there is no quote preceding a list,
1362 the first item of the list is special: it is a command for the computer
1363 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1364 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1365 understands that the @code{+} is an instruction to do something with the
1366 rest of the list: add the numbers that follow.
1369 If you are reading this inside of GNU Emacs in Info, here is how you can
1370 evaluate such a list: place your cursor immediately after the right
1371 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1377 @c use code for the number four, not samp.
1379 You will see the number @code{4} appear in the echo area. (In the
1380 jargon, what you have just done is ``evaluate the list.'' The echo area
1381 is the line at the bottom of the screen that displays or ``echoes''
1382 text.) Now try the same thing with a quoted list: place the cursor
1383 right after the following list and type @kbd{C-x C-e}:
1386 '(this is a quoted list)
1390 You will see @code{(this is a quoted list)} appear in the echo area.
1392 @cindex Lisp interpreter, explained
1393 @cindex Interpreter, Lisp, explained
1394 In both cases, what you are doing is giving a command to the program
1395 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1396 interpreter a command to evaluate the expression. The name of the Lisp
1397 interpreter comes from the word for the task done by a human who comes
1398 up with the meaning of an expression---who ``interprets'' it.
1400 You can also evaluate an atom that is not part of a list---one that is
1401 not surrounded by parentheses; again, the Lisp interpreter translates
1402 from the humanly readable expression to the language of the computer.
1403 But before discussing this (@pxref{Variables}), we will discuss what the
1404 Lisp interpreter does when you make an error.
1407 @section Generate an Error Message
1408 @cindex Generate an error message
1409 @cindex Error message generation
1411 Partly so you won't worry if you do it accidentally, we will now give
1412 a command to the Lisp interpreter that generates an error message.
1413 This is a harmless activity; and indeed, we will often try to generate
1414 error messages intentionally. Once you understand the jargon, error
1415 messages can be informative. Instead of being called ``error''
1416 messages, they should be called ``help'' messages. They are like
1417 signposts to a traveler in a strange country; deciphering them can be
1418 hard, but once understood, they can point the way.
1420 The error message is generated by a built-in GNU Emacs debugger. We
1421 will `enter the debugger'. You get out of the debugger by typing @code{q}.
1423 What we will do is evaluate a list that is not quoted and does not
1424 have a meaningful command as its first element. Here is a list almost
1425 exactly the same as the one we just used, but without the single-quote
1426 in front of it. Position the cursor right after it and type @kbd{C-x
1430 (this is an unquoted list)
1435 What you see depends on which version of Emacs you are running. GNU
1436 Emacs version 22 provides more information than version 20 and before.
1437 First, the more recent result of generating an error; then the
1438 earlier, version 20 result.
1442 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1443 you will see the following in it:
1446 A @file{*Backtrace*} window will open up and you should see the
1451 ---------- Buffer: *Backtrace* ----------
1452 Debugger entered--Lisp error: (void-function this)
1453 (this is an unquoted list)
1454 eval((this is an unquoted list))
1455 eval-last-sexp-1(nil)
1457 call-interactively(eval-last-sexp)
1458 ---------- Buffer: *Backtrace* ----------
1464 Your cursor will be in this window (you may have to wait a few seconds
1465 before it becomes visible). To quit the debugger and make the
1466 debugger window go away, type:
1473 Please type @kbd{q} right now, so you become confident that you can
1474 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1477 @cindex @samp{function} defined
1478 Based on what we already know, we can almost read this error message.
1480 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1481 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1482 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1483 an abbreviation for `evaluate' and @code{sexp} is an abbreviation for
1484 `symbolic expression'. The command means `evaluate last symbolic
1485 expression', which is the expression just before your cursor.
1487 Each line above tells you what the Lisp interpreter evaluated next.
1488 The most recent action is at the top. The buffer is called the
1489 @file{*Backtrace*} buffer because it enables you to track Emacs
1493 At the top of the @file{*Backtrace*} buffer, you see the line:
1496 Debugger entered--Lisp error: (void-function this)
1500 The Lisp interpreter tried to evaluate the first atom of the list, the
1501 word @samp{this}. It is this action that generated the error message
1502 @samp{void-function this}.
1504 The message contains the words @samp{void-function} and @samp{this}.
1506 @cindex @samp{function} defined
1507 The word @samp{function} was mentioned once before. It is a very
1508 important word. For our purposes, we can define it by saying that a
1509 @dfn{function} is a set of instructions to the computer that tell the
1510 computer to do something.
1512 Now we can begin to understand the error message: @samp{void-function
1513 this}. The function (that is, the word @samp{this}) does not have a
1514 definition of any set of instructions for the computer to carry out.
1516 The slightly odd word, @samp{void-function}, is designed to cover the
1517 way Emacs Lisp is implemented, which is that when a symbol does not
1518 have a function definition attached to it, the place that should
1519 contain the instructions is `void'.
1521 On the other hand, since we were able to add 2 plus 2 successfully, by
1522 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1523 have a set of instructions for the computer to obey and those
1524 instructions must be to add the numbers that follow the @code{+}.
1526 It is possible to prevent Emacs entering the debugger in cases like
1527 this. We do not explain how to do that here, but we will mention what
1528 the result looks like, because you may encounter a similar situation
1529 if there is a bug in some Emacs code that you are using. In such
1530 cases, you will see only one line of error message; it will appear in
1531 the echo area and look like this:
1534 Symbol's function definition is void:@: this
1539 (Also, your terminal may beep at you---some do, some don't; and others
1540 blink. This is just a device to get your attention.)
1542 The message goes away as soon as you type a key, even just to
1545 We know the meaning of the word @samp{Symbol}. It refers to the first
1546 atom of the list, the word @samp{this}. The word @samp{function}
1547 refers to the instructions that tell the computer what to do.
1548 (Technically, the symbol tells the computer where to find the
1549 instructions, but this is a complication we can ignore for the
1552 The error message can be understood: @samp{Symbol's function
1553 definition is void:@: this}. The symbol (that is, the word
1554 @samp{this}) lacks instructions for the computer to carry out.
1556 @node Names & Definitions
1557 @section Symbol Names and Function Definitions
1558 @cindex Symbol names
1560 We can articulate another characteristic of Lisp based on what we have
1561 discussed so far---an important characteristic: a symbol, like
1562 @code{+}, is not itself the set of instructions for the computer to
1563 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1564 of locating the definition or set of instructions. What we see is the
1565 name through which the instructions can be found. Names of people
1566 work the same way. I can be referred to as @samp{Bob}; however, I am
1567 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1568 consciousness consistently associated with a particular life-form.
1569 The name is not me, but it can be used to refer to me.
1571 In Lisp, one set of instructions can be attached to several names.
1572 For example, the computer instructions for adding numbers can be
1573 linked to the symbol @code{plus} as well as to the symbol @code{+}
1574 (and are in some dialects of Lisp). Among humans, I can be referred
1575 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1577 On the other hand, a symbol can have only one function definition
1578 attached to it at a time. Otherwise, the computer would be confused as
1579 to which definition to use. If this were the case among people, only
1580 one person in the world could be named @samp{Bob}. However, the function
1581 definition to which the name refers can be changed readily.
1582 (@xref{Install, , Install a Function Definition}.)
1584 Since Emacs Lisp is large, it is customary to name symbols in a way
1585 that identifies the part of Emacs to which the function belongs.
1586 Thus, all the names for functions that deal with Texinfo start with
1587 @samp{texinfo-} and those for functions that deal with reading mail
1588 start with @samp{rmail-}.
1590 @node Lisp Interpreter
1591 @section The Lisp Interpreter
1592 @cindex Lisp interpreter, what it does
1593 @cindex Interpreter, what it does
1595 Based on what we have seen, we can now start to figure out what the
1596 Lisp interpreter does when we command it to evaluate a list.
1597 First, it looks to see whether there is a quote before the list; if
1598 there is, the interpreter just gives us the list. On the other
1599 hand, if there is no quote, the interpreter looks at the first element
1600 in the list and sees whether it has a function definition. If it does,
1601 the interpreter carries out the instructions in the function definition.
1602 Otherwise, the interpreter prints an error message.
1604 This is how Lisp works. Simple. There are added complications which we
1605 will get to in a minute, but these are the fundamentals. Of course, to
1606 write Lisp programs, you need to know how to write function definitions
1607 and attach them to names, and how to do this without confusing either
1608 yourself or the computer.
1611 * Complications:: Variables, Special forms, Lists within.
1612 * Byte Compiling:: Specially processing code for speed.
1617 @unnumberedsubsec Complications
1620 Now, for the first complication. In addition to lists, the Lisp
1621 interpreter can evaluate a symbol that is not quoted and does not have
1622 parentheses around it. The Lisp interpreter will attempt to determine
1623 the symbol's value as a @dfn{variable}. This situation is described
1624 in the section on variables. (@xref{Variables}.)
1626 @cindex Special form
1627 The second complication occurs because some functions are unusual and
1628 do not work in the usual manner. Those that don't are called
1629 @dfn{special forms}. They are used for special jobs, like defining a
1630 function, and there are not many of them. In the next few chapters,
1631 you will be introduced to several of the more important special forms.
1633 As well as special forms, there are also @dfn{macros}. A macro
1634 is a construct defined in Lisp, which differs from a function in that it
1635 translates a Lisp expression into another expression that is to be
1636 evaluated in place of the original expression. (@xref{Lisp macro}.)
1638 For the purposes of this introduction, you do not need to worry too much
1639 about whether something is a special form, macro, or ordinary function.
1640 For example, @code{if} is a special form (@pxref{if}), but @code{when}
1641 is a macro (@pxref{Lisp macro}). In earlier versions of Emacs,
1642 @code{defun} was a special form, but now it is a macro (@pxref{defun}).
1643 It still behaves in the same way.
1645 The final complication is this: if the function that the
1646 Lisp interpreter is looking at is not a special form, and if it is part
1647 of a list, the Lisp interpreter looks to see whether the list has a list
1648 inside of it. If there is an inner list, the Lisp interpreter first
1649 figures out what it should do with the inside list, and then it works on
1650 the outside list. If there is yet another list embedded inside the
1651 inner list, it works on that one first, and so on. It always works on
1652 the innermost list first. The interpreter works on the innermost list
1653 first, to evaluate the result of that list. The result may be
1654 used by the enclosing expression.
1656 Otherwise, the interpreter works left to right, from one expression to
1659 @node Byte Compiling
1660 @subsection Byte Compiling
1661 @cindex Byte compiling
1663 One other aspect of interpreting: the Lisp interpreter is able to
1664 interpret two kinds of entity: humanly readable code, on which we will
1665 focus exclusively, and specially processed code, called @dfn{byte
1666 compiled} code, which is not humanly readable. Byte compiled code
1667 runs faster than humanly readable code.
1669 You can transform humanly readable code into byte compiled code by
1670 running one of the compile commands such as @code{byte-compile-file}.
1671 Byte compiled code is usually stored in a file that ends with a
1672 @file{.elc} extension rather than a @file{.el} extension. You will
1673 see both kinds of file in the @file{emacs/lisp} directory; the files
1674 to read are those with @file{.el} extensions.
1676 As a practical matter, for most things you might do to customize or
1677 extend Emacs, you do not need to byte compile; and I will not discuss
1678 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1679 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1686 When the Lisp interpreter works on an expression, the term for the
1687 activity is called @dfn{evaluation}. We say that the interpreter
1688 `evaluates the expression'. I've used this term several times before.
1689 The word comes from its use in everyday language, `to ascertain the
1690 value or amount of; to appraise', according to @cite{Webster's New
1691 Collegiate Dictionary}.
1694 * How the Interpreter Acts:: Returns and Side Effects...
1695 * Evaluating Inner Lists:: Lists within lists...
1699 @node How the Interpreter Acts
1700 @unnumberedsubsec How the Lisp Interpreter Acts
1703 @cindex @samp{returned value} explained
1704 After evaluating an expression, the Lisp interpreter will most likely
1705 @dfn{return} the value that the computer produces by carrying out the
1706 instructions it found in the function definition, or perhaps it will
1707 give up on that function and produce an error message. (The interpreter
1708 may also find itself tossed, so to speak, to a different function or it
1709 may attempt to repeat continually what it is doing for ever and ever in
1710 what is called an `infinite loop'. These actions are less common; and
1711 we can ignore them.) Most frequently, the interpreter returns a value.
1713 @cindex @samp{side effect} defined
1714 At the same time the interpreter returns a value, it may do something
1715 else as well, such as move a cursor or copy a file; this other kind of
1716 action is called a @dfn{side effect}. Actions that we humans think are
1717 important, such as printing results, are often ``side effects'' to the
1718 Lisp interpreter. The jargon can sound peculiar, but it turns out that
1719 it is fairly easy to learn to use side effects.
1721 In summary, evaluating a symbolic expression most commonly causes the
1722 Lisp interpreter to return a value and perhaps carry out a side effect;
1723 or else produce an error.
1725 @node Evaluating Inner Lists
1726 @subsection Evaluating Inner Lists
1727 @cindex Inner list evaluation
1728 @cindex Evaluating inner lists
1730 If evaluation applies to a list that is inside another list, the outer
1731 list may use the value returned by the first evaluation as information
1732 when the outer list is evaluated. This explains why inner expressions
1733 are evaluated first: the values they return are used by the outer
1737 We can investigate this process by evaluating another addition example.
1738 Place your cursor after the following expression and type @kbd{C-x C-e}:
1745 The number 8 will appear in the echo area.
1747 What happens is that the Lisp interpreter first evaluates the inner
1748 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1749 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1750 returns the value 8. Since there are no more enclosing expressions to
1751 evaluate, the interpreter prints that value in the echo area.
1753 Now it is easy to understand the name of the command invoked by the
1754 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1755 letters @code{sexp} are an abbreviation for `symbolic expression', and
1756 @code{eval} is an abbreviation for `evaluate'. The command means
1757 `evaluate last symbolic expression'.
1759 As an experiment, you can try evaluating the expression by putting the
1760 cursor at the beginning of the next line immediately following the
1761 expression, or inside the expression.
1764 Here is another copy of the expression:
1771 If you place the cursor at the beginning of the blank line that
1772 immediately follows the expression and type @kbd{C-x C-e}, you will
1773 still get the value 8 printed in the echo area. Now try putting the
1774 cursor inside the expression. If you put it right after the next to
1775 last parenthesis (so it appears to sit on top of the last parenthesis),
1776 you will get a 6 printed in the echo area! This is because the command
1777 evaluates the expression @code{(+ 3 3)}.
1779 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1780 you will get the number itself. In Lisp, if you evaluate a number, you
1781 get the number itself---this is how numbers differ from symbols. If you
1782 evaluate a list starting with a symbol like @code{+}, you will get a
1783 value returned that is the result of the computer carrying out the
1784 instructions in the function definition attached to that name. If a
1785 symbol by itself is evaluated, something different happens, as we will
1786 see in the next section.
1792 In Emacs Lisp, a symbol can have a value attached to it just as it can
1793 have a function definition attached to it. The two are different.
1794 The function definition is a set of instructions that a computer will
1795 obey. A value, on the other hand, is something, such as number or a
1796 name, that can vary (which is why such a symbol is called a variable).
1797 The value of a symbol can be any expression in Lisp, such as a symbol,
1798 number, list, or string. A symbol that has a value is often called a
1801 A symbol can have both a function definition and a value attached to
1802 it at the same time. Or it can have just one or the other.
1803 The two are separate. This is somewhat similar
1804 to the way the name Cambridge can refer to the city in Massachusetts
1805 and have some information attached to the name as well, such as
1806 ``great programming center''.
1809 (Incidentally, in Emacs Lisp, a symbol can have two
1810 other things attached to it, too: a property list and a documentation
1811 string; these are discussed later.)
1814 Another way to think about this is to imagine a symbol as being a chest
1815 of drawers. The function definition is put in one drawer, the value in
1816 another, and so on. What is put in the drawer holding the value can be
1817 changed without affecting the contents of the drawer holding the
1818 function definition, and vice-verse.
1821 * fill-column Example::
1822 * Void Function:: The error message for a symbol
1824 * Void Variable:: The error message for a symbol without a value.
1828 @node fill-column Example
1829 @unnumberedsubsec @code{fill-column}, an Example Variable
1832 @findex fill-column, @r{an example variable}
1833 @cindex Example variable, @code{fill-column}
1834 @cindex Variable, example of, @code{fill-column}
1835 The variable @code{fill-column} illustrates a symbol with a value
1836 attached to it: in every GNU Emacs buffer, this symbol is set to some
1837 value, usually 72 or 70, but sometimes to some other value. To find the
1838 value of this symbol, evaluate it by itself. If you are reading this in
1839 Info inside of GNU Emacs, you can do this by putting the cursor after
1840 the symbol and typing @kbd{C-x C-e}:
1847 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1848 area. This is the value for which @code{fill-column} is set for me as I
1849 write this. It may be different for you in your Info buffer. Notice
1850 that the value returned as a variable is printed in exactly the same way
1851 as the value returned by a function carrying out its instructions. From
1852 the point of view of the Lisp interpreter, a value returned is a value
1853 returned. What kind of expression it came from ceases to matter once
1856 A symbol can have any value attached to it or, to use the jargon, we can
1857 @dfn{bind} the variable to a value: to a number, such as 72; to a
1858 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1859 oak)}; we can even bind a variable to a function definition.
1861 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1862 Setting the Value of a Variable}, for information about one way to do
1866 @subsection Error Message for a Symbol Without a Function
1867 @cindex Symbol without function error
1868 @cindex Error for symbol without function
1870 When we evaluated @code{fill-column} to find its value as a variable,
1871 we did not place parentheses around the word. This is because we did
1872 not intend to use it as a function name.
1874 If @code{fill-column} were the first or only element of a list, the
1875 Lisp interpreter would attempt to find the function definition
1876 attached to it. But @code{fill-column} has no function definition.
1877 Try evaluating this:
1885 You will create a @file{*Backtrace*} buffer that says:
1889 ---------- Buffer: *Backtrace* ----------
1890 Debugger entered--Lisp error: (void-function fill-column)
1893 eval-last-sexp-1(nil)
1895 call-interactively(eval-last-sexp)
1896 ---------- Buffer: *Backtrace* ----------
1901 (Remember, to quit the debugger and make the debugger window go away,
1902 type @kbd{q} in the @file{*Backtrace*} buffer.)
1906 In GNU Emacs 20 and before, you will produce an error message that says:
1909 Symbol's function definition is void:@: fill-column
1913 (The message will go away as soon as you move the cursor or type
1918 @subsection Error Message for a Symbol Without a Value
1919 @cindex Symbol without value error
1920 @cindex Error for symbol without value
1922 If you attempt to evaluate a symbol that does not have a value bound to
1923 it, you will receive an error message. You can see this by
1924 experimenting with our 2 plus 2 addition. In the following expression,
1925 put your cursor right after the @code{+}, before the first number 2,
1934 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1939 ---------- Buffer: *Backtrace* ----------
1940 Debugger entered--Lisp error: (void-variable +)
1942 eval-last-sexp-1(nil)
1944 call-interactively(eval-last-sexp)
1945 ---------- Buffer: *Backtrace* ----------
1950 (Again, you can quit the debugger by
1951 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1953 This backtrace is different from the very first error message we saw,
1954 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1955 In this case, the function does not have a value as a variable; while
1956 in the other error message, the function (the word `this') did not
1959 In this experiment with the @code{+}, what we did was cause the Lisp
1960 interpreter to evaluate the @code{+} and look for the value of the
1961 variable instead of the function definition. We did this by placing the
1962 cursor right after the symbol rather than after the parenthesis of the
1963 enclosing list as we did before. As a consequence, the Lisp interpreter
1964 evaluated the preceding s-expression, which in this case was
1967 Since @code{+} does not have a value bound to it, just the function
1968 definition, the error message reported that the symbol's value as a
1973 In GNU Emacs version 20 and before, your error message will say:
1976 Symbol's value as variable is void:@: +
1980 The meaning is the same as in GNU Emacs 22.
1986 @cindex Passing information to functions
1988 To see how information is passed to functions, let's look again at
1989 our old standby, the addition of two plus two. In Lisp, this is written
1996 If you evaluate this expression, the number 4 will appear in your echo
1997 area. What the Lisp interpreter does is add the numbers that follow
2000 @cindex @samp{argument} defined
2001 The numbers added by @code{+} are called the @dfn{arguments} of the
2002 function @code{+}. These numbers are the information that is given to
2003 or @dfn{passed} to the function.
2005 The word `argument' comes from the way it is used in mathematics and
2006 does not refer to a disputation between two people; instead it refers to
2007 the information presented to the function, in this case, to the
2008 @code{+}. In Lisp, the arguments to a function are the atoms or lists
2009 that follow the function. The values returned by the evaluation of
2010 these atoms or lists are passed to the function. Different functions
2011 require different numbers of arguments; some functions require none at
2012 all.@footnote{It is curious to track the path by which the word `argument'
2013 came to have two different meanings, one in mathematics and the other in
2014 everyday English. According to the @cite{Oxford English Dictionary},
2015 the word derives from the Latin for @samp{to make clear, prove}; thus it
2016 came to mean, by one thread of derivation, `the evidence offered as
2017 proof', which is to say, `the information offered', which led to its
2018 meaning in Lisp. But in the other thread of derivation, it came to mean
2019 `to assert in a manner against which others may make counter
2020 assertions', which led to the meaning of the word as a disputation.
2021 (Note here that the English word has two different definitions attached
2022 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
2023 have two different function definitions at the same time.)}
2026 * Data types:: Types of data passed to a function.
2027 * Args as Variable or List:: An argument can be the value
2028 of a variable or list.
2029 * Variable Number of Arguments:: Some functions may take a
2030 variable number of arguments.
2031 * Wrong Type of Argument:: Passing an argument of the wrong type
2033 * message:: A useful function for sending messages.
2037 @subsection Arguments' Data Types
2039 @cindex Types of data
2040 @cindex Arguments' data types
2042 The type of data that should be passed to a function depends on what
2043 kind of information it uses. The arguments to a function such as
2044 @code{+} must have values that are numbers, since @code{+} adds numbers.
2045 Other functions use different kinds of data for their arguments.
2049 For example, the @code{concat} function links together or unites two or
2050 more strings of text to produce a string. The arguments are strings.
2051 Concatenating the two character strings @code{abc}, @code{def} produces
2052 the single string @code{abcdef}. This can be seen by evaluating the
2056 (concat "abc" "def")
2060 The value produced by evaluating this expression is @code{"abcdef"}.
2062 A function such as @code{substring} uses both a string and numbers as
2063 arguments. The function returns a part of the string, a substring of
2064 the first argument. This function takes three arguments. Its first
2065 argument is the string of characters, the second and third arguments are
2066 numbers that indicate the beginning and end of the substring. The
2067 numbers are a count of the number of characters (including spaces and
2068 punctuation) from the beginning of the string.
2071 For example, if you evaluate the following:
2074 (substring "The quick brown fox jumped." 16 19)
2078 you will see @code{"fox"} appear in the echo area. The arguments are the
2079 string and the two numbers.
2081 Note that the string passed to @code{substring} is a single atom even
2082 though it is made up of several words separated by spaces. Lisp counts
2083 everything between the two quotation marks as part of the string,
2084 including the spaces. You can think of the @code{substring} function as
2085 a kind of `atom smasher' since it takes an otherwise indivisible atom
2086 and extracts a part. However, @code{substring} is only able to extract
2087 a substring from an argument that is a string, not from another type of
2088 atom such as a number or symbol.
2090 @node Args as Variable or List
2091 @subsection An Argument as the Value of a Variable or List
2093 An argument can be a symbol that returns a value when it is evaluated.
2094 For example, when the symbol @code{fill-column} by itself is evaluated,
2095 it returns a number. This number can be used in an addition.
2098 Position the cursor after the following expression and type @kbd{C-x
2106 The value will be a number two more than what you get by evaluating
2107 @code{fill-column} alone. For me, this is 74, because my value of
2108 @code{fill-column} is 72.
2110 As we have just seen, an argument can be a symbol that returns a value
2111 when evaluated. In addition, an argument can be a list that returns a
2112 value when it is evaluated. For example, in the following expression,
2113 the arguments to the function @code{concat} are the strings
2114 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2115 @code{(number-to-string (+ 2 fill-column))}.
2117 @c For GNU Emacs 22, need number-to-string
2119 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2123 If you evaluate this expression---and if, as with my Emacs,
2124 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2125 appear in the echo area. (Note that you must put spaces after the
2126 word @samp{The} and before the word @samp{red} so they will appear in
2127 the final string. The function @code{number-to-string} converts the
2128 integer that the addition function returns to a string.
2129 @code{number-to-string} is also known as @code{int-to-string}.)
2131 @node Variable Number of Arguments
2132 @subsection Variable Number of Arguments
2133 @cindex Variable number of arguments
2134 @cindex Arguments, variable number of
2136 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2137 number of arguments. (The @code{*} is the symbol for multiplication.)
2138 This can be seen by evaluating each of the following expressions in
2139 the usual way. What you will see in the echo area is printed in this
2140 text after @samp{@result{}}, which you may read as `evaluates to'.
2143 In the first set, the functions have no arguments:
2154 In this set, the functions have one argument each:
2165 In this set, the functions have three arguments each:
2169 (+ 3 4 5) @result{} 12
2171 (* 3 4 5) @result{} 60
2175 @node Wrong Type of Argument
2176 @subsection Using the Wrong Type Object as an Argument
2177 @cindex Wrong type of argument
2178 @cindex Argument, wrong type of
2180 When a function is passed an argument of the wrong type, the Lisp
2181 interpreter produces an error message. For example, the @code{+}
2182 function expects the values of its arguments to be numbers. As an
2183 experiment we can pass it the quoted symbol @code{hello} instead of a
2184 number. Position the cursor after the following expression and type
2192 When you do this you will generate an error message. What has happened
2193 is that @code{+} has tried to add the 2 to the value returned by
2194 @code{'hello}, but the value returned by @code{'hello} is the symbol
2195 @code{hello}, not a number. Only numbers can be added. So @code{+}
2196 could not carry out its addition.
2199 You will create and enter a @file{*Backtrace*} buffer that says:
2204 ---------- Buffer: *Backtrace* ----------
2205 Debugger entered--Lisp error:
2206 (wrong-type-argument number-or-marker-p hello)
2208 eval((+ 2 (quote hello)))
2209 eval-last-sexp-1(nil)
2211 call-interactively(eval-last-sexp)
2212 ---------- Buffer: *Backtrace* ----------
2217 As usual, the error message tries to be helpful and makes sense after you
2218 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2219 the abbreviation @code{'hello}.}
2221 The first part of the error message is straightforward; it says
2222 @samp{wrong type argument}. Next comes the mysterious jargon word
2223 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2224 kind of argument the @code{+} expected.
2226 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2227 trying to determine whether the information presented it (the value of
2228 the argument) is a number or a marker (a special object representing a
2229 buffer position). What it does is test to see whether the @code{+} is
2230 being given numbers to add. It also tests to see whether the
2231 argument is something called a marker, which is a specific feature of
2232 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2233 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2234 its position is kept as a marker. The mark can be considered a
2235 number---the number of characters the location is from the beginning
2236 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2237 numeric value of marker positions as numbers.
2239 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2240 practice started in the early days of Lisp programming. The @samp{p}
2241 stands for `predicate'. In the jargon used by the early Lisp
2242 researchers, a predicate refers to a function to determine whether some
2243 property is true or false. So the @samp{p} tells us that
2244 @code{number-or-marker-p} is the name of a function that determines
2245 whether it is true or false that the argument supplied is a number or
2246 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2247 a function that tests whether its argument has the value of zero, and
2248 @code{listp}, a function that tests whether its argument is a list.
2250 Finally, the last part of the error message is the symbol @code{hello}.
2251 This is the value of the argument that was passed to @code{+}. If the
2252 addition had been passed the correct type of object, the value passed
2253 would have been a number, such as 37, rather than a symbol like
2254 @code{hello}. But then you would not have got the error message.
2258 In GNU Emacs version 20 and before, the echo area displays an error
2262 Wrong type argument:@: number-or-marker-p, hello
2265 This says, in different words, the same as the top line of the
2266 @file{*Backtrace*} buffer.
2270 @subsection The @code{message} Function
2273 Like @code{+}, the @code{message} function takes a variable number of
2274 arguments. It is used to send messages to the user and is so useful
2275 that we will describe it here.
2278 A message is printed in the echo area. For example, you can print a
2279 message in your echo area by evaluating the following list:
2282 (message "This message appears in the echo area!")
2285 The whole string between double quotation marks is a single argument
2286 and is printed @i{in toto}. (Note that in this example, the message
2287 itself will appear in the echo area within double quotes; that is
2288 because you see the value returned by the @code{message} function. In
2289 most uses of @code{message} in programs that you write, the text will
2290 be printed in the echo area as a side-effect, without the quotes.
2291 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2292 detail}, for an example of this.)
2294 However, if there is a @samp{%s} in the quoted string of characters, the
2295 @code{message} function does not print the @samp{%s} as such, but looks
2296 to the argument that follows the string. It evaluates the second
2297 argument and prints the value at the location in the string where the
2301 You can see this by positioning the cursor after the following
2302 expression and typing @kbd{C-x C-e}:
2305 (message "The name of this buffer is: %s." (buffer-name))
2309 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2310 echo area. The function @code{buffer-name} returns the name of the
2311 buffer as a string, which the @code{message} function inserts in place
2314 To print a value as an integer, use @samp{%d} in the same way as
2315 @samp{%s}. For example, to print a message in the echo area that
2316 states the value of the @code{fill-column}, evaluate the following:
2319 (message "The value of fill-column is %d." fill-column)
2323 On my system, when I evaluate this list, @code{"The value of
2324 fill-column is 72."} appears in my echo area@footnote{Actually, you
2325 can use @code{%s} to print a number. It is non-specific. @code{%d}
2326 prints only the part of a number left of a decimal point, and not
2327 anything that is not a number.}.
2329 If there is more than one @samp{%s} in the quoted string, the value of
2330 the first argument following the quoted string is printed at the
2331 location of the first @samp{%s} and the value of the second argument is
2332 printed at the location of the second @samp{%s}, and so on.
2335 For example, if you evaluate the following,
2339 (message "There are %d %s in the office!"
2340 (- fill-column 14) "pink elephants")
2345 a rather whimsical message will appear in your echo area. On my system
2346 it says, @code{"There are 58 pink elephants in the office!"}.
2348 The expression @code{(- fill-column 14)} is evaluated and the resulting
2349 number is inserted in place of the @samp{%d}; and the string in double
2350 quotes, @code{"pink elephants"}, is treated as a single argument and
2351 inserted in place of the @samp{%s}. (That is to say, a string between
2352 double quotes evaluates to itself, like a number.)
2354 Finally, here is a somewhat complex example that not only illustrates
2355 the computation of a number, but also shows how you can use an
2356 expression within an expression to generate the text that is substituted
2361 (message "He saw %d %s"
2365 "The quick brown foxes jumped." 16 21)
2370 In this example, @code{message} has three arguments: the string,
2371 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2372 the expression beginning with the function @code{concat}. The value
2373 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2374 in place of the @samp{%d}; and the value returned by the expression
2375 beginning with @code{concat} is inserted in place of the @samp{%s}.
2377 When your fill column is 70 and you evaluate the expression, the
2378 message @code{"He saw 38 red foxes leaping."} appears in your echo
2382 @section Setting the Value of a Variable
2383 @cindex Variable, setting value
2384 @cindex Setting value of variable
2386 @cindex @samp{bind} defined
2387 There are several ways by which a variable can be given a value. One of
2388 the ways is to use either the function @code{set} or the function
2389 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2390 jargon for this process is to @dfn{bind} a variable to a value.)
2392 The following sections not only describe how @code{set} and @code{setq}
2393 work but also illustrate how arguments are passed.
2396 * Using set:: Setting values.
2397 * Using setq:: Setting a quoted value.
2398 * Counting:: Using @code{setq} to count.
2402 @subsection Using @code{set}
2405 To set the value of the symbol @code{flowers} to the list @code{'(rose
2406 violet daisy buttercup)}, evaluate the following expression by
2407 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2410 (set 'flowers '(rose violet daisy buttercup))
2414 The list @code{(rose violet daisy buttercup)} will appear in the echo
2415 area. This is what is @emph{returned} by the @code{set} function. As a
2416 side effect, the symbol @code{flowers} is bound to the list; that is,
2417 the symbol @code{flowers}, which can be viewed as a variable, is given
2418 the list as its value. (This process, by the way, illustrates how a
2419 side effect to the Lisp interpreter, setting the value, can be the
2420 primary effect that we humans are interested in. This is because every
2421 Lisp function must return a value if it does not get an error, but it
2422 will only have a side effect if it is designed to have one.)
2424 After evaluating the @code{set} expression, you can evaluate the symbol
2425 @code{flowers} and it will return the value you just set. Here is the
2426 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2433 When you evaluate @code{flowers}, the list
2434 @code{(rose violet daisy buttercup)} appears in the echo area.
2436 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2437 in front of it, what you will see in the echo area is the symbol itself,
2438 @code{flowers}. Here is the quoted symbol, so you can try this:
2444 Note also, that when you use @code{set}, you need to quote both
2445 arguments to @code{set}, unless you want them evaluated. Since we do
2446 not want either argument evaluated, neither the variable
2447 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2448 are quoted. (When you use @code{set} without quoting its first
2449 argument, the first argument is evaluated before anything else is
2450 done. If you did this and @code{flowers} did not have a value
2451 already, you would get an error message that the @samp{Symbol's value
2452 as variable is void}; on the other hand, if @code{flowers} did return
2453 a value after it was evaluated, the @code{set} would attempt to set
2454 the value that was returned. There are situations where this is the
2455 right thing for the function to do; but such situations are rare.)
2458 @subsection Using @code{setq}
2461 As a practical matter, you almost always quote the first argument to
2462 @code{set}. The combination of @code{set} and a quoted first argument
2463 is so common that it has its own name: the special form @code{setq}.
2464 This special form is just like @code{set} except that the first argument
2465 is quoted automatically, so you don't need to type the quote mark
2466 yourself. Also, as an added convenience, @code{setq} permits you to set
2467 several different variables to different values, all in one expression.
2469 To set the value of the variable @code{carnivores} to the list
2470 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2474 (setq carnivores '(lion tiger leopard))
2478 This is exactly the same as using @code{set} except the first argument
2479 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2480 means @code{quote}.)
2483 With @code{set}, the expression would look like this:
2486 (set 'carnivores '(lion tiger leopard))
2489 Also, @code{setq} can be used to assign different values to
2490 different variables. The first argument is bound to the value
2491 of the second argument, the third argument is bound to the value of the
2492 fourth argument, and so on. For example, you could use the following to
2493 assign a list of trees to the symbol @code{trees} and a list of herbivores
2494 to the symbol @code{herbivores}:
2498 (setq trees '(pine fir oak maple)
2499 herbivores '(gazelle antelope zebra))
2504 (The expression could just as well have been on one line, but it might
2505 not have fit on a page; and humans find it easier to read nicely
2508 Although I have been using the term `assign', there is another way of
2509 thinking about the workings of @code{set} and @code{setq}; and that is to
2510 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2511 list. This latter way of thinking is very common and in forthcoming
2512 chapters we shall come upon at least one symbol that has `pointer' as
2513 part of its name. The name is chosen because the symbol has a value,
2514 specifically a list, attached to it; or, expressed another way,
2515 the symbol is set to ``point'' to the list.
2518 @subsection Counting
2521 Here is an example that shows how to use @code{setq} in a counter. You
2522 might use this to count how many times a part of your program repeats
2523 itself. First set a variable to zero; then add one to the number each
2524 time the program repeats itself. To do this, you need a variable that
2525 serves as a counter, and two expressions: an initial @code{setq}
2526 expression that sets the counter variable to zero; and a second
2527 @code{setq} expression that increments the counter each time it is
2532 (setq counter 0) ; @r{Let's call this the initializer.}
2534 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2536 counter ; @r{This is the counter.}
2541 (The text following the @samp{;} are comments. @xref{Change a
2542 defun, , Change a Function Definition}.)
2544 If you evaluate the first of these expressions, the initializer,
2545 @code{(setq counter 0)}, and then evaluate the third expression,
2546 @code{counter}, the number @code{0} will appear in the echo area. If
2547 you then evaluate the second expression, the incrementer, @code{(setq
2548 counter (+ counter 1))}, the counter will get the value 1. So if you
2549 again evaluate @code{counter}, the number @code{1} will appear in the
2550 echo area. Each time you evaluate the second expression, the value of
2551 the counter will be incremented.
2553 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2554 the Lisp interpreter first evaluates the innermost list; this is the
2555 addition. In order to evaluate this list, it must evaluate the variable
2556 @code{counter} and the number @code{1}. When it evaluates the variable
2557 @code{counter}, it receives its current value. It passes this value and
2558 the number @code{1} to the @code{+} which adds them together. The sum
2559 is then returned as the value of the inner list and passed to the
2560 @code{setq} which sets the variable @code{counter} to this new value.
2561 Thus, the value of the variable, @code{counter}, is changed.
2566 Learning Lisp is like climbing a hill in which the first part is the
2567 steepest. You have now climbed the most difficult part; what remains
2568 becomes easier as you progress onwards.
2576 Lisp programs are made up of expressions, which are lists or single atoms.
2579 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2580 surrounded by parentheses. A list can be empty.
2583 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2584 character symbols like @code{+}, strings of characters between double
2585 quotation marks, or numbers.
2588 A number evaluates to itself.
2591 A string between double quotes also evaluates to itself.
2594 When you evaluate a symbol by itself, its value is returned.
2597 When you evaluate a list, the Lisp interpreter looks at the first symbol
2598 in the list and then at the function definition bound to that symbol.
2599 Then the instructions in the function definition are carried out.
2602 A single quotation mark,
2609 , tells the Lisp interpreter that it should
2610 return the following expression as written, and not evaluate it as it
2611 would if the quote were not there.
2614 Arguments are the information passed to a function. The arguments to a
2615 function are computed by evaluating the rest of the elements of the list
2616 of which the function is the first element.
2619 A function always returns a value when it is evaluated (unless it gets
2620 an error); in addition, it may also carry out some action called a
2621 ``side effect''. In many cases, a function's primary purpose is to
2622 create a side effect.
2625 @node Error Message Exercises
2628 A few simple exercises:
2632 Generate an error message by evaluating an appropriate symbol that is
2633 not within parentheses.
2636 Generate an error message by evaluating an appropriate symbol that is
2637 between parentheses.
2640 Create a counter that increments by two rather than one.
2643 Write an expression that prints a message in the echo area when
2647 @node Practicing Evaluation
2648 @chapter Practicing Evaluation
2649 @cindex Practicing evaluation
2650 @cindex Evaluation practice
2652 Before learning how to write a function definition in Emacs Lisp, it is
2653 useful to spend a little time evaluating various expressions that have
2654 already been written. These expressions will be lists with the
2655 functions as their first (and often only) element. Since some of the
2656 functions associated with buffers are both simple and interesting, we
2657 will start with those. In this section, we will evaluate a few of
2658 these. In another section, we will study the code of several other
2659 buffer-related functions, to see how they were written.
2662 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2664 * Buffer Names:: Buffers and files are different.
2665 * Getting Buffers:: Getting a buffer itself, not merely its name.
2666 * Switching Buffers:: How to change to another buffer.
2667 * Buffer Size & Locations:: Where point is located and the size of
2669 * Evaluation Exercise::
2673 @node How to Evaluate
2674 @unnumberedsec How to Evaluate
2677 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2678 command to move the cursor or to scroll the screen, @i{you are evaluating
2679 an expression,} the first element of which is a function. @i{This is
2682 @cindex @samp{interactive function} defined
2683 @cindex @samp{command} defined
2684 When you type keys, you cause the Lisp interpreter to evaluate an
2685 expression and that is how you get your results. Even typing plain text
2686 involves evaluating an Emacs Lisp function, in this case, one that uses
2687 @code{self-insert-command}, which simply inserts the character you
2688 typed. The functions you evaluate by typing keystrokes are called
2689 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2690 interactive will be illustrated in the chapter on how to write function
2691 definitions. @xref{Interactive, , Making a Function Interactive}.
2693 In addition to typing keyboard commands, we have seen a second way to
2694 evaluate an expression: by positioning the cursor after a list and
2695 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2696 section. There are other ways to evaluate an expression as well; these
2697 will be described as we come to them.
2699 Besides being used for practicing evaluation, the functions shown in the
2700 next few sections are important in their own right. A study of these
2701 functions makes clear the distinction between buffers and files, how to
2702 switch to a buffer, and how to determine a location within it.
2705 @section Buffer Names
2707 @findex buffer-file-name
2709 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2710 the difference between a file and a buffer. When you evaluate the
2711 following expression, @code{(buffer-name)}, the name of the buffer
2712 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2713 the name of the file to which the buffer refers appears in the echo
2714 area. Usually, the name returned by @code{(buffer-name)} is the same as
2715 the name of the file to which it refers, and the name returned by
2716 @code{(buffer-file-name)} is the full path-name of the file.
2718 A file and a buffer are two different entities. A file is information
2719 recorded permanently in the computer (unless you delete it). A buffer,
2720 on the other hand, is information inside of Emacs that will vanish at
2721 the end of the editing session (or when you kill the buffer). Usually,
2722 a buffer contains information that you have copied from a file; we say
2723 the buffer is @dfn{visiting} that file. This copy is what you work on
2724 and modify. Changes to the buffer do not change the file, until you
2725 save the buffer. When you save the buffer, the buffer is copied to the file
2726 and is thus saved permanently.
2729 If you are reading this in Info inside of GNU Emacs, you can evaluate
2730 each of the following expressions by positioning the cursor after it and
2731 typing @kbd{C-x C-e}.
2742 When I do this in Info, the value returned by evaluating
2743 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2744 evaluating @code{(buffer-file-name)} is @file{nil}.
2746 On the other hand, while I am writing this document, the value
2747 returned by evaluating @code{(buffer-name)} is
2748 @file{"introduction.texinfo"}, and the value returned by evaluating
2749 @code{(buffer-file-name)} is
2750 @file{"/gnu/work/intro/introduction.texinfo"}.
2752 @cindex @code{nil}, history of word
2753 The former is the name of the buffer and the latter is the name of the
2754 file. In Info, the buffer name is @file{"*info*"}. Info does not
2755 point to any file, so the result of evaluating
2756 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2757 from the Latin word for `nothing'; in this case, it means that the
2758 buffer is not associated with any file. (In Lisp, @code{nil} is also
2759 used to mean `false' and is a synonym for the empty list, @code{()}.)
2761 When I am writing, the name of my buffer is
2762 @file{"introduction.texinfo"}. The name of the file to which it
2763 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2765 (In the expressions, the parentheses tell the Lisp interpreter to
2766 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2767 functions; without the parentheses, the interpreter would attempt to
2768 evaluate the symbols as variables. @xref{Variables}.)
2770 In spite of the distinction between files and buffers, you will often
2771 find that people refer to a file when they mean a buffer and vice-verse.
2772 Indeed, most people say, ``I am editing a file,'' rather than saying,
2773 ``I am editing a buffer which I will soon save to a file.'' It is
2774 almost always clear from context what people mean. When dealing with
2775 computer programs, however, it is important to keep the distinction in mind,
2776 since the computer is not as smart as a person.
2778 @cindex Buffer, history of word
2779 The word `buffer', by the way, comes from the meaning of the word as a
2780 cushion that deadens the force of a collision. In early computers, a
2781 buffer cushioned the interaction between files and the computer's
2782 central processing unit. The drums or tapes that held a file and the
2783 central processing unit were pieces of equipment that were very
2784 different from each other, working at their own speeds, in spurts. The
2785 buffer made it possible for them to work together effectively.
2786 Eventually, the buffer grew from being an intermediary, a temporary
2787 holding place, to being the place where work is done. This
2788 transformation is rather like that of a small seaport that grew into a
2789 great city: once it was merely the place where cargo was warehoused
2790 temporarily before being loaded onto ships; then it became a business
2791 and cultural center in its own right.
2793 Not all buffers are associated with files. For example, a
2794 @file{*scratch*} buffer does not visit any file. Similarly, a
2795 @file{*Help*} buffer is not associated with any file.
2797 In the old days, when you lacked a @file{~/.emacs} file and started an
2798 Emacs session by typing the command @code{emacs} alone, without naming
2799 any files, Emacs started with the @file{*scratch*} buffer visible.
2800 Nowadays, you will see a splash screen. You can follow one of the
2801 commands suggested on the splash screen, visit a file, or press the
2802 spacebar to reach the @file{*scratch*} buffer.
2804 If you switch to the @file{*scratch*} buffer, type
2805 @code{(buffer-name)}, position the cursor after it, and then type
2806 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2807 will be returned and will appear in the echo area. @code{"*scratch*"}
2808 is the name of the buffer. When you type @code{(buffer-file-name)} in
2809 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2810 in the echo area, just as it does when you evaluate
2811 @code{(buffer-file-name)} in Info.
2813 Incidentally, if you are in the @file{*scratch*} buffer and want the
2814 value returned by an expression to appear in the @file{*scratch*}
2815 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2816 instead of @kbd{C-x C-e}. This causes the value returned to appear
2817 after the expression. The buffer will look like this:
2820 (buffer-name)"*scratch*"
2824 You cannot do this in Info since Info is read-only and it will not allow
2825 you to change the contents of the buffer. But you can do this in any
2826 buffer you can edit; and when you write code or documentation (such as
2827 this book), this feature is very useful.
2829 @node Getting Buffers
2830 @section Getting Buffers
2831 @findex current-buffer
2832 @findex other-buffer
2833 @cindex Getting a buffer
2835 The @code{buffer-name} function returns the @emph{name} of the buffer;
2836 to get the buffer @emph{itself}, a different function is needed: the
2837 @code{current-buffer} function. If you use this function in code, what
2838 you get is the buffer itself.
2840 A name and the object or entity to which the name refers are different
2841 from each other. You are not your name. You are a person to whom
2842 others refer by name. If you ask to speak to George and someone hands you
2843 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2844 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2845 not be satisfied. You do not want to speak to the name, but to the
2846 person to whom the name refers. A buffer is similar: the name of the
2847 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2848 get a buffer itself, you need to use a function such as
2849 @code{current-buffer}.
2851 However, there is a slight complication: if you evaluate
2852 @code{current-buffer} in an expression on its own, as we will do here,
2853 what you see is a printed representation of the name of the buffer
2854 without the contents of the buffer. Emacs works this way for two
2855 reasons: the buffer may be thousands of lines long---too long to be
2856 conveniently displayed; and, another buffer may have the same contents
2857 but a different name, and it is important to distinguish between them.
2860 Here is an expression containing the function:
2867 If you evaluate this expression in Info in Emacs in the usual way,
2868 @file{#<buffer *info*>} will appear in the echo area. The special
2869 format indicates that the buffer itself is being returned, rather than
2872 Incidentally, while you can type a number or symbol into a program, you
2873 cannot do that with the printed representation of a buffer: the only way
2874 to get a buffer itself is with a function such as @code{current-buffer}.
2876 A related function is @code{other-buffer}. This returns the most
2877 recently selected buffer other than the one you are in currently, not
2878 a printed representation of its name. If you have recently switched
2879 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2880 will return that buffer.
2883 You can see this by evaluating the expression:
2890 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2891 the name of whatever other buffer you switched back from most
2892 recently@footnote{Actually, by default, if the buffer from which you
2893 just switched is visible to you in another window, @code{other-buffer}
2894 will choose the most recent buffer that you cannot see; this is a
2895 subtlety that I often forget.}.
2897 @node Switching Buffers
2898 @section Switching Buffers
2899 @findex switch-to-buffer
2901 @cindex Switching to a buffer
2903 The @code{other-buffer} function actually provides a buffer when it is
2904 used as an argument to a function that requires one. We can see this
2905 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2908 But first, a brief introduction to the @code{switch-to-buffer}
2909 function. When you switched back and forth from Info to the
2910 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2911 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2912 rather, to save typing, you probably only typed @kbd{RET} if the
2913 default buffer was @file{*scratch*}, or if it was different, then you
2914 typed just part of the name, such as @code{*sc}, pressed your
2915 @kbd{TAB} key to cause it to expand to the full name, and then typed
2916 @kbd{RET}.} when prompted in the minibuffer for the name of
2917 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2918 b}, cause the Lisp interpreter to evaluate the interactive function
2919 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2920 different keystrokes call or run different functions. For example,
2921 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2922 @code{forward-sentence}, and so on.
2924 By writing @code{switch-to-buffer} in an expression, and giving it a
2925 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2929 (switch-to-buffer (other-buffer))
2933 The symbol @code{switch-to-buffer} is the first element of the list,
2934 so the Lisp interpreter will treat it as a function and carry out the
2935 instructions that are attached to it. But before doing that, the
2936 interpreter will note that @code{other-buffer} is inside parentheses
2937 and work on that symbol first. @code{other-buffer} is the first (and
2938 in this case, the only) element of this list, so the Lisp interpreter
2939 calls or runs the function. It returns another buffer. Next, the
2940 interpreter runs @code{switch-to-buffer}, passing to it, as an
2941 argument, the other buffer, which is what Emacs will switch to. If
2942 you are reading this in Info, try this now. Evaluate the expression.
2943 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2944 expression will move you to your most recent other buffer that you
2945 cannot see. If you really want to go to your most recently selected
2946 buffer, even if you can still see it, you need to evaluate the
2947 following more complex expression:
2950 (switch-to-buffer (other-buffer (current-buffer) t))
2954 In this case, the first argument to @code{other-buffer} tells it which
2955 buffer to skip---the current one---and the second argument tells
2956 @code{other-buffer} it is OK to switch to a visible buffer.
2957 In regular use, @code{switch-to-buffer} takes you to an invisible
2958 window since you would most likely use @kbd{C-x o} (@code{other-window})
2959 to go to another visible buffer.}
2961 In the programming examples in later sections of this document, you will
2962 see the function @code{set-buffer} more often than
2963 @code{switch-to-buffer}. This is because of a difference between
2964 computer programs and humans: humans have eyes and expect to see the
2965 buffer on which they are working on their computer terminals. This is
2966 so obvious, it almost goes without saying. However, programs do not
2967 have eyes. When a computer program works on a buffer, that buffer does
2968 not need to be visible on the screen.
2970 @code{switch-to-buffer} is designed for humans and does two different
2971 things: it switches the buffer to which Emacs's attention is directed; and
2972 it switches the buffer displayed in the window to the new buffer.
2973 @code{set-buffer}, on the other hand, does only one thing: it switches
2974 the attention of the computer program to a different buffer. The buffer
2975 on the screen remains unchanged (of course, normally nothing happens
2976 there until the command finishes running).
2978 @cindex @samp{call} defined
2979 Also, we have just introduced another jargon term, the word @dfn{call}.
2980 When you evaluate a list in which the first symbol is a function, you
2981 are calling that function. The use of the term comes from the notion of
2982 the function as an entity that can do something for you if you `call'
2983 it---just as a plumber is an entity who can fix a leak if you call him
2986 @node Buffer Size & Locations
2987 @section Buffer Size and the Location of Point
2988 @cindex Size of buffer
2990 @cindex Point location
2991 @cindex Location of point
2993 Finally, let's look at several rather simple functions,
2994 @code{buffer-size}, @code{point}, @code{point-min}, and
2995 @code{point-max}. These give information about the size of a buffer and
2996 the location of point within it.
2998 The function @code{buffer-size} tells you the size of the current
2999 buffer; that is, the function returns a count of the number of
3000 characters in the buffer.
3007 You can evaluate this in the usual way, by positioning the
3008 cursor after the expression and typing @kbd{C-x C-e}.
3010 @cindex @samp{point} defined
3011 In Emacs, the current position of the cursor is called @dfn{point}.
3012 The expression @code{(point)} returns a number that tells you where the
3013 cursor is located as a count of the number of characters from the
3014 beginning of the buffer up to point.
3017 You can see the character count for point in this buffer by evaluating
3018 the following expression in the usual way:
3025 As I write this, the value of @code{point} is 65724. The @code{point}
3026 function is frequently used in some of the examples later in this
3030 The value of point depends, of course, on its location within the
3031 buffer. If you evaluate point in this spot, the number will be larger:
3038 For me, the value of point in this location is 66043, which means that
3039 there are 319 characters (including spaces) between the two
3040 expressions. (Doubtless, you will see different numbers, since I will
3041 have edited this since I first evaluated point.)
3043 @cindex @samp{narrowing} defined
3044 The function @code{point-min} is somewhat similar to @code{point}, but
3045 it returns the value of the minimum permissible value of point in the
3046 current buffer. This is the number 1 unless @dfn{narrowing} is in
3047 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3048 or a program, to operations on just a part of a buffer.
3049 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3050 function @code{point-max} returns the value of the maximum permissible
3051 value of point in the current buffer.
3053 @node Evaluation Exercise
3056 Find a file with which you are working and move towards its middle.
3057 Find its buffer name, file name, length, and your position in the file.
3059 @node Writing Defuns
3060 @chapter How To Write Function Definitions
3061 @cindex Definition writing
3062 @cindex Function definition writing
3063 @cindex Writing a function definition
3065 When the Lisp interpreter evaluates a list, it looks to see whether the
3066 first symbol on the list has a function definition attached to it; or,
3067 put another way, whether the symbol points to a function definition. If
3068 it does, the computer carries out the instructions in the definition. A
3069 symbol that has a function definition is called, simply, a function
3070 (although, properly speaking, the definition is the function and the
3071 symbol refers to it.)
3074 * Primitive Functions::
3075 * defun:: The @code{defun} macro.
3076 * Install:: Install a function definition.
3077 * Interactive:: Making a function interactive.
3078 * Interactive Options:: Different options for @code{interactive}.
3079 * Permanent Installation:: Installing code permanently.
3080 * let:: Creating and initializing local variables.
3082 * else:: If--then--else expressions.
3083 * Truth & Falsehood:: What Lisp considers false and true.
3084 * save-excursion:: Keeping track of point, mark, and buffer.
3090 @node Primitive Functions
3091 @unnumberedsec An Aside about Primitive Functions
3093 @cindex Primitive functions
3094 @cindex Functions, primitive
3096 @cindex C language primitives
3097 @cindex Primitives written in C
3098 All functions are defined in terms of other functions, except for a few
3099 @dfn{primitive} functions that are written in the C programming
3100 language. When you write functions' definitions, you will write them in
3101 Emacs Lisp and use other functions as your building blocks. Some of the
3102 functions you will use will themselves be written in Emacs Lisp (perhaps
3103 by you) and some will be primitives written in C@. The primitive
3104 functions are used exactly like those written in Emacs Lisp and behave
3105 like them. They are written in C so we can easily run GNU Emacs on any
3106 computer that has sufficient power and can run C.
3108 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3109 distinguish between the use of functions written in C and the use of
3110 functions written in Emacs Lisp. The difference is irrelevant. I
3111 mention the distinction only because it is interesting to know. Indeed,
3112 unless you investigate, you won't know whether an already-written
3113 function is written in Emacs Lisp or C.
3116 @section The @code{defun} Macro
3119 @cindex @samp{function definition} defined
3120 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3121 it that tells the computer what to do when the function is called.
3122 This code is called the @dfn{function definition} and is created by
3123 evaluating a Lisp expression that starts with the symbol @code{defun}
3124 (which is an abbreviation for @emph{define function}).
3126 In subsequent sections, we will look at function definitions from the
3127 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3128 we will describe a simple function definition so you can see how it
3129 looks. This function definition uses arithmetic because it makes for a
3130 simple example. Some people dislike examples using arithmetic; however,
3131 if you are such a person, do not despair. Hardly any of the code we
3132 will study in the remainder of this introduction involves arithmetic or
3133 mathematics. The examples mostly involve text in one way or another.
3135 A function definition has up to five parts following the word
3140 The name of the symbol to which the function definition should be
3144 A list of the arguments that will be passed to the function. If no
3145 arguments will be passed to the function, this is an empty list,
3149 Documentation describing the function. (Technically optional, but
3150 strongly recommended.)
3153 Optionally, an expression to make the function interactive so you can
3154 use it by typing @kbd{M-x} and then the name of the function; or by
3155 typing an appropriate key or keychord.
3157 @cindex @samp{body} defined
3159 The code that instructs the computer what to do: the @dfn{body} of the
3160 function definition.
3163 It is helpful to think of the five parts of a function definition as
3164 being organized in a template, with slots for each part:
3168 (defun @var{function-name} (@var{arguments}@dots{})
3169 "@var{optional-documentation}@dots{}"
3170 (interactive @var{argument-passing-info}) ; @r{optional}
3175 As an example, here is the code for a function that multiplies its
3176 argument by 7. (This example is not interactive. @xref{Interactive,
3177 , Making a Function Interactive}, for that information.)
3181 (defun multiply-by-seven (number)
3182 "Multiply NUMBER by seven."
3187 This definition begins with a parenthesis and the symbol @code{defun},
3188 followed by the name of the function.
3190 @cindex @samp{argument list} defined
3191 The name of the function is followed by a list that contains the
3192 arguments that will be passed to the function. This list is called
3193 the @dfn{argument list}. In this example, the list has only one
3194 element, the symbol, @code{number}. When the function is used, the
3195 symbol will be bound to the value that is used as the argument to the
3198 Instead of choosing the word @code{number} for the name of the argument,
3199 I could have picked any other name. For example, I could have chosen
3200 the word @code{multiplicand}. I picked the word `number' because it
3201 tells what kind of value is intended for this slot; but I could just as
3202 well have chosen the word `multiplicand' to indicate the role that the
3203 value placed in this slot will play in the workings of the function. I
3204 could have called it @code{foogle}, but that would have been a bad
3205 choice because it would not tell humans what it means. The choice of
3206 name is up to the programmer and should be chosen to make the meaning of
3209 Indeed, you can choose any name you wish for a symbol in an argument
3210 list, even the name of a symbol used in some other function: the name
3211 you use in an argument list is private to that particular definition.
3212 In that definition, the name refers to a different entity than any use
3213 of the same name outside the function definition. Suppose you have a
3214 nick-name `Shorty' in your family; when your family members refer to
3215 `Shorty', they mean you. But outside your family, in a movie, for
3216 example, the name `Shorty' refers to someone else. Because a name in an
3217 argument list is private to the function definition, you can change the
3218 value of such a symbol inside the body of a function without changing
3219 its value outside the function. The effect is similar to that produced
3220 by a @code{let} expression. (@xref{let, , @code{let}}.)
3223 Note also that we discuss the word `number' in two different ways: as a
3224 symbol that appears in the code, and as the name of something that will
3225 be replaced by a something else during the evaluation of the function.
3226 In the first case, @code{number} is a symbol, not a number; it happens
3227 that within the function, it is a variable who value is the number in
3228 question, but our primary interest in it is as a symbol. On the other
3229 hand, when we are talking about the function, our interest is that we
3230 will substitute a number for the word @var{number}. To keep this
3231 distinction clear, we use different typography for the two
3232 circumstances. When we talk about this function, or about how it works,
3233 we refer to this number by writing @var{number}. In the function
3234 itself, we refer to it by writing @code{number}.
3237 The argument list is followed by the documentation string that
3238 describes the function. This is what you see when you type
3239 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3240 write a documentation string like this, you should make the first line
3241 a complete sentence since some commands, such as @code{apropos}, print
3242 only the first line of a multi-line documentation string. Also, you
3243 should not indent the second line of a documentation string, if you
3244 have one, because that looks odd when you use @kbd{C-h f}
3245 (@code{describe-function}). The documentation string is optional, but
3246 it is so useful, it should be included in almost every function you
3249 @findex * @r{(multiplication)}
3250 The third line of the example consists of the body of the function
3251 definition. (Most functions' definitions, of course, are longer than
3252 this.) In this function, the body is the list, @code{(* 7 number)}, which
3253 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3254 @code{*} is the function for multiplication, just as @code{+} is the
3255 function for addition.)
3257 When you use the @code{multiply-by-seven} function, the argument
3258 @code{number} evaluates to the actual number you want used. Here is an
3259 example that shows how @code{multiply-by-seven} is used; but don't try
3260 to evaluate this yet!
3263 (multiply-by-seven 3)
3267 The symbol @code{number}, specified in the function definition in the
3268 next section, is given or ``bound to'' the value 3 in the actual use of
3269 the function. Note that although @code{number} was inside parentheses
3270 in the function definition, the argument passed to the
3271 @code{multiply-by-seven} function is not in parentheses. The
3272 parentheses are written in the function definition so the computer can
3273 figure out where the argument list ends and the rest of the function
3276 If you evaluate this example, you are likely to get an error message.
3277 (Go ahead, try it!) This is because we have written the function
3278 definition, but not yet told the computer about the definition---we have
3279 not yet installed (or `loaded') the function definition in Emacs.
3280 Installing a function is the process that tells the Lisp interpreter the
3281 definition of the function. Installation is described in the next
3285 @section Install a Function Definition
3286 @cindex Install a Function Definition
3287 @cindex Definition installation
3288 @cindex Function definition installation
3290 If you are reading this inside of Info in Emacs, you can try out the
3291 @code{multiply-by-seven} function by first evaluating the function
3292 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3293 the function definition follows. Place the cursor after the last
3294 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3295 do this, @code{multiply-by-seven} will appear in the echo area. (What
3296 this means is that when a function definition is evaluated, the value it
3297 returns is the name of the defined function.) At the same time, this
3298 action installs the function definition.
3302 (defun multiply-by-seven (number)
3303 "Multiply NUMBER by seven."
3309 By evaluating this @code{defun}, you have just installed
3310 @code{multiply-by-seven} in Emacs. The function is now just as much a
3311 part of Emacs as @code{forward-word} or any other editing function you
3312 use. (@code{multiply-by-seven} will stay installed until you quit
3313 Emacs. To reload code automatically whenever you start Emacs, see
3314 @ref{Permanent Installation, , Installing Code Permanently}.)
3317 * Effect of installation::
3318 * Change a defun:: How to change a function definition.
3322 @node Effect of installation
3323 @unnumberedsubsec The effect of installation
3326 You can see the effect of installing @code{multiply-by-seven} by
3327 evaluating the following sample. Place the cursor after the following
3328 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3332 (multiply-by-seven 3)
3335 If you wish, you can read the documentation for the function by typing
3336 @kbd{C-h f} (@code{describe-function}) and then the name of the
3337 function, @code{multiply-by-seven}. When you do this, a
3338 @file{*Help*} window will appear on your screen that says:
3342 multiply-by-seven is a Lisp function.
3343 (multiply-by-seven NUMBER)
3345 Multiply NUMBER by seven.
3350 (To return to a single window on your screen, type @kbd{C-x 1}.)
3352 @node Change a defun
3353 @subsection Change a Function Definition
3354 @cindex Changing a function definition
3355 @cindex Function definition, how to change
3356 @cindex Definition, how to change
3358 If you want to change the code in @code{multiply-by-seven}, just rewrite
3359 it. To install the new version in place of the old one, evaluate the
3360 function definition again. This is how you modify code in Emacs. It is
3363 As an example, you can change the @code{multiply-by-seven} function to
3364 add the number to itself seven times instead of multiplying the number
3365 by seven. It produces the same answer, but by a different path. At
3366 the same time, we will add a comment to the code; a comment is text
3367 that the Lisp interpreter ignores, but that a human reader may find
3368 useful or enlightening. The comment is that this is the ``second
3373 (defun multiply-by-seven (number) ; @r{Second version.}
3374 "Multiply NUMBER by seven."
3375 (+ number number number number number number number))
3379 @cindex Comments in Lisp code
3380 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3381 line that follows a semicolon is a comment. The end of the line is the
3382 end of the comment. To stretch a comment over two or more lines, begin
3383 each line with a semicolon.
3385 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3386 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3387 Reference Manual}, for more about comments.
3389 You can install this version of the @code{multiply-by-seven} function by
3390 evaluating it in the same way you evaluated the first function: place
3391 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3393 In summary, this is how you write code in Emacs Lisp: you write a
3394 function; install it; test it; and then make fixes or enhancements and
3398 @section Make a Function Interactive
3399 @cindex Interactive functions
3402 You make a function interactive by placing a list that begins with
3403 the special form @code{interactive} immediately after the
3404 documentation. A user can invoke an interactive function by typing
3405 @kbd{M-x} and then the name of the function; or by typing the keys to
3406 which it is bound, for example, by typing @kbd{C-n} for
3407 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3409 Interestingly, when you call an interactive function interactively,
3410 the value returned is not automatically displayed in the echo area.
3411 This is because you often call an interactive function for its side
3412 effects, such as moving forward by a word or line, and not for the
3413 value returned. If the returned value were displayed in the echo area
3414 each time you typed a key, it would be very distracting.
3417 * Interactive multiply-by-seven:: An overview.
3418 * multiply-by-seven in detail:: The interactive version.
3422 @node Interactive multiply-by-seven
3423 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3426 Both the use of the special form @code{interactive} and one way to
3427 display a value in the echo area can be illustrated by creating an
3428 interactive version of @code{multiply-by-seven}.
3435 (defun multiply-by-seven (number) ; @r{Interactive version.}
3436 "Multiply NUMBER by seven."
3438 (message "The result is %d" (* 7 number)))
3443 You can install this code by placing your cursor after it and typing
3444 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3445 Then, you can use this code by typing @kbd{C-u} and a number and then
3446 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3447 @samp{The result is @dots{}} followed by the product will appear in the
3450 Speaking more generally, you invoke a function like this in either of two
3455 By typing a prefix argument that contains the number to be passed, and
3456 then typing @kbd{M-x} and the name of the function, as with
3457 @kbd{C-u 3 M-x forward-sentence}; or,
3460 By typing whatever key or keychord the function is bound to, as with
3465 Both the examples just mentioned work identically to move point forward
3466 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3467 it could not be used as an example of key binding.)
3469 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3472 A prefix argument is passed to an interactive function by typing the
3473 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3474 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3475 type @kbd{C-u} without a number, it defaults to 4).
3477 @node multiply-by-seven in detail
3478 @subsection An Interactive @code{multiply-by-seven}
3480 Let's look at the use of the special form @code{interactive} and then at
3481 the function @code{message} in the interactive version of
3482 @code{multiply-by-seven}. You will recall that the function definition
3487 (defun multiply-by-seven (number) ; @r{Interactive version.}
3488 "Multiply NUMBER by seven."
3490 (message "The result is %d" (* 7 number)))
3494 In this function, the expression, @code{(interactive "p")}, is a list of
3495 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3496 the function and use its value for the argument of the function.
3499 The argument will be a number. This means that the symbol
3500 @code{number} will be bound to a number in the line:
3503 (message "The result is %d" (* 7 number))
3508 For example, if your prefix argument is 5, the Lisp interpreter will
3509 evaluate the line as if it were:
3512 (message "The result is %d" (* 7 5))
3516 (If you are reading this in GNU Emacs, you can evaluate this expression
3517 yourself.) First, the interpreter will evaluate the inner list, which
3518 is @code{(* 7 5)}. This returns a value of 35. Next, it
3519 will evaluate the outer list, passing the values of the second and
3520 subsequent elements of the list to the function @code{message}.
3522 As we have seen, @code{message} is an Emacs Lisp function especially
3523 designed for sending a one line message to a user. (@xref{message, ,
3524 The @code{message} function}.) In summary, the @code{message}
3525 function prints its first argument in the echo area as is, except for
3526 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3527 which we have not mentioned). When it sees a control sequence, the
3528 function looks to the second or subsequent arguments and prints the
3529 value of the argument in the location in the string where the control
3530 sequence is located.
3532 In the interactive @code{multiply-by-seven} function, the control string
3533 is @samp{%d}, which requires a number, and the value returned by
3534 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3535 is printed in place of the @samp{%d} and the message is @samp{The result
3538 (Note that when you call the function @code{multiply-by-seven}, the
3539 message is printed without quotes, but when you call @code{message}, the
3540 text is printed in double quotes. This is because the value returned by
3541 @code{message} is what appears in the echo area when you evaluate an
3542 expression whose first element is @code{message}; but when embedded in a
3543 function, @code{message} prints the text as a side effect without
3546 @node Interactive Options
3547 @section Different Options for @code{interactive}
3548 @cindex Options for @code{interactive}
3549 @cindex Interactive options
3551 In the example, @code{multiply-by-seven} used @code{"p"} as the
3552 argument to @code{interactive}. This argument told Emacs to interpret
3553 your typing either @kbd{C-u} followed by a number or @key{META}
3554 followed by a number as a command to pass that number to the function
3555 as its argument. Emacs has more than twenty characters predefined for
3556 use with @code{interactive}. In almost every case, one of these
3557 options will enable you to pass the right information interactively to
3558 a function. (@xref{Interactive Codes, , Code Characters for
3559 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3562 Consider the function @code{zap-to-char}. Its interactive expression
3566 (interactive "p\ncZap to char: ")
3569 The first part of the argument to @code{interactive} is @samp{p}, with
3570 which you are already familiar. This argument tells Emacs to
3571 interpret a `prefix', as a number to be passed to the function. You
3572 can specify a prefix either by typing @kbd{C-u} followed by a number
3573 or by typing @key{META} followed by a number. The prefix is the
3574 number of specified characters. Thus, if your prefix is three and the
3575 specified character is @samp{x}, then you will delete all the text up
3576 to and including the third next @samp{x}. If you do not set a prefix,
3577 then you delete all the text up to and including the specified
3578 character, but no more.
3580 The @samp{c} tells the function the name of the character to which to delete.
3582 More formally, a function with two or more arguments can have
3583 information passed to each argument by adding parts to the string that
3584 follows @code{interactive}. When you do this, the information is
3585 passed to each argument in the same order it is specified in the
3586 @code{interactive} list. In the string, each part is separated from
3587 the next part by a @samp{\n}, which is a newline. For example, you
3588 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3589 This causes Emacs to pass the value of the prefix argument (if there
3590 is one) and the character.
3592 In this case, the function definition looks like the following, where
3593 @code{arg} and @code{char} are the symbols to which @code{interactive}
3594 binds the prefix argument and the specified character:
3598 (defun @var{name-of-function} (arg char)
3599 "@var{documentation}@dots{}"
3600 (interactive "p\ncZap to char: ")
3601 @var{body-of-function}@dots{})
3606 (The space after the colon in the prompt makes it look better when you
3607 are prompted. @xref{copy-to-buffer, , The Definition of
3608 @code{copy-to-buffer}}, for an example.)
3610 When a function does not take arguments, @code{interactive} does not
3611 require any. Such a function contains the simple expression
3612 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3615 Alternatively, if the special letter-codes are not right for your
3616 application, you can pass your own arguments to @code{interactive} as
3619 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3620 for an example. @xref{Using Interactive, , Using @code{Interactive},
3621 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3622 explanation about this technique.
3624 @node Permanent Installation
3625 @section Install Code Permanently
3626 @cindex Install code permanently
3627 @cindex Permanent code installation
3628 @cindex Code installation
3630 When you install a function definition by evaluating it, it will stay
3631 installed until you quit Emacs. The next time you start a new session
3632 of Emacs, the function will not be installed unless you evaluate the
3633 function definition again.
3635 At some point, you may want to have code installed automatically
3636 whenever you start a new session of Emacs. There are several ways of
3641 If you have code that is just for yourself, you can put the code for the
3642 function definition in your @file{.emacs} initialization file. When you
3643 start Emacs, your @file{.emacs} file is automatically evaluated and all
3644 the function definitions within it are installed.
3645 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3648 Alternatively, you can put the function definitions that you want
3649 installed in one or more files of their own and use the @code{load}
3650 function to cause Emacs to evaluate and thereby install each of the
3651 functions in the files.
3652 @xref{Loading Files, , Loading Files}.
3655 Thirdly, if you have code that your whole site will use, it is usual
3656 to put it in a file called @file{site-init.el} that is loaded when
3657 Emacs is built. This makes the code available to everyone who uses
3658 your machine. (See the @file{INSTALL} file that is part of the Emacs
3662 Finally, if you have code that everyone who uses Emacs may want, you
3663 can post it on a computer network or send a copy to the Free Software
3664 Foundation. (When you do this, please license the code and its
3665 documentation under a license that permits other people to run, copy,
3666 study, modify, and redistribute the code and which protects you from
3667 having your work taken from you.) If you send a copy of your code to
3668 the Free Software Foundation, and properly protect yourself and
3669 others, it may be included in the next release of Emacs. In large
3670 part, this is how Emacs has grown over the past years, by donations.
3676 The @code{let} expression is a special form in Lisp that you will need
3677 to use in most function definitions.
3679 @code{let} is used to attach or bind a symbol to a value in such a way
3680 that the Lisp interpreter will not confuse the variable with a
3681 variable of the same name that is not part of the function.
3683 To understand why the @code{let} special form is necessary, consider
3684 the situation in which you own a home that you generally refer to as
3685 `the house', as in the sentence, ``The house needs painting.'' If you
3686 are visiting a friend and your host refers to `the house', he is
3687 likely to be referring to @emph{his} house, not yours, that is, to a
3690 If your friend is referring to his house and you think he is referring
3691 to your house, you may be in for some confusion. The same thing could
3692 happen in Lisp if a variable that is used inside of one function has
3693 the same name as a variable that is used inside of another function,
3694 and the two are not intended to refer to the same value. The
3695 @code{let} special form prevents this kind of confusion.
3698 * Prevent confusion::
3699 * Parts of let Expression::
3700 * Sample let Expression::
3701 * Uninitialized let Variables::
3705 @node Prevent confusion
3706 @unnumberedsubsec @code{let} Prevents Confusion
3709 @cindex @samp{local variable} defined
3710 @cindex @samp{variable, local}, defined
3711 The @code{let} special form prevents confusion. @code{let} creates a
3712 name for a @dfn{local variable} that overshadows any use of the same
3713 name outside the @code{let} expression. This is like understanding
3714 that whenever your host refers to `the house', he means his house, not
3715 yours. (Symbols used in argument lists work the same way.
3716 @xref{defun, , The @code{defun} Macro}.)
3718 Local variables created by a @code{let} expression retain their value
3719 @emph{only} within the @code{let} expression itself (and within
3720 expressions called within the @code{let} expression); the local
3721 variables have no effect outside the @code{let} expression.
3723 Another way to think about @code{let} is that it is like a @code{setq}
3724 that is temporary and local. The values set by @code{let} are
3725 automatically undone when the @code{let} is finished. The setting
3726 only affects expressions that are inside the bounds of the @code{let}
3727 expression. In computer science jargon, we would say ``the binding of
3728 a symbol is visible only in functions called in the @code{let} form;
3729 in Emacs Lisp, scoping is dynamic, not lexical.''
3731 @code{let} can create more than one variable at once. Also,
3732 @code{let} gives each variable it creates an initial value, either a
3733 value specified by you, or @code{nil}. (In the jargon, this is called
3734 `binding the variable to the value'.) After @code{let} has created
3735 and bound the variables, it executes the code in the body of the
3736 @code{let}, and returns the value of the last expression in the body,
3737 as the value of the whole @code{let} expression. (`Execute' is a jargon
3738 term that means to evaluate a list; it comes from the use of the word
3739 meaning `to give practical effect to' (@cite{Oxford English
3740 Dictionary}). Since you evaluate an expression to perform an action,
3741 `execute' has evolved as a synonym to `evaluate'.)
3743 @node Parts of let Expression
3744 @subsection The Parts of a @code{let} Expression
3745 @cindex @code{let} expression, parts of
3746 @cindex Parts of @code{let} expression
3748 @cindex @samp{varlist} defined
3749 A @code{let} expression is a list of three parts. The first part is
3750 the symbol @code{let}. The second part is a list, called a
3751 @dfn{varlist}, each element of which is either a symbol by itself or a
3752 two-element list, the first element of which is a symbol. The third
3753 part of the @code{let} expression is the body of the @code{let}. The
3754 body usually consists of one or more lists.
3757 A template for a @code{let} expression looks like this:
3760 (let @var{varlist} @var{body}@dots{})
3764 The symbols in the varlist are the variables that are given initial
3765 values by the @code{let} special form. Symbols by themselves are given
3766 the initial value of @code{nil}; and each symbol that is the first
3767 element of a two-element list is bound to the value that is returned
3768 when the Lisp interpreter evaluates the second element.
3770 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3771 this case, in a @code{let} expression, Emacs binds the symbol
3772 @code{thread} to an initial value of @code{nil}, and binds the symbol
3773 @code{needles} to an initial value of 3.
3775 When you write a @code{let} expression, what you do is put the
3776 appropriate expressions in the slots of the @code{let} expression
3779 If the varlist is composed of two-element lists, as is often the case,
3780 the template for the @code{let} expression looks like this:
3784 (let ((@var{variable} @var{value})
3785 (@var{variable} @var{value})
3791 @node Sample let Expression
3792 @subsection Sample @code{let} Expression
3793 @cindex Sample @code{let} expression
3794 @cindex @code{let} expression sample
3796 The following expression creates and gives initial values
3797 to the two variables @code{zebra} and @code{tiger}. The body of the
3798 @code{let} expression is a list which calls the @code{message} function.
3802 (let ((zebra 'stripes)
3804 (message "One kind of animal has %s and another is %s."
3809 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3811 The two variables are @code{zebra} and @code{tiger}. Each variable is
3812 the first element of a two-element list and each value is the second
3813 element of its two-element list. In the varlist, Emacs binds the
3814 variable @code{zebra} to the value @code{stripes}@footnote{According
3815 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3816 become impossibly dangerous as they grow older'' but the claim here is
3817 that they do not become fierce like a tiger. (1997, W. W. Norton and
3818 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3819 variable @code{tiger} to the value @code{fierce}. In this example,
3820 both values are symbols preceded by a quote. The values could just as
3821 well have been another list or a string. The body of the @code{let}
3822 follows after the list holding the variables. In this example, the
3823 body is a list that uses the @code{message} function to print a string
3827 You may evaluate the example in the usual fashion, by placing the
3828 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3829 this, the following will appear in the echo area:
3832 "One kind of animal has stripes and another is fierce."
3835 As we have seen before, the @code{message} function prints its first
3836 argument, except for @samp{%s}. In this example, the value of the variable
3837 @code{zebra} is printed at the location of the first @samp{%s} and the
3838 value of the variable @code{tiger} is printed at the location of the
3841 @node Uninitialized let Variables
3842 @subsection Uninitialized Variables in a @code{let} Statement
3843 @cindex Uninitialized @code{let} variables
3844 @cindex @code{let} variables uninitialized
3846 If you do not bind the variables in a @code{let} statement to specific
3847 initial values, they will automatically be bound to an initial value of
3848 @code{nil}, as in the following expression:
3857 "Here are %d variables with %s, %s, and %s value."
3858 birch pine fir oak))
3863 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3866 If you evaluate this expression in the usual way, the following will
3867 appear in your echo area:
3870 "Here are 3 variables with nil, nil, and some value."
3874 In this example, Emacs binds the symbol @code{birch} to the number 3,
3875 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3876 the symbol @code{oak} to the value @code{some}.
3878 Note that in the first part of the @code{let}, the variables @code{pine}
3879 and @code{fir} stand alone as atoms that are not surrounded by
3880 parentheses; this is because they are being bound to @code{nil}, the
3881 empty list. But @code{oak} is bound to @code{some} and so is a part of
3882 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3883 number 3 and so is in a list with that number. (Since a number
3884 evaluates to itself, the number does not need to be quoted. Also, the
3885 number is printed in the message using a @samp{%d} rather than a
3886 @samp{%s}.) The four variables as a group are put into a list to
3887 delimit them from the body of the @code{let}.
3890 @section The @code{if} Special Form
3892 @cindex Conditional with @code{if}
3894 A third special form, in addition to @code{defun} and @code{let}, is the
3895 conditional @code{if}. This form is used to instruct the computer to
3896 make decisions. You can write function definitions without using
3897 @code{if}, but it is used often enough, and is important enough, to be
3898 included here. It is used, for example, in the code for the
3899 function @code{beginning-of-buffer}.
3901 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3902 @emph{then} an expression is evaluated.'' If the test is not true, the
3903 expression is not evaluated. For example, you might make a decision
3904 such as, ``if it is warm and sunny, then go to the beach!''
3907 * if in more detail::
3908 * type-of-animal in detail:: An example of an @code{if} expression.
3912 @node if in more detail
3913 @unnumberedsubsec @code{if} in more detail
3916 @cindex @samp{if-part} defined
3917 @cindex @samp{then-part} defined
3918 An @code{if} expression written in Lisp does not use the word `then';
3919 the test and the action are the second and third elements of the list
3920 whose first element is @code{if}. Nonetheless, the test part of an
3921 @code{if} expression is often called the @dfn{if-part} and the second
3922 argument is often called the @dfn{then-part}.
3924 Also, when an @code{if} expression is written, the true-or-false-test
3925 is usually written on the same line as the symbol @code{if}, but the
3926 action to carry out if the test is true, the ``then-part'', is written
3927 on the second and subsequent lines. This makes the @code{if}
3928 expression easier to read.
3932 (if @var{true-or-false-test}
3933 @var{action-to-carry-out-if-test-is-true})
3938 The true-or-false-test will be an expression that
3939 is evaluated by the Lisp interpreter.
3941 Here is an example that you can evaluate in the usual manner. The test
3942 is whether the number 5 is greater than the number 4. Since it is, the
3943 message @samp{5 is greater than 4!} will be printed.
3947 (if (> 5 4) ; @r{if-part}
3948 (message "5 is greater than 4!")) ; @r{then-part}
3953 (The function @code{>} tests whether its first argument is greater than
3954 its second argument and returns true if it is.)
3955 @findex > (greater than)
3957 Of course, in actual use, the test in an @code{if} expression will not
3958 be fixed for all time as it is by the expression @code{(> 5 4)}.
3959 Instead, at least one of the variables used in the test will be bound to
3960 a value that is not known ahead of time. (If the value were known ahead
3961 of time, we would not need to run the test!)
3963 For example, the value may be bound to an argument of a function
3964 definition. In the following function definition, the character of the
3965 animal is a value that is passed to the function. If the value bound to
3966 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3967 tiger!} will be printed; otherwise, @code{nil} will be returned.
3971 (defun type-of-animal (characteristic)
3972 "Print message in echo area depending on CHARACTERISTIC.
3973 If the CHARACTERISTIC is the symbol `fierce',
3974 then warn of a tiger."
3975 (if (equal characteristic 'fierce)
3976 (message "It's a tiger!")))
3982 If you are reading this inside of GNU Emacs, you can evaluate the
3983 function definition in the usual way to install it in Emacs, and then you
3984 can evaluate the following two expressions to see the results:
3988 (type-of-animal 'fierce)
3990 (type-of-animal 'zebra)
3995 @c Following sentences rewritten to prevent overfull hbox.
3997 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
3998 following message printed in the echo area: @code{"It's a tiger!"}; and
3999 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
4000 printed in the echo area.
4002 @node type-of-animal in detail
4003 @subsection The @code{type-of-animal} Function in Detail
4005 Let's look at the @code{type-of-animal} function in detail.
4007 The function definition for @code{type-of-animal} was written by filling
4008 the slots of two templates, one for a function definition as a whole, and
4009 a second for an @code{if} expression.
4012 The template for every function that is not interactive is:
4016 (defun @var{name-of-function} (@var{argument-list})
4017 "@var{documentation}@dots{}"
4023 The parts of the function that match this template look like this:
4027 (defun type-of-animal (characteristic)
4028 "Print message in echo area depending on CHARACTERISTIC.
4029 If the CHARACTERISTIC is the symbol `fierce',
4030 then warn of a tiger."
4031 @var{body: the} @code{if} @var{expression})
4035 The name of function is @code{type-of-animal}; it is passed the value
4036 of one argument. The argument list is followed by a multi-line
4037 documentation string. The documentation string is included in the
4038 example because it is a good habit to write documentation string for
4039 every function definition. The body of the function definition
4040 consists of the @code{if} expression.
4043 The template for an @code{if} expression looks like this:
4047 (if @var{true-or-false-test}
4048 @var{action-to-carry-out-if-the-test-returns-true})
4053 In the @code{type-of-animal} function, the code for the @code{if}
4058 (if (equal characteristic 'fierce)
4059 (message "It's a tiger!")))
4064 Here, the true-or-false-test is the expression:
4067 (equal characteristic 'fierce)
4071 In Lisp, @code{equal} is a function that determines whether its first
4072 argument is equal to its second argument. The second argument is the
4073 quoted symbol @code{'fierce} and the first argument is the value of the
4074 symbol @code{characteristic}---in other words, the argument passed to
4077 In the first exercise of @code{type-of-animal}, the argument
4078 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4079 is equal to @code{fierce}, the expression, @code{(equal characteristic
4080 'fierce)}, returns a value of true. When this happens, the @code{if}
4081 evaluates the second argument or then-part of the @code{if}:
4082 @code{(message "It's tiger!")}.
4084 On the other hand, in the second exercise of @code{type-of-animal}, the
4085 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4086 is not equal to @code{fierce}, so the then-part is not evaluated and
4087 @code{nil} is returned by the @code{if} expression.
4090 @section If--then--else Expressions
4093 An @code{if} expression may have an optional third argument, called
4094 the @dfn{else-part}, for the case when the true-or-false-test returns
4095 false. When this happens, the second argument or then-part of the
4096 overall @code{if} expression is @emph{not} evaluated, but the third or
4097 else-part @emph{is} evaluated. You might think of this as the cloudy
4098 day alternative for the decision ``if it is warm and sunny, then go to
4099 the beach, else read a book!''.
4101 The word ``else'' is not written in the Lisp code; the else-part of an
4102 @code{if} expression comes after the then-part. In the written Lisp, the
4103 else-part is usually written to start on a line of its own and is
4104 indented less than the then-part:
4108 (if @var{true-or-false-test}
4109 @var{action-to-carry-out-if-the-test-returns-true}
4110 @var{action-to-carry-out-if-the-test-returns-false})
4114 For example, the following @code{if} expression prints the message @samp{4
4115 is not greater than 5!} when you evaluate it in the usual way:
4119 (if (> 4 5) ; @r{if-part}
4120 (message "4 falsely greater than 5!") ; @r{then-part}
4121 (message "4 is not greater than 5!")) ; @r{else-part}
4126 Note that the different levels of indentation make it easy to
4127 distinguish the then-part from the else-part. (GNU Emacs has several
4128 commands that automatically indent @code{if} expressions correctly.
4129 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4131 We can extend the @code{type-of-animal} function to include an
4132 else-part by simply incorporating an additional part to the @code{if}
4136 You can see the consequences of doing this if you evaluate the following
4137 version of the @code{type-of-animal} function definition to install it
4138 and then evaluate the two subsequent expressions to pass different
4139 arguments to the function.
4143 (defun type-of-animal (characteristic) ; @r{Second version.}
4144 "Print message in echo area depending on CHARACTERISTIC.
4145 If the CHARACTERISTIC is the symbol `fierce',
4146 then warn of a tiger;
4147 else say it's not fierce."
4148 (if (equal characteristic 'fierce)
4149 (message "It's a tiger!")
4150 (message "It's not fierce!")))
4157 (type-of-animal 'fierce)
4159 (type-of-animal 'zebra)
4164 @c Following sentence rewritten to prevent overfull hbox.
4166 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4167 following message printed in the echo area: @code{"It's a tiger!"}; but
4168 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4169 @code{"It's not fierce!"}.
4171 (Of course, if the @var{characteristic} were @code{ferocious}, the
4172 message @code{"It's not fierce!"} would be printed; and it would be
4173 misleading! When you write code, you need to take into account the
4174 possibility that some such argument will be tested by the @code{if}
4175 and write your program accordingly.)
4177 @node Truth & Falsehood
4178 @section Truth and Falsehood in Emacs Lisp
4179 @cindex Truth and falsehood in Emacs Lisp
4180 @cindex Falsehood and truth in Emacs Lisp
4183 There is an important aspect to the truth test in an @code{if}
4184 expression. So far, we have spoken of `true' and `false' as values of
4185 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4186 `false' is just our old friend @code{nil}. Anything else---anything
4189 The expression that tests for truth is interpreted as @dfn{true}
4190 if the result of evaluating it is a value that is not @code{nil}. In
4191 other words, the result of the test is considered true if the value
4192 returned is a number such as 47, a string such as @code{"hello"}, or a
4193 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4194 long as it is not empty), or even a buffer!
4197 * nil explained:: @code{nil} has two meanings.
4202 @unnumberedsubsec An explanation of @code{nil}
4205 Before illustrating a test for truth, we need an explanation of @code{nil}.
4207 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4208 empty list. Second, it means false and is the value returned when a
4209 true-or-false-test tests false. @code{nil} can be written as an empty
4210 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4211 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4212 to use @code{nil} for false and @code{()} for the empty list.
4214 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4215 list---is considered true. This means that if an evaluation returns
4216 something that is not an empty list, an @code{if} expression will test
4217 true. For example, if a number is put in the slot for the test, it
4218 will be evaluated and will return itself, since that is what numbers
4219 do when evaluated. In this conditional, the @code{if} expression will
4220 test true. The expression tests false only when @code{nil}, an empty
4221 list, is returned by evaluating the expression.
4223 You can see this by evaluating the two expressions in the following examples.
4225 In the first example, the number 4 is evaluated as the test in the
4226 @code{if} expression and returns itself; consequently, the then-part
4227 of the expression is evaluated and returned: @samp{true} appears in
4228 the echo area. In the second example, the @code{nil} indicates false;
4229 consequently, the else-part of the expression is evaluated and
4230 returned: @samp{false} appears in the echo area.
4247 Incidentally, if some other useful value is not available for a test that
4248 returns true, then the Lisp interpreter will return the symbol @code{t}
4249 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4250 when evaluated, as you can see by evaluating it in the usual way:
4258 On the other hand, this function returns @code{nil} if the test is false.
4264 @node save-excursion
4265 @section @code{save-excursion}
4266 @findex save-excursion
4267 @cindex Region, what it is
4268 @cindex Preserving point, mark, and buffer
4269 @cindex Point, mark, buffer preservation
4273 The @code{save-excursion} function is the third and final special form
4274 that we will discuss in this chapter.
4276 In Emacs Lisp programs used for editing, the @code{save-excursion}
4277 function is very common. It saves the location of point and mark,
4278 executes the body of the function, and then restores point and mark to
4279 their previous positions if their locations were changed. Its primary
4280 purpose is to keep the user from being surprised and disturbed by
4281 unexpected movement of point or mark.
4284 * Point and mark:: A review of various locations.
4285 * Template for save-excursion::
4289 @node Point and mark
4290 @unnumberedsubsec Point and Mark
4293 Before discussing @code{save-excursion}, however, it may be useful
4294 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4295 the current location of the cursor. Wherever the cursor
4296 is, that is point. More precisely, on terminals where the cursor
4297 appears to be on top of a character, point is immediately before the
4298 character. In Emacs Lisp, point is an integer. The first character in
4299 a buffer is number one, the second is number two, and so on. The
4300 function @code{point} returns the current position of the cursor as a
4301 number. Each buffer has its own value for point.
4303 The @dfn{mark} is another position in the buffer; its value can be set
4304 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4305 a mark has been set, you can use the command @kbd{C-x C-x}
4306 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4307 and set the mark to be the previous position of point. In addition, if
4308 you set another mark, the position of the previous mark is saved in the
4309 mark ring. Many mark positions can be saved this way. You can jump the
4310 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4313 The part of the buffer between point and mark is called @dfn{the
4314 region}. Numerous commands work on the region, including
4315 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4316 @code{print-region}.
4318 The @code{save-excursion} special form saves the locations of point and
4319 mark and restores those positions after the code within the body of the
4320 special form is evaluated by the Lisp interpreter. Thus, if point were
4321 in the beginning of a piece of text and some code moved point to the end
4322 of the buffer, the @code{save-excursion} would put point back to where
4323 it was before, after the expressions in the body of the function were
4326 In Emacs, a function frequently moves point as part of its internal
4327 workings even though a user would not expect this. For example,
4328 @code{count-lines-region} moves point. To prevent the user from being
4329 bothered by jumps that are both unexpected and (from the user's point of
4330 view) unnecessary, @code{save-excursion} is often used to keep point and
4331 mark in the location expected by the user. The use of
4332 @code{save-excursion} is good housekeeping.
4334 To make sure the house stays clean, @code{save-excursion} restores the
4335 values of point and mark even if something goes wrong in the code inside
4336 of it (or, to be more precise and to use the proper jargon, ``in case of
4337 abnormal exit''). This feature is very helpful.
4339 In addition to recording the values of point and mark,
4340 @code{save-excursion} keeps track of the current buffer, and restores
4341 it, too. This means you can write code that will change the buffer and
4342 have @code{save-excursion} switch you back to the original buffer.
4343 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4344 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4346 @node Template for save-excursion
4347 @subsection Template for a @code{save-excursion} Expression
4350 The template for code using @code{save-excursion} is simple:
4360 The body of the function is one or more expressions that will be
4361 evaluated in sequence by the Lisp interpreter. If there is more than
4362 one expression in the body, the value of the last one will be returned
4363 as the value of the @code{save-excursion} function. The other
4364 expressions in the body are evaluated only for their side effects; and
4365 @code{save-excursion} itself is used only for its side effect (which
4366 is restoring the positions of point and mark).
4369 In more detail, the template for a @code{save-excursion} expression
4375 @var{first-expression-in-body}
4376 @var{second-expression-in-body}
4377 @var{third-expression-in-body}
4379 @var{last-expression-in-body})
4384 An expression, of course, may be a symbol on its own or a list.
4386 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4387 within the body of a @code{let} expression. It looks like this:
4400 In the last few chapters we have introduced a macro and a fair number
4401 of functions and special forms. Here they are described in brief,
4402 along with a few similar functions that have not been mentioned yet.
4405 @item eval-last-sexp
4406 Evaluate the last symbolic expression before the current location of
4407 point. The value is printed in the echo area unless the function is
4408 invoked with an argument; in that case, the output is printed in the
4409 current buffer. This command is normally bound to @kbd{C-x C-e}.
4412 Define function. This macro has up to five parts: the name, a
4413 template for the arguments that will be passed to the function,
4414 documentation, an optional interactive declaration, and the body of
4418 For example, in an early version of Emacs, the function definition was
4419 as follows. (It is slightly more complex now that it seeks the first
4420 non-whitespace character rather than the first visible character.)
4424 (defun back-to-indentation ()
4425 "Move point to first visible character on line."
4427 (beginning-of-line 1)
4428 (skip-chars-forward " \t"))
4435 (defun backward-to-indentation (&optional arg)
4436 "Move backward ARG lines and position at first nonblank character."
4438 (forward-line (- (or arg 1)))
4439 (skip-chars-forward " \t"))
4441 (defun back-to-indentation ()
4442 "Move point to the first non-whitespace character on this line."
4444 (beginning-of-line 1)
4445 (skip-syntax-forward " " (line-end-position))
4446 ;; Move back over chars that have whitespace syntax but have the p flag.
4447 (backward-prefix-chars))
4451 Declare to the interpreter that the function can be used
4452 interactively. This special form may be followed by a string with one
4453 or more parts that pass the information to the arguments of the
4454 function, in sequence. These parts may also tell the interpreter to
4455 prompt for information. Parts of the string are separated by
4456 newlines, @samp{\n}.
4459 Common code characters are:
4463 The name of an existing buffer.
4466 The name of an existing file.
4469 The numeric prefix argument. (Note that this `p' is lower case.)
4472 Point and the mark, as two numeric arguments, smallest first. This
4473 is the only code letter that specifies two successive arguments
4477 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4478 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4482 Declare that a list of variables is for use within the body of the
4483 @code{let} and give them an initial value, either @code{nil} or a
4484 specified value; then evaluate the rest of the expressions in the body
4485 of the @code{let} and return the value of the last one. Inside the
4486 body of the @code{let}, the Lisp interpreter does not see the values of
4487 the variables of the same names that are bound outside of the
4495 (let ((foo (buffer-name))
4496 (bar (buffer-size)))
4498 "This buffer is %s and has %d characters."
4503 @item save-excursion
4504 Record the values of point and mark and the current buffer before
4505 evaluating the body of this special form. Restore the values of point
4506 and mark and buffer afterward.
4513 (message "We are %d characters into this buffer."
4516 (goto-char (point-min)) (point))))
4521 Evaluate the first argument to the function; if it is true, evaluate
4522 the second argument; else evaluate the third argument, if there is one.
4524 The @code{if} special form is called a @dfn{conditional}. There are
4525 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4533 (if (= 22 emacs-major-version)
4534 (message "This is version 22 Emacs")
4535 (message "This is not version 22 Emacs"))
4544 The @code{<} function tests whether its first argument is smaller than
4545 its second argument. A corresponding function, @code{>}, tests whether
4546 the first argument is greater than the second. Likewise, @code{<=}
4547 tests whether the first argument is less than or equal to the second and
4548 @code{>=} tests whether the first argument is greater than or equal to
4549 the second. In all cases, both arguments must be numbers or markers
4550 (markers indicate positions in buffers).
4554 The @code{=} function tests whether two arguments, both numbers or
4560 Test whether two objects are the same. @code{equal} uses one meaning
4561 of the word `same' and @code{eq} uses another: @code{equal} returns
4562 true if the two objects have a similar structure and contents, such as
4563 two copies of the same book. On the other hand, @code{eq}, returns
4564 true if both arguments are actually the same object.
4573 The @code{string-lessp} function tests whether its first argument is
4574 smaller than the second argument. A shorter, alternative name for the
4575 same function (a @code{defalias}) is @code{string<}.
4577 The arguments to @code{string-lessp} must be strings or symbols; the
4578 ordering is lexicographic, so case is significant. The print names of
4579 symbols are used instead of the symbols themselves.
4581 @cindex @samp{empty string} defined
4582 An empty string, @samp{""}, a string with no characters in it, is
4583 smaller than any string of characters.
4585 @code{string-equal} provides the corresponding test for equality. Its
4586 shorter, alternative name is @code{string=}. There are no string test
4587 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4590 Print a message in the echo area. The first argument is a string that
4591 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4592 arguments that follow the string. The argument used by @samp{%s} must
4593 be a string or a symbol; the argument used by @samp{%d} must be a
4594 number. The argument used by @samp{%c} must be an @sc{ascii} code
4595 number; it will be printed as the character with that @sc{ascii} code.
4596 (Various other %-sequences have not been mentioned.)
4600 The @code{setq} function sets the value of its first argument to the
4601 value of the second argument. The first argument is automatically
4602 quoted by @code{setq}. It does the same for succeeding pairs of
4603 arguments. Another function, @code{set}, takes only two arguments and
4604 evaluates both of them before setting the value returned by its first
4605 argument to the value returned by its second argument.
4608 Without an argument, return the name of the buffer, as a string.
4610 @item buffer-file-name
4611 Without an argument, return the name of the file the buffer is
4614 @item current-buffer
4615 Return the buffer in which Emacs is active; it may not be
4616 the buffer that is visible on the screen.
4619 Return the most recently selected buffer (other than the buffer passed
4620 to @code{other-buffer} as an argument and other than the current
4623 @item switch-to-buffer
4624 Select a buffer for Emacs to be active in and display it in the current
4625 window so users can look at it. Usually bound to @kbd{C-x b}.
4628 Switch Emacs's attention to a buffer on which programs will run. Don't
4629 alter what the window is showing.
4632 Return the number of characters in the current buffer.
4635 Return the value of the current position of the cursor, as an
4636 integer counting the number of characters from the beginning of the
4640 Return the minimum permissible value of point in
4641 the current buffer. This is 1, unless narrowing is in effect.
4644 Return the value of the maximum permissible value of point in the
4645 current buffer. This is the end of the buffer, unless narrowing is in
4650 @node defun Exercises
4655 Write a non-interactive function that doubles the value of its
4656 argument, a number. Make that function interactive.
4659 Write a function that tests whether the current value of
4660 @code{fill-column} is greater than the argument passed to the function,
4661 and if so, prints an appropriate message.
4664 @node Buffer Walk Through
4665 @chapter A Few Buffer--Related Functions
4667 In this chapter we study in detail several of the functions used in GNU
4668 Emacs. This is called a ``walk-through''. These functions are used as
4669 examples of Lisp code, but are not imaginary examples; with the
4670 exception of the first, simplified function definition, these functions
4671 show the actual code used in GNU Emacs. You can learn a great deal from
4672 these definitions. The functions described here are all related to
4673 buffers. Later, we will study other functions.
4676 * Finding More:: How to find more information.
4677 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4678 @code{point-min}, and @code{push-mark}.
4679 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4680 * append-to-buffer:: Uses @code{save-excursion} and
4681 @code{insert-buffer-substring}.
4682 * Buffer Related Review:: Review.
4683 * Buffer Exercises::
4687 @section Finding More Information
4689 @findex describe-function, @r{introduced}
4690 @cindex Find function documentation
4691 In this walk-through, I will describe each new function as we come to
4692 it, sometimes in detail and sometimes briefly. If you are interested,
4693 you can get the full documentation of any Emacs Lisp function at any
4694 time by typing @kbd{C-h f} and then the name of the function (and then
4695 @key{RET}). Similarly, you can get the full documentation for a
4696 variable by typing @kbd{C-h v} and then the name of the variable (and
4699 @cindex Find source of function
4700 @c In version 22, tells location both of C and of Emacs Lisp
4701 Also, @code{describe-function} will tell you the location of the
4702 function definition.
4704 Put point into the name of the file that contains the function and
4705 press the @key{RET} key. In this case, @key{RET} means
4706 @code{push-button} rather than `return' or `enter'. Emacs will take
4707 you directly to the function definition.
4712 If you move point over the file name and press
4713 the @key{RET} key, which in this case means @code{help-follow} rather
4714 than `return' or `enter', Emacs will take you directly to the function
4718 More generally, if you want to see a function in its original source
4719 file, you can use the @code{find-tag} function to jump to it.
4720 @code{find-tag} works with a wide variety of languages, not just
4721 Lisp, and C, and it works with non-programming text as well. For
4722 example, @code{find-tag} will jump to the various nodes in the
4723 Texinfo source file of this document.
4724 The @code{find-tag} function depends on `tags tables' that record
4725 the locations of the functions, variables, and other items to which
4726 @code{find-tag} jumps.
4728 To use the @code{find-tag} command, type @kbd{M-.} (i.e., press the
4729 period key while holding down the @key{META} key, or else type the
4730 @key{ESC} key and then type the period key), and then, at the prompt,
4731 type in the name of the function whose source code you want to see,
4732 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4733 switch buffers and display the source code for the function on your
4734 screen. To switch back to your current buffer, type @kbd{C-x b
4735 @key{RET}}. (On some keyboards, the @key{META} key is labeled
4738 @c !!! 22.1.1 tags table location in this paragraph
4739 @cindex TAGS table, specifying
4741 Depending on how the initial default values of your copy of Emacs are
4742 set, you may also need to specify the location of your `tags table',
4743 which is a file called @file{TAGS}. For example, if you are
4744 interested in Emacs sources, the tags table you will most likely want,
4745 if it has already been created for you, will be in a subdirectory of
4746 the @file{/usr/local/share/emacs/} directory; thus you would use the
4747 @code{M-x visit-tags-table} command and specify a pathname such as
4748 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4749 has not already been created, you will have to create it yourself. It
4750 will be in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4753 To create a @file{TAGS} file in a specific directory, switch to that
4754 directory in Emacs using @kbd{M-x cd} command, or list the directory
4755 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4756 @w{@code{etags *.el}} as the command to execute:
4759 M-x compile RET etags *.el RET
4762 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4764 After you become more familiar with Emacs Lisp, you will find that you will
4765 frequently use @code{find-tag} to navigate your way around source code;
4766 and you will create your own @file{TAGS} tables.
4768 @cindex Library, as term for `file'
4769 Incidentally, the files that contain Lisp code are conventionally
4770 called @dfn{libraries}. The metaphor is derived from that of a
4771 specialized library, such as a law library or an engineering library,
4772 rather than a general library. Each library, or file, contains
4773 functions that relate to a particular topic or activity, such as
4774 @file{abbrev.el} for handling abbreviations and other typing
4775 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4776 libraries provide code for a single activity, as the various
4777 @file{rmail@dots{}} files provide code for reading electronic mail.)
4778 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4779 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4780 by topic keywords.''
4782 @node simplified-beginning-of-buffer
4783 @section A Simplified @code{beginning-of-buffer} Definition
4784 @findex simplified-beginning-of-buffer
4786 The @code{beginning-of-buffer} command is a good function to start with
4787 since you are likely to be familiar with it and it is easy to
4788 understand. Used as an interactive command, @code{beginning-of-buffer}
4789 moves the cursor to the beginning of the buffer, leaving the mark at the
4790 previous position. It is generally bound to @kbd{M-<}.
4792 In this section, we will discuss a shortened version of the function
4793 that shows how it is most frequently used. This shortened function
4794 works as written, but it does not contain the code for a complex option.
4795 In another section, we will describe the entire function.
4796 (@xref{beginning-of-buffer, , Complete Definition of
4797 @code{beginning-of-buffer}}.)
4799 Before looking at the code, let's consider what the function
4800 definition has to contain: it must include an expression that makes
4801 the function interactive so it can be called by typing @kbd{M-x
4802 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4803 must include code to leave a mark at the original position in the
4804 buffer; and it must include code to move the cursor to the beginning
4808 Here is the complete text of the shortened version of the function:
4812 (defun simplified-beginning-of-buffer ()
4813 "Move point to the beginning of the buffer;
4814 leave mark at previous position."
4817 (goto-char (point-min)))
4821 Like all function definitions, this definition has five parts following
4822 the macro @code{defun}:
4826 The name: in this example, @code{simplified-beginning-of-buffer}.
4829 A list of the arguments: in this example, an empty list, @code{()},
4832 The documentation string.
4835 The interactive expression.
4842 In this function definition, the argument list is empty; this means that
4843 this function does not require any arguments. (When we look at the
4844 definition for the complete function, we will see that it may be passed
4845 an optional argument.)
4847 The interactive expression tells Emacs that the function is intended to
4848 be used interactively. In this example, @code{interactive} does not have
4849 an argument because @code{simplified-beginning-of-buffer} does not
4853 The body of the function consists of the two lines:
4858 (goto-char (point-min))
4862 The first of these lines is the expression, @code{(push-mark)}. When
4863 this expression is evaluated by the Lisp interpreter, it sets a mark at
4864 the current position of the cursor, wherever that may be. The position
4865 of this mark is saved in the mark ring.
4867 The next line is @code{(goto-char (point-min))}. This expression
4868 jumps the cursor to the minimum point in the buffer, that is, to the
4869 beginning of the buffer (or to the beginning of the accessible portion
4870 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4871 Narrowing and Widening}.)
4873 The @code{push-mark} command sets a mark at the place where the cursor
4874 was located before it was moved to the beginning of the buffer by the
4875 @code{(goto-char (point-min))} expression. Consequently, you can, if
4876 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4878 That is all there is to the function definition!
4880 @findex describe-function
4881 When you are reading code such as this and come upon an unfamiliar
4882 function, such as @code{goto-char}, you can find out what it does by
4883 using the @code{describe-function} command. To use this command, type
4884 @kbd{C-h f} and then type in the name of the function and press
4885 @key{RET}. The @code{describe-function} command will print the
4886 function's documentation string in a @file{*Help*} window. For
4887 example, the documentation for @code{goto-char} is:
4891 Set point to POSITION, a number or marker.
4892 Beginning of buffer is position (point-min), end is (point-max).
4897 The function's one argument is the desired position.
4900 (The prompt for @code{describe-function} will offer you the symbol
4901 under or preceding the cursor, so you can save typing by positioning
4902 the cursor right over or after the function and then typing @kbd{C-h f
4905 The @code{end-of-buffer} function definition is written in the same way as
4906 the @code{beginning-of-buffer} definition except that the body of the
4907 function contains the expression @code{(goto-char (point-max))} in place
4908 of @code{(goto-char (point-min))}.
4910 @node mark-whole-buffer
4911 @section The Definition of @code{mark-whole-buffer}
4912 @findex mark-whole-buffer
4914 The @code{mark-whole-buffer} function is no harder to understand than the
4915 @code{simplified-beginning-of-buffer} function. In this case, however,
4916 we will look at the complete function, not a shortened version.
4918 The @code{mark-whole-buffer} function is not as commonly used as the
4919 @code{beginning-of-buffer} function, but is useful nonetheless: it
4920 marks a whole buffer as a region by putting point at the beginning and
4921 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4925 * mark-whole-buffer overview::
4926 * Body of mark-whole-buffer:: Only three lines of code.
4930 @node mark-whole-buffer overview
4931 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4935 In GNU Emacs 22, the code for the complete function looks like this:
4939 (defun mark-whole-buffer ()
4940 "Put point at beginning and mark at end of buffer.
4941 You probably should not use this function in Lisp programs;
4942 it is usually a mistake for a Lisp function to use any subroutine
4943 that uses or sets the mark."
4946 (push-mark (point-max) nil t)
4947 (goto-char (point-min)))
4952 Like all other functions, the @code{mark-whole-buffer} function fits
4953 into the template for a function definition. The template looks like
4958 (defun @var{name-of-function} (@var{argument-list})
4959 "@var{documentation}@dots{}"
4960 (@var{interactive-expression}@dots{})
4965 Here is how the function works: the name of the function is
4966 @code{mark-whole-buffer}; it is followed by an empty argument list,
4967 @samp{()}, which means that the function does not require arguments.
4968 The documentation comes next.
4970 The next line is an @code{(interactive)} expression that tells Emacs
4971 that the function will be used interactively. These details are similar
4972 to the @code{simplified-beginning-of-buffer} function described in the
4976 @node Body of mark-whole-buffer
4977 @subsection Body of @code{mark-whole-buffer}
4979 The body of the @code{mark-whole-buffer} function consists of three
4986 (push-mark (point-max) nil t)
4987 (goto-char (point-min))
4991 The first of these lines is the expression, @code{(push-mark (point))}.
4993 This line does exactly the same job as the first line of the body of
4994 the @code{simplified-beginning-of-buffer} function, which is written
4995 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
4996 at the current position of the cursor.
4998 I don't know why the expression in @code{mark-whole-buffer} is written
4999 @code{(push-mark (point))} and the expression in
5000 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
5001 whoever wrote the code did not know that the arguments for
5002 @code{push-mark} are optional and that if @code{push-mark} is not
5003 passed an argument, the function automatically sets mark at the
5004 location of point by default. Or perhaps the expression was written
5005 so as to parallel the structure of the next line. In any case, the
5006 line causes Emacs to determine the position of point and set a mark
5009 In earlier versions of GNU Emacs, the next line of
5010 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
5011 expression sets a mark at the point in the buffer that has the highest
5012 number. This will be the end of the buffer (or, if the buffer is
5013 narrowed, the end of the accessible portion of the buffer.
5014 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
5015 narrowing.) After this mark has been set, the previous mark, the one
5016 set at point, is no longer set, but Emacs remembers its position, just
5017 as all other recent marks are always remembered. This means that you
5018 can, if you wish, go back to that position by typing @kbd{C-u
5022 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5026 (push-mark (point-max) nil t)
5030 The expression works nearly the same as before. It sets a mark at the
5031 highest numbered place in the buffer that it can. However, in this
5032 version, @code{push-mark} has two additional arguments. The second
5033 argument to @code{push-mark} is @code{nil}. This tells the function
5034 it @emph{should} display a message that says `Mark set' when it pushes
5035 the mark. The third argument is @code{t}. This tells
5036 @code{push-mark} to activate the mark when Transient Mark mode is
5037 turned on. Transient Mark mode highlights the currently active
5038 region. It is often turned off.
5040 Finally, the last line of the function is @code{(goto-char
5041 (point-min)))}. This is written exactly the same way as it is written
5042 in @code{beginning-of-buffer}. The expression moves the cursor to
5043 the minimum point in the buffer, that is, to the beginning of the buffer
5044 (or to the beginning of the accessible portion of the buffer). As a
5045 result of this, point is placed at the beginning of the buffer and mark
5046 is set at the end of the buffer. The whole buffer is, therefore, the
5049 @node append-to-buffer
5050 @section The Definition of @code{append-to-buffer}
5051 @findex append-to-buffer
5053 The @code{append-to-buffer} command is more complex than the
5054 @code{mark-whole-buffer} command. What it does is copy the region
5055 (that is, the part of the buffer between point and mark) from the
5056 current buffer to a specified buffer.
5059 * append-to-buffer overview::
5060 * append interactive:: A two part interactive expression.
5061 * append-to-buffer body:: Incorporates a @code{let} expression.
5062 * append save-excursion:: How the @code{save-excursion} works.
5066 @node append-to-buffer overview
5067 @unnumberedsubsec An Overview of @code{append-to-buffer}
5070 @findex insert-buffer-substring
5071 The @code{append-to-buffer} command uses the
5072 @code{insert-buffer-substring} function to copy the region.
5073 @code{insert-buffer-substring} is described by its name: it takes a
5074 string of characters from part of a buffer, a ``substring'', and
5075 inserts them into another buffer.
5077 Most of @code{append-to-buffer} is
5078 concerned with setting up the conditions for
5079 @code{insert-buffer-substring} to work: the code must specify both the
5080 buffer to which the text will go, the window it comes from and goes
5081 to, and the region that will be copied.
5084 Here is the complete text of the function:
5088 (defun append-to-buffer (buffer start end)
5089 "Append to specified buffer the text of the region.
5090 It is inserted into that buffer before its point.
5094 When calling from a program, give three arguments:
5095 BUFFER (or buffer name), START and END.
5096 START and END specify the portion of the current buffer to be copied."
5098 (list (read-buffer "Append to buffer: " (other-buffer
5099 (current-buffer) t))
5100 (region-beginning) (region-end)))
5103 (let ((oldbuf (current-buffer)))
5105 (let* ((append-to (get-buffer-create buffer))
5106 (windows (get-buffer-window-list append-to t t))
5108 (set-buffer append-to)
5109 (setq point (point))
5110 (barf-if-buffer-read-only)
5111 (insert-buffer-substring oldbuf start end)
5112 (dolist (window windows)
5113 (when (= (window-point window) point)
5114 (set-window-point window (point))))))))
5118 The function can be understood by looking at it as a series of
5119 filled-in templates.
5121 The outermost template is for the function definition. In this
5122 function, it looks like this (with several slots filled in):
5126 (defun append-to-buffer (buffer start end)
5127 "@var{documentation}@dots{}"
5128 (interactive @dots{})
5133 The first line of the function includes its name and three arguments.
5134 The arguments are the @code{buffer} to which the text will be copied, and
5135 the @code{start} and @code{end} of the region in the current buffer that
5138 The next part of the function is the documentation, which is clear and
5139 complete. As is conventional, the three arguments are written in
5140 upper case so you will notice them easily. Even better, they are
5141 described in the same order as in the argument list.
5143 Note that the documentation distinguishes between a buffer and its
5144 name. (The function can handle either.)
5146 @node append interactive
5147 @subsection The @code{append-to-buffer} Interactive Expression
5149 Since the @code{append-to-buffer} function will be used interactively,
5150 the function must have an @code{interactive} expression. (For a
5151 review of @code{interactive}, see @ref{Interactive, , Making a
5152 Function Interactive}.) The expression reads as follows:
5158 "Append to buffer: "
5159 (other-buffer (current-buffer) t))
5166 This expression is not one with letters standing for parts, as
5167 described earlier. Instead, it starts a list with these parts:
5169 The first part of the list is an expression to read the name of a
5170 buffer and return it as a string. That is @code{read-buffer}. The
5171 function requires a prompt as its first argument, @samp{"Append to
5172 buffer: "}. Its second argument tells the command what value to
5173 provide if you don't specify anything.
5175 In this case that second argument is an expression containing the
5176 function @code{other-buffer}, an exception, and a @samp{t}, standing
5179 The first argument to @code{other-buffer}, the exception, is yet
5180 another function, @code{current-buffer}. That is not going to be
5181 returned. The second argument is the symbol for true, @code{t}. that
5182 tells @code{other-buffer} that it may show visible buffers (except in
5183 this case, it will not show the current buffer, which makes sense).
5186 The expression looks like this:
5189 (other-buffer (current-buffer) t)
5192 The second and third arguments to the @code{list} expression are
5193 @code{(region-beginning)} and @code{(region-end)}. These two
5194 functions specify the beginning and end of the text to be appended.
5197 Originally, the command used the letters @samp{B} and @samp{r}.
5198 The whole @code{interactive} expression looked like this:
5201 (interactive "BAppend to buffer:@: \nr")
5205 But when that was done, the default value of the buffer switched to
5206 was invisible. That was not wanted.
5208 (The prompt was separated from the second argument with a newline,
5209 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5210 two arguments that follow the symbol @code{buffer} in the function's
5211 argument list (that is, @code{start} and @code{end}) to the values of
5212 point and mark. That argument worked fine.)
5214 @node append-to-buffer body
5215 @subsection The Body of @code{append-to-buffer}
5218 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5220 (defun append-to-buffer (buffer start end)
5221 "Append to specified buffer the text of the region.
5222 It is inserted into that buffer before its point.
5224 When calling from a program, give three arguments:
5225 BUFFER (or buffer name), START and END.
5226 START and END specify the portion of the current buffer to be copied."
5228 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5229 (region-beginning) (region-end)))
5230 (let ((oldbuf (current-buffer)))
5232 (let* ((append-to (get-buffer-create buffer))
5233 (windows (get-buffer-window-list append-to t t))
5235 (set-buffer append-to)
5236 (setq point (point))
5237 (barf-if-buffer-read-only)
5238 (insert-buffer-substring oldbuf start end)
5239 (dolist (window windows)
5240 (when (= (window-point window) point)
5241 (set-window-point window (point))))))))
5244 The body of the @code{append-to-buffer} function begins with @code{let}.
5246 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5247 @code{let} expression is to create and give initial values to one or
5248 more variables that will only be used within the body of the
5249 @code{let}. This means that such a variable will not be confused with
5250 any variable of the same name outside the @code{let} expression.
5252 We can see how the @code{let} expression fits into the function as a
5253 whole by showing a template for @code{append-to-buffer} with the
5254 @code{let} expression in outline:
5258 (defun append-to-buffer (buffer start end)
5259 "@var{documentation}@dots{}"
5260 (interactive @dots{})
5261 (let ((@var{variable} @var{value}))
5266 The @code{let} expression has three elements:
5270 The symbol @code{let};
5273 A varlist containing, in this case, a single two-element list,
5274 @code{(@var{variable} @var{value})};
5277 The body of the @code{let} expression.
5281 In the @code{append-to-buffer} function, the varlist looks like this:
5284 (oldbuf (current-buffer))
5288 In this part of the @code{let} expression, the one variable,
5289 @code{oldbuf}, is bound to the value returned by the
5290 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5291 used to keep track of the buffer in which you are working and from
5292 which you will copy.
5294 The element or elements of a varlist are surrounded by a set of
5295 parentheses so the Lisp interpreter can distinguish the varlist from
5296 the body of the @code{let}. As a consequence, the two-element list
5297 within the varlist is surrounded by a circumscribing set of parentheses.
5298 The line looks like this:
5302 (let ((oldbuf (current-buffer)))
5308 The two parentheses before @code{oldbuf} might surprise you if you did
5309 not realize that the first parenthesis before @code{oldbuf} marks the
5310 boundary of the varlist and the second parenthesis marks the beginning
5311 of the two-element list, @code{(oldbuf (current-buffer))}.
5313 @node append save-excursion
5314 @subsection @code{save-excursion} in @code{append-to-buffer}
5316 The body of the @code{let} expression in @code{append-to-buffer}
5317 consists of a @code{save-excursion} expression.
5319 The @code{save-excursion} function saves the locations of point and
5320 mark, and restores them to those positions after the expressions in the
5321 body of the @code{save-excursion} complete execution. In addition,
5322 @code{save-excursion} keeps track of the original buffer, and
5323 restores it. This is how @code{save-excursion} is used in
5324 @code{append-to-buffer}.
5327 @cindex Indentation for formatting
5328 @cindex Formatting convention
5329 Incidentally, it is worth noting here that a Lisp function is normally
5330 formatted so that everything that is enclosed in a multi-line spread is
5331 indented more to the right than the first symbol. In this function
5332 definition, the @code{let} is indented more than the @code{defun}, and
5333 the @code{save-excursion} is indented more than the @code{let}, like
5349 This formatting convention makes it easy to see that the lines in
5350 the body of the @code{save-excursion} are enclosed by the parentheses
5351 associated with @code{save-excursion}, just as the
5352 @code{save-excursion} itself is enclosed by the parentheses associated
5353 with the @code{let}:
5357 (let ((oldbuf (current-buffer)))
5360 (set-buffer @dots{})
5361 (insert-buffer-substring oldbuf start end)
5367 The use of the @code{save-excursion} function can be viewed as a process
5368 of filling in the slots of a template:
5373 @var{first-expression-in-body}
5374 @var{second-expression-in-body}
5376 @var{last-expression-in-body})
5382 In this function, the body of the @code{save-excursion} contains only
5383 one expression, the @code{let*} expression. You know about a
5384 @code{let} function. The @code{let*} function is different. It has a
5385 @samp{*} in its name. It enables Emacs to set each variable in its
5386 varlist in sequence, one after another.
5388 Its critical feature is that variables later in the varlist can make
5389 use of the values to which Emacs set variables earlier in the varlist.
5390 @xref{fwd-para let, , The @code{let*} expression}.
5392 We will skip functions like @code{let*} and focus on two: the
5393 @code{set-buffer} function and the @code{insert-buffer-substring}
5397 In the old days, the @code{set-buffer} expression was simply
5400 (set-buffer (get-buffer-create buffer))
5408 (set-buffer append-to)
5412 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5413 on in the @code{let*} expression. That extra binding would not be
5414 necessary except for that @code{append-to} is used later in the
5415 varlist as an argument to @code{get-buffer-window-list}.
5420 (let ((oldbuf (current-buffer)))
5422 (let* ((append-to (get-buffer-create buffer))
5423 (windows (get-buffer-window-list append-to t t))
5425 (set-buffer append-to)
5426 (setq point (point))
5427 (barf-if-buffer-read-only)
5428 (insert-buffer-substring oldbuf start end)
5429 (dolist (window windows)
5430 (when (= (window-point window) point)
5431 (set-window-point window (point))))))))
5434 The @code{append-to-buffer} function definition inserts text from the
5435 buffer in which you are currently to a named buffer. It happens that
5436 @code{insert-buffer-substring} copies text from another buffer to the
5437 current buffer, just the reverse---that is why the
5438 @code{append-to-buffer} definition starts out with a @code{let} that
5439 binds the local symbol @code{oldbuf} to the value returned by
5440 @code{current-buffer}.
5443 The @code{insert-buffer-substring} expression looks like this:
5446 (insert-buffer-substring oldbuf start end)
5450 The @code{insert-buffer-substring} function copies a string
5451 @emph{from} the buffer specified as its first argument and inserts the
5452 string into the present buffer. In this case, the argument to
5453 @code{insert-buffer-substring} is the value of the variable created
5454 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5455 was the current buffer when you gave the @code{append-to-buffer}
5458 After @code{insert-buffer-substring} has done its work,
5459 @code{save-excursion} will restore the action to the original buffer
5460 and @code{append-to-buffer} will have done its job.
5463 Written in skeletal form, the workings of the body look like this:
5467 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5468 (save-excursion ; @r{Keep track of buffer.}
5470 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5472 @var{change-back-to-original-buffer-when-finished}
5473 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5477 In summary, @code{append-to-buffer} works as follows: it saves the
5478 value of the current buffer in the variable called @code{oldbuf}. It
5479 gets the new buffer (creating one if need be) and switches Emacs's
5480 attention to it. Using the value of @code{oldbuf}, it inserts the
5481 region of text from the old buffer into the new buffer; and then using
5482 @code{save-excursion}, it brings you back to your original buffer.
5484 In looking at @code{append-to-buffer}, you have explored a fairly
5485 complex function. It shows how to use @code{let} and
5486 @code{save-excursion}, and how to change to and come back from another
5487 buffer. Many function definitions use @code{let},
5488 @code{save-excursion}, and @code{set-buffer} this way.
5490 @node Buffer Related Review
5493 Here is a brief summary of the various functions discussed in this chapter.
5496 @item describe-function
5497 @itemx describe-variable
5498 Print the documentation for a function or variable.
5499 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5502 Find the file containing the source for a function or variable and
5503 switch buffers to it, positioning point at the beginning of the item.
5504 Conventionally bound to @kbd{M-.} (that's a period following the
5507 @item save-excursion
5508 Save the location of point and mark and restore their values after the
5509 arguments to @code{save-excursion} have been evaluated. Also, remember
5510 the current buffer and return to it.
5513 Set mark at a location and record the value of the previous mark on the
5514 mark ring. The mark is a location in the buffer that will keep its
5515 relative position even if text is added to or removed from the buffer.
5518 Set point to the location specified by the value of the argument, which
5519 can be a number, a marker, or an expression that returns the number of
5520 a position, such as @code{(point-min)}.
5522 @item insert-buffer-substring
5523 Copy a region of text from a buffer that is passed to the function as
5524 an argument and insert the region into the current buffer.
5526 @item mark-whole-buffer
5527 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5530 Switch the attention of Emacs to another buffer, but do not change the
5531 window being displayed. Used when the program rather than a human is
5532 to work on a different buffer.
5534 @item get-buffer-create
5536 Find a named buffer or create one if a buffer of that name does not
5537 exist. The @code{get-buffer} function returns @code{nil} if the named
5538 buffer does not exist.
5542 @node Buffer Exercises
5547 Write your own @code{simplified-end-of-buffer} function definition;
5548 then test it to see whether it works.
5551 Use @code{if} and @code{get-buffer} to write a function that prints a
5552 message telling you whether a buffer exists.
5555 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5560 @chapter A Few More Complex Functions
5562 In this chapter, we build on what we have learned in previous chapters
5563 by looking at more complex functions. The @code{copy-to-buffer}
5564 function illustrates use of two @code{save-excursion} expressions in
5565 one definition, while the @code{insert-buffer} function illustrates
5566 use of an asterisk in an @code{interactive} expression, use of
5567 @code{or}, and the important distinction between a name and the object
5568 to which the name refers.
5571 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5572 * insert-buffer:: Read-only, and with @code{or}.
5573 * beginning-of-buffer:: Shows @code{goto-char},
5574 @code{point-min}, and @code{push-mark}.
5575 * Second Buffer Related Review::
5576 * optional Exercise::
5579 @node copy-to-buffer
5580 @section The Definition of @code{copy-to-buffer}
5581 @findex copy-to-buffer
5583 After understanding how @code{append-to-buffer} works, it is easy to
5584 understand @code{copy-to-buffer}. This function copies text into a
5585 buffer, but instead of adding to the second buffer, it replaces all the
5586 previous text in the second buffer.
5589 The body of @code{copy-to-buffer} looks like this,
5594 (interactive "BCopy to buffer: \nr")
5595 (let ((oldbuf (current-buffer)))
5596 (with-current-buffer (get-buffer-create buffer)
5597 (barf-if-buffer-read-only)
5600 (insert-buffer-substring oldbuf start end)))))
5604 The @code{copy-to-buffer} function has a simpler @code{interactive}
5605 expression than @code{append-to-buffer}.
5608 The definition then says
5611 (with-current-buffer (get-buffer-create buffer) @dots{}
5614 First, look at the earliest inner expression; that is evaluated first.
5615 That expression starts with @code{get-buffer-create buffer}. The
5616 function tells the computer to use the buffer with the name specified
5617 as the one to which you are copying, or if such a buffer does not
5618 exist, to create it. Then, the @code{with-current-buffer} function
5619 evaluates its body with that buffer temporarily current.
5621 (This demonstrates another way to shift the computer's attention but
5622 not the user's. The @code{append-to-buffer} function showed how to do
5623 the same with @code{save-excursion} and @code{set-buffer}.
5624 @code{with-current-buffer} is a newer, and arguably easier,
5627 The @code{barf-if-buffer-read-only} function sends you an error
5628 message saying the buffer is read-only if you cannot modify it.
5630 The next line has the @code{erase-buffer} function as its sole
5631 contents. That function erases the buffer.
5633 Finally, the last two lines contain the @code{save-excursion}
5634 expression with @code{insert-buffer-substring} as its body.
5635 The @code{insert-buffer-substring} expression copies the text from
5636 the buffer you are in (and you have not seen the computer shift its
5637 attention, so you don't know that that buffer is now called
5640 Incidentally, this is what is meant by `replacement'. To replace text,
5641 Emacs erases the previous text and then inserts new text.
5644 In outline, the body of @code{copy-to-buffer} looks like this:
5648 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5649 (@var{with-the-buffer-you-are-copying-to}
5650 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5653 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5658 @section The Definition of @code{insert-buffer}
5659 @findex insert-buffer
5661 @code{insert-buffer} is yet another buffer-related function. This
5662 command copies another buffer @emph{into} the current buffer. It is the
5663 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5664 copy a region of text @emph{from} the current buffer to another buffer.
5666 Here is a discussion based on the original code. The code was
5667 simplified in 2003 and is harder to understand.
5669 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5670 a discussion of the new body.)
5672 In addition, this code illustrates the use of @code{interactive} with a
5673 buffer that might be @dfn{read-only} and the important distinction
5674 between the name of an object and the object actually referred to.
5677 * insert-buffer code::
5678 * insert-buffer interactive:: When you can read, but not write.
5679 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5680 * if & or:: Using an @code{if} instead of an @code{or}.
5681 * Insert or:: How the @code{or} expression works.
5682 * Insert let:: Two @code{save-excursion} expressions.
5683 * New insert-buffer::
5687 @node insert-buffer code
5688 @unnumberedsubsec The Code for @code{insert-buffer}
5692 Here is the earlier code:
5696 (defun insert-buffer (buffer)
5697 "Insert after point the contents of BUFFER.
5698 Puts mark after the inserted text.
5699 BUFFER may be a buffer or a buffer name."
5700 (interactive "*bInsert buffer:@: ")
5703 (or (bufferp buffer)
5704 (setq buffer (get-buffer buffer)))
5705 (let (start end newmark)
5709 (setq start (point-min) end (point-max)))
5712 (insert-buffer-substring buffer start end)
5713 (setq newmark (point)))
5714 (push-mark newmark)))
5719 As with other function definitions, you can use a template to see an
5720 outline of the function:
5724 (defun insert-buffer (buffer)
5725 "@var{documentation}@dots{}"
5726 (interactive "*bInsert buffer:@: ")
5731 @node insert-buffer interactive
5732 @subsection The Interactive Expression in @code{insert-buffer}
5733 @findex interactive, @r{example use of}
5735 In @code{insert-buffer}, the argument to the @code{interactive}
5736 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5740 * Read-only buffer:: When a buffer cannot be modified.
5741 * b for interactive:: An existing buffer or else its name.
5744 @node Read-only buffer
5745 @unnumberedsubsubsec A Read-only Buffer
5746 @cindex Read-only buffer
5747 @cindex Asterisk for read-only buffer
5748 @findex * @r{for read-only buffer}
5750 The asterisk is for the situation when the current buffer is a
5751 read-only buffer---a buffer that cannot be modified. If
5752 @code{insert-buffer} is called when the current buffer is read-only, a
5753 message to this effect is printed in the echo area and the terminal
5754 may beep or blink at you; you will not be permitted to insert anything
5755 into current buffer. The asterisk does not need to be followed by a
5756 newline to separate it from the next argument.
5758 @node b for interactive
5759 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5761 The next argument in the interactive expression starts with a lower
5762 case @samp{b}. (This is different from the code for
5763 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5764 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5765 The lower-case @samp{b} tells the Lisp interpreter that the argument
5766 for @code{insert-buffer} should be an existing buffer or else its
5767 name. (The upper-case @samp{B} option provides for the possibility
5768 that the buffer does not exist.) Emacs will prompt you for the name
5769 of the buffer, offering you a default buffer, with name completion
5770 enabled. If the buffer does not exist, you receive a message that
5771 says ``No match''; your terminal may beep at you as well.
5773 The new and simplified code generates a list for @code{interactive}.
5774 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5775 functions with which we are already familiar and the @code{progn}
5776 special form with which we are not. (It will be described later.)
5778 @node insert-buffer body
5779 @subsection The Body of the @code{insert-buffer} Function
5781 The body of the @code{insert-buffer} function has two major parts: an
5782 @code{or} expression and a @code{let} expression. The purpose of the
5783 @code{or} expression is to ensure that the argument @code{buffer} is
5784 bound to a buffer and not just the name of a buffer. The body of the
5785 @code{let} expression contains the code which copies the other buffer
5786 into the current buffer.
5789 In outline, the two expressions fit into the @code{insert-buffer}
5794 (defun insert-buffer (buffer)
5795 "@var{documentation}@dots{}"
5796 (interactive "*bInsert buffer:@: ")
5801 (let (@var{varlist})
5802 @var{body-of-}@code{let}@dots{} )
5806 To understand how the @code{or} expression ensures that the argument
5807 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5808 is first necessary to understand the @code{or} function.
5810 Before doing this, let me rewrite this part of the function using
5811 @code{if} so that you can see what is done in a manner that will be familiar.
5814 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5816 The job to be done is to make sure the value of @code{buffer} is a
5817 buffer itself and not the name of a buffer. If the value is the name,
5818 then the buffer itself must be got.
5820 You can imagine yourself at a conference where an usher is wandering
5821 around holding a list with your name on it and looking for you: the
5822 usher is ``bound'' to your name, not to you; but when the usher finds
5823 you and takes your arm, the usher becomes ``bound'' to you.
5826 In Lisp, you might describe this situation like this:
5830 (if (not (holding-on-to-guest))
5831 (find-and-take-arm-of-guest))
5835 We want to do the same thing with a buffer---if we do not have the
5836 buffer itself, we want to get it.
5839 Using a predicate called @code{bufferp} that tells us whether we have a
5840 buffer (rather than its name), we can write the code like this:
5844 (if (not (bufferp buffer)) ; @r{if-part}
5845 (setq buffer (get-buffer buffer))) ; @r{then-part}
5850 Here, the true-or-false-test of the @code{if} expression is
5851 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5852 @w{@code{(setq buffer (get-buffer buffer))}}.
5854 In the test, the function @code{bufferp} returns true if its argument is
5855 a buffer---but false if its argument is the name of the buffer. (The
5856 last character of the function name @code{bufferp} is the character
5857 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5858 indicates that the function is a predicate, which is a term that means
5859 that the function will determine whether some property is true or false.
5860 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5864 The function @code{not} precedes the expression @code{(bufferp buffer)},
5865 so the true-or-false-test looks like this:
5868 (not (bufferp buffer))
5872 @code{not} is a function that returns true if its argument is false
5873 and false if its argument is true. So if @code{(bufferp buffer)}
5874 returns true, the @code{not} expression returns false and vice-verse:
5875 what is ``not true'' is false and what is ``not false'' is true.
5877 Using this test, the @code{if} expression works as follows: when the
5878 value of the variable @code{buffer} is actually a buffer rather than
5879 its name, the true-or-false-test returns false and the @code{if}
5880 expression does not evaluate the then-part. This is fine, since we do
5881 not need to do anything to the variable @code{buffer} if it really is
5884 On the other hand, when the value of @code{buffer} is not a buffer
5885 itself, but the name of a buffer, the true-or-false-test returns true
5886 and the then-part of the expression is evaluated. In this case, the
5887 then-part is @code{(setq buffer (get-buffer buffer))}. This
5888 expression uses the @code{get-buffer} function to return an actual
5889 buffer itself, given its name. The @code{setq} then sets the variable
5890 @code{buffer} to the value of the buffer itself, replacing its previous
5891 value (which was the name of the buffer).
5894 @subsection The @code{or} in the Body
5896 The purpose of the @code{or} expression in the @code{insert-buffer}
5897 function is to ensure that the argument @code{buffer} is bound to a
5898 buffer and not just to the name of a buffer. The previous section shows
5899 how the job could have been done using an @code{if} expression.
5900 However, the @code{insert-buffer} function actually uses @code{or}.
5901 To understand this, it is necessary to understand how @code{or} works.
5904 An @code{or} function can have any number of arguments. It evaluates
5905 each argument in turn and returns the value of the first of its
5906 arguments that is not @code{nil}. Also, and this is a crucial feature
5907 of @code{or}, it does not evaluate any subsequent arguments after
5908 returning the first non-@code{nil} value.
5911 The @code{or} expression looks like this:
5915 (or (bufferp buffer)
5916 (setq buffer (get-buffer buffer)))
5921 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5922 This expression returns true (a non-@code{nil} value) if the buffer is
5923 actually a buffer, and not just the name of a buffer. In the @code{or}
5924 expression, if this is the case, the @code{or} expression returns this
5925 true value and does not evaluate the next expression---and this is fine
5926 with us, since we do not want to do anything to the value of
5927 @code{buffer} if it really is a buffer.
5929 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5930 which it will be if the value of @code{buffer} is the name of a buffer,
5931 the Lisp interpreter evaluates the next element of the @code{or}
5932 expression. This is the expression @code{(setq buffer (get-buffer
5933 buffer))}. This expression returns a non-@code{nil} value, which
5934 is the value to which it sets the variable @code{buffer}---and this
5935 value is a buffer itself, not the name of a buffer.
5937 The result of all this is that the symbol @code{buffer} is always
5938 bound to a buffer itself rather than to the name of a buffer. All
5939 this is necessary because the @code{set-buffer} function in a
5940 following line only works with a buffer itself, not with the name to a
5944 Incidentally, using @code{or}, the situation with the usher would be
5948 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5952 @subsection The @code{let} Expression in @code{insert-buffer}
5954 After ensuring that the variable @code{buffer} refers to a buffer itself
5955 and not just to the name of a buffer, the @code{insert-buffer function}
5956 continues with a @code{let} expression. This specifies three local
5957 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5958 to the initial value @code{nil}. These variables are used inside the
5959 remainder of the @code{let} and temporarily hide any other occurrence of
5960 variables of the same name in Emacs until the end of the @code{let}.
5963 The body of the @code{let} contains two @code{save-excursion}
5964 expressions. First, we will look at the inner @code{save-excursion}
5965 expression in detail. The expression looks like this:
5971 (setq start (point-min) end (point-max)))
5976 The expression @code{(set-buffer buffer)} changes Emacs's attention
5977 from the current buffer to the one from which the text will copied.
5978 In that buffer, the variables @code{start} and @code{end} are set to
5979 the beginning and end of the buffer, using the commands
5980 @code{point-min} and @code{point-max}. Note that we have here an
5981 illustration of how @code{setq} is able to set two variables in the
5982 same expression. The first argument of @code{setq} is set to the
5983 value of its second, and its third argument is set to the value of its
5986 After the body of the inner @code{save-excursion} is evaluated, the
5987 @code{save-excursion} restores the original buffer, but @code{start} and
5988 @code{end} remain set to the values of the beginning and end of the
5989 buffer from which the text will be copied.
5992 The outer @code{save-excursion} expression looks like this:
5997 (@var{inner-}@code{save-excursion}@var{-expression}
5998 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
5999 (insert-buffer-substring buffer start end)
6000 (setq newmark (point)))
6005 The @code{insert-buffer-substring} function copies the text
6006 @emph{into} the current buffer @emph{from} the region indicated by
6007 @code{start} and @code{end} in @code{buffer}. Since the whole of the
6008 second buffer lies between @code{start} and @code{end}, the whole of
6009 the second buffer is copied into the buffer you are editing. Next,
6010 the value of point, which will be at the end of the inserted text, is
6011 recorded in the variable @code{newmark}.
6013 After the body of the outer @code{save-excursion} is evaluated, point
6014 and mark are relocated to their original places.
6016 However, it is convenient to locate a mark at the end of the newly
6017 inserted text and locate point at its beginning. The @code{newmark}
6018 variable records the end of the inserted text. In the last line of
6019 the @code{let} expression, the @code{(push-mark newmark)} expression
6020 function sets a mark to this location. (The previous location of the
6021 mark is still accessible; it is recorded on the mark ring and you can
6022 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6023 located at the beginning of the inserted text, which is where it was
6024 before you called the insert function, the position of which was saved
6025 by the first @code{save-excursion}.
6028 The whole @code{let} expression looks like this:
6032 (let (start end newmark)
6036 (setq start (point-min) end (point-max)))
6037 (insert-buffer-substring buffer start end)
6038 (setq newmark (point)))
6039 (push-mark newmark))
6043 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6044 function uses @code{let}, @code{save-excursion}, and
6045 @code{set-buffer}. In addition, the function illustrates one way to
6046 use @code{or}. All these functions are building blocks that we will
6047 find and use again and again.
6049 @node New insert-buffer
6050 @subsection New Body for @code{insert-buffer}
6051 @findex insert-buffer, new version body
6052 @findex new version body for insert-buffer
6054 The body in the GNU Emacs 22 version is more confusing than the original.
6057 It consists of two expressions,
6063 (insert-buffer-substring (get-buffer buffer))
6071 except, and this is what confuses novices, very important work is done
6072 inside the @code{push-mark} expression.
6074 The @code{get-buffer} function returns a buffer with the name
6075 provided. You will note that the function is @emph{not} called
6076 @code{get-buffer-create}; it does not create a buffer if one does not
6077 already exist. The buffer returned by @code{get-buffer}, an existing
6078 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6079 whole of the buffer (since you did not specify anything else).
6081 The location into which the buffer is inserted is recorded by
6082 @code{push-mark}. Then the function returns @code{nil}, the value of
6083 its last command. Put another way, the @code{insert-buffer} function
6084 exists only to produce a side effect, inserting another buffer, not to
6087 @node beginning-of-buffer
6088 @section Complete Definition of @code{beginning-of-buffer}
6089 @findex beginning-of-buffer
6091 The basic structure of the @code{beginning-of-buffer} function has
6092 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6093 Simplified @code{beginning-of-buffer} Definition}.)
6094 This section describes the complex part of the definition.
6096 As previously described, when invoked without an argument,
6097 @code{beginning-of-buffer} moves the cursor to the beginning of the
6098 buffer (in truth, the beginning of the accessible portion of the
6099 buffer), leaving the mark at the previous position. However, when the
6100 command is invoked with a number between one and ten, the function
6101 considers that number to be a fraction of the length of the buffer,
6102 measured in tenths, and Emacs moves the cursor that fraction of the
6103 way from the beginning of the buffer. Thus, you can either call this
6104 function with the key command @kbd{M-<}, which will move the cursor to
6105 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6106 M-<} which will move the cursor to a point 70% of the way through the
6107 buffer. If a number bigger than ten is used for the argument, it
6108 moves to the end of the buffer.
6110 The @code{beginning-of-buffer} function can be called with or without an
6111 argument. The use of the argument is optional.
6114 * Optional Arguments::
6115 * beginning-of-buffer opt arg:: Example with optional argument.
6116 * beginning-of-buffer complete::
6119 @node Optional Arguments
6120 @subsection Optional Arguments
6122 Unless told otherwise, Lisp expects that a function with an argument in
6123 its function definition will be called with a value for that argument.
6124 If that does not happen, you get an error and a message that says
6125 @samp{Wrong number of arguments}.
6127 @cindex Optional arguments
6130 However, optional arguments are a feature of Lisp: a particular
6131 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6132 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6133 @samp{optional} is part of the keyword.) In a function definition, if
6134 an argument follows the keyword @code{&optional}, no value need be
6135 passed to that argument when the function is called.
6138 The first line of the function definition of @code{beginning-of-buffer}
6139 therefore looks like this:
6142 (defun beginning-of-buffer (&optional arg)
6146 In outline, the whole function looks like this:
6150 (defun beginning-of-buffer (&optional arg)
6151 "@var{documentation}@dots{}"
6153 (or (@var{is-the-argument-a-cons-cell} arg)
6154 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6156 (let (@var{determine-size-and-set-it})
6158 (@var{if-there-is-an-argument}
6159 @var{figure-out-where-to-go}
6166 The function is similar to the @code{simplified-beginning-of-buffer}
6167 function except that the @code{interactive} expression has @code{"P"}
6168 as an argument and the @code{goto-char} function is followed by an
6169 if-then-else expression that figures out where to put the cursor if
6170 there is an argument that is not a cons cell.
6172 (Since I do not explain a cons cell for many more chapters, please
6173 consider ignoring the function @code{consp}. @xref{List
6174 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6175 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6178 The @code{"P"} in the @code{interactive} expression tells Emacs to
6179 pass a prefix argument, if there is one, to the function in raw form.
6180 A prefix argument is made by typing the @key{META} key followed by a
6181 number, or by typing @kbd{C-u} and then a number. (If you don't type
6182 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6183 @code{"p"} in the @code{interactive} expression causes the function to
6184 convert a prefix arg to a number.)
6186 The true-or-false-test of the @code{if} expression looks complex, but
6187 it is not: it checks whether @code{arg} has a value that is not
6188 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6189 does; it checks whether its argument is a cons cell.) If @code{arg}
6190 has a value that is not @code{nil} (and is not a cons cell), which
6191 will be the case if @code{beginning-of-buffer} is called with a
6192 numeric argument, then this true-or-false-test will return true and
6193 the then-part of the @code{if} expression will be evaluated. On the
6194 other hand, if @code{beginning-of-buffer} is not called with an
6195 argument, the value of @code{arg} will be @code{nil} and the else-part
6196 of the @code{if} expression will be evaluated. The else-part is
6197 simply @code{point-min}, and when this is the outcome, the whole
6198 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6199 is how we saw the @code{beginning-of-buffer} function in its
6202 @node beginning-of-buffer opt arg
6203 @subsection @code{beginning-of-buffer} with an Argument
6205 When @code{beginning-of-buffer} is called with an argument, an
6206 expression is evaluated which calculates what value to pass to
6207 @code{goto-char}. This expression is rather complicated at first sight.
6208 It includes an inner @code{if} expression and much arithmetic. It looks
6213 (if (> (buffer-size) 10000)
6214 ;; @r{Avoid overflow for large buffer sizes!}
6215 (* (prefix-numeric-value arg)
6220 size (prefix-numeric-value arg))) 10)))
6225 * Disentangle beginning-of-buffer::
6226 * Large buffer case::
6227 * Small buffer case::
6231 @node Disentangle beginning-of-buffer
6232 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6235 Like other complex-looking expressions, the conditional expression
6236 within @code{beginning-of-buffer} can be disentangled by looking at it
6237 as parts of a template, in this case, the template for an if-then-else
6238 expression. In skeletal form, the expression looks like this:
6242 (if (@var{buffer-is-large}
6243 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6244 @var{else-use-alternate-calculation}
6248 The true-or-false-test of this inner @code{if} expression checks the
6249 size of the buffer. The reason for this is that the old version 18
6250 Emacs used numbers that are no bigger than eight million or so and in
6251 the computation that followed, the programmer feared that Emacs might
6252 try to use over-large numbers if the buffer were large. The term
6253 `overflow', mentioned in the comment, means numbers that are over
6254 large. More recent versions of Emacs use larger numbers, but this
6255 code has not been touched, if only because people now look at buffers
6256 that are far, far larger than ever before.
6258 There are two cases: if the buffer is large and if it is not.
6260 @node Large buffer case
6261 @unnumberedsubsubsec What happens in a large buffer
6263 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6264 whether the size of the buffer is greater than 10,000 characters. To do
6265 this, it uses the @code{>} function and the computation of @code{size}
6266 that comes from the let expression.
6268 In the old days, the function @code{buffer-size} was used. Not only
6269 was that function called several times, it gave the size of the whole
6270 buffer, not the accessible part. The computation makes much more
6271 sense when it handles just the accessible part. (@xref{Narrowing &
6272 Widening, , Narrowing and Widening}, for more information on focusing
6273 attention to an `accessible' part.)
6276 The line looks like this:
6284 When the buffer is large, the then-part of the @code{if} expression is
6285 evaluated. It reads like this (after formatting for easy reading):
6290 (prefix-numeric-value arg)
6296 This expression is a multiplication, with two arguments to the function
6299 The first argument is @code{(prefix-numeric-value arg)}. When
6300 @code{"P"} is used as the argument for @code{interactive}, the value
6301 passed to the function as its argument is passed a ``raw prefix
6302 argument'', and not a number. (It is a number in a list.) To perform
6303 the arithmetic, a conversion is necessary, and
6304 @code{prefix-numeric-value} does the job.
6306 @findex / @r{(division)}
6308 The second argument is @code{(/ size 10)}. This expression divides
6309 the numeric value by ten---the numeric value of the size of the
6310 accessible portion of the buffer. This produces a number that tells
6311 how many characters make up one tenth of the buffer size. (In Lisp,
6312 @code{/} is used for division, just as @code{*} is used for
6316 In the multiplication expression as a whole, this amount is multiplied
6317 by the value of the prefix argument---the multiplication looks like this:
6321 (* @var{numeric-value-of-prefix-arg}
6322 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6327 If, for example, the prefix argument is @samp{7}, the one-tenth value
6328 will be multiplied by 7 to give a position 70% of the way through.
6331 The result of all this is that if the accessible portion of the buffer
6332 is large, the @code{goto-char} expression reads like this:
6336 (goto-char (* (prefix-numeric-value arg)
6341 This puts the cursor where we want it.
6343 @node Small buffer case
6344 @unnumberedsubsubsec What happens in a small buffer
6346 If the buffer contains fewer than 10,000 characters, a slightly
6347 different computation is performed. You might think this is not
6348 necessary, since the first computation could do the job. However, in
6349 a small buffer, the first method may not put the cursor on exactly the
6350 desired line; the second method does a better job.
6353 The code looks like this:
6355 @c Keep this on one line.
6357 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6362 This is code in which you figure out what happens by discovering how the
6363 functions are embedded in parentheses. It is easier to read if you
6364 reformat it with each expression indented more deeply than its
6365 enclosing expression:
6373 (prefix-numeric-value arg)))
6380 Looking at parentheses, we see that the innermost operation is
6381 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6382 a number. In the following expression, this number is multiplied by
6383 the size of the accessible portion of the buffer:
6386 (* size (prefix-numeric-value arg))
6390 This multiplication creates a number that may be larger than the size of
6391 the buffer---seven times larger if the argument is 7, for example. Ten
6392 is then added to this number and finally the large number is divided by
6393 ten to provide a value that is one character larger than the percentage
6394 position in the buffer.
6396 The number that results from all this is passed to @code{goto-char} and
6397 the cursor is moved to that point.
6400 @node beginning-of-buffer complete
6401 @subsection The Complete @code{beginning-of-buffer}
6404 Here is the complete text of the @code{beginning-of-buffer} function:
6410 (defun beginning-of-buffer (&optional arg)
6411 "Move point to the beginning of the buffer;
6412 leave mark at previous position.
6413 With \\[universal-argument] prefix,
6414 do not set mark at previous position.
6416 put point N/10 of the way from the beginning.
6418 If the buffer is narrowed,
6419 this command uses the beginning and size
6420 of the accessible part of the buffer.
6424 Don't use this command in Lisp programs!
6425 \(goto-char (point-min)) is faster
6426 and avoids clobbering the mark."
6429 (and transient-mark-mode mark-active)
6433 (let ((size (- (point-max) (point-min))))
6434 (goto-char (if (and arg (not (consp arg)))
6437 ;; Avoid overflow for large buffer sizes!
6438 (* (prefix-numeric-value arg)
6440 (/ (+ 10 (* size (prefix-numeric-value arg)))
6443 (if arg (forward-line 1)))
6448 From before GNU Emacs 22
6451 (defun beginning-of-buffer (&optional arg)
6452 "Move point to the beginning of the buffer;
6453 leave mark at previous position.
6454 With arg N, put point N/10 of the way
6455 from the true beginning.
6458 Don't use this in Lisp programs!
6459 \(goto-char (point-min)) is faster
6460 and does not set the mark."
6467 (if (> (buffer-size) 10000)
6468 ;; @r{Avoid overflow for large buffer sizes!}
6469 (* (prefix-numeric-value arg)
6470 (/ (buffer-size) 10))
6473 (/ (+ 10 (* (buffer-size)
6474 (prefix-numeric-value arg)))
6477 (if arg (forward-line 1)))
6483 Except for two small points, the previous discussion shows how this
6484 function works. The first point deals with a detail in the
6485 documentation string, and the second point concerns the last line of
6489 In the documentation string, there is reference to an expression:
6492 \\[universal-argument]
6496 A @samp{\\} is used before the first square bracket of this
6497 expression. This @samp{\\} tells the Lisp interpreter to substitute
6498 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6499 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6500 be different. (@xref{Documentation Tips, , Tips for Documentation
6501 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6505 Finally, the last line of the @code{beginning-of-buffer} command says
6506 to move point to the beginning of the next line if the command is
6507 invoked with an argument:
6510 (if arg (forward-line 1)))
6514 This puts the cursor at the beginning of the first line after the
6515 appropriate tenths position in the buffer. This is a flourish that
6516 means that the cursor is always located @emph{at least} the requested
6517 tenths of the way through the buffer, which is a nicety that is,
6518 perhaps, not necessary, but which, if it did not occur, would be sure
6521 On the other hand, it also means that if you specify the command with
6522 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6523 argument' is simply a cons cell, then the command puts you at the
6524 beginning of the second line @dots{} I don't know whether this is
6525 intended or whether no one has dealt with the code to avoid this
6528 @node Second Buffer Related Review
6531 Here is a brief summary of some of the topics covered in this chapter.
6535 Evaluate each argument in sequence, and return the value of the first
6536 argument that is not @code{nil}; if none return a value that is not
6537 @code{nil}, return @code{nil}. In brief, return the first true value
6538 of the arguments; return a true value if one @emph{or} any of the
6542 Evaluate each argument in sequence, and if any are @code{nil}, return
6543 @code{nil}; if none are @code{nil}, return the value of the last
6544 argument. In brief, return a true value only if all the arguments are
6545 true; return a true value if one @emph{and} each of the others is
6549 A keyword used to indicate that an argument to a function definition
6550 is optional; this means that the function can be evaluated without the
6551 argument, if desired.
6553 @item prefix-numeric-value
6554 Convert the `raw prefix argument' produced by @code{(interactive
6555 "P")} to a numeric value.
6558 Move point forward to the beginning of the next line, or if the argument
6559 is greater than one, forward that many lines. If it can't move as far
6560 forward as it is supposed to, @code{forward-line} goes forward as far as
6561 it can and then returns a count of the number of additional lines it was
6562 supposed to move but couldn't.
6565 Delete the entire contents of the current buffer.
6568 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6571 @node optional Exercise
6572 @section @code{optional} Argument Exercise
6574 Write an interactive function with an optional argument that tests
6575 whether its argument, a number, is greater than or equal to, or else,
6576 less than the value of @code{fill-column}, and tells you which, in a
6577 message. However, if you do not pass an argument to the function, use
6578 56 as a default value.
6580 @node Narrowing & Widening
6581 @chapter Narrowing and Widening
6582 @cindex Focusing attention (narrowing)
6586 Narrowing is a feature of Emacs that makes it possible for you to focus
6587 on a specific part of a buffer, and work without accidentally changing
6588 other parts. Narrowing is normally disabled since it can confuse
6592 * Narrowing advantages:: The advantages of narrowing
6593 * save-restriction:: The @code{save-restriction} special form.
6594 * what-line:: The number of the line that point is on.
6599 @node Narrowing advantages
6600 @unnumberedsec The Advantages of Narrowing
6603 With narrowing, the rest of a buffer is made invisible, as if it weren't
6604 there. This is an advantage if, for example, you want to replace a word
6605 in one part of a buffer but not in another: you narrow to the part you want
6606 and the replacement is carried out only in that section, not in the rest
6607 of the buffer. Searches will only work within a narrowed region, not
6608 outside of one, so if you are fixing a part of a document, you can keep
6609 yourself from accidentally finding parts you do not need to fix by
6610 narrowing just to the region you want.
6611 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6613 However, narrowing does make the rest of the buffer invisible, which
6614 can scare people who inadvertently invoke narrowing and think they
6615 have deleted a part of their file. Moreover, the @code{undo} command
6616 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6617 (nor should it), so people can become quite desperate if they do not
6618 know that they can return the rest of a buffer to visibility with the
6619 @code{widen} command.
6620 (The key binding for @code{widen} is @kbd{C-x n w}.)
6622 Narrowing is just as useful to the Lisp interpreter as to a human.
6623 Often, an Emacs Lisp function is designed to work on just part of a
6624 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6625 buffer that has been narrowed. The @code{what-line} function, for
6626 example, removes the narrowing from a buffer, if it has any narrowing
6627 and when it has finished its job, restores the narrowing to what it was.
6628 On the other hand, the @code{count-lines} function
6629 uses narrowing to restrict itself to just that portion
6630 of the buffer in which it is interested and then restores the previous
6633 @node save-restriction
6634 @section The @code{save-restriction} Special Form
6635 @findex save-restriction
6637 In Emacs Lisp, you can use the @code{save-restriction} special form to
6638 keep track of whatever narrowing is in effect, if any. When the Lisp
6639 interpreter meets with @code{save-restriction}, it executes the code
6640 in the body of the @code{save-restriction} expression, and then undoes
6641 any changes to narrowing that the code caused. If, for example, the
6642 buffer is narrowed and the code that follows @code{save-restriction}
6643 gets rid of the narrowing, @code{save-restriction} returns the buffer
6644 to its narrowed region afterwards. In the @code{what-line} command,
6645 any narrowing the buffer may have is undone by the @code{widen}
6646 command that immediately follows the @code{save-restriction} command.
6647 Any original narrowing is restored just before the completion of the
6651 The template for a @code{save-restriction} expression is simple:
6661 The body of the @code{save-restriction} is one or more expressions that
6662 will be evaluated in sequence by the Lisp interpreter.
6664 Finally, a point to note: when you use both @code{save-excursion} and
6665 @code{save-restriction}, one right after the other, you should use
6666 @code{save-excursion} outermost. If you write them in reverse order,
6667 you may fail to record narrowing in the buffer to which Emacs switches
6668 after calling @code{save-excursion}. Thus, when written together,
6669 @code{save-excursion} and @code{save-restriction} should be written
6680 In other circumstances, when not written together, the
6681 @code{save-excursion} and @code{save-restriction} special forms must
6682 be written in the order appropriate to the function.
6698 /usr/local/src/emacs/lisp/simple.el
6701 "Print the current buffer line number and narrowed line number of point."
6703 (let ((start (point-min))
6704 (n (line-number-at-pos)))
6706 (message "Line %d" n)
6710 (message "line %d (narrowed line %d)"
6711 (+ n (line-number-at-pos start) -1) n))))))
6713 (defun line-number-at-pos (&optional pos)
6714 "Return (narrowed) buffer line number at position POS.
6715 If POS is nil, use current buffer location.
6716 Counting starts at (point-min), so the value refers
6717 to the contents of the accessible portion of the buffer."
6718 (let ((opoint (or pos (point))) start)
6720 (goto-char (point-min))
6721 (setq start (point))
6724 (1+ (count-lines start (point))))))
6726 (defun count-lines (start end)
6727 "Return number of lines between START and END.
6728 This is usually the number of newlines between them,
6729 but can be one more if START is not equal to END
6730 and the greater of them is not at the start of a line."
6733 (narrow-to-region start end)
6734 (goto-char (point-min))
6735 (if (eq selective-display t)
6738 (while (re-search-forward "[\n\C-m]" nil t 40)
6739 (setq done (+ 40 done)))
6740 (while (re-search-forward "[\n\C-m]" nil t 1)
6741 (setq done (+ 1 done)))
6742 (goto-char (point-max))
6743 (if (and (/= start end)
6747 (- (buffer-size) (forward-line (buffer-size)))))))
6751 @section @code{what-line}
6753 @cindex Widening, example of
6755 The @code{what-line} command tells you the number of the line in which
6756 the cursor is located. The function illustrates the use of the
6757 @code{save-restriction} and @code{save-excursion} commands. Here is the
6758 original text of the function:
6763 "Print the current line number (in the buffer) of point."
6770 (1+ (count-lines 1 (point)))))))
6774 (In recent versions of GNU Emacs, the @code{what-line} function has
6775 been expanded to tell you your line number in a narrowed buffer as
6776 well as your line number in a widened buffer. The recent version is
6777 more complex than the version shown here. If you feel adventurous,
6778 you might want to look at it after figuring out how this version
6779 works. You will probably need to use @kbd{C-h f}
6780 (@code{describe-function}). The newer version uses a conditional to
6781 determine whether the buffer has been narrowed.
6783 (Also, it uses @code{line-number-at-pos}, which among other simple
6784 expressions, such as @code{(goto-char (point-min))}, moves point to
6785 the beginning of the current line with @code{(forward-line 0)} rather
6786 than @code{beginning-of-line}.)
6788 The @code{what-line} function as shown here has a documentation line
6789 and is interactive, as you would expect. The next two lines use the
6790 functions @code{save-restriction} and @code{widen}.
6792 The @code{save-restriction} special form notes whatever narrowing is in
6793 effect, if any, in the current buffer and restores that narrowing after
6794 the code in the body of the @code{save-restriction} has been evaluated.
6796 The @code{save-restriction} special form is followed by @code{widen}.
6797 This function undoes any narrowing the current buffer may have had
6798 when @code{what-line} was called. (The narrowing that was there is
6799 the narrowing that @code{save-restriction} remembers.) This widening
6800 makes it possible for the line counting commands to count from the
6801 beginning of the buffer. Otherwise, they would have been limited to
6802 counting within the accessible region. Any original narrowing is
6803 restored just before the completion of the function by the
6804 @code{save-restriction} special form.
6806 The call to @code{widen} is followed by @code{save-excursion}, which
6807 saves the location of the cursor (i.e., of point) and of the mark, and
6808 restores them after the code in the body of the @code{save-excursion}
6809 uses the @code{beginning-of-line} function to move point.
6811 (Note that the @code{(widen)} expression comes between the
6812 @code{save-restriction} and @code{save-excursion} special forms. When
6813 you write the two @code{save- @dots{}} expressions in sequence, write
6814 @code{save-excursion} outermost.)
6817 The last two lines of the @code{what-line} function are functions to
6818 count the number of lines in the buffer and then print the number in the
6824 (1+ (count-lines 1 (point)))))))
6828 The @code{message} function prints a one-line message at the bottom of
6829 the Emacs screen. The first argument is inside of quotation marks and
6830 is printed as a string of characters. However, it may contain a
6831 @samp{%d} expression to print a following argument. @samp{%d} prints
6832 the argument as a decimal, so the message will say something such as
6836 The number that is printed in place of the @samp{%d} is computed by the
6837 last line of the function:
6840 (1+ (count-lines 1 (point)))
6846 (defun count-lines (start end)
6847 "Return number of lines between START and END.
6848 This is usually the number of newlines between them,
6849 but can be one more if START is not equal to END
6850 and the greater of them is not at the start of a line."
6853 (narrow-to-region start end)
6854 (goto-char (point-min))
6855 (if (eq selective-display t)
6858 (while (re-search-forward "[\n\C-m]" nil t 40)
6859 (setq done (+ 40 done)))
6860 (while (re-search-forward "[\n\C-m]" nil t 1)
6861 (setq done (+ 1 done)))
6862 (goto-char (point-max))
6863 (if (and (/= start end)
6867 (- (buffer-size) (forward-line (buffer-size)))))))
6871 What this does is count the lines from the first position of the
6872 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6873 one to that number. (The @code{1+} function adds one to its
6874 argument.) We add one to it because line 2 has only one line before
6875 it, and @code{count-lines} counts only the lines @emph{before} the
6878 After @code{count-lines} has done its job, and the message has been
6879 printed in the echo area, the @code{save-excursion} restores point and
6880 mark to their original positions; and @code{save-restriction} restores
6881 the original narrowing, if any.
6883 @node narrow Exercise
6884 @section Exercise with Narrowing
6886 Write a function that will display the first 60 characters of the
6887 current buffer, even if you have narrowed the buffer to its latter
6888 half so that the first line is inaccessible. Restore point, mark, and
6889 narrowing. For this exercise, you need to use a whole potpourri of
6890 functions, including @code{save-restriction}, @code{widen},
6891 @code{goto-char}, @code{point-min}, @code{message}, and
6892 @code{buffer-substring}.
6894 @cindex Properties, mention of @code{buffer-substring-no-properties}
6895 (@code{buffer-substring} is a previously unmentioned function you will
6896 have to investigate yourself; or perhaps you will have to use
6897 @code{buffer-substring-no-properties} or
6898 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6899 properties are a feature otherwise not discussed here. @xref{Text
6900 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6903 Additionally, do you really need @code{goto-char} or @code{point-min}?
6904 Or can you write the function without them?
6906 @node car cdr & cons
6907 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6908 @findex car, @r{introduced}
6909 @findex cdr, @r{introduced}
6911 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6912 functions. The @code{cons} function is used to construct lists, and
6913 the @code{car} and @code{cdr} functions are used to take them apart.
6915 In the walk through of the @code{copy-region-as-kill} function, we
6916 will see @code{cons} as well as two variants on @code{cdr},
6917 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6920 * Strange Names:: An historical aside: why the strange names?
6921 * car & cdr:: Functions for extracting part of a list.
6922 * cons:: Constructing a list.
6923 * nthcdr:: Calling @code{cdr} repeatedly.
6925 * setcar:: Changing the first element of a list.
6926 * setcdr:: Changing the rest of a list.
6932 @unnumberedsec Strange Names
6935 The name of the @code{cons} function is not unreasonable: it is an
6936 abbreviation of the word `construct'. The origins of the names for
6937 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6938 is an acronym from the phrase `Contents of the Address part of the
6939 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6940 the phrase `Contents of the Decrement part of the Register'. These
6941 phrases refer to specific pieces of hardware on the very early
6942 computer on which the original Lisp was developed. Besides being
6943 obsolete, the phrases have been completely irrelevant for more than 25
6944 years to anyone thinking about Lisp. Nonetheless, although a few
6945 brave scholars have begun to use more reasonable names for these
6946 functions, the old terms are still in use. In particular, since the
6947 terms are used in the Emacs Lisp source code, we will use them in this
6951 @section @code{car} and @code{cdr}
6953 The @sc{car} of a list is, quite simply, the first item in the list.
6954 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
6958 If you are reading this in Info in GNU Emacs, you can see this by
6959 evaluating the following:
6962 (car '(rose violet daisy buttercup))
6966 After evaluating the expression, @code{rose} will appear in the echo
6969 Clearly, a more reasonable name for the @code{car} function would be
6970 @code{first} and this is often suggested.
6972 @code{car} does not remove the first item from the list; it only reports
6973 what it is. After @code{car} has been applied to a list, the list is
6974 still the same as it was. In the jargon, @code{car} is
6975 `non-destructive'. This feature turns out to be important.
6977 The @sc{cdr} of a list is the rest of the list, that is, the
6978 @code{cdr} function returns the part of the list that follows the
6979 first item. Thus, while the @sc{car} of the list @code{'(rose violet
6980 daisy buttercup)} is @code{rose}, the rest of the list, the value
6981 returned by the @code{cdr} function, is @code{(violet daisy
6985 You can see this by evaluating the following in the usual way:
6988 (cdr '(rose violet daisy buttercup))
6992 When you evaluate this, @code{(violet daisy buttercup)} will appear in
6995 Like @code{car}, @code{cdr} does not remove any elements from the
6996 list---it just returns a report of what the second and subsequent
6999 Incidentally, in the example, the list of flowers is quoted. If it were
7000 not, the Lisp interpreter would try to evaluate the list by calling
7001 @code{rose} as a function. In this example, we do not want to do that.
7003 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
7005 (There is a lesson here: when you name new functions, consider very
7006 carefully what you are doing, since you may be stuck with the names
7007 for far longer than you expect. The reason this document perpetuates
7008 these names is that the Emacs Lisp source code uses them, and if I did
7009 not use them, you would have a hard time reading the code; but do,
7010 please, try to avoid using these terms yourself. The people who come
7011 after you will be grateful to you.)
7013 When @code{car} and @code{cdr} are applied to a list made up of symbols,
7014 such as the list @code{(pine fir oak maple)}, the element of the list
7015 returned by the function @code{car} is the symbol @code{pine} without
7016 any parentheses around it. @code{pine} is the first element in the
7017 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7018 oak maple)}, as you can see by evaluating the following expressions in
7023 (car '(pine fir oak maple))
7025 (cdr '(pine fir oak maple))
7029 On the other hand, in a list of lists, the first element is itself a
7030 list. @code{car} returns this first element as a list. For example,
7031 the following list contains three sub-lists, a list of carnivores, a
7032 list of herbivores and a list of sea mammals:
7036 (car '((lion tiger cheetah)
7037 (gazelle antelope zebra)
7038 (whale dolphin seal)))
7043 In this example, the first element or @sc{car} of the list is the list of
7044 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7045 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7049 (cdr '((lion tiger cheetah)
7050 (gazelle antelope zebra)
7051 (whale dolphin seal)))
7055 It is worth saying again that @code{car} and @code{cdr} are
7056 non-destructive---that is, they do not modify or change lists to which
7057 they are applied. This is very important for how they are used.
7059 Also, in the first chapter, in the discussion about atoms, I said that
7060 in Lisp, ``certain kinds of atom, such as an array, can be separated
7061 into parts; but the mechanism for doing this is different from the
7062 mechanism for splitting a list. As far as Lisp is concerned, the
7063 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7064 @code{car} and @code{cdr} functions are used for splitting lists and
7065 are considered fundamental to Lisp. Since they cannot split or gain
7066 access to the parts of an array, an array is considered an atom.
7067 Conversely, the other fundamental function, @code{cons}, can put
7068 together or construct a list, but not an array. (Arrays are handled
7069 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7070 Emacs Lisp Reference Manual}.)
7073 @section @code{cons}
7074 @findex cons, @r{introduced}
7076 The @code{cons} function constructs lists; it is the inverse of
7077 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7078 a four element list from the three element list, @code{(fir oak maple)}:
7081 (cons 'pine '(fir oak maple))
7086 After evaluating this list, you will see
7089 (pine fir oak maple)
7093 appear in the echo area. @code{cons} causes the creation of a new
7094 list in which the element is followed by the elements of the original
7097 We often say that `@code{cons} puts a new element at the beginning of
7098 a list; it attaches or pushes elements onto the list', but this
7099 phrasing can be misleading, since @code{cons} does not change an
7100 existing list, but creates a new one.
7102 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7106 * length:: How to find the length of a list.
7111 @unnumberedsubsec Build a list
7114 @code{cons} must have a list to attach to.@footnote{Actually, you can
7115 @code{cons} an element to an atom to produce a dotted pair. Dotted
7116 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7117 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7118 cannot start from absolutely nothing. If you are building a list, you
7119 need to provide at least an empty list at the beginning. Here is a
7120 series of @code{cons} expressions that build up a list of flowers. If
7121 you are reading this in Info in GNU Emacs, you can evaluate each of
7122 the expressions in the usual way; the value is printed in this text
7123 after @samp{@result{}}, which you may read as `evaluates to'.
7127 (cons 'buttercup ())
7128 @result{} (buttercup)
7132 (cons 'daisy '(buttercup))
7133 @result{} (daisy buttercup)
7137 (cons 'violet '(daisy buttercup))
7138 @result{} (violet daisy buttercup)
7142 (cons 'rose '(violet daisy buttercup))
7143 @result{} (rose violet daisy buttercup)
7148 In the first example, the empty list is shown as @code{()} and a list
7149 made up of @code{buttercup} followed by the empty list is constructed.
7150 As you can see, the empty list is not shown in the list that was
7151 constructed. All that you see is @code{(buttercup)}. The empty list is
7152 not counted as an element of a list because there is nothing in an empty
7153 list. Generally speaking, an empty list is invisible.
7155 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7156 two element list by putting @code{daisy} in front of @code{buttercup};
7157 and the third example constructs a three element list by putting
7158 @code{violet} in front of @code{daisy} and @code{buttercup}.
7161 @subsection Find the Length of a List: @code{length}
7164 You can find out how many elements there are in a list by using the Lisp
7165 function @code{length}, as in the following examples:
7169 (length '(buttercup))
7174 (length '(daisy buttercup))
7179 (length (cons 'violet '(daisy buttercup)))
7185 In the third example, the @code{cons} function is used to construct a
7186 three element list which is then passed to the @code{length} function as
7190 We can also use @code{length} to count the number of elements in an
7201 As you would expect, the number of elements in an empty list is zero.
7203 An interesting experiment is to find out what happens if you try to find
7204 the length of no list at all; that is, if you try to call @code{length}
7205 without giving it an argument, not even an empty list:
7213 What you see, if you evaluate this, is the error message
7216 Lisp error: (wrong-number-of-arguments length 0)
7220 This means that the function receives the wrong number of
7221 arguments, zero, when it expects some other number of arguments. In
7222 this case, one argument is expected, the argument being a list whose
7223 length the function is measuring. (Note that @emph{one} list is
7224 @emph{one} argument, even if the list has many elements inside it.)
7226 The part of the error message that says @samp{length} is the name of
7230 @code{length} is still a subroutine, but you need C-h f to discover that.
7232 In an earlier version:
7233 This is written with a special notation, @samp{#<subr},
7234 that indicates that the function @code{length} is one of the primitive
7235 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7236 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7237 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7242 @section @code{nthcdr}
7245 The @code{nthcdr} function is associated with the @code{cdr} function.
7246 What it does is take the @sc{cdr} of a list repeatedly.
7248 If you take the @sc{cdr} of the list @code{(pine fir
7249 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7250 repeat this on what was returned, you will be returned the list
7251 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7252 list will just give you the original @sc{cdr} since the function does
7253 not change the list. You need to evaluate the @sc{cdr} of the
7254 @sc{cdr} and so on.) If you continue this, eventually you will be
7255 returned an empty list, which in this case, instead of being shown as
7256 @code{()} is shown as @code{nil}.
7259 For review, here is a series of repeated @sc{cdr}s, the text following
7260 the @samp{@result{}} shows what is returned.
7264 (cdr '(pine fir oak maple))
7265 @result{}(fir oak maple)
7269 (cdr '(fir oak maple))
7270 @result{} (oak maple)
7295 You can also do several @sc{cdr}s without printing the values in
7300 (cdr (cdr '(pine fir oak maple)))
7301 @result{} (oak maple)
7306 In this example, the Lisp interpreter evaluates the innermost list first.
7307 The innermost list is quoted, so it just passes the list as it is to the
7308 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7309 second and subsequent elements of the list to the outermost @code{cdr},
7310 which produces a list composed of the third and subsequent elements of
7311 the original list. In this example, the @code{cdr} function is repeated
7312 and returns a list that consists of the original list without its
7315 The @code{nthcdr} function does the same as repeating the call to
7316 @code{cdr}. In the following example, the argument 2 is passed to the
7317 function @code{nthcdr}, along with the list, and the value returned is
7318 the list without its first two items, which is exactly the same
7319 as repeating @code{cdr} twice on the list:
7323 (nthcdr 2 '(pine fir oak maple))
7324 @result{} (oak maple)
7329 Using the original four element list, we can see what happens when
7330 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7335 ;; @r{Leave the list as it was.}
7336 (nthcdr 0 '(pine fir oak maple))
7337 @result{} (pine fir oak maple)
7341 ;; @r{Return a copy without the first element.}
7342 (nthcdr 1 '(pine fir oak maple))
7343 @result{} (fir oak maple)
7347 ;; @r{Return a copy of the list without three elements.}
7348 (nthcdr 3 '(pine fir oak maple))
7353 ;; @r{Return a copy lacking all four elements.}
7354 (nthcdr 4 '(pine fir oak maple))
7359 ;; @r{Return a copy lacking all elements.}
7360 (nthcdr 5 '(pine fir oak maple))
7369 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7370 The @code{nth} function takes the @sc{car} of the result returned by
7371 @code{nthcdr}. It returns the Nth element of the list.
7374 Thus, if it were not defined in C for speed, the definition of
7375 @code{nth} would be:
7380 "Returns the Nth element of LIST.
7381 N counts from zero. If LIST is not that long, nil is returned."
7382 (car (nthcdr n list)))
7387 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7388 but its definition was redone in C in the 1980s.)
7390 The @code{nth} function returns a single element of a list.
7391 This can be very convenient.
7393 Note that the elements are numbered from zero, not one. That is to
7394 say, the first element of a list, its @sc{car} is the zeroth element.
7395 This is called `zero-based' counting and often bothers people who
7396 are accustomed to the first element in a list being number one, which
7404 (nth 0 '("one" "two" "three"))
7407 (nth 1 '("one" "two" "three"))
7412 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7413 @code{cdr}, does not change the original list---the function is
7414 non-destructive. This is in sharp contrast to the @code{setcar} and
7415 @code{setcdr} functions.
7418 @section @code{setcar}
7421 As you might guess from their names, the @code{setcar} and @code{setcdr}
7422 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7423 They actually change the original list, unlike @code{car} and @code{cdr}
7424 which leave the original list as it was. One way to find out how this
7425 works is to experiment. We will start with the @code{setcar} function.
7428 First, we can make a list and then set the value of a variable to the
7429 list, using the @code{setq} function. Here is a list of animals:
7432 (setq animals '(antelope giraffe lion tiger))
7436 If you are reading this in Info inside of GNU Emacs, you can evaluate
7437 this expression in the usual fashion, by positioning the cursor after
7438 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7439 as I write this. This is one of the advantages of having the
7440 interpreter built into the computing environment. Incidentally, when
7441 there is nothing on the line after the final parentheses, such as a
7442 comment, point can be on the next line. Thus, if your cursor is in
7443 the first column of the next line, you do not need to move it.
7444 Indeed, Emacs permits any amount of white space after the final
7448 When we evaluate the variable @code{animals}, we see that it is bound to
7449 the list @code{(antelope giraffe lion tiger)}:
7454 @result{} (antelope giraffe lion tiger)
7459 Put another way, the variable @code{animals} points to the list
7460 @code{(antelope giraffe lion tiger)}.
7462 Next, evaluate the function @code{setcar} while passing it two
7463 arguments, the variable @code{animals} and the quoted symbol
7464 @code{hippopotamus}; this is done by writing the three element list
7465 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7469 (setcar animals 'hippopotamus)
7474 After evaluating this expression, evaluate the variable @code{animals}
7475 again. You will see that the list of animals has changed:
7480 @result{} (hippopotamus giraffe lion tiger)
7485 The first element on the list, @code{antelope} is replaced by
7486 @code{hippopotamus}.
7488 So we can see that @code{setcar} did not add a new element to the list
7489 as @code{cons} would have; it replaced @code{antelope} with
7490 @code{hippopotamus}; it @emph{changed} the list.
7493 @section @code{setcdr}
7496 The @code{setcdr} function is similar to the @code{setcar} function,
7497 except that the function replaces the second and subsequent elements of
7498 a list rather than the first element.
7500 (To see how to change the last element of a list, look ahead to
7501 @ref{kill-new function, , The @code{kill-new} function}, which uses
7502 the @code{nthcdr} and @code{setcdr} functions.)
7505 To see how this works, set the value of the variable to a list of
7506 domesticated animals by evaluating the following expression:
7509 (setq domesticated-animals '(horse cow sheep goat))
7514 If you now evaluate the list, you will be returned the list
7515 @code{(horse cow sheep goat)}:
7519 domesticated-animals
7520 @result{} (horse cow sheep goat)
7525 Next, evaluate @code{setcdr} with two arguments, the name of the
7526 variable which has a list as its value, and the list to which the
7527 @sc{cdr} of the first list will be set;
7530 (setcdr domesticated-animals '(cat dog))
7534 If you evaluate this expression, the list @code{(cat dog)} will appear
7535 in the echo area. This is the value returned by the function. The
7536 result we are interested in is the ``side effect'', which we can see by
7537 evaluating the variable @code{domesticated-animals}:
7541 domesticated-animals
7542 @result{} (horse cat dog)
7547 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7548 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7549 @code{(cow sheep goat)} to @code{(cat dog)}.
7554 Construct a list of four birds by evaluating several expressions with
7555 @code{cons}. Find out what happens when you @code{cons} a list onto
7556 itself. Replace the first element of the list of four birds with a
7557 fish. Replace the rest of that list with a list of other fish.
7559 @node Cutting & Storing Text
7560 @chapter Cutting and Storing Text
7561 @cindex Cutting and storing text
7562 @cindex Storing and cutting text
7563 @cindex Killing text
7564 @cindex Clipping text
7565 @cindex Erasing text
7566 @cindex Deleting text
7568 Whenever you cut or clip text out of a buffer with a `kill' command in
7569 GNU Emacs, it is stored in a list and you can bring it back with a
7572 (The use of the word `kill' in Emacs for processes which specifically
7573 @emph{do not} destroy the values of the entities is an unfortunate
7574 historical accident. A much more appropriate word would be `clip' since
7575 that is what the kill commands do; they clip text out of a buffer and
7576 put it into storage from which it can be brought back. I have often
7577 been tempted to replace globally all occurrences of `kill' in the Emacs
7578 sources with `clip' and all occurrences of `killed' with `clipped'.)
7581 * Storing Text:: Text is stored in a list.
7582 * zap-to-char:: Cutting out text up to a character.
7583 * kill-region:: Cutting text out of a region.
7584 * copy-region-as-kill:: A definition for copying text.
7585 * Digression into C:: Minor note on C programming language macros.
7586 * defvar:: How to give a variable an initial value.
7587 * cons & search-fwd Review::
7588 * search Exercises::
7593 @unnumberedsec Storing Text in a List
7596 When text is cut out of a buffer, it is stored on a list. Successive
7597 pieces of text are stored on the list successively, so the list might
7601 ("a piece of text" "previous piece")
7606 The function @code{cons} can be used to create a new list from a piece
7607 of text (an `atom', to use the jargon) and an existing list, like
7612 (cons "another piece"
7613 '("a piece of text" "previous piece"))
7619 If you evaluate this expression, a list of three elements will appear in
7623 ("another piece" "a piece of text" "previous piece")
7626 With the @code{car} and @code{nthcdr} functions, you can retrieve
7627 whichever piece of text you want. For example, in the following code,
7628 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7629 and the @code{car} returns the first element of that remainder---the
7630 second element of the original list:
7634 (car (nthcdr 1 '("another piece"
7637 @result{} "a piece of text"
7641 The actual functions in Emacs are more complex than this, of course.
7642 The code for cutting and retrieving text has to be written so that
7643 Emacs can figure out which element in the list you want---the first,
7644 second, third, or whatever. In addition, when you get to the end of
7645 the list, Emacs should give you the first element of the list, rather
7646 than nothing at all.
7648 The list that holds the pieces of text is called the @dfn{kill ring}.
7649 This chapter leads up to a description of the kill ring and how it is
7650 used by first tracing how the @code{zap-to-char} function works. This
7651 function uses (or `calls') a function that invokes a function that
7652 manipulates the kill ring. Thus, before reaching the mountains, we
7653 climb the foothills.
7655 A subsequent chapter describes how text that is cut from the buffer is
7656 retrieved. @xref{Yanking, , Yanking Text Back}.
7659 @section @code{zap-to-char}
7662 @c FIXME remove obsolete stuff
7663 The @code{zap-to-char} function changed little between GNU Emacs
7664 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7665 calls another function, @code{kill-region}, which enjoyed a major
7668 The @code{kill-region} function in Emacs 19 is complex, but does not
7669 use code that is important at this time. We will skip it.
7671 The @code{kill-region} function in Emacs 22 is easier to read than the
7672 same function in Emacs 19 and introduces a very important concept,
7673 that of error handling. We will walk through the function.
7675 But first, let us look at the interactive @code{zap-to-char} function.
7678 * Complete zap-to-char:: The complete implementation.
7679 * zap-to-char interactive:: A three part interactive expression.
7680 * zap-to-char body:: A short overview.
7681 * search-forward:: How to search for a string.
7682 * progn:: The @code{progn} special form.
7683 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7687 @node Complete zap-to-char
7688 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7691 The @code{zap-to-char} function removes the text in the region between
7692 the location of the cursor (i.e., of point) up to and including the
7693 next occurrence of a specified character. The text that
7694 @code{zap-to-char} removes is put in the kill ring; and it can be
7695 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7696 the command is given an argument, it removes text through that number
7697 of occurrences. Thus, if the cursor were at the beginning of this
7698 sentence and the character were @samp{s}, @samp{Thus} would be
7699 removed. If the argument were two, @samp{Thus, if the curs} would be
7700 removed, up to and including the @samp{s} in @samp{cursor}.
7702 If the specified character is not found, @code{zap-to-char} will say
7703 ``Search failed'', tell you the character you typed, and not remove
7706 In order to determine how much text to remove, @code{zap-to-char} uses
7707 a search function. Searches are used extensively in code that
7708 manipulates text, and we will focus attention on them as well as on the
7712 @c GNU Emacs version 19
7713 (defun zap-to-char (arg char) ; version 19 implementation
7714 "Kill up to and including ARG'th occurrence of CHAR.
7715 Goes backward if ARG is negative; error if CHAR not found."
7716 (interactive "*p\ncZap to char: ")
7717 (kill-region (point)
7720 (char-to-string char) nil nil arg)
7725 Here is the complete text of the version 22 implementation of the function:
7730 (defun zap-to-char (arg char)
7731 "Kill up to and including ARG'th occurrence of CHAR.
7732 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7733 Goes backward if ARG is negative; error if CHAR not found."
7734 (interactive "p\ncZap to char: ")
7735 (if (char-table-p translation-table-for-input)
7736 (setq char (or (aref translation-table-for-input char) char)))
7737 (kill-region (point) (progn
7738 (search-forward (char-to-string char)
7744 The documentation is thorough. You do need to know the jargon meaning
7747 @node zap-to-char interactive
7748 @subsection The @code{interactive} Expression
7751 The interactive expression in the @code{zap-to-char} command looks like
7755 (interactive "p\ncZap to char: ")
7758 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7759 two different things. First, and most simply, is the @samp{p}.
7760 This part is separated from the next part by a newline, @samp{\n}.
7761 The @samp{p} means that the first argument to the function will be
7762 passed the value of a `processed prefix'. The prefix argument is
7763 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7764 the function is called interactively without a prefix, 1 is passed to
7767 The second part of @code{"p\ncZap to char:@: "} is
7768 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7769 indicates that @code{interactive} expects a prompt and that the
7770 argument will be a character. The prompt follows the @samp{c} and is
7771 the string @samp{Zap to char:@: } (with a space after the colon to
7774 What all this does is prepare the arguments to @code{zap-to-char} so they
7775 are of the right type, and give the user a prompt.
7777 In a read-only buffer, the @code{zap-to-char} function copies the text
7778 to the kill ring, but does not remove it. The echo area displays a
7779 message saying that the buffer is read-only. Also, the terminal may
7780 beep or blink at you.
7782 @node zap-to-char body
7783 @subsection The Body of @code{zap-to-char}
7785 The body of the @code{zap-to-char} function contains the code that
7786 kills (that is, removes) the text in the region from the current
7787 position of the cursor up to and including the specified character.
7789 The first part of the code looks like this:
7792 (if (char-table-p translation-table-for-input)
7793 (setq char (or (aref translation-table-for-input char) char)))
7794 (kill-region (point) (progn
7795 (search-forward (char-to-string char) nil nil arg)
7800 @code{char-table-p} is an hitherto unseen function. It determines
7801 whether its argument is a character table. When it is, it sets the
7802 character passed to @code{zap-to-char} to one of them, if that
7803 character exists, or to the character itself. (This becomes important
7804 for certain characters in non-European languages. The @code{aref}
7805 function extracts an element from an array. It is an array-specific
7806 function that is not described in this document. @xref{Arrays, ,
7807 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7810 @code{(point)} is the current position of the cursor.
7812 The next part of the code is an expression using @code{progn}. The body
7813 of the @code{progn} consists of calls to @code{search-forward} and
7816 It is easier to understand how @code{progn} works after learning about
7817 @code{search-forward}, so we will look at @code{search-forward} and
7818 then at @code{progn}.
7820 @node search-forward
7821 @subsection The @code{search-forward} Function
7822 @findex search-forward
7824 The @code{search-forward} function is used to locate the
7825 zapped-for-character in @code{zap-to-char}. If the search is
7826 successful, @code{search-forward} leaves point immediately after the
7827 last character in the target string. (In @code{zap-to-char}, the
7828 target string is just one character long. @code{zap-to-char} uses the
7829 function @code{char-to-string} to ensure that the computer treats that
7830 character as a string.) If the search is backwards,
7831 @code{search-forward} leaves point just before the first character in
7832 the target. Also, @code{search-forward} returns @code{t} for true.
7833 (Moving point is therefore a `side effect'.)
7836 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7839 (search-forward (char-to-string char) nil nil arg)
7842 The @code{search-forward} function takes four arguments:
7846 The first argument is the target, what is searched for. This must be a
7847 string, such as @samp{"z"}.
7849 As it happens, the argument passed to @code{zap-to-char} is a single
7850 character. Because of the way computers are built, the Lisp
7851 interpreter may treat a single character as being different from a
7852 string of characters. Inside the computer, a single character has a
7853 different electronic format than a string of one character. (A single
7854 character can often be recorded in the computer using exactly one
7855 byte; but a string may be longer, and the computer needs to be ready
7856 for this.) Since the @code{search-forward} function searches for a
7857 string, the character that the @code{zap-to-char} function receives as
7858 its argument must be converted inside the computer from one format to
7859 the other; otherwise the @code{search-forward} function will fail.
7860 The @code{char-to-string} function is used to make this conversion.
7863 The second argument bounds the search; it is specified as a position in
7864 the buffer. In this case, the search can go to the end of the buffer,
7865 so no bound is set and the second argument is @code{nil}.
7868 The third argument tells the function what it should do if the search
7869 fails---it can signal an error (and print a message) or it can return
7870 @code{nil}. A @code{nil} as the third argument causes the function to
7871 signal an error when the search fails.
7874 The fourth argument to @code{search-forward} is the repeat count---how
7875 many occurrences of the string to look for. This argument is optional
7876 and if the function is called without a repeat count, this argument is
7877 passed the value 1. If this argument is negative, the search goes
7882 In template form, a @code{search-forward} expression looks like this:
7886 (search-forward "@var{target-string}"
7887 @var{limit-of-search}
7888 @var{what-to-do-if-search-fails}
7893 We will look at @code{progn} next.
7896 @subsection The @code{progn} Special Form
7899 @code{progn} is a special form that causes each of its arguments to be
7900 evaluated in sequence and then returns the value of the last one. The
7901 preceding expressions are evaluated only for the side effects they
7902 perform. The values produced by them are discarded.
7905 The template for a @code{progn} expression is very simple:
7914 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7915 put point in exactly the right position; and return the location of
7916 point so that @code{kill-region} will know how far to kill to.
7918 The first argument to the @code{progn} is @code{search-forward}. When
7919 @code{search-forward} finds the string, the function leaves point
7920 immediately after the last character in the target string. (In this
7921 case the target string is just one character long.) If the search is
7922 backwards, @code{search-forward} leaves point just before the first
7923 character in the target. The movement of point is a side effect.
7925 The second and last argument to @code{progn} is the expression
7926 @code{(point)}. This expression returns the value of point, which in
7927 this case will be the location to which it has been moved by
7928 @code{search-forward}. (In the source, a line that tells the function
7929 to go to the previous character, if it is going forward, was commented
7930 out in 1999; I don't remember whether that feature or mis-feature was
7931 ever a part of the distributed source.) The value of @code{point} is
7932 returned by the @code{progn} expression and is passed to
7933 @code{kill-region} as @code{kill-region}'s second argument.
7935 @node Summing up zap-to-char
7936 @subsection Summing up @code{zap-to-char}
7938 Now that we have seen how @code{search-forward} and @code{progn} work,
7939 we can see how the @code{zap-to-char} function works as a whole.
7941 The first argument to @code{kill-region} is the position of the cursor
7942 when the @code{zap-to-char} command is given---the value of point at
7943 that time. Within the @code{progn}, the search function then moves
7944 point to just after the zapped-to-character and @code{point} returns the
7945 value of this location. The @code{kill-region} function puts together
7946 these two values of point, the first one as the beginning of the region
7947 and the second one as the end of the region, and removes the region.
7949 The @code{progn} special form is necessary because the
7950 @code{kill-region} command takes two arguments; and it would fail if
7951 @code{search-forward} and @code{point} expressions were written in
7952 sequence as two additional arguments. The @code{progn} expression is
7953 a single argument to @code{kill-region} and returns the one value that
7954 @code{kill-region} needs for its second argument.
7957 @section @code{kill-region}
7960 The @code{zap-to-char} function uses the @code{kill-region} function.
7961 This function clips text from a region and copies that text to
7962 the kill ring, from which it may be retrieved.
7967 (defun kill-region (beg end &optional yank-handler)
7968 "Kill (\"cut\") text between point and mark.
7969 This deletes the text from the buffer and saves it in the kill ring.
7970 The command \\[yank] can retrieve it from there.
7971 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
7973 If you want to append the killed region to the last killed text,
7974 use \\[append-next-kill] before \\[kill-region].
7976 If the buffer is read-only, Emacs will beep and refrain from deleting
7977 the text, but put the text in the kill ring anyway. This means that
7978 you can use the killing commands to copy text from a read-only buffer.
7980 This is the primitive for programs to kill text (as opposed to deleting it).
7981 Supply two arguments, character positions indicating the stretch of text
7983 Any command that calls this function is a \"kill command\".
7984 If the previous command was also a kill command,
7985 the text killed this time appends to the text killed last time
7986 to make one entry in the kill ring.
7988 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
7989 specifies the yank-handler text property to be set on the killed
7990 text. See `insert-for-yank'."
7991 ;; Pass point first, then mark, because the order matters
7992 ;; when calling kill-append.
7993 (interactive (list (point) (mark)))
7994 (unless (and beg end)
7995 (error "The mark is not set now, so there is no region"))
7997 (let ((string (filter-buffer-substring beg end t)))
7998 (when string ;STRING is nil if BEG = END
7999 ;; Add that string to the kill ring, one way or another.
8000 (if (eq last-command 'kill-region)
8001 (kill-append string (< end beg) yank-handler)
8002 (kill-new string nil yank-handler)))
8003 (when (or string (eq last-command 'kill-region))
8004 (setq this-command 'kill-region))
8006 ((buffer-read-only text-read-only)
8007 ;; The code above failed because the buffer, or some of the characters
8008 ;; in the region, are read-only.
8009 ;; We should beep, in case the user just isn't aware of this.
8010 ;; However, there's no harm in putting
8011 ;; the region's text in the kill ring, anyway.
8012 (copy-region-as-kill beg end)
8013 ;; Set this-command now, so it will be set even if we get an error.
8014 (setq this-command 'kill-region)
8015 ;; This should barf, if appropriate, and give us the correct error.
8016 (if kill-read-only-ok
8017 (progn (message "Read only text copied to kill ring") nil)
8018 ;; Signal an error if the buffer is read-only.
8019 (barf-if-buffer-read-only)
8020 ;; If the buffer isn't read-only, the text is.
8021 (signal 'text-read-only (list (current-buffer)))))))
8024 The Emacs 22 version of that function uses @code{condition-case} and
8025 @code{copy-region-as-kill}, both of which we will explain.
8026 @code{condition-case} is an important special form.
8028 In essence, the @code{kill-region} function calls
8029 @code{condition-case}, which takes three arguments. In this function,
8030 the first argument does nothing. The second argument contains the
8031 code that does the work when all goes well. The third argument
8032 contains the code that is called in the event of an error.
8035 * Complete kill-region:: The function definition.
8036 * condition-case:: Dealing with a problem.
8041 @node Complete kill-region
8042 @unnumberedsubsec The Complete @code{kill-region} Definition
8046 We will go through the @code{condition-case} code in a moment. First,
8047 let us look at the definition of @code{kill-region}, with comments
8053 (defun kill-region (beg end)
8054 "Kill (\"cut\") text between point and mark.
8055 This deletes the text from the buffer and saves it in the kill ring.
8056 The command \\[yank] can retrieve it from there. @dots{} "
8060 ;; @bullet{} Since order matters, pass point first.
8061 (interactive (list (point) (mark)))
8062 ;; @bullet{} And tell us if we cannot cut the text.
8063 ;; `unless' is an `if' without a then-part.
8064 (unless (and beg end)
8065 (error "The mark is not set now, so there is no region"))
8069 ;; @bullet{} `condition-case' takes three arguments.
8070 ;; If the first argument is nil, as it is here,
8071 ;; information about the error signal is not
8072 ;; stored for use by another function.
8077 ;; @bullet{} The second argument to `condition-case' tells the
8078 ;; Lisp interpreter what to do when all goes well.
8082 ;; It starts with a `let' function that extracts the string
8083 ;; and tests whether it exists. If so (that is what the
8084 ;; `when' checks), it calls an `if' function that determines
8085 ;; whether the previous command was another call to
8086 ;; `kill-region'; if it was, then the new text is appended to
8087 ;; the previous text; if not, then a different function,
8088 ;; `kill-new', is called.
8092 ;; The `kill-append' function concatenates the new string and
8093 ;; the old. The `kill-new' function inserts text into a new
8094 ;; item in the kill ring.
8098 ;; `when' is an `if' without an else-part. The second `when'
8099 ;; again checks whether the current string exists; in
8100 ;; addition, it checks whether the previous command was
8101 ;; another call to `kill-region'. If one or the other
8102 ;; condition is true, then it sets the current command to
8103 ;; be `kill-region'.
8106 (let ((string (filter-buffer-substring beg end t)))
8107 (when string ;STRING is nil if BEG = END
8108 ;; Add that string to the kill ring, one way or another.
8109 (if (eq last-command 'kill-region)
8112 ;; @minus{} `yank-handler' is an optional argument to
8113 ;; `kill-region' that tells the `kill-append' and
8114 ;; `kill-new' functions how deal with properties
8115 ;; added to the text, such as `bold' or `italics'.
8116 (kill-append string (< end beg) yank-handler)
8117 (kill-new string nil yank-handler)))
8118 (when (or string (eq last-command 'kill-region))
8119 (setq this-command 'kill-region))
8124 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8125 ;; what to do with an error.
8128 ;; The third argument has a conditions part and a body part.
8129 ;; If the conditions are met (in this case,
8130 ;; if text or buffer are read-only)
8131 ;; then the body is executed.
8134 ;; The first part of the third argument is the following:
8135 ((buffer-read-only text-read-only) ;; the if-part
8136 ;; @dots{} the then-part
8137 (copy-region-as-kill beg end)
8140 ;; Next, also as part of the then-part, set this-command, so
8141 ;; it will be set in an error
8142 (setq this-command 'kill-region)
8143 ;; Finally, in the then-part, send a message if you may copy
8144 ;; the text to the kill ring without signaling an error, but
8145 ;; don't if you may not.
8148 (if kill-read-only-ok
8149 (progn (message "Read only text copied to kill ring") nil)
8150 (barf-if-buffer-read-only)
8151 ;; If the buffer isn't read-only, the text is.
8152 (signal 'text-read-only (list (current-buffer)))))
8160 (defun kill-region (beg end)
8161 "Kill between point and mark.
8162 The text is deleted but saved in the kill ring."
8167 ;; 1. `condition-case' takes three arguments.
8168 ;; If the first argument is nil, as it is here,
8169 ;; information about the error signal is not
8170 ;; stored for use by another function.
8175 ;; 2. The second argument to `condition-case'
8176 ;; tells the Lisp interpreter what to do when all goes well.
8180 ;; The `delete-and-extract-region' function usually does the
8181 ;; work. If the beginning and ending of the region are both
8182 ;; the same, then the variable `string' will be empty, or nil
8183 (let ((string (delete-and-extract-region beg end)))
8187 ;; `when' is an `if' clause that cannot take an `else-part'.
8188 ;; Emacs normally sets the value of `last-command' to the
8189 ;; previous command.
8192 ;; `kill-append' concatenates the new string and the old.
8193 ;; `kill-new' inserts text into a new item in the kill ring.
8195 (if (eq last-command 'kill-region)
8196 ;; if true, prepend string
8197 (kill-append string (< end beg))
8199 (setq this-command 'kill-region))
8203 ;; 3. The third argument to `condition-case' tells the interpreter
8204 ;; what to do with an error.
8207 ;; The third argument has a conditions part and a body part.
8208 ;; If the conditions are met (in this case,
8209 ;; if text or buffer are read-only)
8210 ;; then the body is executed.
8213 ((buffer-read-only text-read-only) ;; this is the if-part
8215 (copy-region-as-kill beg end)
8218 (if kill-read-only-ok ;; usually this variable is nil
8219 (message "Read only text copied to kill ring")
8220 ;; or else, signal an error if the buffer is read-only;
8221 (barf-if-buffer-read-only)
8222 ;; and, in any case, signal that the text is read-only.
8223 (signal 'text-read-only (list (current-buffer)))))))
8228 @node condition-case
8229 @subsection @code{condition-case}
8230 @findex condition-case
8232 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8233 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8234 expression, it provides you with help; in the jargon, this is called
8235 ``signaling an error''. Usually, the computer stops the program and
8236 shows you a message.
8238 However, some programs undertake complicated actions. They should not
8239 simply stop on an error. In the @code{kill-region} function, the most
8240 likely error is that you will try to kill text that is read-only and
8241 cannot be removed. So the @code{kill-region} function contains code
8242 to handle this circumstance. This code, which makes up the body of
8243 the @code{kill-region} function, is inside of a @code{condition-case}
8247 The template for @code{condition-case} looks like this:
8254 @var{error-handler}@dots{})
8258 The second argument, @var{bodyform}, is straightforward. The
8259 @code{condition-case} special form causes the Lisp interpreter to
8260 evaluate the code in @var{bodyform}. If no error occurs, the special
8261 form returns the code's value and produces the side-effects, if any.
8263 In short, the @var{bodyform} part of a @code{condition-case}
8264 expression determines what should happen when everything works
8267 However, if an error occurs, among its other actions, the function
8268 generating the error signal will define one or more error condition
8271 An error handler is the third argument to @code{condition case}.
8272 An error handler has two parts, a @var{condition-name} and a
8273 @var{body}. If the @var{condition-name} part of an error handler
8274 matches a condition name generated by an error, then the @var{body}
8275 part of the error handler is run.
8277 As you will expect, the @var{condition-name} part of an error handler
8278 may be either a single condition name or a list of condition names.
8280 Also, a complete @code{condition-case} expression may contain more
8281 than one error handler. When an error occurs, the first applicable
8284 Lastly, the first argument to the @code{condition-case} expression,
8285 the @var{var} argument, is sometimes bound to a variable that
8286 contains information about the error. However, if that argument is
8287 nil, as is the case in @code{kill-region}, that information is
8291 In brief, in the @code{kill-region} function, the code
8292 @code{condition-case} works like this:
8296 @var{If no errors}, @var{run only this code}
8297 @var{but}, @var{if errors}, @var{run this other code}.
8304 copy-region-as-kill is short, 12 lines, and uses
8305 filter-buffer-substring, which is longer, 39 lines
8306 and has delete-and-extract-region in it.
8307 delete-and-extract-region is written in C.
8309 see Initializing a Variable with @code{defvar}
8311 Initializing a Variable with @code{defvar} includes line 8350
8315 @subsection Lisp macro
8319 The part of the @code{condition-case} expression that is evaluated in
8320 the expectation that all goes well has a @code{when}. The code uses
8321 @code{when} to determine whether the @code{string} variable points to
8324 A @code{when} expression is simply a programmers' convenience. It is
8325 an @code{if} without the possibility of an else clause. In your mind,
8326 you can replace @code{when} with @code{if} and understand what goes
8327 on. That is what the Lisp interpreter does.
8329 Technically speaking, @code{when} is a Lisp macro. A Lisp macro
8330 enables you to define new control constructs and other language
8331 features. It tells the interpreter how to compute another Lisp
8332 expression which will in turn compute the value. In this case, the
8333 `other expression' is an @code{if} expression.
8335 The @code{kill-region} function definition also has an @code{unless}
8336 macro; it is the converse of @code{when}. The @code{unless} macro is
8337 an @code{if} without a then clause
8339 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8340 Emacs Lisp Reference Manual}. The C programming language also
8341 provides macros. These are different, but also useful.
8344 We will briefly look at C macros in
8345 @ref{Digression into C}.
8349 Regarding the @code{when} macro, in the @code{condition-case}
8350 expression, when the string has content, then another conditional
8351 expression is executed. This is an @code{if} with both a then-part
8356 (if (eq last-command 'kill-region)
8357 (kill-append string (< end beg) yank-handler)
8358 (kill-new string nil yank-handler))
8362 The then-part is evaluated if the previous command was another call to
8363 @code{kill-region}; if not, the else-part is evaluated.
8365 @code{yank-handler} is an optional argument to @code{kill-region} that
8366 tells the @code{kill-append} and @code{kill-new} functions how deal
8367 with properties added to the text, such as `bold' or `italics'.
8369 @code{last-command} is a variable that comes with Emacs that we have
8370 not seen before. Normally, whenever a function is executed, Emacs
8371 sets the value of @code{last-command} to the previous command.
8374 In this segment of the definition, the @code{if} expression checks
8375 whether the previous command was @code{kill-region}. If it was,
8378 (kill-append string (< end beg) yank-handler)
8382 concatenates a copy of the newly clipped text to the just previously
8383 clipped text in the kill ring.
8385 @node copy-region-as-kill
8386 @section @code{copy-region-as-kill}
8387 @findex copy-region-as-kill
8390 The @code{copy-region-as-kill} function copies a region of text from a
8391 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8392 in the @code{kill-ring}.
8394 If you call @code{copy-region-as-kill} immediately after a
8395 @code{kill-region} command, Emacs appends the newly copied text to the
8396 previously copied text. This means that if you yank back the text, you
8397 get it all, from both this and the previous operation. On the other
8398 hand, if some other command precedes the @code{copy-region-as-kill},
8399 the function copies the text into a separate entry in the kill ring.
8402 * Complete copy-region-as-kill:: The complete function definition.
8403 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8407 @node Complete copy-region-as-kill
8408 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8412 Here is the complete text of the version 22 @code{copy-region-as-kill}
8417 (defun copy-region-as-kill (beg end)
8418 "Save the region as if killed, but don't kill it.
8419 In Transient Mark mode, deactivate the mark.
8420 If `interprogram-cut-function' is non-nil, also save the text for a window
8421 system cut and paste."
8425 (if (eq last-command 'kill-region)
8426 (kill-append (filter-buffer-substring beg end) (< end beg))
8427 (kill-new (filter-buffer-substring beg end)))
8430 (if transient-mark-mode
8431 (setq deactivate-mark t))
8437 As usual, this function can be divided into its component parts:
8441 (defun copy-region-as-kill (@var{argument-list})
8442 "@var{documentation}@dots{}"
8448 The arguments are @code{beg} and @code{end} and the function is
8449 interactive with @code{"r"}, so the two arguments must refer to the
8450 beginning and end of the region. If you have been reading though this
8451 document from the beginning, understanding these parts of a function is
8452 almost becoming routine.
8454 The documentation is somewhat confusing unless you remember that the
8455 word `kill' has a meaning different from usual. The `Transient Mark'
8456 and @code{interprogram-cut-function} comments explain certain
8459 After you once set a mark, a buffer always contains a region. If you
8460 wish, you can use Transient Mark mode to highlight the region
8461 temporarily. (No one wants to highlight the region all the time, so
8462 Transient Mark mode highlights it only at appropriate times. Many
8463 people turn off Transient Mark mode, so the region is never
8466 Also, a windowing system allows you to copy, cut, and paste among
8467 different programs. In the X windowing system, for example, the
8468 @code{interprogram-cut-function} function is @code{x-select-text},
8469 which works with the windowing system's equivalent of the Emacs kill
8472 The body of the @code{copy-region-as-kill} function starts with an
8473 @code{if} clause. What this clause does is distinguish between two
8474 different situations: whether or not this command is executed
8475 immediately after a previous @code{kill-region} command. In the first
8476 case, the new region is appended to the previously copied text.
8477 Otherwise, it is inserted into the beginning of the kill ring as a
8478 separate piece of text from the previous piece.
8480 The last two lines of the function prevent the region from lighting up
8481 if Transient Mark mode is turned on.
8483 The body of @code{copy-region-as-kill} merits discussion in detail.
8485 @node copy-region-as-kill body
8486 @subsection The Body of @code{copy-region-as-kill}
8488 The @code{copy-region-as-kill} function works in much the same way as
8489 the @code{kill-region} function. Both are written so that two or more
8490 kills in a row combine their text into a single entry. If you yank
8491 back the text from the kill ring, you get it all in one piece.
8492 Moreover, kills that kill forward from the current position of the
8493 cursor are added to the end of the previously copied text and commands
8494 that copy text backwards add it to the beginning of the previously
8495 copied text. This way, the words in the text stay in the proper
8498 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8499 use of the @code{last-command} variable that keeps track of the
8500 previous Emacs command.
8503 * last-command & this-command::
8504 * kill-append function::
8505 * kill-new function::
8509 @node last-command & this-command
8510 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8513 Normally, whenever a function is executed, Emacs sets the value of
8514 @code{this-command} to the function being executed (which in this case
8515 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8516 the value of @code{last-command} to the previous value of
8517 @code{this-command}.
8519 In the first part of the body of the @code{copy-region-as-kill}
8520 function, an @code{if} expression determines whether the value of
8521 @code{last-command} is @code{kill-region}. If so, the then-part of
8522 the @code{if} expression is evaluated; it uses the @code{kill-append}
8523 function to concatenate the text copied at this call to the function
8524 with the text already in the first element (the @sc{car}) of the kill
8525 ring. On the other hand, if the value of @code{last-command} is not
8526 @code{kill-region}, then the @code{copy-region-as-kill} function
8527 attaches a new element to the kill ring using the @code{kill-new}
8531 The @code{if} expression reads as follows; it uses @code{eq}:
8535 (if (eq last-command 'kill-region)
8537 (kill-append (filter-buffer-substring beg end) (< end beg))
8539 (kill-new (filter-buffer-substring beg end)))
8543 @findex filter-buffer-substring
8544 (The @code{filter-buffer-substring} function returns a filtered
8545 substring of the buffer, if any. Optionally---the arguments are not
8546 here, so neither is done---the function may delete the initial text or
8547 return the text without its properties; this function is a replacement
8548 for the older @code{buffer-substring} function, which came before text
8549 properties were implemented.)
8551 @findex eq @r{(example of use)}
8553 The @code{eq} function tests whether its first argument is the same Lisp
8554 object as its second argument. The @code{eq} function is similar to the
8555 @code{equal} function in that it is used to test for equality, but
8556 differs in that it determines whether two representations are actually
8557 the same object inside the computer, but with different names.
8558 @code{equal} determines whether the structure and contents of two
8559 expressions are the same.
8561 If the previous command was @code{kill-region}, then the Emacs Lisp
8562 interpreter calls the @code{kill-append} function
8564 @node kill-append function
8565 @unnumberedsubsubsec The @code{kill-append} function
8569 The @code{kill-append} function looks like this:
8574 (defun kill-append (string before-p &optional yank-handler)
8575 "Append STRING to the end of the latest kill in the kill ring.
8576 If BEFORE-P is non-nil, prepend STRING to the kill.
8578 (let* ((cur (car kill-ring)))
8579 (kill-new (if before-p (concat string cur) (concat cur string))
8580 (or (= (length cur) 0)
8582 (get-text-property 0 'yank-handler cur)))
8589 (defun kill-append (string before-p)
8590 "Append STRING to the end of the latest kill in the kill ring.
8591 If BEFORE-P is non-nil, prepend STRING to the kill.
8592 If `interprogram-cut-function' is set, pass the resulting kill to
8594 (kill-new (if before-p
8595 (concat string (car kill-ring))
8596 (concat (car kill-ring) string))
8601 The @code{kill-append} function is fairly straightforward. It uses
8602 the @code{kill-new} function, which we will discuss in more detail in
8605 (Also, the function provides an optional argument called
8606 @code{yank-handler}; when invoked, this argument tells the function
8607 how to deal with properties added to the text, such as `bold' or
8610 @c !!! bug in GNU Emacs 22 version of kill-append ?
8611 It has a @code{let*} function to set the value of the first element of
8612 the kill ring to @code{cur}. (I do not know why the function does not
8613 use @code{let} instead; only one value is set in the expression.
8614 Perhaps this is a bug that produces no problems?)
8616 Consider the conditional that is one of the two arguments to
8617 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8618 the @sc{car} of the kill ring. Whether it prepends or appends the
8619 text depends on the results of an @code{if} expression:
8623 (if before-p ; @r{if-part}
8624 (concat string cur) ; @r{then-part}
8625 (concat cur string)) ; @r{else-part}
8630 If the region being killed is before the region that was killed in the
8631 last command, then it should be prepended before the material that was
8632 saved in the previous kill; and conversely, if the killed text follows
8633 what was just killed, it should be appended after the previous text.
8634 The @code{if} expression depends on the predicate @code{before-p} to
8635 decide whether the newly saved text should be put before or after the
8636 previously saved text.
8638 The symbol @code{before-p} is the name of one of the arguments to
8639 @code{kill-append}. When the @code{kill-append} function is
8640 evaluated, it is bound to the value returned by evaluating the actual
8641 argument. In this case, this is the expression @code{(< end beg)}.
8642 This expression does not directly determine whether the killed text in
8643 this command is located before or after the kill text of the last
8644 command; what it does is determine whether the value of the variable
8645 @code{end} is less than the value of the variable @code{beg}. If it
8646 is, it means that the user is most likely heading towards the
8647 beginning of the buffer. Also, the result of evaluating the predicate
8648 expression, @code{(< end beg)}, will be true and the text will be
8649 prepended before the previous text. On the other hand, if the value of
8650 the variable @code{end} is greater than the value of the variable
8651 @code{beg}, the text will be appended after the previous text.
8654 When the newly saved text will be prepended, then the string with the new
8655 text will be concatenated before the old text:
8663 But if the text will be appended, it will be concatenated
8667 (concat cur string))
8670 To understand how this works, we first need to review the
8671 @code{concat} function. The @code{concat} function links together or
8672 unites two strings of text. The result is a string. For example:
8676 (concat "abc" "def")
8682 (car '("first element" "second element")))
8683 @result{} "new first element"
8686 '("first element" "second element")) " modified")
8687 @result{} "first element modified"
8691 We can now make sense of @code{kill-append}: it modifies the contents
8692 of the kill ring. The kill ring is a list, each element of which is
8693 saved text. The @code{kill-append} function uses the @code{kill-new}
8694 function which in turn uses the @code{setcar} function.
8696 @node kill-new function
8697 @unnumberedsubsubsec The @code{kill-new} function
8700 @c in GNU Emacs 22, additional documentation to kill-new:
8702 Optional third arguments YANK-HANDLER controls how the STRING is later
8703 inserted into a buffer; see `insert-for-yank' for details.
8704 When a yank handler is specified, STRING must be non-empty (the yank
8705 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8707 When the yank handler has a non-nil PARAM element, the original STRING
8708 argument is not used by `insert-for-yank'. However, since Lisp code
8709 may access and use elements from the kill ring directly, the STRING
8710 argument should still be a \"useful\" string for such uses."
8713 The @code{kill-new} function looks like this:
8717 (defun kill-new (string &optional replace yank-handler)
8718 "Make STRING the latest kill in the kill ring.
8719 Set `kill-ring-yank-pointer' to point to it.
8721 If `interprogram-cut-function' is non-nil, apply it to STRING.
8722 Optional second argument REPLACE non-nil means that STRING will replace
8723 the front of the kill ring, rather than being added to the list.
8727 (if (> (length string) 0)
8729 (put-text-property 0 (length string)
8730 'yank-handler yank-handler string))
8732 (signal 'args-out-of-range
8733 (list string "yank-handler specified for empty string"))))
8736 (if (fboundp 'menu-bar-update-yank-menu)
8737 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8740 (if (and replace kill-ring)
8741 (setcar kill-ring string)
8742 (push string kill-ring)
8743 (if (> (length kill-ring) kill-ring-max)
8744 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8747 (setq kill-ring-yank-pointer kill-ring)
8748 (if interprogram-cut-function
8749 (funcall interprogram-cut-function string (not replace))))
8754 (defun kill-new (string &optional replace)
8755 "Make STRING the latest kill in the kill ring.
8756 Set the kill-ring-yank pointer to point to it.
8757 If `interprogram-cut-function' is non-nil, apply it to STRING.
8758 Optional second argument REPLACE non-nil means that STRING will replace
8759 the front of the kill ring, rather than being added to the list."
8760 (and (fboundp 'menu-bar-update-yank-menu)
8761 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8762 (if (and replace kill-ring)
8763 (setcar kill-ring string)
8764 (setq kill-ring (cons string kill-ring))
8765 (if (> (length kill-ring) kill-ring-max)
8766 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8767 (setq kill-ring-yank-pointer kill-ring)
8768 (if interprogram-cut-function
8769 (funcall interprogram-cut-function string (not replace))))
8772 (Notice that the function is not interactive.)
8774 As usual, we can look at this function in parts.
8776 The function definition has an optional @code{yank-handler} argument,
8777 which when invoked tells the function how to deal with properties
8778 added to the text, such as `bold' or `italics'. We will skip that.
8781 The first line of the documentation makes sense:
8784 Make STRING the latest kill in the kill ring.
8788 Let's skip over the rest of the documentation for the moment.
8791 Also, let's skip over the initial @code{if} expression and those lines
8792 of code involving @code{menu-bar-update-yank-menu}. We will explain
8796 The critical lines are these:
8800 (if (and replace kill-ring)
8802 (setcar kill-ring string)
8806 (push string kill-ring)
8809 (setq kill-ring (cons string kill-ring))
8810 (if (> (length kill-ring) kill-ring-max)
8811 ;; @r{avoid overly long kill ring}
8812 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8815 (setq kill-ring-yank-pointer kill-ring)
8816 (if interprogram-cut-function
8817 (funcall interprogram-cut-function string (not replace))))
8821 The conditional test is @w{@code{(and replace kill-ring)}}.
8822 This will be true when two conditions are met: the kill ring has
8823 something in it, and the @code{replace} variable is true.
8826 When the @code{kill-append} function sets @code{replace} to be true
8827 and when the kill ring has at least one item in it, the @code{setcar}
8828 expression is executed:
8831 (setcar kill-ring string)
8834 The @code{setcar} function actually changes the first element of the
8835 @code{kill-ring} list to the value of @code{string}. It replaces the
8839 On the other hand, if the kill ring is empty, or replace is false, the
8840 else-part of the condition is executed:
8843 (push string kill-ring)
8848 @code{push} puts its first argument onto the second. It is similar to
8852 (setq kill-ring (cons string kill-ring))
8860 (add-to-list kill-ring string)
8864 When it is false, the expression first constructs a new version of the
8865 kill ring by prepending @code{string} to the existing kill ring as a
8866 new element (that is what the @code{push} does). Then it executes a
8867 second @code{if} clause. This second @code{if} clause keeps the kill
8868 ring from growing too long.
8870 Let's look at these two expressions in order.
8872 The @code{push} line of the else-part sets the new value of the kill
8873 ring to what results from adding the string being killed to the old
8876 We can see how this works with an example.
8882 (setq example-list '("here is a clause" "another clause"))
8887 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8888 @code{example-list} and see what it returns:
8893 @result{} ("here is a clause" "another clause")
8899 Now, we can add a new element on to this list by evaluating the
8900 following expression:
8901 @findex push, @r{example}
8904 (push "a third clause" example-list)
8909 When we evaluate @code{example-list}, we find its value is:
8914 @result{} ("a third clause" "here is a clause" "another clause")
8919 Thus, the third clause is added to the list by @code{push}.
8922 Now for the second part of the @code{if} clause. This expression
8923 keeps the kill ring from growing too long. It looks like this:
8927 (if (> (length kill-ring) kill-ring-max)
8928 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8932 The code checks whether the length of the kill ring is greater than
8933 the maximum permitted length. This is the value of
8934 @code{kill-ring-max} (which is 60, by default). If the length of the
8935 kill ring is too long, then this code sets the last element of the
8936 kill ring to @code{nil}. It does this by using two functions,
8937 @code{nthcdr} and @code{setcdr}.
8939 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
8940 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
8941 @sc{car} of a list. In this case, however, @code{setcdr} will not be
8942 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
8943 function is used to cause it to set the @sc{cdr} of the next to last
8944 element of the kill ring---this means that since the @sc{cdr} of the
8945 next to last element is the last element of the kill ring, it will set
8946 the last element of the kill ring.
8948 @findex nthcdr, @r{example}
8949 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
8950 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
8951 @dots{} It does this @var{N} times and returns the results.
8952 (@xref{nthcdr, , @code{nthcdr}}.)
8954 @findex setcdr, @r{example}
8955 Thus, if we had a four element list that was supposed to be three
8956 elements long, we could set the @sc{cdr} of the next to last element
8957 to @code{nil}, and thereby shorten the list. (If you set the last
8958 element to some other value than @code{nil}, which you could do, then
8959 you would not have shortened the list. @xref{setcdr, ,
8962 You can see shortening by evaluating the following three expressions
8963 in turn. First set the value of @code{trees} to @code{(maple oak pine
8964 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
8965 and then find the value of @code{trees}:
8969 (setq trees '(maple oak pine birch))
8970 @result{} (maple oak pine birch)
8974 (setcdr (nthcdr 2 trees) nil)
8978 @result{} (maple oak pine)
8983 (The value returned by the @code{setcdr} expression is @code{nil} since
8984 that is what the @sc{cdr} is set to.)
8986 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
8987 @sc{cdr} a number of times that is one less than the maximum permitted
8988 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
8989 element (which will be the rest of the elements in the kill ring) to
8990 @code{nil}. This prevents the kill ring from growing too long.
8993 The next to last expression in the @code{kill-new} function is
8996 (setq kill-ring-yank-pointer kill-ring)
8999 The @code{kill-ring-yank-pointer} is a global variable that is set to be
9000 the @code{kill-ring}.
9002 Even though the @code{kill-ring-yank-pointer} is called a
9003 @samp{pointer}, it is a variable just like the kill ring. However, the
9004 name has been chosen to help humans understand how the variable is used.
9007 Now, to return to an early expression in the body of the function:
9011 (if (fboundp 'menu-bar-update-yank-menu)
9012 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9017 It starts with an @code{if} expression
9019 In this case, the expression tests first to see whether
9020 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9021 calls it. The @code{fboundp} function returns true if the symbol it
9022 is testing has a function definition that `is not void'. If the
9023 symbol's function definition were void, we would receive an error
9024 message, as we did when we created errors intentionally (@pxref{Making
9025 Errors, , Generate an Error Message}).
9028 The then-part contains an expression whose first element is the
9029 function @code{and}.
9032 The @code{and} special form evaluates each of its arguments until one
9033 of the arguments returns a value of @code{nil}, in which case the
9034 @code{and} expression returns @code{nil}; however, if none of the
9035 arguments returns a value of @code{nil}, the value resulting from
9036 evaluating the last argument is returned. (Since such a value is not
9037 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9038 @code{and} expression returns a true value only if all its arguments
9039 are true. (@xref{Second Buffer Related Review}.)
9041 The expression determines whether the second argument to
9042 @code{menu-bar-update-yank-menu} is true or not.
9044 ;; If we're supposed to be extending an existing string, and that
9045 ;; string really is at the front of the menu, then update it in place.
9048 @code{menu-bar-update-yank-menu} is one of the functions that make it
9049 possible to use the `Select and Paste' menu in the Edit item of a menu
9050 bar; using a mouse, you can look at the various pieces of text you
9051 have saved and select one piece to paste.
9053 The last expression in the @code{kill-new} function adds the newly
9054 copied string to whatever facility exists for copying and pasting
9055 among different programs running in a windowing system. In the X
9056 Windowing system, for example, the @code{x-select-text} function takes
9057 the string and stores it in memory operated by X@. You can paste the
9058 string in another program, such as an Xterm.
9061 The expression looks like this:
9065 (if interprogram-cut-function
9066 (funcall interprogram-cut-function string (not replace))))
9070 If an @code{interprogram-cut-function} exists, then Emacs executes
9071 @code{funcall}, which in turn calls its first argument as a function
9072 and passes the remaining arguments to it. (Incidentally, as far as I
9073 can see, this @code{if} expression could be replaced by an @code{and}
9074 expression similar to the one in the first part of the function.)
9076 We are not going to discuss windowing systems and other programs
9077 further, but merely note that this is a mechanism that enables GNU
9078 Emacs to work easily and well with other programs.
9080 This code for placing text in the kill ring, either concatenated with
9081 an existing element or as a new element, leads us to the code for
9082 bringing back text that has been cut out of the buffer---the yank
9083 commands. However, before discussing the yank commands, it is better
9084 to learn how lists are implemented in a computer. This will make
9085 clear such mysteries as the use of the term `pointer'. But before
9086 that, we will digress into C.
9089 @c is this true in Emacs 22? Does not seems to be
9091 (If the @w{@code{(< end beg))}}
9092 expression is true, @code{kill-append} prepends the string to the just
9093 previously clipped text. For a detailed discussion, see
9094 @ref{kill-append function, , The @code{kill-append} function}.)
9096 If you then yank back the text, i.e., `paste' it, you get both
9097 pieces of text at once. That way, if you delete two words in a row,
9098 and then yank them back, you get both words, in their proper order,
9099 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9102 On the other hand, if the previous command is not @code{kill-region},
9103 then the @code{kill-new} function is called, which adds the text to
9104 the kill ring as the latest item, and sets the
9105 @code{kill-ring-yank-pointer} variable to point to it.
9109 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9110 @c use the delete-and-extract-region function
9112 2006 Oct 26, the Digression into C is now OK but should come after
9113 copy-region-as-kill and filter-buffer-substring
9117 copy-region-as-kill is short, 12 lines, and uses
9118 filter-buffer-substring, which is longer, 39 lines
9119 and has delete-and-extract-region in it.
9120 delete-and-extract-region is written in C.
9122 see Initializing a Variable with @code{defvar}
9125 @node Digression into C
9126 @section Digression into C
9127 @findex delete-and-extract-region
9128 @cindex C, a digression into
9129 @cindex Digression into C
9131 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9132 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9133 function, which in turn uses the @code{delete-and-extract-region}
9134 function. It removes the contents of a region and you cannot get them
9137 Unlike the other code discussed here, the
9138 @code{delete-and-extract-region} function is not written in Emacs
9139 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9140 system. Since it is very simple, I will digress briefly from Lisp and
9143 @c GNU Emacs 24 in src/editfns.c
9144 @c the DEFUN for delete-and-extract-region
9147 Like many of the other Emacs primitives,
9148 @code{delete-and-extract-region} is written as an instance of a C
9149 macro, a macro being a template for code. The complete macro looks
9154 DEFUN ("delete-and-extract-region", Fdelete_and_extract_region,
9155 Sdelete_and_extract_region, 2, 2, 0,
9156 doc: /* Delete the text between START and END and return it. */)
9157 (Lisp_Object start, Lisp_Object end)
9159 validate_region (&start, &end);
9160 if (XINT (start) == XINT (end))
9161 return empty_unibyte_string;
9162 return del_range_1 (XINT (start), XINT (end), 1, 1);
9167 Without going into the details of the macro writing process, let me
9168 point out that this macro starts with the word @code{DEFUN}. The word
9169 @code{DEFUN} was chosen since the code serves the same purpose as
9170 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9171 @file{emacs/src/lisp.h}.)
9173 The word @code{DEFUN} is followed by seven parts inside of
9178 The first part is the name given to the function in Lisp,
9179 @code{delete-and-extract-region}.
9182 The second part is the name of the function in C,
9183 @code{Fdelete_and_extract_region}. By convention, it starts with
9184 @samp{F}. Since C does not use hyphens in names, underscores are used
9188 The third part is the name for the C constant structure that records
9189 information on this function for internal use. It is the name of the
9190 function in C but begins with an @samp{S} instead of an @samp{F}.
9193 The fourth and fifth parts specify the minimum and maximum number of
9194 arguments the function can have. This function demands exactly 2
9198 The sixth part is nearly like the argument that follows the
9199 @code{interactive} declaration in a function written in Lisp: a letter
9200 followed, perhaps, by a prompt. The only difference from the Lisp is
9201 when the macro is called with no arguments. Then you write a @code{0}
9202 (which is a `null string'), as in this macro.
9204 If you were to specify arguments, you would place them between
9205 quotation marks. The C macro for @code{goto-char} includes
9206 @code{"NGoto char: "} in this position to indicate that the function
9207 expects a raw prefix, in this case, a numerical location in a buffer,
9208 and provides a prompt.
9211 The seventh part is a documentation string, just like the one for a
9212 function written in Emacs Lisp. This is written as a C comment. (When
9213 you build Emacs, the program @command{lib-src/make-docfile} extracts
9214 these comments and uses them to make the ``real'' documentation.)
9218 In a C macro, the formal parameters come next, with a statement of
9219 what kind of object they are, followed by what might be called the `body'
9220 of the macro. For @code{delete-and-extract-region} the `body'
9221 consists of the following four lines:
9225 validate_region (&start, &end);
9226 if (XINT (start) == XINT (end))
9227 return empty_unibyte_string;
9228 return del_range_1 (XINT (start), XINT (end), 1, 1);
9232 The @code{validate_region} function checks whether the values
9233 passed as the beginning and end of the region are the proper type and
9234 are within range. If the beginning and end positions are the same,
9235 then return an empty string.
9237 The @code{del_range_1} function actually deletes the text. It is a
9238 complex function we will not look into. It updates the buffer and
9239 does other things. However, it is worth looking at the two arguments
9240 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9241 @w{@code{XINT (end)}}.
9243 As far as the C language is concerned, @code{start} and @code{end} are
9244 two integers that mark the beginning and end of the region to be
9245 deleted@footnote{More precisely, and requiring more expert knowledge
9246 to understand, the two integers are of type `Lisp_Object', which can
9247 also be a C union instead of an integer type.}.
9249 In early versions of Emacs, these two numbers were thirty-two bits
9250 long, but the code is slowly being generalized to handle other
9251 lengths. Three of the available bits are used to specify the type of
9252 information; the remaining bits are used as `content'.
9254 @samp{XINT} is a C macro that extracts the relevant number from the
9255 longer collection of bits; the three other bits are discarded.
9258 The command in @code{delete-and-extract-region} looks like this:
9261 del_range_1 (XINT (start), XINT (end), 1, 1);
9265 It deletes the region between the beginning position, @code{start},
9266 and the ending position, @code{end}.
9268 From the point of view of the person writing Lisp, Emacs is all very
9269 simple; but hidden underneath is a great deal of complexity to make it
9273 @section Initializing a Variable with @code{defvar}
9275 @cindex Initializing a variable
9276 @cindex Variable initialization
9281 copy-region-as-kill is short, 12 lines, and uses
9282 filter-buffer-substring, which is longer, 39 lines
9283 and has delete-and-extract-region in it.
9284 delete-and-extract-region is written in C.
9286 see Initializing a Variable with @code{defvar}
9290 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9291 functions within it, @code{kill-append} and @code{kill-new}, copy a
9292 region in a buffer and save it in a variable called the
9293 @code{kill-ring}. This section describes how the @code{kill-ring}
9294 variable is created and initialized using the @code{defvar} special
9297 (Again we note that the term @code{kill-ring} is a misnomer. The text
9298 that is clipped out of the buffer can be brought back; it is not a ring
9299 of corpses, but a ring of resurrectable text.)
9301 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9302 given an initial value by using the @code{defvar} special form. The
9303 name comes from ``define variable''.
9305 The @code{defvar} special form is similar to @code{setq} in that it sets
9306 the value of a variable. It is unlike @code{setq} in two ways: first,
9307 it only sets the value of the variable if the variable does not already
9308 have a value. If the variable already has a value, @code{defvar} does
9309 not override the existing value. Second, @code{defvar} has a
9310 documentation string.
9312 (There is a related macro, @code{defcustom}, designed for variables
9313 that people customize. It has more features than @code{defvar}.
9314 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9317 * See variable current value::
9318 * defvar and asterisk::
9322 @node See variable current value
9323 @unnumberedsubsec Seeing the Current Value of a Variable
9326 You can see the current value of a variable, any variable, by using
9327 the @code{describe-variable} function, which is usually invoked by
9328 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9329 (followed by @key{RET}) when prompted, you will see what is in your
9330 current kill ring---this may be quite a lot! Conversely, if you have
9331 been doing nothing this Emacs session except read this document, you
9332 may have nothing in it. Also, you will see the documentation for
9338 List of killed text sequences.
9339 Since the kill ring is supposed to interact nicely with cut-and-paste
9340 facilities offered by window systems, use of this variable should
9343 interact nicely with `interprogram-cut-function' and
9344 `interprogram-paste-function'. The functions `kill-new',
9345 `kill-append', and `current-kill' are supposed to implement this
9346 interaction; you may want to use them instead of manipulating the kill
9352 The kill ring is defined by a @code{defvar} in the following way:
9356 (defvar kill-ring nil
9357 "List of killed text sequences.
9363 In this variable definition, the variable is given an initial value of
9364 @code{nil}, which makes sense, since if you have saved nothing, you want
9365 nothing back if you give a @code{yank} command. The documentation
9366 string is written just like the documentation string of a @code{defun}.
9367 As with the documentation string of the @code{defun}, the first line of
9368 the documentation should be a complete sentence, since some commands,
9369 like @code{apropos}, print only the first line of documentation.
9370 Succeeding lines should not be indented; otherwise they look odd when
9371 you use @kbd{C-h v} (@code{describe-variable}).
9373 @node defvar and asterisk
9374 @subsection @code{defvar} and an asterisk
9375 @findex defvar @r{for a user customizable variable}
9376 @findex defvar @r{with an asterisk}
9378 In the past, Emacs used the @code{defvar} special form both for
9379 internal variables that you would not expect a user to change and for
9380 variables that you do expect a user to change. Although you can still
9381 use @code{defvar} for user customizable variables, please use
9382 @code{defcustom} instead, since it provides a path into
9383 the Customization commands. (@xref{defcustom, , Specifying Variables
9384 using @code{defcustom}}.)
9386 When you specified a variable using the @code{defvar} special form,
9387 you could distinguish a variable that a user might want to change from
9388 others by typing an asterisk, @samp{*}, in the first column of its
9389 documentation string. For example:
9393 (defvar shell-command-default-error-buffer nil
9394 "*Buffer name for `shell-command' @dots{} error output.
9399 @findex set-variable
9401 You could (and still can) use the @code{set-variable} command to
9402 change the value of @code{shell-command-default-error-buffer}
9403 temporarily. However, options set using @code{set-variable} are set
9404 only for the duration of your editing session. The new values are not
9405 saved between sessions. Each time Emacs starts, it reads the original
9406 value, unless you change the value within your @file{.emacs} file,
9407 either by setting it manually or by using @code{customize}.
9408 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9410 For me, the major use of the @code{set-variable} command is to suggest
9411 variables that I might want to set in my @file{.emacs} file. There
9412 are now more than 700 such variables, far too many to remember
9413 readily. Fortunately, you can press @key{TAB} after calling the
9414 @code{M-x set-variable} command to see the list of variables.
9415 (@xref{Examining, , Examining and Setting Variables, emacs,
9416 The GNU Emacs Manual}.)
9419 @node cons & search-fwd Review
9422 Here is a brief summary of some recently introduced functions.
9427 @code{car} returns the first element of a list; @code{cdr} returns the
9428 second and subsequent elements of a list.
9435 (car '(1 2 3 4 5 6 7))
9437 (cdr '(1 2 3 4 5 6 7))
9438 @result{} (2 3 4 5 6 7)
9443 @code{cons} constructs a list by prepending its first argument to its
9457 @code{funcall} evaluates its first argument as a function. It passes
9458 its remaining arguments to its first argument.
9461 Return the result of taking @sc{cdr} `n' times on a list.
9469 The `rest of the rest', as it were.
9476 (nthcdr 3 '(1 2 3 4 5 6 7))
9483 @code{setcar} changes the first element of a list; @code{setcdr}
9484 changes the second and subsequent elements of a list.
9491 (setq triple '(1 2 3))
9498 (setcdr triple '("foo" "bar"))
9501 @result{} (37 "foo" "bar")
9506 Evaluate each argument in sequence and then return the value of the
9519 @item save-restriction
9520 Record whatever narrowing is in effect in the current buffer, if any,
9521 and restore that narrowing after evaluating the arguments.
9523 @item search-forward
9524 Search for a string, and if the string is found, move point. With a
9525 regular expression, use the similar @code{re-search-forward}.
9526 (@xref{Regexp Search, , Regular Expression Searches}, for an
9527 explanation of regular expression patterns and searches.)
9531 @code{search-forward} and @code{re-search-forward} take four
9536 The string or regular expression to search for.
9539 Optionally, the limit of the search.
9542 Optionally, what to do if the search fails, return @code{nil} or an
9546 Optionally, how many times to repeat the search; if negative, the
9547 search goes backwards.
9551 @itemx delete-and-extract-region
9552 @itemx copy-region-as-kill
9554 @code{kill-region} cuts the text between point and mark from the
9555 buffer and stores that text in the kill ring, so you can get it back
9558 @code{copy-region-as-kill} copies the text between point and mark into
9559 the kill ring, from which you can get it by yanking. The function
9560 does not cut or remove the text from the buffer.
9563 @code{delete-and-extract-region} removes the text between point and
9564 mark from the buffer and throws it away. You cannot get it back.
9565 (This is not an interactive command.)
9568 @node search Exercises
9569 @section Searching Exercises
9573 Write an interactive function that searches for a string. If the
9574 search finds the string, leave point after it and display a message
9575 that says ``Found!''. (Do not use @code{search-forward} for the name
9576 of this function; if you do, you will overwrite the existing version of
9577 @code{search-forward} that comes with Emacs. Use a name such as
9578 @code{test-search} instead.)
9581 Write a function that prints the third element of the kill ring in the
9582 echo area, if any; if the kill ring does not contain a third element,
9583 print an appropriate message.
9586 @node List Implementation
9587 @chapter How Lists are Implemented
9588 @cindex Lists in a computer
9590 In Lisp, atoms are recorded in a straightforward fashion; if the
9591 implementation is not straightforward in practice, it is, nonetheless,
9592 straightforward in theory. The atom @samp{rose}, for example, is
9593 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9594 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9595 is equally simple, but it takes a moment to get used to the idea. A
9596 list is kept using a series of pairs of pointers. In the series, the
9597 first pointer in each pair points to an atom or to another list, and the
9598 second pointer in each pair points to the next pair, or to the symbol
9599 @code{nil}, which marks the end of the list.
9601 A pointer itself is quite simply the electronic address of what is
9602 pointed to. Hence, a list is kept as a series of electronic addresses.
9605 * Lists diagrammed::
9606 * Symbols as Chest:: Exploring a powerful metaphor.
9611 @node Lists diagrammed
9612 @unnumberedsec Lists diagrammed
9615 For example, the list @code{(rose violet buttercup)} has three elements,
9616 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9617 electronic address of @samp{rose} is recorded in a segment of computer
9618 memory along with the address that gives the electronic address of where
9619 the atom @samp{violet} is located; and that address (the one that tells
9620 where @samp{violet} is located) is kept along with an address that tells
9621 where the address for the atom @samp{buttercup} is located.
9624 This sounds more complicated than it is and is easier seen in a diagram:
9626 @c clear print-postscript-figures
9627 @c !!! cons-cell-diagram #1
9631 ___ ___ ___ ___ ___ ___
9632 |___|___|--> |___|___|--> |___|___|--> nil
9635 --> rose --> violet --> buttercup
9639 @ifset print-postscript-figures
9642 @center @image{cons-1}
9643 %%%% old method of including an image
9644 % \input /usr/local/lib/tex/inputs/psfig.tex
9645 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9650 @ifclear print-postscript-figures
9654 ___ ___ ___ ___ ___ ___
9655 |___|___|--> |___|___|--> |___|___|--> nil
9658 --> rose --> violet --> buttercup
9665 In the diagram, each box represents a word of computer memory that
9666 holds a Lisp object, usually in the form of a memory address. The boxes,
9667 i.e., the addresses, are in pairs. Each arrow points to what the address
9668 is the address of, either an atom or another pair of addresses. The
9669 first box is the electronic address of @samp{rose} and the arrow points
9670 to @samp{rose}; the second box is the address of the next pair of boxes,
9671 the first part of which is the address of @samp{violet} and the second
9672 part of which is the address of the next pair. The very last box
9673 points to the symbol @code{nil}, which marks the end of the list.
9676 When a variable is set to a list with a function such as @code{setq},
9677 it stores the address of the first box in the variable. Thus,
9678 evaluation of the expression
9681 (setq bouquet '(rose violet buttercup))
9686 creates a situation like this:
9688 @c cons-cell-diagram #2
9694 | ___ ___ ___ ___ ___ ___
9695 --> |___|___|--> |___|___|--> |___|___|--> nil
9698 --> rose --> violet --> buttercup
9702 @ifset print-postscript-figures
9705 @center @image{cons-2}
9706 %%%% old method of including an image
9707 % \input /usr/local/lib/tex/inputs/psfig.tex
9708 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9713 @ifclear print-postscript-figures
9719 | ___ ___ ___ ___ ___ ___
9720 --> |___|___|--> |___|___|--> |___|___|--> nil
9723 --> rose --> violet --> buttercup
9730 In this example, the symbol @code{bouquet} holds the address of the first
9734 This same list can be illustrated in a different sort of box notation
9737 @c cons-cell-diagram #2a
9743 | -------------- --------------- ----------------
9744 | | car | cdr | | car | cdr | | car | cdr |
9745 -->| rose | o------->| violet | o------->| butter- | nil |
9746 | | | | | | | cup | |
9747 -------------- --------------- ----------------
9751 @ifset print-postscript-figures
9754 @center @image{cons-2a}
9755 %%%% old method of including an image
9756 % \input /usr/local/lib/tex/inputs/psfig.tex
9757 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9762 @ifclear print-postscript-figures
9768 | -------------- --------------- ----------------
9769 | | car | cdr | | car | cdr | | car | cdr |
9770 -->| rose | o------->| violet | o------->| butter- | nil |
9771 | | | | | | | cup | |
9772 -------------- --------------- ----------------
9778 (Symbols consist of more than pairs of addresses, but the structure of
9779 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9780 consists of a group of address-boxes, one of which is the address of
9781 the printed word @samp{bouquet}, a second of which is the address of a
9782 function definition attached to the symbol, if any, a third of which
9783 is the address of the first pair of address-boxes for the list
9784 @code{(rose violet buttercup)}, and so on. Here we are showing that
9785 the symbol's third address-box points to the first pair of
9786 address-boxes for the list.)
9788 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9789 changed; the symbol simply has an address further down the list. (In
9790 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9791 evaluation of the following expression
9794 (setq flowers (cdr bouquet))
9801 @c cons-cell-diagram #3
9808 | ___ ___ | ___ ___ ___ ___
9809 --> | | | --> | | | | | |
9810 |___|___|----> |___|___|--> |___|___|--> nil
9813 --> rose --> violet --> buttercup
9818 @ifset print-postscript-figures
9821 @center @image{cons-3}
9822 %%%% old method of including an image
9823 % \input /usr/local/lib/tex/inputs/psfig.tex
9824 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9829 @ifclear print-postscript-figures
9836 | ___ ___ | ___ ___ ___ ___
9837 --> | | | --> | | | | | |
9838 |___|___|----> |___|___|--> |___|___|--> nil
9841 --> rose --> violet --> buttercup
9849 The value of @code{flowers} is @code{(violet buttercup)}, which is
9850 to say, the symbol @code{flowers} holds the address of the pair of
9851 address-boxes, the first of which holds the address of @code{violet},
9852 and the second of which holds the address of @code{buttercup}.
9854 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9855 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9856 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9857 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9858 information about cons cells and dotted pairs.
9861 The function @code{cons} adds a new pair of addresses to the front of
9862 a series of addresses like that shown above. For example, evaluating
9866 (setq bouquet (cons 'lily bouquet))
9873 @c cons-cell-diagram #4
9880 | ___ ___ ___ ___ | ___ ___ ___ ___
9881 --> | | | | | | --> | | | | | |
9882 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9885 --> lily --> rose --> violet --> buttercup
9890 @ifset print-postscript-figures
9893 @center @image{cons-4}
9894 %%%% old method of including an image
9895 % \input /usr/local/lib/tex/inputs/psfig.tex
9896 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9901 @ifclear print-postscript-figures
9908 | ___ ___ ___ ___ | ___ ___ ___ ___
9909 --> | | | | | | --> | | | | | |
9910 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9913 --> lily --> rose --> violet --> buttercup
9922 However, this does not change the value of the symbol
9923 @code{flowers}, as you can see by evaluating the following,
9926 (eq (cdr (cdr bouquet)) flowers)
9930 which returns @code{t} for true.
9932 Until it is reset, @code{flowers} still has the value
9933 @code{(violet buttercup)}; that is, it has the address of the cons
9934 cell whose first address is of @code{violet}. Also, this does not
9935 alter any of the pre-existing cons cells; they are all still there.
9937 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
9938 of the next cons cell in the series; to get the @sc{car} of a list,
9939 you get the address of the first element of the list; to @code{cons} a
9940 new element on a list, you add a new cons cell to the front of the list.
9941 That is all there is to it! The underlying structure of Lisp is
9944 And what does the last address in a series of cons cells refer to? It
9945 is the address of the empty list, of @code{nil}.
9947 In summary, when a Lisp variable is set to a value, it is provided with
9948 the address of the list to which the variable refers.
9950 @node Symbols as Chest
9951 @section Symbols as a Chest of Drawers
9952 @cindex Symbols as a Chest of Drawers
9953 @cindex Chest of Drawers, metaphor for a symbol
9954 @cindex Drawers, Chest of, metaphor for a symbol
9956 In an earlier section, I suggested that you might imagine a symbol as
9957 being a chest of drawers. The function definition is put in one
9958 drawer, the value in another, and so on. What is put in the drawer
9959 holding the value can be changed without affecting the contents of the
9960 drawer holding the function definition, and vice-verse.
9962 Actually, what is put in each drawer is the address of the value or
9963 function definition. It is as if you found an old chest in the attic,
9964 and in one of its drawers you found a map giving you directions to
9965 where the buried treasure lies.
9967 (In addition to its name, symbol definition, and variable value, a
9968 symbol has a `drawer' for a @dfn{property list} which can be used to
9969 record other information. Property lists are not discussed here; see
9970 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
9974 Here is a fanciful representation:
9976 @c chest-of-drawers diagram
9981 Chest of Drawers Contents of Drawers
9985 ---------------------
9986 | directions to | [map to]
9987 | symbol name | bouquet
9989 +---------------------+
9991 | symbol definition | [none]
9993 +---------------------+
9994 | directions to | [map to]
9995 | variable value | (rose violet buttercup)
9997 +---------------------+
9999 | property list | [not described here]
10001 +---------------------+
10007 @ifset print-postscript-figures
10010 @center @image{drawers}
10011 %%%% old method of including an image
10012 % \input /usr/local/lib/tex/inputs/psfig.tex
10013 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10018 @ifclear print-postscript-figures
10023 Chest of Drawers Contents of Drawers
10027 ---------------------
10028 | directions to | [map to]
10029 | symbol name | bouquet
10031 +---------------------+
10033 | symbol definition | [none]
10035 +---------------------+
10036 | directions to | [map to]
10037 | variable value | (rose violet buttercup)
10039 +---------------------+
10041 | property list | [not described here]
10043 +---------------------+
10051 @node List Exercise
10054 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10055 more flowers on to this list and set this new list to
10056 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10057 What does the @code{more-flowers} list now contain?
10060 @chapter Yanking Text Back
10062 @cindex Text retrieval
10063 @cindex Retrieving text
10064 @cindex Pasting text
10066 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10067 you can bring it back with a `yank' command. The text that is cut out of
10068 the buffer is put in the kill ring and the yank commands insert the
10069 appropriate contents of the kill ring back into a buffer (not necessarily
10070 the original buffer).
10072 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10073 the kill ring into the current buffer. If the @kbd{C-y} command is
10074 followed immediately by @kbd{M-y}, the first element is replaced by
10075 the second element. Successive @kbd{M-y} commands replace the second
10076 element with the third, fourth, or fifth element, and so on. When the
10077 last element in the kill ring is reached, it is replaced by the first
10078 element and the cycle is repeated. (Thus the kill ring is called a
10079 `ring' rather than just a `list'. However, the actual data structure
10080 that holds the text is a list.
10081 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10082 list is handled as a ring.)
10085 * Kill Ring Overview::
10086 * kill-ring-yank-pointer:: The kill ring is a list.
10087 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
10090 @node Kill Ring Overview
10091 @section Kill Ring Overview
10092 @cindex Kill ring overview
10094 The kill ring is a list of textual strings. This is what it looks like:
10097 ("some text" "a different piece of text" "yet more text")
10100 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10101 string of characters saying @samp{some text} would be inserted in this
10102 buffer where my cursor is located.
10104 The @code{yank} command is also used for duplicating text by copying it.
10105 The copied text is not cut from the buffer, but a copy of it is put on the
10106 kill ring and is inserted by yanking it back.
10108 Three functions are used for bringing text back from the kill ring:
10109 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10110 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10111 which is used by the two other functions.
10113 These functions refer to the kill ring through a variable called the
10114 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10115 @code{yank} and @code{yank-pop} functions is:
10118 (insert (car kill-ring-yank-pointer))
10122 (Well, no more. In GNU Emacs 22, the function has been replaced by
10123 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10124 repetitively for each @code{yank-handler} segment. In turn,
10125 @code{insert-for-yank-1} strips text properties from the inserted text
10126 according to @code{yank-excluded-properties}. Otherwise, it is just
10127 like @code{insert}. We will stick with plain @code{insert} since it
10128 is easier to understand.)
10130 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10131 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10133 @node kill-ring-yank-pointer
10134 @section The @code{kill-ring-yank-pointer} Variable
10136 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10137 a variable. It points to something by being bound to the value of what
10138 it points to, like any other Lisp variable.
10141 Thus, if the value of the kill ring is:
10144 ("some text" "a different piece of text" "yet more text")
10149 and the @code{kill-ring-yank-pointer} points to the second clause, the
10150 value of @code{kill-ring-yank-pointer} is:
10153 ("a different piece of text" "yet more text")
10156 As explained in the previous chapter (@pxref{List Implementation}), the
10157 computer does not keep two different copies of the text being pointed to
10158 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10159 words ``a different piece of text'' and ``yet more text'' are not
10160 duplicated. Instead, the two Lisp variables point to the same pieces of
10161 text. Here is a diagram:
10163 @c cons-cell-diagram #5
10167 kill-ring kill-ring-yank-pointer
10169 | ___ ___ | ___ ___ ___ ___
10170 ---> | | | --> | | | | | |
10171 |___|___|----> |___|___|--> |___|___|--> nil
10174 | | --> "yet more text"
10176 | --> "a different piece of text"
10183 @ifset print-postscript-figures
10186 @center @image{cons-5}
10187 %%%% old method of including an image
10188 % \input /usr/local/lib/tex/inputs/psfig.tex
10189 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10194 @ifclear print-postscript-figures
10198 kill-ring kill-ring-yank-pointer
10200 | ___ ___ | ___ ___ ___ ___
10201 ---> | | | --> | | | | | |
10202 |___|___|----> |___|___|--> |___|___|--> nil
10205 | | --> "yet more text"
10207 | --> "a different piece of text
10216 Both the variable @code{kill-ring} and the variable
10217 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10218 usually described as if it were actually what it is composed of. The
10219 @code{kill-ring} is spoken of as if it were the list rather than that it
10220 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10221 spoken of as pointing to a list.
10223 These two ways of talking about the same thing sound confusing at first but
10224 make sense on reflection. The kill ring is generally thought of as the
10225 complete structure of data that holds the information of what has recently
10226 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10227 on the other hand, serves to indicate---that is, to `point to'---that part
10228 of the kill ring of which the first element (the @sc{car}) will be
10232 In GNU Emacs 22, the @code{kill-new} function calls
10234 @code{(setq kill-ring-yank-pointer kill-ring)}
10236 (defun rotate-yank-pointer (arg)
10237 "Rotate the yanking point in the kill ring.
10238 With argument, rotate that many kills forward (or backward, if negative)."
10240 (current-kill arg))
10242 (defun current-kill (n &optional do-not-move)
10243 "Rotate the yanking point by N places, and then return that kill.
10244 If N is zero, `interprogram-paste-function' is set, and calling it
10245 returns a string, then that string is added to the front of the
10246 kill ring and returned as the latest kill.
10247 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10248 yanking point; just return the Nth kill forward."
10249 (let ((interprogram-paste (and (= n 0)
10250 interprogram-paste-function
10251 (funcall interprogram-paste-function))))
10252 (if interprogram-paste
10254 ;; Disable the interprogram cut function when we add the new
10255 ;; text to the kill ring, so Emacs doesn't try to own the
10256 ;; selection, with identical text.
10257 (let ((interprogram-cut-function nil))
10258 (kill-new interprogram-paste))
10259 interprogram-paste)
10260 (or kill-ring (error "Kill ring is empty"))
10261 (let ((ARGth-kill-element
10262 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10263 (length kill-ring))
10266 (setq kill-ring-yank-pointer ARGth-kill-element))
10267 (car ARGth-kill-element)))))
10272 @node yank nthcdr Exercises
10273 @section Exercises with @code{yank} and @code{nthcdr}
10277 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10278 your kill ring. Add several items to your kill ring; look at its
10279 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10280 around the kill ring. How many items were in your kill ring? Find
10281 the value of @code{kill-ring-max}. Was your kill ring full, or could
10282 you have kept more blocks of text within it?
10285 Using @code{nthcdr} and @code{car}, construct a series of expressions
10286 to return the first, second, third, and fourth elements of a list.
10289 @node Loops & Recursion
10290 @chapter Loops and Recursion
10291 @cindex Loops and recursion
10292 @cindex Recursion and loops
10293 @cindex Repetition (loops)
10295 Emacs Lisp has two primary ways to cause an expression, or a series of
10296 expressions, to be evaluated repeatedly: one uses a @code{while}
10297 loop, and the other uses @dfn{recursion}.
10299 Repetition can be very valuable. For example, to move forward four
10300 sentences, you need only write a program that will move forward one
10301 sentence and then repeat the process four times. Since a computer does
10302 not get bored or tired, such repetitive action does not have the
10303 deleterious effects that excessive or the wrong kinds of repetition can
10306 People mostly write Emacs Lisp functions using @code{while} loops and
10307 their kin; but you can use recursion, which provides a very powerful
10308 way to think about and then to solve problems@footnote{You can write
10309 recursive functions to be frugal or wasteful of mental or computer
10310 resources; as it happens, methods that people find easy---that are
10311 frugal of `mental resources'---sometimes use considerable computer
10312 resources. Emacs was designed to run on machines that we now consider
10313 limited and its default settings are conservative. You may want to
10314 increase the values of @code{max-specpdl-size} and
10315 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10316 15 and 30 times their default value.}.
10319 * while:: Causing a stretch of code to repeat.
10321 * Recursion:: Causing a function to call itself.
10322 * Looping exercise::
10326 @section @code{while}
10330 The @code{while} special form tests whether the value returned by
10331 evaluating its first argument is true or false. This is similar to what
10332 the Lisp interpreter does with an @code{if}; what the interpreter does
10333 next, however, is different.
10335 In a @code{while} expression, if the value returned by evaluating the
10336 first argument is false, the Lisp interpreter skips the rest of the
10337 expression (the @dfn{body} of the expression) and does not evaluate it.
10338 However, if the value is true, the Lisp interpreter evaluates the body
10339 of the expression and then again tests whether the first argument to
10340 @code{while} is true or false. If the value returned by evaluating the
10341 first argument is again true, the Lisp interpreter again evaluates the
10342 body of the expression.
10345 The template for a @code{while} expression looks like this:
10349 (while @var{true-or-false-test}
10355 * Looping with while:: Repeat so long as test returns true.
10356 * Loop Example:: A @code{while} loop that uses a list.
10357 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10358 * Incrementing Loop:: A loop with an incrementing counter.
10359 * Incrementing Loop Details::
10360 * Decrementing Loop:: A loop with a decrementing counter.
10364 @node Looping with while
10365 @unnumberedsubsec Looping with @code{while}
10368 So long as the true-or-false-test of the @code{while} expression
10369 returns a true value when it is evaluated, the body is repeatedly
10370 evaluated. This process is called a loop since the Lisp interpreter
10371 repeats the same thing again and again, like an airplane doing a loop.
10372 When the result of evaluating the true-or-false-test is false, the
10373 Lisp interpreter does not evaluate the rest of the @code{while}
10374 expression and `exits the loop'.
10376 Clearly, if the value returned by evaluating the first argument to
10377 @code{while} is always true, the body following will be evaluated
10378 again and again @dots{} and again @dots{} forever. Conversely, if the
10379 value returned is never true, the expressions in the body will never
10380 be evaluated. The craft of writing a @code{while} loop consists of
10381 choosing a mechanism such that the true-or-false-test returns true
10382 just the number of times that you want the subsequent expressions to
10383 be evaluated, and then have the test return false.
10385 The value returned by evaluating a @code{while} is the value of the
10386 true-or-false-test. An interesting consequence of this is that a
10387 @code{while} loop that evaluates without error will return @code{nil}
10388 or false regardless of whether it has looped 1 or 100 times or none at
10389 all. A @code{while} expression that evaluates successfully never
10390 returns a true value! What this means is that @code{while} is always
10391 evaluated for its side effects, which is to say, the consequences of
10392 evaluating the expressions within the body of the @code{while} loop.
10393 This makes sense. It is not the mere act of looping that is desired,
10394 but the consequences of what happens when the expressions in the loop
10395 are repeatedly evaluated.
10398 @subsection A @code{while} Loop and a List
10400 A common way to control a @code{while} loop is to test whether a list
10401 has any elements. If it does, the loop is repeated; but if it does not,
10402 the repetition is ended. Since this is an important technique, we will
10403 create a short example to illustrate it.
10405 A simple way to test whether a list has elements is to evaluate the
10406 list: if it has no elements, it is an empty list and will return the
10407 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10408 the other hand, a list with elements will return those elements when it
10409 is evaluated. Since Emacs Lisp considers as true any value that is not
10410 @code{nil}, a list that returns elements will test true in a
10414 For example, you can set the variable @code{empty-list} to @code{nil} by
10415 evaluating the following @code{setq} expression:
10418 (setq empty-list ())
10422 After evaluating the @code{setq} expression, you can evaluate the
10423 variable @code{empty-list} in the usual way, by placing the cursor after
10424 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10431 On the other hand, if you set a variable to be a list with elements, the
10432 list will appear when you evaluate the variable, as you can see by
10433 evaluating the following two expressions:
10437 (setq animals '(gazelle giraffe lion tiger))
10443 Thus, to create a @code{while} loop that tests whether there are any
10444 items in the list @code{animals}, the first part of the loop will be
10455 When the @code{while} tests its first argument, the variable
10456 @code{animals} is evaluated. It returns a list. So long as the list
10457 has elements, the @code{while} considers the results of the test to be
10458 true; but when the list is empty, it considers the results of the test
10461 To prevent the @code{while} loop from running forever, some mechanism
10462 needs to be provided to empty the list eventually. An oft-used
10463 technique is to have one of the subsequent forms in the @code{while}
10464 expression set the value of the list to be the @sc{cdr} of the list.
10465 Each time the @code{cdr} function is evaluated, the list will be made
10466 shorter, until eventually only the empty list will be left. At this
10467 point, the test of the @code{while} loop will return false, and the
10468 arguments to the @code{while} will no longer be evaluated.
10470 For example, the list of animals bound to the variable @code{animals}
10471 can be set to be the @sc{cdr} of the original list with the
10472 following expression:
10475 (setq animals (cdr animals))
10479 If you have evaluated the previous expressions and then evaluate this
10480 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10481 area. If you evaluate the expression again, @code{(lion tiger)} will
10482 appear in the echo area. If you evaluate it again and yet again,
10483 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10485 A template for a @code{while} loop that uses the @code{cdr} function
10486 repeatedly to cause the true-or-false-test eventually to test false
10491 (while @var{test-whether-list-is-empty}
10493 @var{set-list-to-cdr-of-list})
10497 This test and use of @code{cdr} can be put together in a function that
10498 goes through a list and prints each element of the list on a line of its
10501 @node print-elements-of-list
10502 @subsection An Example: @code{print-elements-of-list}
10503 @findex print-elements-of-list
10505 The @code{print-elements-of-list} function illustrates a @code{while}
10508 @cindex @file{*scratch*} buffer
10509 The function requires several lines for its output. If you are
10510 reading this in a recent instance of GNU Emacs,
10511 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10512 you can evaluate the following expression inside of Info, as usual.
10514 If you are using an earlier version of Emacs, you need to copy the
10515 necessary expressions to your @file{*scratch*} buffer and evaluate
10516 them there. This is because the echo area had only one line in the
10519 You can copy the expressions by marking the beginning of the region
10520 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10521 the end of the region and then copying the region using @kbd{M-w}
10522 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10523 then provides visual feedback). In the @file{*scratch*}
10524 buffer, you can yank the expressions back by typing @kbd{C-y}
10527 After you have copied the expressions to the @file{*scratch*} buffer,
10528 evaluate each expression in turn. Be sure to evaluate the last
10529 expression, @code{(print-elements-of-list animals)}, by typing
10530 @kbd{C-u C-x C-e}, that is, by giving an argument to
10531 @code{eval-last-sexp}. This will cause the result of the evaluation
10532 to be printed in the @file{*scratch*} buffer instead of being printed
10533 in the echo area. (Otherwise you will see something like this in your
10534 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10535 each @samp{^J} stands for a `newline'.)
10538 In a recent instance of GNU Emacs, you can evaluate these expressions
10539 directly in the Info buffer, and the echo area will grow to show the
10544 (setq animals '(gazelle giraffe lion tiger))
10546 (defun print-elements-of-list (list)
10547 "Print each element of LIST on a line of its own."
10550 (setq list (cdr list))))
10552 (print-elements-of-list animals)
10558 When you evaluate the three expressions in sequence, you will see
10574 Each element of the list is printed on a line of its own (that is what
10575 the function @code{print} does) and then the value returned by the
10576 function is printed. Since the last expression in the function is the
10577 @code{while} loop, and since @code{while} loops always return
10578 @code{nil}, a @code{nil} is printed after the last element of the list.
10580 @node Incrementing Loop
10581 @subsection A Loop with an Incrementing Counter
10583 A loop is not useful unless it stops when it ought. Besides
10584 controlling a loop with a list, a common way of stopping a loop is to
10585 write the first argument as a test that returns false when the correct
10586 number of repetitions are complete. This means that the loop must
10587 have a counter---an expression that counts how many times the loop
10591 @node Incrementing Loop Details
10592 @unnumberedsubsec Details of an Incrementing Loop
10595 The test for a loop with an incrementing counter can be an expression
10596 such as @code{(< count desired-number)} which returns @code{t} for
10597 true if the value of @code{count} is less than the
10598 @code{desired-number} of repetitions and @code{nil} for false if the
10599 value of @code{count} is equal to or is greater than the
10600 @code{desired-number}. The expression that increments the count can
10601 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10602 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10603 argument. (The expression @w{@code{(1+ count)}} has the same result
10604 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10607 The template for a @code{while} loop controlled by an incrementing
10608 counter looks like this:
10612 @var{set-count-to-initial-value}
10613 (while (< count desired-number) ; @r{true-or-false-test}
10615 (setq count (1+ count))) ; @r{incrementer}
10620 Note that you need to set the initial value of @code{count}; usually it
10624 * Incrementing Example:: Counting pebbles in a triangle.
10625 * Inc Example parts:: The parts of the function definition.
10626 * Inc Example altogether:: Putting the function definition together.
10629 @node Incrementing Example
10630 @unnumberedsubsubsec Example with incrementing counter
10632 Suppose you are playing on the beach and decide to make a triangle of
10633 pebbles, putting one pebble in the first row, two in the second row,
10634 three in the third row and so on, like this:
10652 @bullet{} @bullet{}
10653 @bullet{} @bullet{} @bullet{}
10654 @bullet{} @bullet{} @bullet{} @bullet{}
10661 (About 2500 years ago, Pythagoras and others developed the beginnings of
10662 number theory by considering questions such as this.)
10664 Suppose you want to know how many pebbles you will need to make a
10665 triangle with 7 rows?
10667 Clearly, what you need to do is add up the numbers from 1 to 7. There
10668 are two ways to do this; start with the smallest number, one, and add up
10669 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10670 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10671 mechanisms illustrate common ways of writing @code{while} loops, we will
10672 create two examples, one counting up and the other counting down. In
10673 this first example, we will start with 1 and add 2, 3, 4 and so on.
10675 If you are just adding up a short list of numbers, the easiest way to do
10676 it is to add up all the numbers at once. However, if you do not know
10677 ahead of time how many numbers your list will have, or if you want to be
10678 prepared for a very long list, then you need to design your addition so
10679 that what you do is repeat a simple process many times instead of doing
10680 a more complex process once.
10682 For example, instead of adding up all the pebbles all at once, what you
10683 can do is add the number of pebbles in the first row, 1, to the number
10684 in the second row, 2, and then add the total of those two rows to the
10685 third row, 3. Then you can add the number in the fourth row, 4, to the
10686 total of the first three rows; and so on.
10688 The critical characteristic of the process is that each repetitive
10689 action is simple. In this case, at each step we add only two numbers,
10690 the number of pebbles in the row and the total already found. This
10691 process of adding two numbers is repeated again and again until the last
10692 row has been added to the total of all the preceding rows. In a more
10693 complex loop the repetitive action might not be so simple, but it will
10694 be simpler than doing everything all at once.
10696 @node Inc Example parts
10697 @unnumberedsubsubsec The parts of the function definition
10699 The preceding analysis gives us the bones of our function definition:
10700 first, we will need a variable that we can call @code{total} that will
10701 be the total number of pebbles. This will be the value returned by
10704 Second, we know that the function will require an argument: this
10705 argument will be the total number of rows in the triangle. It can be
10706 called @code{number-of-rows}.
10708 Finally, we need a variable to use as a counter. We could call this
10709 variable @code{counter}, but a better name is @code{row-number}. That
10710 is because what the counter does in this function is count rows, and a
10711 program should be written to be as understandable as possible.
10713 When the Lisp interpreter first starts evaluating the expressions in the
10714 function, the value of @code{total} should be set to zero, since we have
10715 not added anything to it. Then the function should add the number of
10716 pebbles in the first row to the total, and then add the number of
10717 pebbles in the second to the total, and then add the number of
10718 pebbles in the third row to the total, and so on, until there are no
10719 more rows left to add.
10721 Both @code{total} and @code{row-number} are used only inside the
10722 function, so they can be declared as local variables with @code{let}
10723 and given initial values. Clearly, the initial value for @code{total}
10724 should be 0. The initial value of @code{row-number} should be 1,
10725 since we start with the first row. This means that the @code{let}
10726 statement will look like this:
10736 After the internal variables are declared and bound to their initial
10737 values, we can begin the @code{while} loop. The expression that serves
10738 as the test should return a value of @code{t} for true so long as the
10739 @code{row-number} is less than or equal to the @code{number-of-rows}.
10740 (If the expression tests true only so long as the row number is less
10741 than the number of rows in the triangle, the last row will never be
10742 added to the total; hence the row number has to be either less than or
10743 equal to the number of rows.)
10746 @findex <= @r{(less than or equal)}
10747 Lisp provides the @code{<=} function that returns true if the value of
10748 its first argument is less than or equal to the value of its second
10749 argument and false otherwise. So the expression that the @code{while}
10750 will evaluate as its test should look like this:
10753 (<= row-number number-of-rows)
10756 The total number of pebbles can be found by repeatedly adding the number
10757 of pebbles in a row to the total already found. Since the number of
10758 pebbles in the row is equal to the row number, the total can be found by
10759 adding the row number to the total. (Clearly, in a more complex
10760 situation, the number of pebbles in the row might be related to the row
10761 number in a more complicated way; if this were the case, the row number
10762 would be replaced by the appropriate expression.)
10765 (setq total (+ total row-number))
10769 What this does is set the new value of @code{total} to be equal to the
10770 sum of adding the number of pebbles in the row to the previous total.
10772 After setting the value of @code{total}, the conditions need to be
10773 established for the next repetition of the loop, if there is one. This
10774 is done by incrementing the value of the @code{row-number} variable,
10775 which serves as a counter. After the @code{row-number} variable has
10776 been incremented, the true-or-false-test at the beginning of the
10777 @code{while} loop tests whether its value is still less than or equal to
10778 the value of the @code{number-of-rows} and if it is, adds the new value
10779 of the @code{row-number} variable to the @code{total} of the previous
10780 repetition of the loop.
10783 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10784 @code{row-number} variable can be incremented with this expression:
10787 (setq row-number (1+ row-number))
10790 @node Inc Example altogether
10791 @unnumberedsubsubsec Putting the function definition together
10793 We have created the parts for the function definition; now we need to
10797 First, the contents of the @code{while} expression:
10801 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10802 (setq total (+ total row-number))
10803 (setq row-number (1+ row-number))) ; @r{incrementer}
10807 Along with the @code{let} expression varlist, this very nearly
10808 completes the body of the function definition. However, it requires
10809 one final element, the need for which is somewhat subtle.
10811 The final touch is to place the variable @code{total} on a line by
10812 itself after the @code{while} expression. Otherwise, the value returned
10813 by the whole function is the value of the last expression that is
10814 evaluated in the body of the @code{let}, and this is the value
10815 returned by the @code{while}, which is always @code{nil}.
10817 This may not be evident at first sight. It almost looks as if the
10818 incrementing expression is the last expression of the whole function.
10819 But that expression is part of the body of the @code{while}; it is the
10820 last element of the list that starts with the symbol @code{while}.
10821 Moreover, the whole of the @code{while} loop is a list within the body
10825 In outline, the function will look like this:
10829 (defun @var{name-of-function} (@var{argument-list})
10830 "@var{documentation}@dots{}"
10831 (let (@var{varlist})
10832 (while (@var{true-or-false-test})
10833 @var{body-of-while}@dots{} )
10834 @dots{} )) ; @r{Need final expression here.}
10838 The result of evaluating the @code{let} is what is going to be returned
10839 by the @code{defun} since the @code{let} is not embedded within any
10840 containing list, except for the @code{defun} as a whole. However, if
10841 the @code{while} is the last element of the @code{let} expression, the
10842 function will always return @code{nil}. This is not what we want!
10843 Instead, what we want is the value of the variable @code{total}. This
10844 is returned by simply placing the symbol as the last element of the list
10845 starting with @code{let}. It gets evaluated after the preceding
10846 elements of the list are evaluated, which means it gets evaluated after
10847 it has been assigned the correct value for the total.
10849 It may be easier to see this by printing the list starting with
10850 @code{let} all on one line. This format makes it evident that the
10851 @var{varlist} and @code{while} expressions are the second and third
10852 elements of the list starting with @code{let}, and the @code{total} is
10857 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10862 Putting everything together, the @code{triangle} function definition
10867 (defun triangle (number-of-rows) ; @r{Version with}
10868 ; @r{ incrementing counter.}
10869 "Add up the number of pebbles in a triangle.
10870 The first row has one pebble, the second row two pebbles,
10871 the third row three pebbles, and so on.
10872 The argument is NUMBER-OF-ROWS."
10877 (while (<= row-number number-of-rows)
10878 (setq total (+ total row-number))
10879 (setq row-number (1+ row-number)))
10885 After you have installed @code{triangle} by evaluating the function, you
10886 can try it out. Here are two examples:
10897 The sum of the first four numbers is 10 and the sum of the first seven
10900 @node Decrementing Loop
10901 @subsection Loop with a Decrementing Counter
10903 Another common way to write a @code{while} loop is to write the test
10904 so that it determines whether a counter is greater than zero. So long
10905 as the counter is greater than zero, the loop is repeated. But when
10906 the counter is equal to or less than zero, the loop is stopped. For
10907 this to work, the counter has to start out greater than zero and then
10908 be made smaller and smaller by a form that is evaluated
10911 The test will be an expression such as @code{(> counter 0)} which
10912 returns @code{t} for true if the value of @code{counter} is greater
10913 than zero, and @code{nil} for false if the value of @code{counter} is
10914 equal to or less than zero. The expression that makes the number
10915 smaller and smaller can be a simple @code{setq} such as @code{(setq
10916 counter (1- counter))}, where @code{1-} is a built-in function in
10917 Emacs Lisp that subtracts 1 from its argument.
10920 The template for a decrementing @code{while} loop looks like this:
10924 (while (> counter 0) ; @r{true-or-false-test}
10926 (setq counter (1- counter))) ; @r{decrementer}
10931 * Decrementing Example:: More pebbles on the beach.
10932 * Dec Example parts:: The parts of the function definition.
10933 * Dec Example altogether:: Putting the function definition together.
10936 @node Decrementing Example
10937 @unnumberedsubsubsec Example with decrementing counter
10939 To illustrate a loop with a decrementing counter, we will rewrite the
10940 @code{triangle} function so the counter decreases to zero.
10942 This is the reverse of the earlier version of the function. In this
10943 case, to find out how many pebbles are needed to make a triangle with
10944 3 rows, add the number of pebbles in the third row, 3, to the number
10945 in the preceding row, 2, and then add the total of those two rows to
10946 the row that precedes them, which is 1.
10948 Likewise, to find the number of pebbles in a triangle with 7 rows, add
10949 the number of pebbles in the seventh row, 7, to the number in the
10950 preceding row, which is 6, and then add the total of those two rows to
10951 the row that precedes them, which is 5, and so on. As in the previous
10952 example, each addition only involves adding two numbers, the total of
10953 the rows already added up and the number of pebbles in the row that is
10954 being added to the total. This process of adding two numbers is
10955 repeated again and again until there are no more pebbles to add.
10957 We know how many pebbles to start with: the number of pebbles in the
10958 last row is equal to the number of rows. If the triangle has seven
10959 rows, the number of pebbles in the last row is 7. Likewise, we know how
10960 many pebbles are in the preceding row: it is one less than the number in
10963 @node Dec Example parts
10964 @unnumberedsubsubsec The parts of the function definition
10966 We start with three variables: the total number of rows in the
10967 triangle; the number of pebbles in a row; and the total number of
10968 pebbles, which is what we want to calculate. These variables can be
10969 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
10970 @code{total}, respectively.
10972 Both @code{total} and @code{number-of-pebbles-in-row} are used only
10973 inside the function and are declared with @code{let}. The initial
10974 value of @code{total} should, of course, be zero. However, the
10975 initial value of @code{number-of-pebbles-in-row} should be equal to
10976 the number of rows in the triangle, since the addition will start with
10980 This means that the beginning of the @code{let} expression will look
10986 (number-of-pebbles-in-row number-of-rows))
10991 The total number of pebbles can be found by repeatedly adding the number
10992 of pebbles in a row to the total already found, that is, by repeatedly
10993 evaluating the following expression:
10996 (setq total (+ total number-of-pebbles-in-row))
11000 After the @code{number-of-pebbles-in-row} is added to the @code{total},
11001 the @code{number-of-pebbles-in-row} should be decremented by one, since
11002 the next time the loop repeats, the preceding row will be
11003 added to the total.
11005 The number of pebbles in a preceding row is one less than the number of
11006 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
11007 used to compute the number of pebbles in the preceding row. This can be
11008 done with the following expression:
11012 (setq number-of-pebbles-in-row
11013 (1- number-of-pebbles-in-row))
11017 Finally, we know that the @code{while} loop should stop making repeated
11018 additions when there are no pebbles in a row. So the test for
11019 the @code{while} loop is simply:
11022 (while (> number-of-pebbles-in-row 0)
11025 @node Dec Example altogether
11026 @unnumberedsubsubsec Putting the function definition together
11028 We can put these expressions together to create a function definition
11029 that works. However, on examination, we find that one of the local
11030 variables is unneeded!
11033 The function definition looks like this:
11037 ;;; @r{First subtractive version.}
11038 (defun triangle (number-of-rows)
11039 "Add up the number of pebbles in a triangle."
11041 (number-of-pebbles-in-row number-of-rows))
11042 (while (> number-of-pebbles-in-row 0)
11043 (setq total (+ total number-of-pebbles-in-row))
11044 (setq number-of-pebbles-in-row
11045 (1- number-of-pebbles-in-row)))
11050 As written, this function works.
11052 However, we do not need @code{number-of-pebbles-in-row}.
11054 @cindex Argument as local variable
11055 When the @code{triangle} function is evaluated, the symbol
11056 @code{number-of-rows} will be bound to a number, giving it an initial
11057 value. That number can be changed in the body of the function as if
11058 it were a local variable, without any fear that such a change will
11059 effect the value of the variable outside of the function. This is a
11060 very useful characteristic of Lisp; it means that the variable
11061 @code{number-of-rows} can be used anywhere in the function where
11062 @code{number-of-pebbles-in-row} is used.
11065 Here is a second version of the function written a bit more cleanly:
11069 (defun triangle (number) ; @r{Second version.}
11070 "Return sum of numbers 1 through NUMBER inclusive."
11072 (while (> number 0)
11073 (setq total (+ total number))
11074 (setq number (1- number)))
11079 In brief, a properly written @code{while} loop will consist of three parts:
11083 A test that will return false after the loop has repeated itself the
11084 correct number of times.
11087 An expression the evaluation of which will return the value desired
11088 after being repeatedly evaluated.
11091 An expression to change the value passed to the true-or-false-test so
11092 that the test returns false after the loop has repeated itself the right
11096 @node dolist dotimes
11097 @section Save your time: @code{dolist} and @code{dotimes}
11099 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11100 provide for looping. Sometimes these are quicker to write than the
11101 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11102 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11104 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11105 list': @code{dolist} automatically shortens the list each time it
11106 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11107 each shorter version of the list to the first of its arguments.
11109 @code{dotimes} loops a specific number of times: you specify the number.
11117 @unnumberedsubsec The @code{dolist} Macro
11120 Suppose, for example, you want to reverse a list, so that
11121 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11124 In practice, you would use the @code{reverse} function, like this:
11128 (setq animals '(gazelle giraffe lion tiger))
11136 Here is how you could reverse the list using a @code{while} loop:
11140 (setq animals '(gazelle giraffe lion tiger))
11142 (defun reverse-list-with-while (list)
11143 "Using while, reverse the order of LIST."
11144 (let (value) ; make sure list starts empty
11146 (setq value (cons (car list) value))
11147 (setq list (cdr list)))
11150 (reverse-list-with-while animals)
11156 And here is how you could use the @code{dolist} macro:
11160 (setq animals '(gazelle giraffe lion tiger))
11162 (defun reverse-list-with-dolist (list)
11163 "Using dolist, reverse the order of LIST."
11164 (let (value) ; make sure list starts empty
11165 (dolist (element list value)
11166 (setq value (cons element value)))))
11168 (reverse-list-with-dolist animals)
11174 In Info, you can place your cursor after the closing parenthesis of
11175 each expression and type @kbd{C-x C-e}; in each case, you should see
11178 (tiger lion giraffe gazelle)
11184 For this example, the existing @code{reverse} function is obviously best.
11185 The @code{while} loop is just like our first example (@pxref{Loop
11186 Example, , A @code{while} Loop and a List}). The @code{while} first
11187 checks whether the list has elements; if so, it constructs a new list
11188 by adding the first element of the list to the existing list (which in
11189 the first iteration of the loop is @code{nil}). Since the second
11190 element is prepended in front of the first element, and the third
11191 element is prepended in front of the second element, the list is reversed.
11193 In the expression using a @code{while} loop,
11194 the @w{@code{(setq list (cdr list))}}
11195 expression shortens the list, so the @code{while} loop eventually
11196 stops. In addition, it provides the @code{cons} expression with a new
11197 first element by creating a new and shorter list at each repetition of
11200 The @code{dolist} expression does very much the same as the
11201 @code{while} expression, except that the @code{dolist} macro does some
11202 of the work you have to do when writing a @code{while} expression.
11204 Like a @code{while} loop, a @code{dolist} loops. What is different is
11205 that it automatically shortens the list each time it loops---it
11206 `@sc{cdr}s down the list' on its own---and it automatically binds
11207 the @sc{car} of each shorter version of the list to the first of its
11210 In the example, the @sc{car} of each shorter version of the list is
11211 referred to using the symbol @samp{element}, the list itself is called
11212 @samp{list}, and the value returned is called @samp{value}. The
11213 remainder of the @code{dolist} expression is the body.
11215 The @code{dolist} expression binds the @sc{car} of each shorter
11216 version of the list to @code{element} and then evaluates the body of
11217 the expression; and repeats the loop. The result is returned in
11221 @unnumberedsubsec The @code{dotimes} Macro
11224 The @code{dotimes} macro is similar to @code{dolist}, except that it
11225 loops a specific number of times.
11227 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11228 and so forth each time around the loop, and the value of the third
11229 argument is returned. You need to provide the value of the second
11230 argument, which is how many times the macro loops.
11233 For example, the following binds the numbers from 0 up to, but not
11234 including, the number 3 to the first argument, @var{number}, and then
11235 constructs a list of the three numbers. (The first number is 0, the
11236 second number is 1, and the third number is 2; this makes a total of
11237 three numbers in all, starting with zero as the first number.)
11241 (let (value) ; otherwise a value is a void variable
11242 (dotimes (number 3 value)
11243 (setq value (cons number value))))
11250 @code{dotimes} returns @code{value}, so the way to use
11251 @code{dotimes} is to operate on some expression @var{number} number of
11252 times and then return the result, either as a list or an atom.
11255 Here is an example of a @code{defun} that uses @code{dotimes} to add
11256 up the number of pebbles in a triangle.
11260 (defun triangle-using-dotimes (number-of-rows)
11261 "Using dotimes, add up the number of pebbles in a triangle."
11262 (let ((total 0)) ; otherwise a total is a void variable
11263 (dotimes (number number-of-rows total)
11264 (setq total (+ total (1+ number))))))
11266 (triangle-using-dotimes 4)
11274 A recursive function contains code that tells the Lisp interpreter to
11275 call a program that runs exactly like itself, but with slightly
11276 different arguments. The code runs exactly the same because it has
11277 the same name. However, even though the program has the same name, it
11278 is not the same entity. It is different. In the jargon, it is a
11279 different `instance'.
11281 Eventually, if the program is written correctly, the `slightly
11282 different arguments' will become sufficiently different from the first
11283 arguments that the final instance will stop.
11286 * Building Robots:: Same model, different serial number ...
11287 * Recursive Definition Parts:: Walk until you stop ...
11288 * Recursion with list:: Using a list as the test whether to recurse.
11289 * Recursive triangle function::
11290 * Recursion with cond::
11291 * Recursive Patterns:: Often used templates.
11292 * No Deferment:: Don't store up work ...
11293 * No deferment solution::
11296 @node Building Robots
11297 @subsection Building Robots: Extending the Metaphor
11298 @cindex Building robots
11299 @cindex Robots, building
11301 It is sometimes helpful to think of a running program as a robot that
11302 does a job. In doing its job, a recursive function calls on a second
11303 robot to help it. The second robot is identical to the first in every
11304 way, except that the second robot helps the first and has been
11305 passed different arguments than the first.
11307 In a recursive function, the second robot may call a third; and the
11308 third may call a fourth, and so on. Each of these is a different
11309 entity; but all are clones.
11311 Since each robot has slightly different instructions---the arguments
11312 will differ from one robot to the next---the last robot should know
11315 Let's expand on the metaphor in which a computer program is a robot.
11317 A function definition provides the blueprints for a robot. When you
11318 install a function definition, that is, when you evaluate a
11319 @code{defun} macro, you install the necessary equipment to build
11320 robots. It is as if you were in a factory, setting up an assembly
11321 line. Robots with the same name are built according to the same
11322 blueprints. So they have, as it were, the same `model number', but a
11323 different `serial number'.
11325 We often say that a recursive function `calls itself'. What we mean
11326 is that the instructions in a recursive function cause the Lisp
11327 interpreter to run a different function that has the same name and
11328 does the same job as the first, but with different arguments.
11330 It is important that the arguments differ from one instance to the
11331 next; otherwise, the process will never stop.
11333 @node Recursive Definition Parts
11334 @subsection The Parts of a Recursive Definition
11335 @cindex Parts of a Recursive Definition
11336 @cindex Recursive Definition Parts
11338 A recursive function typically contains a conditional expression which
11343 A true-or-false-test that determines whether the function is called
11344 again, here called the @dfn{do-again-test}.
11347 The name of the function. When this name is called, a new instance of
11348 the function---a new robot, as it were---is created and told what to do.
11351 An expression that returns a different value each time the function is
11352 called, here called the @dfn{next-step-expression}. Consequently, the
11353 argument (or arguments) passed to the new instance of the function
11354 will be different from that passed to the previous instance. This
11355 causes the conditional expression, the @dfn{do-again-test}, to test
11356 false after the correct number of repetitions.
11359 Recursive functions can be much simpler than any other kind of
11360 function. Indeed, when people first start to use them, they often look
11361 so mysteriously simple as to be incomprehensible. Like riding a
11362 bicycle, reading a recursive function definition takes a certain knack
11363 which is hard at first but then seems simple.
11366 There are several different common recursive patterns. A very simple
11367 pattern looks like this:
11371 (defun @var{name-of-recursive-function} (@var{argument-list})
11372 "@var{documentation}@dots{}"
11373 (if @var{do-again-test}
11375 (@var{name-of-recursive-function}
11376 @var{next-step-expression})))
11380 Each time a recursive function is evaluated, a new instance of it is
11381 created and told what to do. The arguments tell the instance what to do.
11383 An argument is bound to the value of the next-step-expression. Each
11384 instance runs with a different value of the next-step-expression.
11386 The value in the next-step-expression is used in the do-again-test.
11388 The value returned by the next-step-expression is passed to the new
11389 instance of the function, which evaluates it (or some
11390 transmogrification of it) to determine whether to continue or stop.
11391 The next-step-expression is designed so that the do-again-test returns
11392 false when the function should no longer be repeated.
11394 The do-again-test is sometimes called the @dfn{stop condition},
11395 since it stops the repetitions when it tests false.
11397 @node Recursion with list
11398 @subsection Recursion with a List
11400 The example of a @code{while} loop that printed the elements of a list
11401 of numbers can be written recursively. Here is the code, including
11402 an expression to set the value of the variable @code{animals} to a list.
11404 If you are reading this in Info in Emacs, you can evaluate this
11405 expression directly in Info. Otherwise, you must copy the example
11406 to the @file{*scratch*} buffer and evaluate each expression there.
11407 Use @kbd{C-u C-x C-e} to evaluate the
11408 @code{(print-elements-recursively animals)} expression so that the
11409 results are printed in the buffer; otherwise the Lisp interpreter will
11410 try to squeeze the results into the one line of the echo area.
11412 Also, place your cursor immediately after the last closing parenthesis
11413 of the @code{print-elements-recursively} function, before the comment.
11414 Otherwise, the Lisp interpreter will try to evaluate the comment.
11416 @findex print-elements-recursively
11419 (setq animals '(gazelle giraffe lion tiger))
11421 (defun print-elements-recursively (list)
11422 "Print each element of LIST on a line of its own.
11424 (when list ; @r{do-again-test}
11425 (print (car list)) ; @r{body}
11426 (print-elements-recursively ; @r{recursive call}
11427 (cdr list)))) ; @r{next-step-expression}
11429 (print-elements-recursively animals)
11433 The @code{print-elements-recursively} function first tests whether
11434 there is any content in the list; if there is, the function prints the
11435 first element of the list, the @sc{car} of the list. Then the
11436 function `invokes itself', but gives itself as its argument, not the
11437 whole list, but the second and subsequent elements of the list, the
11438 @sc{cdr} of the list.
11440 Put another way, if the list is not empty, the function invokes
11441 another instance of code that is similar to the initial code, but is a
11442 different thread of execution, with different arguments than the first
11445 Put in yet another way, if the list is not empty, the first robot
11446 assembles a second robot and tells it what to do; the second robot is
11447 a different individual from the first, but is the same model.
11449 When the second evaluation occurs, the @code{when} expression is
11450 evaluated and if true, prints the first element of the list it
11451 receives as its argument (which is the second element of the original
11452 list). Then the function `calls itself' with the @sc{cdr} of the list
11453 it is invoked with, which (the second time around) is the @sc{cdr} of
11454 the @sc{cdr} of the original list.
11456 Note that although we say that the function `calls itself', what we
11457 mean is that the Lisp interpreter assembles and instructs a new
11458 instance of the program. The new instance is a clone of the first,
11459 but is a separate individual.
11461 Each time the function `invokes itself', it invokes itself on a
11462 shorter version of the original list. It creates a new instance that
11463 works on a shorter list.
11465 Eventually, the function invokes itself on an empty list. It creates
11466 a new instance whose argument is @code{nil}. The conditional expression
11467 tests the value of @code{list}. Since the value of @code{list} is
11468 @code{nil}, the @code{when} expression tests false so the then-part is
11469 not evaluated. The function as a whole then returns @code{nil}.
11472 When you evaluate the expression @code{(print-elements-recursively
11473 animals)} in the @file{*scratch*} buffer, you see this result:
11489 @node Recursive triangle function
11490 @subsection Recursion in Place of a Counter
11491 @findex triangle-recursively
11494 The @code{triangle} function described in a previous section can also
11495 be written recursively. It looks like this:
11499 (defun triangle-recursively (number)
11500 "Return the sum of the numbers 1 through NUMBER inclusive.
11502 (if (= number 1) ; @r{do-again-test}
11504 (+ number ; @r{else-part}
11505 (triangle-recursively ; @r{recursive call}
11506 (1- number))))) ; @r{next-step-expression}
11508 (triangle-recursively 7)
11513 You can install this function by evaluating it and then try it by
11514 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11515 cursor immediately after the last parenthesis of the function
11516 definition, before the comment.) The function evaluates to 28.
11518 To understand how this function works, let's consider what happens in the
11519 various cases when the function is passed 1, 2, 3, or 4 as the value of
11523 * Recursive Example arg of 1 or 2::
11524 * Recursive Example arg of 3 or 4::
11528 @node Recursive Example arg of 1 or 2
11529 @unnumberedsubsubsec An argument of 1 or 2
11532 First, what happens if the value of the argument is 1?
11534 The function has an @code{if} expression after the documentation
11535 string. It tests whether the value of @code{number} is equal to 1; if
11536 so, Emacs evaluates the then-part of the @code{if} expression, which
11537 returns the number 1 as the value of the function. (A triangle with
11538 one row has one pebble in it.)
11540 Suppose, however, that the value of the argument is 2. In this case,
11541 Emacs evaluates the else-part of the @code{if} expression.
11544 The else-part consists of an addition, the recursive call to
11545 @code{triangle-recursively} and a decrementing action; and it looks like
11549 (+ number (triangle-recursively (1- number)))
11552 When Emacs evaluates this expression, the innermost expression is
11553 evaluated first; then the other parts in sequence. Here are the steps
11557 @item Step 1 @w{ } Evaluate the innermost expression.
11559 The innermost expression is @code{(1- number)} so Emacs decrements the
11560 value of @code{number} from 2 to 1.
11562 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11564 The Lisp interpreter creates an individual instance of
11565 @code{triangle-recursively}. It does not matter that this function is
11566 contained within itself. Emacs passes the result Step 1 as the
11567 argument used by this instance of the @code{triangle-recursively}
11570 In this case, Emacs evaluates @code{triangle-recursively} with an
11571 argument of 1. This means that this evaluation of
11572 @code{triangle-recursively} returns 1.
11574 @item Step 3 @w{ } Evaluate the value of @code{number}.
11576 The variable @code{number} is the second element of the list that
11577 starts with @code{+}; its value is 2.
11579 @item Step 4 @w{ } Evaluate the @code{+} expression.
11581 The @code{+} expression receives two arguments, the first
11582 from the evaluation of @code{number} (Step 3) and the second from the
11583 evaluation of @code{triangle-recursively} (Step 2).
11585 The result of the addition is the sum of 2 plus 1, and the number 3 is
11586 returned, which is correct. A triangle with two rows has three
11590 @node Recursive Example arg of 3 or 4
11591 @unnumberedsubsubsec An argument of 3 or 4
11593 Suppose that @code{triangle-recursively} is called with an argument of
11597 @item Step 1 @w{ } Evaluate the do-again-test.
11599 The @code{if} expression is evaluated first. This is the do-again
11600 test and returns false, so the else-part of the @code{if} expression
11601 is evaluated. (Note that in this example, the do-again-test causes
11602 the function to call itself when it tests false, not when it tests
11605 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11607 The innermost expression of the else-part is evaluated, which decrements
11608 3 to 2. This is the next-step-expression.
11610 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11612 The number 2 is passed to the @code{triangle-recursively} function.
11614 We already know what happens when Emacs evaluates @code{triangle-recursively} with
11615 an argument of 2. After going through the sequence of actions described
11616 earlier, it returns a value of 3. So that is what will happen here.
11618 @item Step 4 @w{ } Evaluate the addition.
11620 3 will be passed as an argument to the addition and will be added to the
11621 number with which the function was called, which is 3.
11625 The value returned by the function as a whole will be 6.
11627 Now that we know what will happen when @code{triangle-recursively} is
11628 called with an argument of 3, it is evident what will happen if it is
11629 called with an argument of 4:
11633 In the recursive call, the evaluation of
11636 (triangle-recursively (1- 4))
11641 will return the value of evaluating
11644 (triangle-recursively 3)
11648 which is 6 and this value will be added to 4 by the addition in the
11653 The value returned by the function as a whole will be 10.
11655 Each time @code{triangle-recursively} is evaluated, it evaluates a
11656 version of itself---a different instance of itself---with a smaller
11657 argument, until the argument is small enough so that it does not
11660 Note that this particular design for a recursive function
11661 requires that operations be deferred.
11663 Before @code{(triangle-recursively 7)} can calculate its answer, it
11664 must call @code{(triangle-recursively 6)}; and before
11665 @code{(triangle-recursively 6)} can calculate its answer, it must call
11666 @code{(triangle-recursively 5)}; and so on. That is to say, the
11667 calculation that @code{(triangle-recursively 7)} makes must be
11668 deferred until @code{(triangle-recursively 6)} makes its calculation;
11669 and @code{(triangle-recursively 6)} must defer until
11670 @code{(triangle-recursively 5)} completes; and so on.
11672 If each of these instances of @code{triangle-recursively} are thought
11673 of as different robots, the first robot must wait for the second to
11674 complete its job, which must wait until the third completes, and so
11677 There is a way around this kind of waiting, which we will discuss in
11678 @ref{No Deferment, , Recursion without Deferments}.
11680 @node Recursion with cond
11681 @subsection Recursion Example Using @code{cond}
11684 The version of @code{triangle-recursively} described earlier is written
11685 with the @code{if} special form. It can also be written using another
11686 special form called @code{cond}. The name of the special form
11687 @code{cond} is an abbreviation of the word @samp{conditional}.
11689 Although the @code{cond} special form is not used as often in the
11690 Emacs Lisp sources as @code{if}, it is used often enough to justify
11694 The template for a @code{cond} expression looks like this:
11704 where the @var{body} is a series of lists.
11707 Written out more fully, the template looks like this:
11712 (@var{first-true-or-false-test} @var{first-consequent})
11713 (@var{second-true-or-false-test} @var{second-consequent})
11714 (@var{third-true-or-false-test} @var{third-consequent})
11719 When the Lisp interpreter evaluates the @code{cond} expression, it
11720 evaluates the first element (the @sc{car} or true-or-false-test) of
11721 the first expression in a series of expressions within the body of the
11724 If the true-or-false-test returns @code{nil} the rest of that
11725 expression, the consequent, is skipped and the true-or-false-test of the
11726 next expression is evaluated. When an expression is found whose
11727 true-or-false-test returns a value that is not @code{nil}, the
11728 consequent of that expression is evaluated. The consequent can be one
11729 or more expressions. If the consequent consists of more than one
11730 expression, the expressions are evaluated in sequence and the value of
11731 the last one is returned. If the expression does not have a consequent,
11732 the value of the true-or-false-test is returned.
11734 If none of the true-or-false-tests test true, the @code{cond} expression
11735 returns @code{nil}.
11738 Written using @code{cond}, the @code{triangle} function looks like this:
11742 (defun triangle-using-cond (number)
11743 (cond ((<= number 0) 0)
11746 (+ number (triangle-using-cond (1- number))))))
11751 In this example, the @code{cond} returns 0 if the number is less than or
11752 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11753 number (triangle-using-cond (1- number)))} if the number is greater than
11756 @node Recursive Patterns
11757 @subsection Recursive Patterns
11758 @cindex Recursive Patterns
11760 Here are three common recursive patterns. Each involves a list.
11761 Recursion does not need to involve lists, but Lisp is designed for lists
11762 and this provides a sense of its primal capabilities.
11771 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11772 @cindex Every, type of recursive pattern
11773 @cindex Recursive pattern: every
11775 In the @code{every} recursive pattern, an action is performed on every
11779 The basic pattern is:
11783 If a list be empty, return @code{nil}.
11785 Else, act on the beginning of the list (the @sc{car} of the list)
11788 through a recursive call by the function on the rest (the
11789 @sc{cdr}) of the list,
11791 and, optionally, combine the acted-on element, using @code{cons},
11792 with the results of acting on the rest.
11801 (defun square-each (numbers-list)
11802 "Square each of a NUMBERS LIST, recursively."
11803 (if (not numbers-list) ; do-again-test
11806 (* (car numbers-list) (car numbers-list))
11807 (square-each (cdr numbers-list))))) ; next-step-expression
11811 (square-each '(1 2 3))
11818 If @code{numbers-list} is empty, do nothing. But if it has content,
11819 construct a list combining the square of the first number in the list
11820 with the result of the recursive call.
11822 (The example follows the pattern exactly: @code{nil} is returned if
11823 the numbers' list is empty. In practice, you would write the
11824 conditional so it carries out the action when the numbers' list is not
11827 The @code{print-elements-recursively} function (@pxref{Recursion with
11828 list, , Recursion with a List}) is another example of an @code{every}
11829 pattern, except in this case, rather than bring the results together
11830 using @code{cons}, we print each element of output.
11833 The @code{print-elements-recursively} function looks like this:
11837 (setq animals '(gazelle giraffe lion tiger))
11841 (defun print-elements-recursively (list)
11842 "Print each element of LIST on a line of its own.
11844 (when list ; @r{do-again-test}
11845 (print (car list)) ; @r{body}
11846 (print-elements-recursively ; @r{recursive call}
11847 (cdr list)))) ; @r{next-step-expression}
11849 (print-elements-recursively animals)
11854 The pattern for @code{print-elements-recursively} is:
11858 When the list is empty, do nothing.
11860 But when the list has at least one element,
11863 act on the beginning of the list (the @sc{car} of the list),
11865 and make a recursive call on the rest (the @sc{cdr}) of the list.
11870 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11871 @cindex Accumulate, type of recursive pattern
11872 @cindex Recursive pattern: accumulate
11874 Another recursive pattern is called the @code{accumulate} pattern. In
11875 the @code{accumulate} recursive pattern, an action is performed on
11876 every element of a list and the result of that action is accumulated
11877 with the results of performing the action on the other elements.
11879 This is very like the `every' pattern using @code{cons}, except that
11880 @code{cons} is not used, but some other combiner.
11887 If a list be empty, return zero or some other constant.
11889 Else, act on the beginning of the list (the @sc{car} of the list),
11892 and combine that acted-on element, using @code{+} or
11893 some other combining function, with
11895 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11900 Here is an example:
11904 (defun add-elements (numbers-list)
11905 "Add the elements of NUMBERS-LIST together."
11906 (if (not numbers-list)
11908 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
11912 (add-elements '(1 2 3 4))
11917 @xref{Files List, , Making a List of Files}, for an example of the
11918 accumulate pattern.
11921 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
11922 @cindex Keep, type of recursive pattern
11923 @cindex Recursive pattern: keep
11925 A third recursive pattern is called the @code{keep} pattern.
11926 In the @code{keep} recursive pattern, each element of a list is tested;
11927 the element is acted on and the results are kept only if the element
11930 Again, this is very like the `every' pattern, except the element is
11931 skipped unless it meets a criterion.
11934 The pattern has three parts:
11938 If a list be empty, return @code{nil}.
11940 Else, if the beginning of the list (the @sc{car} of the list) passes
11944 act on that element and combine it, using @code{cons} with
11946 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11949 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
11953 skip on that element,
11955 and, recursively call the function on the rest (the @sc{cdr}) of the list.
11960 Here is an example that uses @code{cond}:
11964 (defun keep-three-letter-words (word-list)
11965 "Keep three letter words in WORD-LIST."
11967 ;; First do-again-test: stop-condition
11968 ((not word-list) nil)
11970 ;; Second do-again-test: when to act
11971 ((eq 3 (length (symbol-name (car word-list))))
11972 ;; combine acted-on element with recursive call on shorter list
11973 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
11975 ;; Third do-again-test: when to skip element;
11976 ;; recursively call shorter list with next-step expression
11977 (t (keep-three-letter-words (cdr word-list)))))
11981 (keep-three-letter-words '(one two three four five six))
11982 @result{} (one two six)
11986 It goes without saying that you need not use @code{nil} as the test for
11987 when to stop; and you can, of course, combine these patterns.
11990 @subsection Recursion without Deferments
11991 @cindex Deferment in recursion
11992 @cindex Recursion without Deferments
11994 Let's consider again what happens with the @code{triangle-recursively}
11995 function. We will find that the intermediate calculations are
11996 deferred until all can be done.
11999 Here is the function definition:
12003 (defun triangle-recursively (number)
12004 "Return the sum of the numbers 1 through NUMBER inclusive.
12006 (if (= number 1) ; @r{do-again-test}
12008 (+ number ; @r{else-part}
12009 (triangle-recursively ; @r{recursive call}
12010 (1- number))))) ; @r{next-step-expression}
12014 What happens when we call this function with a argument of 7?
12016 The first instance of the @code{triangle-recursively} function adds
12017 the number 7 to the value returned by a second instance of
12018 @code{triangle-recursively}, an instance that has been passed an
12019 argument of 6. That is to say, the first calculation is:
12022 (+ 7 (triangle-recursively 6))
12026 The first instance of @code{triangle-recursively}---you may want to
12027 think of it as a little robot---cannot complete its job. It must hand
12028 off the calculation for @code{(triangle-recursively 6)} to a second
12029 instance of the program, to a second robot. This second individual is
12030 completely different from the first one; it is, in the jargon, a
12031 `different instantiation'. Or, put another way, it is a different
12032 robot. It is the same model as the first; it calculates triangle
12033 numbers recursively; but it has a different serial number.
12035 And what does @code{(triangle-recursively 6)} return? It returns the
12036 number 6 added to the value returned by evaluating
12037 @code{triangle-recursively} with an argument of 5. Using the robot
12038 metaphor, it asks yet another robot to help it.
12044 (+ 7 6 (triangle-recursively 5))
12048 And what happens next?
12051 (+ 7 6 5 (triangle-recursively 4))
12054 Each time @code{triangle-recursively} is called, except for the last
12055 time, it creates another instance of the program---another robot---and
12056 asks it to make a calculation.
12059 Eventually, the full addition is set up and performed:
12065 This design for the function defers the calculation of the first step
12066 until the second can be done, and defers that until the third can be
12067 done, and so on. Each deferment means the computer must remember what
12068 is being waited on. This is not a problem when there are only a few
12069 steps, as in this example. But it can be a problem when there are
12072 @node No deferment solution
12073 @subsection No Deferment Solution
12074 @cindex No deferment solution
12075 @cindex Defermentless solution
12076 @cindex Solution without deferment
12078 The solution to the problem of deferred operations is to write in a
12079 manner that does not defer operations@footnote{The phrase @dfn{tail
12080 recursive} is used to describe such a process, one that uses
12081 `constant space'.}. This requires
12082 writing to a different pattern, often one that involves writing two
12083 function definitions, an `initialization' function and a `helper'
12086 The `initialization' function sets up the job; the `helper' function
12090 Here are the two function definitions for adding up numbers. They are
12091 so simple, I find them hard to understand.
12095 (defun triangle-initialization (number)
12096 "Return the sum of the numbers 1 through NUMBER inclusive.
12097 This is the `initialization' component of a two function
12098 duo that uses recursion."
12099 (triangle-recursive-helper 0 0 number))
12105 (defun triangle-recursive-helper (sum counter number)
12106 "Return SUM, using COUNTER, through NUMBER inclusive.
12107 This is the `helper' component of a two function duo
12108 that uses recursion."
12109 (if (> counter number)
12111 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12112 (1+ counter) ; @r{counter}
12113 number))) ; @r{number}
12118 Install both function definitions by evaluating them, then call
12119 @code{triangle-initialization} with 2 rows:
12123 (triangle-initialization 2)
12128 The `initialization' function calls the first instance of the `helper'
12129 function with three arguments: zero, zero, and a number which is the
12130 number of rows in the triangle.
12132 The first two arguments passed to the `helper' function are
12133 initialization values. These values are changed when
12134 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12135 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12136 process that is iterative in a procedure that is recursive. The
12137 process is called iterative because the computer need only record the
12138 three values, @code{sum}, @code{counter}, and @code{number}; the
12139 procedure is recursive because the function `calls itself'. On the
12140 other hand, both the process and the procedure used by
12141 @code{triangle-recursively} are called recursive. The word
12142 `recursive' has different meanings in the two contexts.}
12144 Let's see what happens when we have a triangle that has one row. (This
12145 triangle will have one pebble in it!)
12148 @code{triangle-initialization} will call its helper with
12149 the arguments @w{@code{0 0 1}}. That function will run the conditional
12150 test whether @code{(> counter number)}:
12158 and find that the result is false, so it will invoke
12159 the else-part of the @code{if} clause:
12163 (triangle-recursive-helper
12164 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12165 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12166 number) ; @r{number stays the same}
12172 which will first compute:
12176 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12177 (1+ 0) ; @r{counter}
12181 (triangle-recursive-helper 0 1 1)
12185 Again, @code{(> counter number)} will be false, so again, the Lisp
12186 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12187 new instance with new arguments.
12190 This new instance will be;
12194 (triangle-recursive-helper
12195 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12196 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12197 number) ; @r{number stays the same}
12201 (triangle-recursive-helper 1 2 1)
12205 In this case, the @code{(> counter number)} test will be true! So the
12206 instance will return the value of the sum, which will be 1, as
12209 Now, let's pass @code{triangle-initialization} an argument
12210 of 2, to find out how many pebbles there are in a triangle with two rows.
12212 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12215 In stages, the instances called will be:
12219 @r{sum counter number}
12220 (triangle-recursive-helper 0 1 2)
12222 (triangle-recursive-helper 1 2 2)
12224 (triangle-recursive-helper 3 3 2)
12228 When the last instance is called, the @code{(> counter number)} test
12229 will be true, so the instance will return the value of @code{sum},
12232 This kind of pattern helps when you are writing functions that can use
12233 many resources in a computer.
12236 @node Looping exercise
12237 @section Looping Exercise
12241 Write a function similar to @code{triangle} in which each row has a
12242 value which is the square of the row number. Use a @code{while} loop.
12245 Write a function similar to @code{triangle} that multiplies instead of
12249 Rewrite these two functions recursively. Rewrite these functions
12252 @c comma in printed title causes problem in Info cross reference
12254 Write a function for Texinfo mode that creates an index entry at the
12255 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12256 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12257 written in Texinfo.)
12259 Many of the functions you will need are described in two of the
12260 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12261 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12262 @code{forward-paragraph} to put the index entry at the beginning of
12263 the paragraph, you will have to use @w{@kbd{C-h f}}
12264 (@code{describe-function}) to find out how to make the command go
12267 For more information, see
12269 @ref{Indicating, , Indicating Definitions, texinfo}.
12272 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12273 a Texinfo manual in the current directory. Or, if you are on the
12275 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12278 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12279 Documentation Format}.
12283 @node Regexp Search
12284 @chapter Regular Expression Searches
12285 @cindex Searches, illustrating
12286 @cindex Regular expression searches
12287 @cindex Patterns, searching for
12288 @cindex Motion by sentence and paragraph
12289 @cindex Sentences, movement by
12290 @cindex Paragraphs, movement by
12292 Regular expression searches are used extensively in GNU Emacs. The
12293 two functions, @code{forward-sentence} and @code{forward-paragraph},
12294 illustrate these searches well. They use regular expressions to find
12295 where to move point. The phrase `regular expression' is often written
12298 Regular expression searches are described in @ref{Regexp Search, ,
12299 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12300 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12301 Manual}. In writing this chapter, I am presuming that you have at
12302 least a mild acquaintance with them. The major point to remember is
12303 that regular expressions permit you to search for patterns as well as
12304 for literal strings of characters. For example, the code in
12305 @code{forward-sentence} searches for the pattern of possible
12306 characters that could mark the end of a sentence, and moves point to
12309 Before looking at the code for the @code{forward-sentence} function, it
12310 is worth considering what the pattern that marks the end of a sentence
12311 must be. The pattern is discussed in the next section; following that
12312 is a description of the regular expression search function,
12313 @code{re-search-forward}. The @code{forward-sentence} function
12314 is described in the section following. Finally, the
12315 @code{forward-paragraph} function is described in the last section of
12316 this chapter. @code{forward-paragraph} is a complex function that
12317 introduces several new features.
12320 * sentence-end:: The regular expression for @code{sentence-end}.
12321 * re-search-forward:: Very similar to @code{search-forward}.
12322 * forward-sentence:: A straightforward example of regexp search.
12323 * forward-paragraph:: A somewhat complex example.
12324 * etags:: How to create your own @file{TAGS} table.
12326 * re-search Exercises::
12330 @section The Regular Expression for @code{sentence-end}
12331 @findex sentence-end
12333 The symbol @code{sentence-end} is bound to the pattern that marks the
12334 end of a sentence. What should this regular expression be?
12336 Clearly, a sentence may be ended by a period, a question mark, or an
12337 exclamation mark. Indeed, in English, only clauses that end with one
12338 of those three characters should be considered the end of a sentence.
12339 This means that the pattern should include the character set:
12345 However, we do not want @code{forward-sentence} merely to jump to a
12346 period, a question mark, or an exclamation mark, because such a character
12347 might be used in the middle of a sentence. A period, for example, is
12348 used after abbreviations. So other information is needed.
12350 According to convention, you type two spaces after every sentence, but
12351 only one space after a period, a question mark, or an exclamation mark in
12352 the body of a sentence. So a period, a question mark, or an exclamation
12353 mark followed by two spaces is a good indicator of an end of sentence.
12354 However, in a file, the two spaces may instead be a tab or the end of a
12355 line. This means that the regular expression should include these three
12356 items as alternatives.
12359 This group of alternatives will look like this:
12370 Here, @samp{$} indicates the end of the line, and I have pointed out
12371 where the tab and two spaces are inserted in the expression. Both are
12372 inserted by putting the actual characters into the expression.
12374 Two backslashes, @samp{\\}, are required before the parentheses and
12375 vertical bars: the first backslash quotes the following backslash in
12376 Emacs; and the second indicates that the following character, the
12377 parenthesis or the vertical bar, is special.
12380 Also, a sentence may be followed by one or more carriage returns, like
12391 Like tabs and spaces, a carriage return is inserted into a regular
12392 expression by inserting it literally. The asterisk indicates that the
12393 @key{RET} is repeated zero or more times.
12395 But a sentence end does not consist only of a period, a question mark or
12396 an exclamation mark followed by appropriate space: a closing quotation
12397 mark or a closing brace of some kind may precede the space. Indeed more
12398 than one such mark or brace may precede the space. These require a
12399 expression that looks like this:
12405 In this expression, the first @samp{]} is the first character in the
12406 expression; the second character is @samp{"}, which is preceded by a
12407 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12408 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12410 All this suggests what the regular expression pattern for matching the
12411 end of a sentence should be; and, indeed, if we evaluate
12412 @code{sentence-end} we find that it returns the following value:
12417 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12423 (Well, not in GNU Emacs 22; that is because of an effort to make the
12424 process simpler and to handle more glyphs and languages. When the
12425 value of @code{sentence-end} is @code{nil}, then use the value defined
12426 by the function @code{sentence-end}. (Here is a use of the difference
12427 between a value and a function in Emacs Lisp.) The function returns a
12428 value constructed from the variables @code{sentence-end-base},
12429 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12430 and @code{sentence-end-without-space}. The critical variable is
12431 @code{sentence-end-base}; its global value is similar to the one
12432 described above but it also contains two additional quotation marks.
12433 These have differing degrees of curliness. The
12434 @code{sentence-end-without-period} variable, when true, tells Emacs
12435 that a sentence may end without a period, such as text in Thai.)
12439 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12440 literally in the pattern.)
12442 This regular expression can be deciphered as follows:
12446 The first part of the pattern is the three characters, a period, a question
12447 mark and an exclamation mark, within square brackets. The pattern must
12448 begin with one or other of these characters.
12451 The second part of the pattern is the group of closing braces and
12452 quotation marks, which can appear zero or more times. These may follow
12453 the period, question mark or exclamation mark. In a regular expression,
12454 the backslash, @samp{\}, followed by the double quotation mark,
12455 @samp{"}, indicates the class of string-quote characters. Usually, the
12456 double quotation mark is the only character in this class. The
12457 asterisk, @samp{*}, indicates that the items in the previous group (the
12458 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12461 @item \\($\\| \\| \\)
12462 The third part of the pattern is one or other of: either the end of a
12463 line, or two blank spaces, or a tab. The double back-slashes are used
12464 to prevent Emacs from reading the parentheses and vertical bars as part
12465 of the search pattern; the parentheses are used to mark the group and
12466 the vertical bars are used to indicated that the patterns to either side
12467 of them are alternatives. The dollar sign is used to indicate the end
12468 of a line and both the two spaces and the tab are each inserted as is to
12469 indicate what they are.
12472 Finally, the last part of the pattern indicates that the end of the line
12473 or the whitespace following the period, question mark or exclamation
12474 mark may, but need not, be followed by one or more carriage returns. In
12475 the pattern, the carriage return is inserted as an actual carriage
12476 return between square brackets but here it is shown as @key{RET}.
12480 @node re-search-forward
12481 @section The @code{re-search-forward} Function
12482 @findex re-search-forward
12484 The @code{re-search-forward} function is very like the
12485 @code{search-forward} function. (@xref{search-forward, , The
12486 @code{search-forward} Function}.)
12488 @code{re-search-forward} searches for a regular expression. If the
12489 search is successful, it leaves point immediately after the last
12490 character in the target. If the search is backwards, it leaves point
12491 just before the first character in the target. You may tell
12492 @code{re-search-forward} to return @code{t} for true. (Moving point
12493 is therefore a `side effect'.)
12495 Like @code{search-forward}, the @code{re-search-forward} function takes
12500 The first argument is the regular expression that the function searches
12501 for. The regular expression will be a string between quotation marks.
12504 The optional second argument limits how far the function will search; it is a
12505 bound, which is specified as a position in the buffer.
12508 The optional third argument specifies how the function responds to
12509 failure: @code{nil} as the third argument causes the function to
12510 signal an error (and print a message) when the search fails; any other
12511 value causes it to return @code{nil} if the search fails and @code{t}
12512 if the search succeeds.
12515 The optional fourth argument is the repeat count. A negative repeat
12516 count causes @code{re-search-forward} to search backwards.
12520 The template for @code{re-search-forward} looks like this:
12524 (re-search-forward "@var{regular-expression}"
12525 @var{limit-of-search}
12526 @var{what-to-do-if-search-fails}
12527 @var{repeat-count})
12531 The second, third, and fourth arguments are optional. However, if you
12532 want to pass a value to either or both of the last two arguments, you
12533 must also pass a value to all the preceding arguments. Otherwise, the
12534 Lisp interpreter will mistake which argument you are passing the value
12538 In the @code{forward-sentence} function, the regular expression will be
12539 the value of the variable @code{sentence-end}. In simple form, that is:
12543 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12549 The limit of the search will be the end of the paragraph (since a
12550 sentence cannot go beyond a paragraph). If the search fails, the
12551 function will return @code{nil}; and the repeat count will be provided
12552 by the argument to the @code{forward-sentence} function.
12554 @node forward-sentence
12555 @section @code{forward-sentence}
12556 @findex forward-sentence
12558 The command to move the cursor forward a sentence is a straightforward
12559 illustration of how to use regular expression searches in Emacs Lisp.
12560 Indeed, the function looks longer and more complicated than it is; this
12561 is because the function is designed to go backwards as well as forwards;
12562 and, optionally, over more than one sentence. The function is usually
12563 bound to the key command @kbd{M-e}.
12566 * Complete forward-sentence::
12567 * fwd-sentence while loops:: Two @code{while} loops.
12568 * fwd-sentence re-search:: A regular expression search.
12572 @node Complete forward-sentence
12573 @unnumberedsubsec Complete @code{forward-sentence} function definition
12577 Here is the code for @code{forward-sentence}:
12582 (defun forward-sentence (&optional arg)
12583 "Move forward to next `sentence-end'. With argument, repeat.
12584 With negative argument, move backward repeatedly to `sentence-beginning'.
12586 The variable `sentence-end' is a regular expression that matches ends of
12587 sentences. Also, every paragraph boundary terminates sentences as well."
12591 (or arg (setq arg 1))
12592 (let ((opoint (point))
12593 (sentence-end (sentence-end)))
12595 (let ((pos (point))
12596 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12597 (if (and (re-search-backward sentence-end par-beg t)
12598 (or (< (match-end 0) pos)
12599 (re-search-backward sentence-end par-beg t)))
12600 (goto-char (match-end 0))
12601 (goto-char par-beg)))
12602 (setq arg (1+ arg)))
12606 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12607 (if (re-search-forward sentence-end par-end t)
12608 (skip-chars-backward " \t\n")
12609 (goto-char par-end)))
12610 (setq arg (1- arg)))
12611 (constrain-to-field nil opoint t)))
12619 (defun forward-sentence (&optional arg)
12620 "Move forward to next sentence-end. With argument, repeat.
12621 With negative argument, move backward repeatedly to sentence-beginning.
12622 Sentence ends are identified by the value of sentence-end
12623 treated as a regular expression. Also, every paragraph boundary
12624 terminates sentences as well."
12628 (or arg (setq arg 1))
12631 (save-excursion (start-of-paragraph-text) (point))))
12632 (if (re-search-backward
12633 (concat sentence-end "[^ \t\n]") par-beg t)
12634 (goto-char (1- (match-end 0)))
12635 (goto-char par-beg)))
12636 (setq arg (1+ arg)))
12639 (save-excursion (end-of-paragraph-text) (point))))
12640 (if (re-search-forward sentence-end par-end t)
12641 (skip-chars-backward " \t\n")
12642 (goto-char par-end)))
12643 (setq arg (1- arg))))
12648 The function looks long at first sight and it is best to look at its
12649 skeleton first, and then its muscle. The way to see the skeleton is to
12650 look at the expressions that start in the left-most columns:
12654 (defun forward-sentence (&optional arg)
12655 "@var{documentation}@dots{}"
12657 (or arg (setq arg 1))
12658 (let ((opoint (point)) (sentence-end (sentence-end)))
12660 (let ((pos (point))
12661 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12662 @var{rest-of-body-of-while-loop-when-going-backwards}
12664 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12665 @var{rest-of-body-of-while-loop-when-going-forwards}
12666 @var{handle-forms-and-equivalent}
12670 This looks much simpler! The function definition consists of
12671 documentation, an @code{interactive} expression, an @code{or}
12672 expression, a @code{let} expression, and @code{while} loops.
12674 Let's look at each of these parts in turn.
12676 We note that the documentation is thorough and understandable.
12678 The function has an @code{interactive "p"} declaration. This means
12679 that the processed prefix argument, if any, is passed to the
12680 function as its argument. (This will be a number.) If the function
12681 is not passed an argument (it is optional) then the argument
12682 @code{arg} will be bound to 1.
12684 When @code{forward-sentence} is called non-interactively without an
12685 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12686 handles this. What it does is either leave the value of @code{arg} as
12687 it is, but only if @code{arg} is bound to a value; or it sets the
12688 value of @code{arg} to 1, in the case when @code{arg} is bound to
12691 Next is a @code{let}. That specifies the values of two local
12692 variables, @code{point} and @code{sentence-end}. The local value of
12693 point, from before the search, is used in the
12694 @code{constrain-to-field} function which handles forms and
12695 equivalents. The @code{sentence-end} variable is set by the
12696 @code{sentence-end} function.
12698 @node fwd-sentence while loops
12699 @unnumberedsubsec The @code{while} loops
12701 Two @code{while} loops follow. The first @code{while} has a
12702 true-or-false-test that tests true if the prefix argument for
12703 @code{forward-sentence} is a negative number. This is for going
12704 backwards. The body of this loop is similar to the body of the second
12705 @code{while} clause, but it is not exactly the same. We will skip
12706 this @code{while} loop and concentrate on the second @code{while}
12710 The second @code{while} loop is for moving point forward. Its skeleton
12715 (while (> arg 0) ; @r{true-or-false-test}
12717 (if (@var{true-or-false-test})
12720 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12724 The @code{while} loop is of the decrementing kind.
12725 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12726 has a true-or-false-test that tests true so long as the counter (in
12727 this case, the variable @code{arg}) is greater than zero; and it has a
12728 decrementer that subtracts 1 from the value of the counter every time
12731 If no prefix argument is given to @code{forward-sentence}, which is
12732 the most common way the command is used, this @code{while} loop will
12733 run once, since the value of @code{arg} will be 1.
12735 The body of the @code{while} loop consists of a @code{let} expression,
12736 which creates and binds a local variable, and has, as its body, an
12737 @code{if} expression.
12740 The body of the @code{while} loop looks like this:
12745 (save-excursion (end-of-paragraph-text) (point))))
12746 (if (re-search-forward sentence-end par-end t)
12747 (skip-chars-backward " \t\n")
12748 (goto-char par-end)))
12752 The @code{let} expression creates and binds the local variable
12753 @code{par-end}. As we shall see, this local variable is designed to
12754 provide a bound or limit to the regular expression search. If the
12755 search fails to find a proper sentence ending in the paragraph, it will
12756 stop on reaching the end of the paragraph.
12758 But first, let us examine how @code{par-end} is bound to the value of
12759 the end of the paragraph. What happens is that the @code{let} sets the
12760 value of @code{par-end} to the value returned when the Lisp interpreter
12761 evaluates the expression
12765 (save-excursion (end-of-paragraph-text) (point))
12770 In this expression, @code{(end-of-paragraph-text)} moves point to the
12771 end of the paragraph, @code{(point)} returns the value of point, and then
12772 @code{save-excursion} restores point to its original position. Thus,
12773 the @code{let} binds @code{par-end} to the value returned by the
12774 @code{save-excursion} expression, which is the position of the end of
12775 the paragraph. (The @code{end-of-paragraph-text} function uses
12776 @code{forward-paragraph}, which we will discuss shortly.)
12779 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12780 expression that looks like this:
12784 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12785 (skip-chars-backward " \t\n") ; @r{then-part}
12786 (goto-char par-end))) ; @r{else-part}
12790 The @code{if} tests whether its first argument is true and if so,
12791 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12792 evaluates the else-part. The true-or-false-test of the @code{if}
12793 expression is the regular expression search.
12795 It may seem odd to have what looks like the `real work' of
12796 the @code{forward-sentence} function buried here, but this is a common
12797 way this kind of operation is carried out in Lisp.
12799 @node fwd-sentence re-search
12800 @unnumberedsubsec The regular expression search
12802 The @code{re-search-forward} function searches for the end of the
12803 sentence, that is, for the pattern defined by the @code{sentence-end}
12804 regular expression. If the pattern is found---if the end of the sentence is
12805 found---then the @code{re-search-forward} function does two things:
12809 The @code{re-search-forward} function carries out a side effect, which
12810 is to move point to the end of the occurrence found.
12813 The @code{re-search-forward} function returns a value of true. This is
12814 the value received by the @code{if}, and means that the search was
12819 The side effect, the movement of point, is completed before the
12820 @code{if} function is handed the value returned by the successful
12821 conclusion of the search.
12823 When the @code{if} function receives the value of true from a successful
12824 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12825 which is the expression @code{(skip-chars-backward " \t\n")}. This
12826 expression moves backwards over any blank spaces, tabs or carriage
12827 returns until a printed character is found and then leaves point after
12828 the character. Since point has already been moved to the end of the
12829 pattern that marks the end of the sentence, this action leaves point
12830 right after the closing printed character of the sentence, which is
12833 On the other hand, if the @code{re-search-forward} function fails to
12834 find a pattern marking the end of the sentence, the function returns
12835 false. The false then causes the @code{if} to evaluate its third
12836 argument, which is @code{(goto-char par-end)}: it moves point to the
12837 end of the paragraph.
12839 (And if the text is in a form or equivalent, and point may not move
12840 fully, then the @code{constrain-to-field} function comes into play.)
12842 Regular expression searches are exceptionally useful and the pattern
12843 illustrated by @code{re-search-forward}, in which the search is the
12844 test of an @code{if} expression, is handy. You will see or write code
12845 incorporating this pattern often.
12847 @node forward-paragraph
12848 @section @code{forward-paragraph}: a Goldmine of Functions
12849 @findex forward-paragraph
12853 (defun forward-paragraph (&optional arg)
12854 "Move forward to end of paragraph.
12855 With argument ARG, do it ARG times;
12856 a negative argument ARG = -N means move backward N paragraphs.
12858 A line which `paragraph-start' matches either separates paragraphs
12859 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12860 A paragraph end is the beginning of a line which is not part of the paragraph
12861 to which the end of the previous line belongs, or the end of the buffer.
12862 Returns the count of paragraphs left to move."
12864 (or arg (setq arg 1))
12865 (let* ((opoint (point))
12866 (fill-prefix-regexp
12867 (and fill-prefix (not (equal fill-prefix ""))
12868 (not paragraph-ignore-fill-prefix)
12869 (regexp-quote fill-prefix)))
12870 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12871 ;; These regexps shouldn't be anchored, because we look for them
12872 ;; starting at the left-margin. This allows paragraph commands to
12873 ;; work normally with indented text.
12874 ;; This hack will not find problem cases like "whatever\\|^something".
12875 (parstart (if (and (not (equal "" paragraph-start))
12876 (equal ?^ (aref paragraph-start 0)))
12877 (substring paragraph-start 1)
12879 (parsep (if (and (not (equal "" paragraph-separate))
12880 (equal ?^ (aref paragraph-separate 0)))
12881 (substring paragraph-separate 1)
12882 paragraph-separate))
12884 (if fill-prefix-regexp
12885 (concat parsep "\\|"
12886 fill-prefix-regexp "[ \t]*$")
12888 ;; This is used for searching.
12889 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
12891 (while (and (< arg 0) (not (bobp)))
12892 (if (and (not (looking-at parsep))
12893 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
12894 (looking-at parsep))
12895 (setq arg (1+ arg))
12896 (setq start (point))
12897 ;; Move back over paragraph-separating lines.
12898 (forward-char -1) (beginning-of-line)
12899 (while (and (not (bobp))
12900 (progn (move-to-left-margin)
12901 (looking-at parsep)))
12905 (setq arg (1+ arg))
12906 ;; Go to end of the previous (non-separating) line.
12908 ;; Search back for line that starts or separates paragraphs.
12909 (if (if fill-prefix-regexp
12910 ;; There is a fill prefix; it overrides parstart.
12911 (let (multiple-lines)
12912 (while (and (progn (beginning-of-line) (not (bobp)))
12913 (progn (move-to-left-margin)
12914 (not (looking-at parsep)))
12915 (looking-at fill-prefix-regexp))
12916 (unless (= (point) start)
12917 (setq multiple-lines t))
12919 (move-to-left-margin)
12920 ;; This deleted code caused a long hanging-indent line
12921 ;; not to be filled together with the following lines.
12922 ;; ;; Don't move back over a line before the paragraph
12923 ;; ;; which doesn't start with fill-prefix
12924 ;; ;; unless that is the only line we've moved over.
12925 ;; (and (not (looking-at fill-prefix-regexp))
12927 ;; (forward-line 1))
12929 (while (and (re-search-backward sp-parstart nil 1)
12930 (setq found-start t)
12931 ;; Found a candidate, but need to check if it is a
12933 (progn (setq start (point))
12934 (move-to-left-margin)
12935 (not (looking-at parsep)))
12936 (not (and (looking-at parstart)
12937 (or (not use-hard-newlines)
12940 (1- start) 'hard)))))
12941 (setq found-start nil)
12946 ;; Move forward over paragraph separators.
12947 ;; We know this cannot reach the place we started
12948 ;; because we know we moved back over a non-separator.
12949 (while (and (not (eobp))
12950 (progn (move-to-left-margin)
12951 (looking-at parsep)))
12953 ;; If line before paragraph is just margin, back up to there.
12955 (if (> (current-column) (current-left-margin))
12957 (skip-chars-backward " \t")
12959 (forward-line 1))))
12960 ;; No starter or separator line => use buffer beg.
12961 (goto-char (point-min))))))
12963 (while (and (> arg 0) (not (eobp)))
12964 ;; Move forward over separator lines...
12965 (while (and (not (eobp))
12966 (progn (move-to-left-margin) (not (eobp)))
12967 (looking-at parsep))
12969 (unless (eobp) (setq arg (1- arg)))
12970 ;; ... and one more line.
12972 (if fill-prefix-regexp
12973 ;; There is a fill prefix; it overrides parstart.
12974 (while (and (not (eobp))
12975 (progn (move-to-left-margin) (not (eobp)))
12976 (not (looking-at parsep))
12977 (looking-at fill-prefix-regexp))
12979 (while (and (re-search-forward sp-parstart nil 1)
12980 (progn (setq start (match-beginning 0))
12983 (progn (move-to-left-margin)
12984 (not (looking-at parsep)))
12985 (or (not (looking-at parstart))
12986 (and use-hard-newlines
12987 (not (get-text-property (1- start) 'hard)))))
12989 (if (< (point) (point-max))
12990 (goto-char start))))
12991 (constrain-to-field nil opoint t)
12992 ;; Return the number of steps that could not be done.
12996 The @code{forward-paragraph} function moves point forward to the end
12997 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
12998 number of functions that are important in themselves, including
12999 @code{let*}, @code{match-beginning}, and @code{looking-at}.
13001 The function definition for @code{forward-paragraph} is considerably
13002 longer than the function definition for @code{forward-sentence}
13003 because it works with a paragraph, each line of which may begin with a
13006 A fill prefix consists of a string of characters that are repeated at
13007 the beginning of each line. For example, in Lisp code, it is a
13008 convention to start each line of a paragraph-long comment with
13009 @samp{;;; }. In Text mode, four blank spaces make up another common
13010 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
13011 emacs, The GNU Emacs Manual}, for more information about fill
13014 The existence of a fill prefix means that in addition to being able to
13015 find the end of a paragraph whose lines begin on the left-most
13016 column, the @code{forward-paragraph} function must be able to find the
13017 end of a paragraph when all or many of the lines in the buffer begin
13018 with the fill prefix.
13020 Moreover, it is sometimes practical to ignore a fill prefix that
13021 exists, especially when blank lines separate paragraphs.
13022 This is an added complication.
13025 * forward-paragraph in brief:: Key parts of the function definition.
13026 * fwd-para let:: The @code{let*} expression.
13027 * fwd-para while:: The forward motion @code{while} loop.
13031 @node forward-paragraph in brief
13032 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13035 Rather than print all of the @code{forward-paragraph} function, we
13036 will only print parts of it. Read without preparation, the function
13040 In outline, the function looks like this:
13044 (defun forward-paragraph (&optional arg)
13045 "@var{documentation}@dots{}"
13047 (or arg (setq arg 1))
13050 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13052 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13057 The first parts of the function are routine: the function's argument
13058 list consists of one optional argument. Documentation follows.
13060 The lower case @samp{p} in the @code{interactive} declaration means
13061 that the processed prefix argument, if any, is passed to the function.
13062 This will be a number, and is the repeat count of how many paragraphs
13063 point will move. The @code{or} expression in the next line handles
13064 the common case when no argument is passed to the function, which occurs
13065 if the function is called from other code rather than interactively.
13066 This case was described earlier. (@xref{forward-sentence, The
13067 @code{forward-sentence} function}.) Now we reach the end of the
13068 familiar part of this function.
13071 @unnumberedsubsec The @code{let*} expression
13073 The next line of the @code{forward-paragraph} function begins a
13074 @code{let*} expression. This is a different than @code{let}. The
13075 symbol is @code{let*} not @code{let}.
13077 The @code{let*} special form is like @code{let} except that Emacs sets
13078 each variable in sequence, one after another, and variables in the
13079 latter part of the varlist can make use of the values to which Emacs
13080 set variables in the earlier part of the varlist.
13083 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13086 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13088 In the @code{let*} expression in this function, Emacs binds a total of
13089 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13090 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13091 @code{found-start}.
13093 The variable @code{parsep} appears twice, first, to remove instances
13094 of @samp{^}, and second, to handle fill prefixes.
13096 The variable @code{opoint} is just the value of @code{point}. As you
13097 can guess, it is used in a @code{constrain-to-field} expression, just
13098 as in @code{forward-sentence}.
13100 The variable @code{fill-prefix-regexp} is set to the value returned by
13101 evaluating the following list:
13106 (not (equal fill-prefix ""))
13107 (not paragraph-ignore-fill-prefix)
13108 (regexp-quote fill-prefix))
13113 This is an expression whose first element is the @code{and} special form.
13115 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13116 function}), the @code{and} special form evaluates each of its
13117 arguments until one of the arguments returns a value of @code{nil}, in
13118 which case the @code{and} expression returns @code{nil}; however, if
13119 none of the arguments returns a value of @code{nil}, the value
13120 resulting from evaluating the last argument is returned. (Since such
13121 a value is not @code{nil}, it is considered true in Lisp.) In other
13122 words, an @code{and} expression returns a true value only if all its
13123 arguments are true.
13126 In this case, the variable @code{fill-prefix-regexp} is bound to a
13127 non-@code{nil} value only if the following four expressions produce a
13128 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13129 @code{fill-prefix-regexp} is bound to @code{nil}.
13133 When this variable is evaluated, the value of the fill prefix, if any,
13134 is returned. If there is no fill prefix, this variable returns
13137 @item (not (equal fill-prefix "")
13138 This expression checks whether an existing fill prefix is an empty
13139 string, that is, a string with no characters in it. An empty string is
13140 not a useful fill prefix.
13142 @item (not paragraph-ignore-fill-prefix)
13143 This expression returns @code{nil} if the variable
13144 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13145 true value such as @code{t}.
13147 @item (regexp-quote fill-prefix)
13148 This is the last argument to the @code{and} special form. If all the
13149 arguments to the @code{and} are true, the value resulting from
13150 evaluating this expression will be returned by the @code{and} expression
13151 and bound to the variable @code{fill-prefix-regexp},
13154 @findex regexp-quote
13156 The result of evaluating this @code{and} expression successfully is that
13157 @code{fill-prefix-regexp} will be bound to the value of
13158 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13159 What @code{regexp-quote} does is read a string and return a regular
13160 expression that will exactly match the string and match nothing else.
13161 This means that @code{fill-prefix-regexp} will be set to a value that
13162 will exactly match the fill prefix if the fill prefix exists.
13163 Otherwise, the variable will be set to @code{nil}.
13165 The next two local variables in the @code{let*} expression are
13166 designed to remove instances of @samp{^} from @code{parstart} and
13167 @code{parsep}, the local variables which indicate the paragraph start
13168 and the paragraph separator. The next expression sets @code{parsep}
13169 again. That is to handle fill prefixes.
13171 This is the setting that requires the definition call @code{let*}
13172 rather than @code{let}. The true-or-false-test for the @code{if}
13173 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13174 @code{nil} or some other value.
13176 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13177 the else-part of the @code{if} expression and binds @code{parsep} to
13178 its local value. (@code{parsep} is a regular expression that matches
13179 what separates paragraphs.)
13181 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13182 the then-part of the @code{if} expression and binds @code{parsep} to a
13183 regular expression that includes the @code{fill-prefix-regexp} as part
13186 Specifically, @code{parsep} is set to the original value of the
13187 paragraph separate regular expression concatenated with an alternative
13188 expression that consists of the @code{fill-prefix-regexp} followed by
13189 optional whitespace to the end of the line. The whitespace is defined
13190 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13191 regexp as an alternative to @code{parsep}.
13193 According to a comment in the code, the next local variable,
13194 @code{sp-parstart}, is used for searching, and then the final two,
13195 @code{start} and @code{found-start}, are set to @code{nil}.
13197 Now we get into the body of the @code{let*}. The first part of the body
13198 of the @code{let*} deals with the case when the function is given a
13199 negative argument and is therefore moving backwards. We will skip this
13202 @node fwd-para while
13203 @unnumberedsubsec The forward motion @code{while} loop
13205 The second part of the body of the @code{let*} deals with forward
13206 motion. It is a @code{while} loop that repeats itself so long as the
13207 value of @code{arg} is greater than zero. In the most common use of
13208 the function, the value of the argument is 1, so the body of the
13209 @code{while} loop is evaluated exactly once, and the cursor moves
13210 forward one paragraph.
13213 (while (and (> arg 0) (not (eobp)))
13215 ;; Move forward over separator lines...
13216 (while (and (not (eobp))
13217 (progn (move-to-left-margin) (not (eobp)))
13218 (looking-at parsep))
13220 (unless (eobp) (setq arg (1- arg)))
13221 ;; ... and one more line.
13224 (if fill-prefix-regexp
13225 ;; There is a fill prefix; it overrides parstart.
13226 (while (and (not (eobp))
13227 (progn (move-to-left-margin) (not (eobp)))
13228 (not (looking-at parsep))
13229 (looking-at fill-prefix-regexp))
13232 (while (and (re-search-forward sp-parstart nil 1)
13233 (progn (setq start (match-beginning 0))
13236 (progn (move-to-left-margin)
13237 (not (looking-at parsep)))
13238 (or (not (looking-at parstart))
13239 (and use-hard-newlines
13240 (not (get-text-property (1- start) 'hard)))))
13243 (if (< (point) (point-max))
13244 (goto-char start))))
13247 This part handles three situations: when point is between paragraphs,
13248 when there is a fill prefix and when there is no fill prefix.
13251 The @code{while} loop looks like this:
13255 ;; @r{going forwards and not at the end of the buffer}
13256 (while (and (> arg 0) (not (eobp)))
13258 ;; @r{between paragraphs}
13259 ;; Move forward over separator lines...
13260 (while (and (not (eobp))
13261 (progn (move-to-left-margin) (not (eobp)))
13262 (looking-at parsep))
13264 ;; @r{This decrements the loop}
13265 (unless (eobp) (setq arg (1- arg)))
13266 ;; ... and one more line.
13271 (if fill-prefix-regexp
13272 ;; There is a fill prefix; it overrides parstart;
13273 ;; we go forward line by line
13274 (while (and (not (eobp))
13275 (progn (move-to-left-margin) (not (eobp)))
13276 (not (looking-at parsep))
13277 (looking-at fill-prefix-regexp))
13282 ;; There is no fill prefix;
13283 ;; we go forward character by character
13284 (while (and (re-search-forward sp-parstart nil 1)
13285 (progn (setq start (match-beginning 0))
13288 (progn (move-to-left-margin)
13289 (not (looking-at parsep)))
13290 (or (not (looking-at parstart))
13291 (and use-hard-newlines
13292 (not (get-text-property (1- start) 'hard)))))
13297 ;; and if there is no fill prefix and if we are not at the end,
13298 ;; go to whatever was found in the regular expression search
13300 (if (< (point) (point-max))
13301 (goto-char start))))
13306 We can see that this is a decrementing counter @code{while} loop,
13307 using the expression @code{(setq arg (1- arg))} as the decrementer.
13308 That expression is not far from the @code{while}, but is hidden in
13309 another Lisp macro, an @code{unless} macro. Unless we are at the end
13310 of the buffer---that is what the @code{eobp} function determines; it
13311 is an abbreviation of @samp{End Of Buffer P}---we decrease the value
13312 of @code{arg} by one.
13314 (If we are at the end of the buffer, we cannot go forward any more and
13315 the next loop of the @code{while} expression will test false since the
13316 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13317 function means exactly as you expect; it is another name for
13318 @code{null}, a function that returns true when its argument is false.)
13320 Interestingly, the loop count is not decremented until we leave the
13321 space between paragraphs, unless we come to the end of buffer or stop
13322 seeing the local value of the paragraph separator.
13324 That second @code{while} also has a @code{(move-to-left-margin)}
13325 expression. The function is self-explanatory. It is inside a
13326 @code{progn} expression and not the last element of its body, so it is
13327 only invoked for its side effect, which is to move point to the left
13328 margin of the current line.
13331 The @code{looking-at} function is also self-explanatory; it returns
13332 true if the text after point matches the regular expression given as
13335 The rest of the body of the loop looks difficult at first, but makes
13336 sense as you come to understand it.
13339 First consider what happens if there is a fill prefix:
13343 (if fill-prefix-regexp
13344 ;; There is a fill prefix; it overrides parstart;
13345 ;; we go forward line by line
13346 (while (and (not (eobp))
13347 (progn (move-to-left-margin) (not (eobp)))
13348 (not (looking-at parsep))
13349 (looking-at fill-prefix-regexp))
13355 This expression moves point forward line by line so long
13356 as four conditions are true:
13360 Point is not at the end of the buffer.
13363 We can move to the left margin of the text and are
13364 not at the end of the buffer.
13367 The text following point does not separate paragraphs.
13370 The pattern following point is the fill prefix regular expression.
13373 The last condition may be puzzling, until you remember that point was
13374 moved to the beginning of the line early in the @code{forward-paragraph}
13375 function. This means that if the text has a fill prefix, the
13376 @code{looking-at} function will see it.
13379 Consider what happens when there is no fill prefix.
13383 (while (and (re-search-forward sp-parstart nil 1)
13384 (progn (setq start (match-beginning 0))
13387 (progn (move-to-left-margin)
13388 (not (looking-at parsep)))
13389 (or (not (looking-at parstart))
13390 (and use-hard-newlines
13391 (not (get-text-property (1- start) 'hard)))))
13397 This @code{while} loop has us searching forward for
13398 @code{sp-parstart}, which is the combination of possible whitespace
13399 with a the local value of the start of a paragraph or of a paragraph
13400 separator. (The latter two are within an expression starting
13401 @code{\(?:} so that they are not referenced by the
13402 @code{match-beginning} function.)
13405 The two expressions,
13409 (setq start (match-beginning 0))
13415 mean go to the start of the text matched by the regular expression
13418 The @code{(match-beginning 0)} expression is new. It returns a number
13419 specifying the location of the start of the text that was matched by
13422 The @code{match-beginning} function is used here because of a
13423 characteristic of a forward search: a successful forward search,
13424 regardless of whether it is a plain search or a regular expression
13425 search, moves point to the end of the text that is found. In this
13426 case, a successful search moves point to the end of the pattern for
13427 @code{sp-parstart}.
13429 However, we want to put point at the end of the current paragraph, not
13430 somewhere else. Indeed, since the search possibly includes the
13431 paragraph separator, point may end up at the beginning of the next one
13432 unless we use an expression that includes @code{match-beginning}.
13434 @findex match-beginning
13435 When given an argument of 0, @code{match-beginning} returns the
13436 position that is the start of the text matched by the most recent
13437 search. In this case, the most recent search looks for
13438 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13439 the beginning position of that pattern, rather than the end position
13442 (Incidentally, when passed a positive number as an argument, the
13443 @code{match-beginning} function returns the location of point at that
13444 parenthesized expression in the last search unless that parenthesized
13445 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13446 appears here since the argument is 0.)
13449 The last expression when there is no fill prefix is
13453 (if (< (point) (point-max))
13454 (goto-char start))))
13459 This says that if there is no fill prefix and if we are not at the
13460 end, point should move to the beginning of whatever was found by the
13461 regular expression search for @code{sp-parstart}.
13463 The full definition for the @code{forward-paragraph} function not only
13464 includes code for going forwards, but also code for going backwards.
13466 If you are reading this inside of GNU Emacs and you want to see the
13467 whole function, you can type @kbd{C-h f} (@code{describe-function})
13468 and the name of the function. This gives you the function
13469 documentation and the name of the library containing the function's
13470 source. Place point over the name of the library and press the RET
13471 key; you will be taken directly to the source. (Be sure to install
13472 your sources! Without them, you are like a person who tries to drive
13473 a car with his eyes shut!)
13476 @section Create Your Own @file{TAGS} File
13478 @cindex @file{TAGS} file, create own
13480 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13481 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13482 name of the function when prompted for it. This is a good habit to
13483 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13484 to the source for a function, variable, or node. The function depends
13485 on tags tables to tell it where to go.
13487 If the @code{find-tag} function first asks you for the name of a
13488 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13489 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13490 @file{TAGS} file depends on how your copy of Emacs was installed. I
13491 just told you the location that provides both my C and my Emacs Lisp
13494 You can also create your own @file{TAGS} file for directories that
13497 You often need to build and install tags tables yourself. They are
13498 not built automatically. A tags table is called a @file{TAGS} file;
13499 the name is in upper case letters.
13501 You can create a @file{TAGS} file by calling the @code{etags} program
13502 that comes as a part of the Emacs distribution. Usually, @code{etags}
13503 is compiled and installed when Emacs is built. (@code{etags} is not
13504 an Emacs Lisp function or a part of Emacs; it is a C program.)
13507 To create a @file{TAGS} file, first switch to the directory in which
13508 you want to create the file. In Emacs you can do this with the
13509 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13510 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13511 compile command, with @w{@code{etags *.el}} as the command to execute
13514 M-x compile RET etags *.el RET
13518 to create a @file{TAGS} file for Emacs Lisp.
13520 For example, if you have a large number of files in your
13521 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13522 of which I load 12---you can create a @file{TAGS} file for the Emacs
13523 Lisp files in that directory.
13526 The @code{etags} program takes all the usual shell `wildcards'. For
13527 example, if you have two directories for which you want a single
13528 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13529 @file{../elisp/} is the second directory:
13532 M-x compile RET etags *.el ../elisp/*.el RET
13539 M-x compile RET etags --help RET
13543 to see a list of the options accepted by @code{etags} as well as a
13544 list of supported languages.
13546 The @code{etags} program handles more than 20 languages, including
13547 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13548 LaTeX, Pascal, Perl, PostScript, Python, TeX, Texinfo, makefiles, and
13549 most assemblers. The program has no switches for specifying the
13550 language; it recognizes the language in an input file according to its
13551 file name and contents.
13553 @file{etags} is very helpful when you are writing code yourself and
13554 want to refer back to functions you have already written. Just run
13555 @code{etags} again at intervals as you write new functions, so they
13556 become part of the @file{TAGS} file.
13558 If you think an appropriate @file{TAGS} file already exists for what
13559 you want, but do not know where it is, you can use the @code{locate}
13560 program to attempt to find it.
13562 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13563 for you the full path names of all your @file{TAGS} files. On my
13564 system, this command lists 34 @file{TAGS} files. On the other hand, a
13565 `plain vanilla' system I recently installed did not contain any
13568 If the tags table you want has been created, you can use the @code{M-x
13569 visit-tags-table} command to specify it. Otherwise, you will need to
13570 create the tag table yourself and then use @code{M-x
13573 @subsubheading Building Tags in the Emacs sources
13574 @cindex Building Tags in the Emacs sources
13575 @cindex Tags in the Emacs sources
13578 The GNU Emacs sources come with a @file{Makefile} that contains a
13579 sophisticated @code{etags} command that creates, collects, and merges
13580 tags tables from all over the Emacs sources and puts the information
13581 into one @file{TAGS} file in the @file{src/} directory. (The
13582 @file{src/} directory is below the top level of your Emacs directory.)
13585 To build this @file{TAGS} file, go to the top level of your Emacs
13586 source directory and run the compile command @code{make tags}:
13589 M-x compile RET make tags RET
13593 (The @code{make tags} command works well with the GNU Emacs sources,
13594 as well as with some other source packages.)
13596 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13599 @node Regexp Review
13602 Here is a brief summary of some recently introduced functions.
13606 Repeatedly evaluate the body of the expression so long as the first
13607 element of the body tests true. Then return @code{nil}. (The
13608 expression is evaluated only for its side effects.)
13617 (insert (format "foo is %d.\n" foo))
13618 (setq foo (1- foo))))
13620 @result{} foo is 2.
13627 (The @code{insert} function inserts its arguments at point; the
13628 @code{format} function returns a string formatted from its arguments
13629 the way @code{message} formats its arguments; @code{\n} produces a new
13632 @item re-search-forward
13633 Search for a pattern, and if the pattern is found, move point to rest
13637 Takes four arguments, like @code{search-forward}:
13641 A regular expression that specifies the pattern to search for.
13642 (Remember to put quotation marks around this argument!)
13645 Optionally, the limit of the search.
13648 Optionally, what to do if the search fails, return @code{nil} or an
13652 Optionally, how many times to repeat the search; if negative, the
13653 search goes backwards.
13657 Bind some variables locally to particular values,
13658 and then evaluate the remaining arguments, returning the value of the
13659 last one. While binding the local variables, use the local values of
13660 variables bound earlier, if any.
13669 (message "`bar' is %d." bar))
13670 @result{} `bar' is 21.
13674 @item match-beginning
13675 Return the position of the start of the text found by the last regular
13679 Return @code{t} for true if the text after point matches the argument,
13680 which should be a regular expression.
13683 Return @code{t} for true if point is at the end of the accessible part
13684 of a buffer. The end of the accessible part is the end of the buffer
13685 if the buffer is not narrowed; it is the end of the narrowed part if
13686 the buffer is narrowed.
13690 @node re-search Exercises
13691 @section Exercises with @code{re-search-forward}
13695 Write a function to search for a regular expression that matches two
13696 or more blank lines in sequence.
13699 Write a function to search for duplicated words, such as `the the'.
13700 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13701 Manual}, for information on how to write a regexp (a regular
13702 expression) to match a string that is composed of two identical
13703 halves. You can devise several regexps; some are better than others.
13704 The function I use is described in an appendix, along with several
13705 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13708 @node Counting Words
13709 @chapter Counting: Repetition and Regexps
13710 @cindex Repetition for word counting
13711 @cindex Regular expressions for word counting
13713 Repetition and regular expression searches are powerful tools that you
13714 often use when you write code in Emacs Lisp. This chapter illustrates
13715 the use of regular expression searches through the construction of
13716 word count commands using @code{while} loops and recursion.
13719 * Why Count Words::
13720 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
13721 * recursive-count-words:: Start with case of no words in region.
13722 * Counting Exercise::
13726 @node Why Count Words
13727 @unnumberedsec Counting words
13730 The standard Emacs distribution contains functions for counting the
13731 number of lines and words within a region.
13733 Certain types of writing ask you to count words. Thus, if you write
13734 an essay, you may be limited to 800 words; if you write a novel, you
13735 may discipline yourself to write 1000 words a day. It seems odd, but
13736 for a long time, Emacs lacked a word count command. Perhaps people used
13737 Emacs mostly for code or types of documentation that did not require
13738 word counts; or perhaps they restricted themselves to the operating
13739 system word count command, @code{wc}. Alternatively, people may have
13740 followed the publishers' convention and computed a word count by
13741 dividing the number of characters in a document by five.
13743 There are many ways to implement a command to count words. Here are
13744 some examples, which you may wish to compare with the standard Emacs
13745 command, @code{count-words-region}.
13747 @node @value{COUNT-WORDS}
13748 @section The @code{@value{COUNT-WORDS}} Function
13749 @findex @value{COUNT-WORDS}
13751 A word count command could count words in a line, paragraph, region,
13752 or buffer. What should the command cover? You could design the
13753 command to count the number of words in a complete buffer. However,
13754 the Emacs tradition encourages flexibility---you may want to count
13755 words in just a section, rather than all of a buffer. So it makes
13756 more sense to design the command to count the number of words in a
13757 region. Once you have a command to count words in a region, you can,
13758 if you wish, count words in a whole buffer by marking it with
13759 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13761 Clearly, counting words is a repetitive act: starting from the
13762 beginning of the region, you count the first word, then the second
13763 word, then the third word, and so on, until you reach the end of the
13764 region. This means that word counting is ideally suited to recursion
13765 or to a @code{while} loop.
13768 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
13769 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
13773 @node Design @value{COUNT-WORDS}
13774 @unnumberedsubsec Designing @code{@value{COUNT-WORDS}}
13777 First, we will implement the word count command with a @code{while}
13778 loop, then with recursion. The command will, of course, be
13782 The template for an interactive function definition is, as always:
13786 (defun @var{name-of-function} (@var{argument-list})
13787 "@var{documentation}@dots{}"
13788 (@var{interactive-expression}@dots{})
13793 What we need to do is fill in the slots.
13795 The name of the function should be self-explanatory and similar to the
13796 existing @code{count-lines-region} name. This makes the name easier
13797 to remember. @code{count-words-region} is the obvious choice. Since
13798 that name is now used for the standard Emacs command to count words, we
13799 will name our implementation @code{@value{COUNT-WORDS}}.
13801 The function counts words within a region. This means that the
13802 argument list must contain symbols that are bound to the two
13803 positions, the beginning and end of the region. These two positions
13804 can be called @samp{beginning} and @samp{end} respectively. The first
13805 line of the documentation should be a single sentence, since that is
13806 all that is printed as documentation by a command such as
13807 @code{apropos}. The interactive expression will be of the form
13808 @samp{(interactive "r")}, since that will cause Emacs to pass the
13809 beginning and end of the region to the function's argument list. All
13812 The body of the function needs to be written to do three tasks:
13813 first, to set up conditions under which the @code{while} loop can
13814 count words, second, to run the @code{while} loop, and third, to send
13815 a message to the user.
13817 When a user calls @code{@value{COUNT-WORDS}}, point may be at the
13818 beginning or the end of the region. However, the counting process
13819 must start at the beginning of the region. This means we will want
13820 to put point there if it is not already there. Executing
13821 @code{(goto-char beginning)} ensures this. Of course, we will want to
13822 return point to its expected position when the function finishes its
13823 work. For this reason, the body must be enclosed in a
13824 @code{save-excursion} expression.
13826 The central part of the body of the function consists of a
13827 @code{while} loop in which one expression jumps point forward word by
13828 word, and another expression counts those jumps. The true-or-false-test
13829 of the @code{while} loop should test true so long as point should jump
13830 forward, and false when point is at the end of the region.
13832 We could use @code{(forward-word 1)} as the expression for moving point
13833 forward word by word, but it is easier to see what Emacs identifies as a
13834 `word' if we use a regular expression search.
13836 A regular expression search that finds the pattern for which it is
13837 searching leaves point after the last character matched. This means
13838 that a succession of successful word searches will move point forward
13841 As a practical matter, we want the regular expression search to jump
13842 over whitespace and punctuation between words as well as over the
13843 words themselves. A regexp that refuses to jump over interword
13844 whitespace would never jump more than one word! This means that
13845 the regexp should include the whitespace and punctuation that follows
13846 a word, if any, as well as the word itself. (A word may end a buffer
13847 and not have any following whitespace or punctuation, so that part of
13848 the regexp must be optional.)
13850 Thus, what we want for the regexp is a pattern defining one or more
13851 word constituent characters followed, optionally, by one or more
13852 characters that are not word constituents. The regular expression for
13860 The buffer's syntax table determines which characters are and are not
13861 word constituents. For more information about syntax,
13862 @pxref{Syntax Tables, , Syntax Tables, elisp, The GNU Emacs Lisp
13866 The search expression looks like this:
13869 (re-search-forward "\\w+\\W*")
13873 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13874 single backslash has special meaning to the Emacs Lisp interpreter.
13875 It indicates that the following character is interpreted differently
13876 than usual. For example, the two characters, @samp{\n}, stand for
13877 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13878 backslashes in a row stand for an ordinary, `unspecial' backslash, so
13879 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13880 letter. So it discovers the letter is special.)
13882 We need a counter to count how many words there are; this variable
13883 must first be set to 0 and then incremented each time Emacs goes
13884 around the @code{while} loop. The incrementing expression is simply:
13887 (setq count (1+ count))
13890 Finally, we want to tell the user how many words there are in the
13891 region. The @code{message} function is intended for presenting this
13892 kind of information to the user. The message has to be phrased so
13893 that it reads properly regardless of how many words there are in the
13894 region: we don't want to say that ``there are 1 words in the region''.
13895 The conflict between singular and plural is ungrammatical. We can
13896 solve this problem by using a conditional expression that evaluates
13897 different messages depending on the number of words in the region.
13898 There are three possibilities: no words in the region, one word in the
13899 region, and more than one word. This means that the @code{cond}
13900 special form is appropriate.
13903 All this leads to the following function definition:
13907 ;;; @r{First version; has bugs!}
13908 (defun @value{COUNT-WORDS} (beginning end)
13909 "Print number of words in the region.
13910 Words are defined as at least one word-constituent
13911 character followed by at least one character that
13912 is not a word-constituent. The buffer's syntax
13913 table determines which characters these are."
13915 (message "Counting words in region ... ")
13919 ;;; @r{1. Set up appropriate conditions.}
13921 (goto-char beginning)
13926 ;;; @r{2. Run the} while @r{loop.}
13927 (while (< (point) end)
13928 (re-search-forward "\\w+\\W*")
13929 (setq count (1+ count)))
13933 ;;; @r{3. Send a message to the user.}
13934 (cond ((zerop count)
13936 "The region does NOT have any words."))
13939 "The region has 1 word."))
13942 "The region has %d words." count))))))
13947 As written, the function works, but not in all circumstances.
13949 @node Whitespace Bug
13950 @subsection The Whitespace Bug in @code{@value{COUNT-WORDS}}
13952 The @code{@value{COUNT-WORDS}} command described in the preceding
13953 section has two bugs, or rather, one bug with two manifestations.
13954 First, if you mark a region containing only whitespace in the middle
13955 of some text, the @code{@value{COUNT-WORDS}} command tells you that the
13956 region contains one word! Second, if you mark a region containing
13957 only whitespace at the end of the buffer or the accessible portion of
13958 a narrowed buffer, the command displays an error message that looks
13962 Search failed: "\\w+\\W*"
13965 If you are reading this in Info in GNU Emacs, you can test for these
13968 First, evaluate the function in the usual manner to install it.
13970 Here is a copy of the definition. Place your cursor after the closing
13971 parenthesis and type @kbd{C-x C-e} to install it.
13975 ;; @r{First version; has bugs!}
13976 (defun @value{COUNT-WORDS} (beginning end)
13977 "Print number of words in the region.
13978 Words are defined as at least one word-constituent character followed
13979 by at least one character that is not a word-constituent. The buffer's
13980 syntax table determines which characters these are."
13984 (message "Counting words in region ... ")
13988 ;;; @r{1. Set up appropriate conditions.}
13990 (goto-char beginning)
13995 ;;; @r{2. Run the} while @r{loop.}
13996 (while (< (point) end)
13997 (re-search-forward "\\w+\\W*")
13998 (setq count (1+ count)))
14002 ;;; @r{3. Send a message to the user.}
14003 (cond ((zerop count)
14004 (message "The region does NOT have any words."))
14005 ((= 1 count) (message "The region has 1 word."))
14006 (t (message "The region has %d words." count))))))
14012 If you wish, you can also install this keybinding by evaluating it:
14015 (global-set-key "\C-c=" '@value{COUNT-WORDS})
14018 To conduct the first test, set mark and point to the beginning and end
14019 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14020 @value{COUNT-WORDS}} if you have not bound @kbd{C-c =}):
14027 Emacs will tell you, correctly, that the region has three words.
14029 Repeat the test, but place mark at the beginning of the line and place
14030 point just @emph{before} the word @samp{one}. Again type the command
14031 @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}). Emacs should tell you
14032 that the region has no words, since it is composed only of the
14033 whitespace at the beginning of the line. But instead Emacs tells you
14034 that the region has one word!
14036 For the third test, copy the sample line to the end of the
14037 @file{*scratch*} buffer and then type several spaces at the end of the
14038 line. Place mark right after the word @samp{three} and point at the
14039 end of line. (The end of the line will be the end of the buffer.)
14040 Type @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}) as you did before.
14041 Again, Emacs should tell you that the region has no words, since it is
14042 composed only of the whitespace at the end of the line. Instead,
14043 Emacs displays an error message saying @samp{Search failed}.
14045 The two bugs stem from the same problem.
14047 Consider the first manifestation of the bug, in which the command
14048 tells you that the whitespace at the beginning of the line contains
14049 one word. What happens is this: The @code{M-x @value{COUNT-WORDS}}
14050 command moves point to the beginning of the region. The @code{while}
14051 tests whether the value of point is smaller than the value of
14052 @code{end}, which it is. Consequently, the regular expression search
14053 looks for and finds the first word. It leaves point after the word.
14054 @code{count} is set to one. The @code{while} loop repeats; but this
14055 time the value of point is larger than the value of @code{end}, the
14056 loop is exited; and the function displays a message saying the number
14057 of words in the region is one. In brief, the regular expression
14058 search looks for and finds the word even though it is outside
14061 In the second manifestation of the bug, the region is whitespace at
14062 the end of the buffer. Emacs says @samp{Search failed}. What happens
14063 is that the true-or-false-test in the @code{while} loop tests true, so
14064 the search expression is executed. But since there are no more words
14065 in the buffer, the search fails.
14067 In both manifestations of the bug, the search extends or attempts to
14068 extend outside of the region.
14070 The solution is to limit the search to the region---this is a fairly
14071 simple action, but as you may have come to expect, it is not quite as
14072 simple as you might think.
14074 As we have seen, the @code{re-search-forward} function takes a search
14075 pattern as its first argument. But in addition to this first,
14076 mandatory argument, it accepts three optional arguments. The optional
14077 second argument bounds the search. The optional third argument, if
14078 @code{t}, causes the function to return @code{nil} rather than signal
14079 an error if the search fails. The optional fourth argument is a
14080 repeat count. (In Emacs, you can see a function's documentation by
14081 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14083 In the @code{@value{COUNT-WORDS}} definition, the value of the end of
14084 the region is held by the variable @code{end} which is passed as an
14085 argument to the function. Thus, we can add @code{end} as an argument
14086 to the regular expression search expression:
14089 (re-search-forward "\\w+\\W*" end)
14092 However, if you make only this change to the @code{@value{COUNT-WORDS}}
14093 definition and then test the new version of the definition on a
14094 stretch of whitespace, you will receive an error message saying
14095 @samp{Search failed}.
14097 What happens is this: the search is limited to the region, and fails
14098 as you expect because there are no word-constituent characters in the
14099 region. Since it fails, we receive an error message. But we do not
14100 want to receive an error message in this case; we want to receive the
14101 message that "The region does NOT have any words."
14103 The solution to this problem is to provide @code{re-search-forward}
14104 with a third argument of @code{t}, which causes the function to return
14105 @code{nil} rather than signal an error if the search fails.
14107 However, if you make this change and try it, you will see the message
14108 ``Counting words in region ... '' and @dots{} you will keep on seeing
14109 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14111 Here is what happens: the search is limited to the region, as before,
14112 and it fails because there are no word-constituent characters in the
14113 region, as expected. Consequently, the @code{re-search-forward}
14114 expression returns @code{nil}. It does nothing else. In particular,
14115 it does not move point, which it does as a side effect if it finds the
14116 search target. After the @code{re-search-forward} expression returns
14117 @code{nil}, the next expression in the @code{while} loop is evaluated.
14118 This expression increments the count. Then the loop repeats. The
14119 true-or-false-test tests true because the value of point is still less
14120 than the value of end, since the @code{re-search-forward} expression
14121 did not move point. @dots{} and the cycle repeats @dots{}
14123 The @code{@value{COUNT-WORDS}} definition requires yet another
14124 modification, to cause the true-or-false-test of the @code{while} loop
14125 to test false if the search fails. Put another way, there are two
14126 conditions that must be satisfied in the true-or-false-test before the
14127 word count variable is incremented: point must still be within the
14128 region and the search expression must have found a word to count.
14130 Since both the first condition and the second condition must be true
14131 together, the two expressions, the region test and the search
14132 expression, can be joined with an @code{and} special form and embedded in
14133 the @code{while} loop as the true-or-false-test, like this:
14136 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14139 @c colon in printed section title causes problem in Info cross reference
14140 @c also trouble with an overfull hbox
14143 (For information about @code{and}, see
14144 @ref{kill-new function, , The @code{kill-new} function}.)
14148 (@xref{kill-new function, , The @code{kill-new} function}, for
14149 information about @code{and}.)
14152 The @code{re-search-forward} expression returns @code{t} if the search
14153 succeeds and as a side effect moves point. Consequently, as words are
14154 found, point is moved through the region. When the search expression
14155 fails to find another word, or when point reaches the end of the
14156 region, the true-or-false-test tests false, the @code{while} loop
14157 exits, and the @code{@value{COUNT-WORDS}} function displays one or
14158 other of its messages.
14160 After incorporating these final changes, the @code{@value{COUNT-WORDS}}
14161 works without bugs (or at least, without bugs that I have found!).
14162 Here is what it looks like:
14166 ;;; @r{Final version:} @code{while}
14167 (defun @value{COUNT-WORDS} (beginning end)
14168 "Print number of words in the region."
14170 (message "Counting words in region ... ")
14174 ;;; @r{1. Set up appropriate conditions.}
14177 (goto-char beginning)
14181 ;;; @r{2. Run the} while @r{loop.}
14182 (while (and (< (point) end)
14183 (re-search-forward "\\w+\\W*" end t))
14184 (setq count (1+ count)))
14188 ;;; @r{3. Send a message to the user.}
14189 (cond ((zerop count)
14191 "The region does NOT have any words."))
14194 "The region has 1 word."))
14197 "The region has %d words." count))))))
14201 @node recursive-count-words
14202 @section Count Words Recursively
14203 @cindex Count words recursively
14204 @cindex Recursively counting words
14205 @cindex Words, counted recursively
14207 You can write the function for counting words recursively as well as
14208 with a @code{while} loop. Let's see how this is done.
14210 First, we need to recognize that the @code{@value{COUNT-WORDS}}
14211 function has three jobs: it sets up the appropriate conditions for
14212 counting to occur; it counts the words in the region; and it sends a
14213 message to the user telling how many words there are.
14215 If we write a single recursive function to do everything, we will
14216 receive a message for every recursive call. If the region contains 13
14217 words, we will receive thirteen messages, one right after the other.
14218 We don't want this! Instead, we must write two functions to do the
14219 job, one of which (the recursive function) will be used inside of the
14220 other. One function will set up the conditions and display the
14221 message; the other will return the word count.
14223 Let us start with the function that causes the message to be displayed.
14224 We can continue to call this @code{@value{COUNT-WORDS}}.
14226 This is the function that the user will call. It will be interactive.
14227 Indeed, it will be similar to our previous versions of this
14228 function, except that it will call @code{recursive-count-words} to
14229 determine how many words are in the region.
14232 We can readily construct a template for this function, based on our
14237 ;; @r{Recursive version; uses regular expression search}
14238 (defun @value{COUNT-WORDS} (beginning end)
14239 "@var{documentation}@dots{}"
14240 (@var{interactive-expression}@dots{})
14244 ;;; @r{1. Set up appropriate conditions.}
14245 (@var{explanatory message})
14246 (@var{set-up functions}@dots{}
14250 ;;; @r{2. Count the words.}
14251 @var{recursive call}
14255 ;;; @r{3. Send a message to the user.}
14256 @var{message providing word count}))
14260 The definition looks straightforward, except that somehow the count
14261 returned by the recursive call must be passed to the message
14262 displaying the word count. A little thought suggests that this can be
14263 done by making use of a @code{let} expression: we can bind a variable
14264 in the varlist of a @code{let} expression to the number of words in
14265 the region, as returned by the recursive call; and then the
14266 @code{cond} expression, using binding, can display the value to the
14269 Often, one thinks of the binding within a @code{let} expression as
14270 somehow secondary to the `primary' work of a function. But in this
14271 case, what you might consider the `primary' job of the function,
14272 counting words, is done within the @code{let} expression.
14275 Using @code{let}, the function definition looks like this:
14279 (defun @value{COUNT-WORDS} (beginning end)
14280 "Print number of words in the region."
14285 ;;; @r{1. Set up appropriate conditions.}
14286 (message "Counting words in region ... ")
14288 (goto-char beginning)
14292 ;;; @r{2. Count the words.}
14293 (let ((count (recursive-count-words end)))
14297 ;;; @r{3. Send a message to the user.}
14298 (cond ((zerop count)
14300 "The region does NOT have any words."))
14303 "The region has 1 word."))
14306 "The region has %d words." count))))))
14310 Next, we need to write the recursive counting function.
14312 A recursive function has at least three parts: the `do-again-test', the
14313 `next-step-expression', and the recursive call.
14315 The do-again-test determines whether the function will or will not be
14316 called again. Since we are counting words in a region and can use a
14317 function that moves point forward for every word, the do-again-test
14318 can check whether point is still within the region. The do-again-test
14319 should find the value of point and determine whether point is before,
14320 at, or after the value of the end of the region. We can use the
14321 @code{point} function to locate point. Clearly, we must pass the
14322 value of the end of the region to the recursive counting function as an
14325 In addition, the do-again-test should also test whether the search finds a
14326 word. If it does not, the function should not call itself again.
14328 The next-step-expression changes a value so that when the recursive
14329 function is supposed to stop calling itself, it stops. More
14330 precisely, the next-step-expression changes a value so that at the
14331 right time, the do-again-test stops the recursive function from
14332 calling itself again. In this case, the next-step-expression can be
14333 the expression that moves point forward, word by word.
14335 The third part of a recursive function is the recursive call.
14337 Somewhere, also, we also need a part that does the `work' of the
14338 function, a part that does the counting. A vital part!
14341 But already, we have an outline of the recursive counting function:
14345 (defun recursive-count-words (region-end)
14346 "@var{documentation}@dots{}"
14347 @var{do-again-test}
14348 @var{next-step-expression}
14349 @var{recursive call})
14353 Now we need to fill in the slots. Let's start with the simplest cases
14354 first: if point is at or beyond the end of the region, there cannot
14355 be any words in the region, so the function should return zero.
14356 Likewise, if the search fails, there are no words to count, so the
14357 function should return zero.
14359 On the other hand, if point is within the region and the search
14360 succeeds, the function should call itself again.
14363 Thus, the do-again-test should look like this:
14367 (and (< (point) region-end)
14368 (re-search-forward "\\w+\\W*" region-end t))
14372 Note that the search expression is part of the do-again-test---the
14373 function returns @code{t} if its search succeeds and @code{nil} if it
14374 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14375 @code{@value{COUNT-WORDS}}}, for an explanation of how
14376 @code{re-search-forward} works.)
14378 The do-again-test is the true-or-false test of an @code{if} clause.
14379 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14380 clause should call the function again; but if it fails, the else-part
14381 should return zero since either point is outside the region or the
14382 search failed because there were no words to find.
14384 But before considering the recursive call, we need to consider the
14385 next-step-expression. What is it? Interestingly, it is the search
14386 part of the do-again-test.
14388 In addition to returning @code{t} or @code{nil} for the
14389 do-again-test, @code{re-search-forward} moves point forward as a side
14390 effect of a successful search. This is the action that changes the
14391 value of point so that the recursive function stops calling itself
14392 when point completes its movement through the region. Consequently,
14393 the @code{re-search-forward} expression is the next-step-expression.
14396 In outline, then, the body of the @code{recursive-count-words}
14397 function looks like this:
14401 (if @var{do-again-test-and-next-step-combined}
14403 @var{recursive-call-returning-count}
14409 How to incorporate the mechanism that counts?
14411 If you are not used to writing recursive functions, a question like
14412 this can be troublesome. But it can and should be approached
14415 We know that the counting mechanism should be associated in some way
14416 with the recursive call. Indeed, since the next-step-expression moves
14417 point forward by one word, and since a recursive call is made for
14418 each word, the counting mechanism must be an expression that adds one
14419 to the value returned by a call to @code{recursive-count-words}.
14422 Consider several cases:
14426 If there are two words in the region, the function should return
14427 a value resulting from adding one to the value returned when it counts
14428 the first word, plus the number returned when it counts the remaining
14429 words in the region, which in this case is one.
14432 If there is one word in the region, the function should return
14433 a value resulting from adding one to the value returned when it counts
14434 that word, plus the number returned when it counts the remaining
14435 words in the region, which in this case is zero.
14438 If there are no words in the region, the function should return zero.
14441 From the sketch we can see that the else-part of the @code{if} returns
14442 zero for the case of no words. This means that the then-part of the
14443 @code{if} must return a value resulting from adding one to the value
14444 returned from a count of the remaining words.
14447 The expression will look like this, where @code{1+} is a function that
14448 adds one to its argument.
14451 (1+ (recursive-count-words region-end))
14455 The whole @code{recursive-count-words} function will then look like
14460 (defun recursive-count-words (region-end)
14461 "@var{documentation}@dots{}"
14463 ;;; @r{1. do-again-test}
14464 (if (and (< (point) region-end)
14465 (re-search-forward "\\w+\\W*" region-end t))
14469 ;;; @r{2. then-part: the recursive call}
14470 (1+ (recursive-count-words region-end))
14472 ;;; @r{3. else-part}
14478 Let's examine how this works:
14480 If there are no words in the region, the else part of the @code{if}
14481 expression is evaluated and consequently the function returns zero.
14483 If there is one word in the region, the value of point is less than
14484 the value of @code{region-end} and the search succeeds. In this case,
14485 the true-or-false-test of the @code{if} expression tests true, and the
14486 then-part of the @code{if} expression is evaluated. The counting
14487 expression is evaluated. This expression returns a value (which will
14488 be the value returned by the whole function) that is the sum of one
14489 added to the value returned by a recursive call.
14491 Meanwhile, the next-step-expression has caused point to jump over the
14492 first (and in this case only) word in the region. This means that
14493 when @code{(recursive-count-words region-end)} is evaluated a second
14494 time, as a result of the recursive call, the value of point will be
14495 equal to or greater than the value of region end. So this time,
14496 @code{recursive-count-words} will return zero. The zero will be added
14497 to one, and the original evaluation of @code{recursive-count-words}
14498 will return one plus zero, which is one, which is the correct amount.
14500 Clearly, if there are two words in the region, the first call to
14501 @code{recursive-count-words} returns one added to the value returned
14502 by calling @code{recursive-count-words} on a region containing the
14503 remaining word---that is, it adds one to one, producing two, which is
14504 the correct amount.
14506 Similarly, if there are three words in the region, the first call to
14507 @code{recursive-count-words} returns one added to the value returned
14508 by calling @code{recursive-count-words} on a region containing the
14509 remaining two words---and so on and so on.
14513 With full documentation the two functions look like this:
14517 The recursive function:
14519 @findex recursive-count-words
14522 (defun recursive-count-words (region-end)
14523 "Number of words between point and REGION-END."
14527 ;;; @r{1. do-again-test}
14528 (if (and (< (point) region-end)
14529 (re-search-forward "\\w+\\W*" region-end t))
14533 ;;; @r{2. then-part: the recursive call}
14534 (1+ (recursive-count-words region-end))
14536 ;;; @r{3. else-part}
14547 ;;; @r{Recursive version}
14548 (defun @value{COUNT-WORDS} (beginning end)
14549 "Print number of words in the region.
14553 Words are defined as at least one word-constituent
14554 character followed by at least one character that is
14555 not a word-constituent. The buffer's syntax table
14556 determines which characters these are."
14560 (message "Counting words in region ... ")
14562 (goto-char beginning)
14563 (let ((count (recursive-count-words end)))
14566 (cond ((zerop count)
14568 "The region does NOT have any words."))
14572 (message "The region has 1 word."))
14575 "The region has %d words." count))))))
14579 @node Counting Exercise
14580 @section Exercise: Counting Punctuation
14582 Using a @code{while} loop, write a function to count the number of
14583 punctuation marks in a region---period, comma, semicolon, colon,
14584 exclamation mark, and question mark. Do the same using recursion.
14586 @node Words in a defun
14587 @chapter Counting Words in a @code{defun}
14588 @cindex Counting words in a @code{defun}
14589 @cindex Word counting in a @code{defun}
14591 Our next project is to count the number of words in a function
14592 definition. Clearly, this can be done using some variant of
14593 @code{@value{COUNT-WORDS}}. @xref{Counting Words, , Counting Words:
14594 Repetition and Regexps}. If we are just going to count the words in
14595 one definition, it is easy enough to mark the definition with the
14596 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14597 @code{@value{COUNT-WORDS}}.
14599 However, I am more ambitious: I want to count the words and symbols in
14600 every definition in the Emacs sources and then print a graph that
14601 shows how many functions there are of each length: how many contain 40
14602 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14603 and so on. I have often been curious how long a typical function is,
14604 and this will tell.
14607 * Divide and Conquer::
14608 * Words and Symbols:: What to count?
14609 * Syntax:: What constitutes a word or symbol?
14610 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
14611 * Several defuns:: Counting several defuns in a file.
14612 * Find a File:: Do you want to look at a file?
14613 * lengths-list-file:: A list of the lengths of many definitions.
14614 * Several files:: Counting in definitions in different files.
14615 * Several files recursively:: Recursively counting in different files.
14616 * Prepare the data:: Prepare the data for display in a graph.
14620 @node Divide and Conquer
14621 @unnumberedsec Divide and Conquer
14624 Described in one phrase, the histogram project is daunting; but
14625 divided into numerous small steps, each of which we can take one at a
14626 time, the project becomes less fearsome. Let us consider what the
14631 First, write a function to count the words in one definition. This
14632 includes the problem of handling symbols as well as words.
14635 Second, write a function to list the numbers of words in each function
14636 in a file. This function can use the @code{count-words-in-defun}
14640 Third, write a function to list the numbers of words in each function
14641 in each of several files. This entails automatically finding the
14642 various files, switching to them, and counting the words in the
14643 definitions within them.
14646 Fourth, write a function to convert the list of numbers that we
14647 created in step three to a form that will be suitable for printing as
14651 Fifth, write a function to print the results as a graph.
14654 This is quite a project! But if we take each step slowly, it will not
14657 @node Words and Symbols
14658 @section What to Count?
14659 @cindex Words and symbols in defun
14661 When we first start thinking about how to count the words in a
14662 function definition, the first question is (or ought to be) what are
14663 we going to count? When we speak of `words' with respect to a Lisp
14664 function definition, we are actually speaking, in large part, of
14665 `symbols'. For example, the following @code{multiply-by-seven}
14666 function contains the five symbols @code{defun},
14667 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14668 addition, in the documentation string, it contains the four words
14669 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14670 symbol @samp{number} is repeated, so the definition contains a total
14671 of ten words and symbols.
14675 (defun multiply-by-seven (number)
14676 "Multiply NUMBER by seven."
14682 However, if we mark the @code{multiply-by-seven} definition with
14683 @kbd{C-M-h} (@code{mark-defun}), and then call
14684 @code{@value{COUNT-WORDS}} on it, we will find that
14685 @code{@value{COUNT-WORDS}} claims the definition has eleven words, not
14686 ten! Something is wrong!
14688 The problem is twofold: @code{@value{COUNT-WORDS}} does not count the
14689 @samp{*} as a word, and it counts the single symbol,
14690 @code{multiply-by-seven}, as containing three words. The hyphens are
14691 treated as if they were interword spaces rather than intraword
14692 connectors: @samp{multiply-by-seven} is counted as if it were written
14693 @samp{multiply by seven}.
14695 The cause of this confusion is the regular expression search within
14696 the @code{@value{COUNT-WORDS}} definition that moves point forward word
14697 by word. In the canonical version of @code{@value{COUNT-WORDS}}, the
14705 This regular expression is a pattern defining one or more word
14706 constituent characters possibly followed by one or more characters
14707 that are not word constituents. What is meant by `word constituent
14708 characters' brings us to the issue of syntax, which is worth a section
14712 @section What Constitutes a Word or Symbol?
14713 @cindex Syntax categories and tables
14715 Emacs treats different characters as belonging to different
14716 @dfn{syntax categories}. For example, the regular expression,
14717 @samp{\\w+}, is a pattern specifying one or more @emph{word
14718 constituent} characters. Word constituent characters are members of
14719 one syntax category. Other syntax categories include the class of
14720 punctuation characters, such as the period and the comma, and the
14721 class of whitespace characters, such as the blank space and the tab
14722 character. (For more information, @pxref{Syntax Tables, , Syntax
14723 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14725 Syntax tables specify which characters belong to which categories.
14726 Usually, a hyphen is not specified as a `word constituent character'.
14727 Instead, it is specified as being in the `class of characters that are
14728 part of symbol names but not words.' This means that the
14729 @code{@value{COUNT-WORDS}} function treats it in the same way it treats
14730 an interword white space, which is why @code{@value{COUNT-WORDS}}
14731 counts @samp{multiply-by-seven} as three words.
14733 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14734 one symbol: modify the syntax table or modify the regular expression.
14736 We could redefine a hyphen as a word constituent character by
14737 modifying the syntax table that Emacs keeps for each mode. This
14738 action would serve our purpose, except that a hyphen is merely the
14739 most common character within symbols that is not typically a word
14740 constituent character; there are others, too.
14742 Alternatively, we can redefine the regexp used in the
14743 @code{@value{COUNT-WORDS}} definition so as to include symbols. This
14744 procedure has the merit of clarity, but the task is a little tricky.
14747 The first part is simple enough: the pattern must match ``at least one
14748 character that is a word or symbol constituent''. Thus:
14751 "\\(\\w\\|\\s_\\)+"
14755 The @samp{\\(} is the first part of the grouping construct that
14756 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14757 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14758 character and the @samp{\\s_} matches any character that is part of a
14759 symbol name but not a word-constituent character. The @samp{+}
14760 following the group indicates that the word or symbol constituent
14761 characters must be matched at least once.
14763 However, the second part of the regexp is more difficult to design.
14764 What we want is to follow the first part with ``optionally one or more
14765 characters that are not constituents of a word or symbol''. At first,
14766 I thought I could define this with the following:
14769 "\\(\\W\\|\\S_\\)*"
14773 The upper case @samp{W} and @samp{S} match characters that are
14774 @emph{not} word or symbol constituents. Unfortunately, this
14775 expression matches any character that is either not a word constituent
14776 or not a symbol constituent. This matches any character!
14778 I then noticed that every word or symbol in my test region was
14779 followed by white space (blank space, tab, or newline). So I tried
14780 placing a pattern to match one or more blank spaces after the pattern
14781 for one or more word or symbol constituents. This failed, too. Words
14782 and symbols are often separated by whitespace, but in actual code
14783 parentheses may follow symbols and punctuation may follow words. So
14784 finally, I designed a pattern in which the word or symbol constituents
14785 are followed optionally by characters that are not white space and
14786 then followed optionally by white space.
14789 Here is the full regular expression:
14792 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14795 @node count-words-in-defun
14796 @section The @code{count-words-in-defun} Function
14797 @cindex Counting words in a @code{defun}
14799 We have seen that there are several ways to write a
14800 @code{count-words-region} function. To write a
14801 @code{count-words-in-defun}, we need merely adapt one of these
14804 The version that uses a @code{while} loop is easy to understand, so I
14805 am going to adapt that. Because @code{count-words-in-defun} will be
14806 part of a more complex program, it need not be interactive and it need
14807 not display a message but just return the count. These considerations
14808 simplify the definition a little.
14810 On the other hand, @code{count-words-in-defun} will be used within a
14811 buffer that contains function definitions. Consequently, it is
14812 reasonable to ask that the function determine whether it is called
14813 when point is within a function definition, and if it is, to return
14814 the count for that definition. This adds complexity to the
14815 definition, but saves us from needing to pass arguments to the
14819 These considerations lead us to prepare the following template:
14823 (defun count-words-in-defun ()
14824 "@var{documentation}@dots{}"
14825 (@var{set up}@dots{}
14826 (@var{while loop}@dots{})
14827 @var{return count})
14832 As usual, our job is to fill in the slots.
14836 We are presuming that this function will be called within a buffer
14837 containing function definitions. Point will either be within a
14838 function definition or not. For @code{count-words-in-defun} to work,
14839 point must move to the beginning of the definition, a counter must
14840 start at zero, and the counting loop must stop when point reaches the
14841 end of the definition.
14843 The @code{beginning-of-defun} function searches backwards for an
14844 opening delimiter such as a @samp{(} at the beginning of a line, and
14845 moves point to that position, or else to the limit of the search. In
14846 practice, this means that @code{beginning-of-defun} moves point to the
14847 beginning of an enclosing or preceding function definition, or else to
14848 the beginning of the buffer. We can use @code{beginning-of-defun} to
14849 place point where we wish to start.
14851 The @code{while} loop requires a counter to keep track of the words or
14852 symbols being counted. A @code{let} expression can be used to create
14853 a local variable for this purpose, and bind it to an initial value of zero.
14855 The @code{end-of-defun} function works like @code{beginning-of-defun}
14856 except that it moves point to the end of the definition.
14857 @code{end-of-defun} can be used as part of an expression that
14858 determines the position of the end of the definition.
14860 The set up for @code{count-words-in-defun} takes shape rapidly: first
14861 we move point to the beginning of the definition, then we create a
14862 local variable to hold the count, and finally, we record the position
14863 of the end of the definition so the @code{while} loop will know when to stop
14867 The code looks like this:
14871 (beginning-of-defun)
14873 (end (save-excursion (end-of-defun) (point))))
14878 The code is simple. The only slight complication is likely to concern
14879 @code{end}: it is bound to the position of the end of the definition
14880 by a @code{save-excursion} expression that returns the value of point
14881 after @code{end-of-defun} temporarily moves it to the end of the
14884 The second part of the @code{count-words-in-defun}, after the set up,
14885 is the @code{while} loop.
14887 The loop must contain an expression that jumps point forward word by
14888 word and symbol by symbol, and another expression that counts the
14889 jumps. The true-or-false-test for the @code{while} loop should test
14890 true so long as point should jump forward, and false when point is at
14891 the end of the definition. We have already redefined the regular
14892 expression for this, so the loop is straightforward:
14896 (while (and (< (point) end)
14898 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t))
14899 (setq count (1+ count)))
14903 The third part of the function definition returns the count of words
14904 and symbols. This part is the last expression within the body of the
14905 @code{let} expression, and can be, very simply, the local variable
14906 @code{count}, which when evaluated returns the count.
14909 Put together, the @code{count-words-in-defun} definition looks like this:
14911 @findex count-words-in-defun
14914 (defun count-words-in-defun ()
14915 "Return the number of words and symbols in a defun."
14916 (beginning-of-defun)
14918 (end (save-excursion (end-of-defun) (point))))
14922 (and (< (point) end)
14924 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14926 (setq count (1+ count)))
14931 How to test this? The function is not interactive, but it is easy to
14932 put a wrapper around the function to make it interactive; we can use
14933 almost the same code as for the recursive version of
14934 @code{@value{COUNT-WORDS}}:
14938 ;;; @r{Interactive version.}
14939 (defun count-words-defun ()
14940 "Number of words and symbols in a function definition."
14943 "Counting words and symbols in function definition ... ")
14946 (let ((count (count-words-in-defun)))
14950 "The definition does NOT have any words or symbols."))
14955 "The definition has 1 word or symbol."))
14958 "The definition has %d words or symbols." count)))))
14964 Let's re-use @kbd{C-c =} as a convenient keybinding:
14967 (global-set-key "\C-c=" 'count-words-defun)
14970 Now we can try out @code{count-words-defun}: install both
14971 @code{count-words-in-defun} and @code{count-words-defun}, and set the
14972 keybinding, and then place the cursor within the following definition:
14976 (defun multiply-by-seven (number)
14977 "Multiply NUMBER by seven."
14984 Success! The definition has 10 words and symbols.
14986 The next problem is to count the numbers of words and symbols in
14987 several definitions within a single file.
14989 @node Several defuns
14990 @section Count Several @code{defuns} Within a File
14992 A file such as @file{simple.el} may have a hundred or more function
14993 definitions within it. Our long term goal is to collect statistics on
14994 many files, but as a first step, our immediate goal is to collect
14995 statistics on one file.
14997 The information will be a series of numbers, each number being the
14998 length of a function definition. We can store the numbers in a list.
15000 We know that we will want to incorporate the information regarding one
15001 file with information about many other files; this means that the
15002 function for counting definition lengths within one file need only
15003 return the list of lengths. It need not and should not display any
15006 The word count commands contain one expression to jump point forward
15007 word by word and another expression to count the jumps. The function
15008 to return the lengths of definitions can be designed to work the same
15009 way, with one expression to jump point forward definition by
15010 definition and another expression to construct the lengths' list.
15012 This statement of the problem makes it elementary to write the
15013 function definition. Clearly, we will start the count at the
15014 beginning of the file, so the first command will be @code{(goto-char
15015 (point-min))}. Next, we start the @code{while} loop; and the
15016 true-or-false test of the loop can be a regular expression search for
15017 the next function definition---so long as the search succeeds, point
15018 is moved forward and then the body of the loop is evaluated. The body
15019 needs an expression that constructs the lengths' list. @code{cons},
15020 the list construction command, can be used to create the list. That
15021 is almost all there is to it.
15024 Here is what this fragment of code looks like:
15028 (goto-char (point-min))
15029 (while (re-search-forward "^(defun" nil t)
15031 (cons (count-words-in-defun) lengths-list)))
15035 What we have left out is the mechanism for finding the file that
15036 contains the function definitions.
15038 In previous examples, we either used this, the Info file, or we
15039 switched back and forth to some other buffer, such as the
15040 @file{*scratch*} buffer.
15042 Finding a file is a new process that we have not yet discussed.
15045 @section Find a File
15046 @cindex Find a File
15048 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15049 command. This command is almost, but not quite right for the lengths
15053 Let's look at the source for @code{find-file}:
15057 (defun find-file (filename)
15058 "Edit file FILENAME.
15059 Switch to a buffer visiting file FILENAME,
15060 creating one if none already exists."
15061 (interactive "FFind file: ")
15062 (switch-to-buffer (find-file-noselect filename)))
15067 (The most recent version of the @code{find-file} function definition
15068 permits you to specify optional wildcards to visit multiple files; that
15069 makes the definition more complex and we will not discuss it here,
15070 since it is not relevant. You can see its source using either
15071 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15075 (defun find-file (filename &optional wildcards)
15076 "Edit file FILENAME.
15077 Switch to a buffer visiting file FILENAME,
15078 creating one if none already exists.
15079 Interactively, the default if you just type RET is the current directory,
15080 but the visited file name is available through the minibuffer history:
15081 type M-n to pull it into the minibuffer.
15083 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15084 expand wildcards (if any) and visit multiple files. You can
15085 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15087 To visit a file without any kind of conversion and without
15088 automatically choosing a major mode, use \\[find-file-literally]."
15089 (interactive (find-file-read-args "Find file: " nil))
15090 (let ((value (find-file-noselect filename nil nil wildcards)))
15092 (mapcar 'switch-to-buffer (nreverse value))
15093 (switch-to-buffer value))))
15096 The definition I am showing possesses short but complete documentation
15097 and an interactive specification that prompts you for a file name when
15098 you use the command interactively. The body of the definition
15099 contains two functions, @code{find-file-noselect} and
15100 @code{switch-to-buffer}.
15102 According to its documentation as shown by @kbd{C-h f} (the
15103 @code{describe-function} command), the @code{find-file-noselect}
15104 function reads the named file into a buffer and returns the buffer.
15105 (Its most recent version includes an optional wildcards argument,
15106 too, as well as another to read a file literally and an other you
15107 suppress warning messages. These optional arguments are irrelevant.)
15109 However, the @code{find-file-noselect} function does not select the
15110 buffer in which it puts the file. Emacs does not switch its attention
15111 (or yours if you are using @code{find-file-noselect}) to the selected
15112 buffer. That is what @code{switch-to-buffer} does: it switches the
15113 buffer to which Emacs attention is directed; and it switches the
15114 buffer displayed in the window to the new buffer. We have discussed
15115 buffer switching elsewhere. (@xref{Switching Buffers}.)
15117 In this histogram project, we do not need to display each file on the
15118 screen as the program determines the length of each definition within
15119 it. Instead of employing @code{switch-to-buffer}, we can work with
15120 @code{set-buffer}, which redirects the attention of the computer
15121 program to a different buffer but does not redisplay it on the screen.
15122 So instead of calling on @code{find-file} to do the job, we must write
15123 our own expression.
15125 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15127 @node lengths-list-file
15128 @section @code{lengths-list-file} in Detail
15130 The core of the @code{lengths-list-file} function is a @code{while}
15131 loop containing a function to move point forward `defun by defun' and
15132 a function to count the number of words and symbols in each defun.
15133 This core must be surrounded by functions that do various other tasks,
15134 including finding the file, and ensuring that point starts out at the
15135 beginning of the file. The function definition looks like this:
15136 @findex lengths-list-file
15140 (defun lengths-list-file (filename)
15141 "Return list of definitions' lengths within FILE.
15142 The returned list is a list of numbers.
15143 Each number is the number of words or
15144 symbols in one function definition."
15147 (message "Working on `%s' ... " filename)
15149 (let ((buffer (find-file-noselect filename))
15151 (set-buffer buffer)
15152 (setq buffer-read-only t)
15154 (goto-char (point-min))
15155 (while (re-search-forward "^(defun" nil t)
15157 (cons (count-words-in-defun) lengths-list)))
15158 (kill-buffer buffer)
15164 The function is passed one argument, the name of the file on which it
15165 will work. It has four lines of documentation, but no interactive
15166 specification. Since people worry that a computer is broken if they
15167 don't see anything going on, the first line of the body is a
15170 The next line contains a @code{save-excursion} that returns Emacs's
15171 attention to the current buffer when the function completes. This is
15172 useful in case you embed this function in another function that
15173 presumes point is restored to the original buffer.
15175 In the varlist of the @code{let} expression, Emacs finds the file and
15176 binds the local variable @code{buffer} to the buffer containing the
15177 file. At the same time, Emacs creates @code{lengths-list} as a local
15180 Next, Emacs switches its attention to the buffer.
15182 In the following line, Emacs makes the buffer read-only. Ideally,
15183 this line is not necessary. None of the functions for counting words
15184 and symbols in a function definition should change the buffer.
15185 Besides, the buffer is not going to be saved, even if it were changed.
15186 This line is entirely the consequence of great, perhaps excessive,
15187 caution. The reason for the caution is that this function and those
15188 it calls work on the sources for Emacs and it is inconvenient if they
15189 are inadvertently modified. It goes without saying that I did not
15190 realize a need for this line until an experiment went awry and started
15191 to modify my Emacs source files @dots{}
15193 Next comes a call to widen the buffer if it is narrowed. This
15194 function is usually not needed---Emacs creates a fresh buffer if none
15195 already exists; but if a buffer visiting the file already exists Emacs
15196 returns that one. In this case, the buffer may be narrowed and must
15197 be widened. If we wanted to be fully `user-friendly', we would
15198 arrange to save the restriction and the location of point, but we
15201 The @code{(goto-char (point-min))} expression moves point to the
15202 beginning of the buffer.
15204 Then comes a @code{while} loop in which the `work' of the function is
15205 carried out. In the loop, Emacs determines the length of each
15206 definition and constructs a lengths' list containing the information.
15208 Emacs kills the buffer after working through it. This is to save
15209 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15210 source files of interest; GNU Emacs 22 contains over a thousand source
15211 files. Another function will apply @code{lengths-list-file} to each
15214 Finally, the last expression within the @code{let} expression is the
15215 @code{lengths-list} variable; its value is returned as the value of
15216 the whole function.
15218 You can try this function by installing it in the usual fashion. Then
15219 place your cursor after the following expression and type @kbd{C-x
15220 C-e} (@code{eval-last-sexp}).
15222 @c !!! 22.1.1 lisp sources location here
15225 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15229 (You may need to change the pathname of the file; the one here is for
15230 GNU Emacs version 22.1.1. To change the expression, copy it to
15231 the @file{*scratch*} buffer and edit it.
15235 (Also, to see the full length of the list, rather than a truncated
15236 version, you may have to evaluate the following:
15239 (custom-set-variables '(eval-expression-print-length nil))
15243 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15244 Then evaluate the @code{lengths-list-file} expression.)
15247 The lengths' list for @file{debug.el} takes less than a second to
15248 produce and looks like this in GNU Emacs 22:
15251 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15255 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15256 took seven seconds to produce and looked like this:
15259 (75 41 80 62 20 45 44 68 45 12 34 235)
15262 (The newer version of @file{debug.el} contains more defuns than the
15263 earlier one; and my new machine is much faster than the old one.)
15265 Note that the length of the last definition in the file is first in
15268 @node Several files
15269 @section Count Words in @code{defuns} in Different Files
15271 In the previous section, we created a function that returns a list of
15272 the lengths of each definition in a file. Now, we want to define a
15273 function to return a master list of the lengths of the definitions in
15276 Working on each of a list of files is a repetitious act, so we can use
15277 either a @code{while} loop or recursion.
15280 * lengths-list-many-files:: Return a list of the lengths of defuns.
15281 * append:: Attach one list to another.
15285 @node lengths-list-many-files
15286 @unnumberedsubsec Determine the lengths of @code{defuns}
15289 The design using a @code{while} loop is routine. The argument passed
15290 the function is a list of files. As we saw earlier (@pxref{Loop
15291 Example}), you can write a @code{while} loop so that the body of the
15292 loop is evaluated if such a list contains elements, but to exit the
15293 loop if the list is empty. For this design to work, the body of the
15294 loop must contain an expression that shortens the list each time the
15295 body is evaluated, so that eventually the list is empty. The usual
15296 technique is to set the value of the list to the value of the @sc{cdr}
15297 of the list each time the body is evaluated.
15300 The template looks like this:
15304 (while @var{test-whether-list-is-empty}
15306 @var{set-list-to-cdr-of-list})
15310 Also, we remember that a @code{while} loop returns @code{nil} (the
15311 result of evaluating the true-or-false-test), not the result of any
15312 evaluation within its body. (The evaluations within the body of the
15313 loop are done for their side effects.) However, the expression that
15314 sets the lengths' list is part of the body---and that is the value
15315 that we want returned by the function as a whole. To do this, we
15316 enclose the @code{while} loop within a @code{let} expression, and
15317 arrange that the last element of the @code{let} expression contains
15318 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15319 Example with an Incrementing Counter}.)
15321 @findex lengths-list-many-files
15323 These considerations lead us directly to the function itself:
15327 ;;; @r{Use @code{while} loop.}
15328 (defun lengths-list-many-files (list-of-files)
15329 "Return list of lengths of defuns in LIST-OF-FILES."
15332 (let (lengths-list)
15334 ;;; @r{true-or-false-test}
15335 (while list-of-files
15340 ;;; @r{Generate a lengths' list.}
15342 (expand-file-name (car list-of-files)))))
15346 ;;; @r{Make files' list shorter.}
15347 (setq list-of-files (cdr list-of-files)))
15349 ;;; @r{Return final value of lengths' list.}
15354 @code{expand-file-name} is a built-in function that converts a file
15355 name to the absolute, long, path name form. The function employs the
15356 name of the directory in which the function is called.
15358 @c !!! 22.1.1 lisp sources location here
15360 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15361 Emacs is visiting the
15362 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15372 @c !!! 22.1.1 lisp sources location here
15374 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15377 The only other new element of this function definition is the as yet
15378 unstudied function @code{append}, which merits a short section for
15382 @subsection The @code{append} Function
15385 The @code{append} function attaches one list to another. Thus,
15388 (append '(1 2 3 4) '(5 6 7 8))
15399 This is exactly how we want to attach two lengths' lists produced by
15400 @code{lengths-list-file} to each other. The results contrast with
15404 (cons '(1 2 3 4) '(5 6 7 8))
15409 which constructs a new list in which the first argument to @code{cons}
15410 becomes the first element of the new list:
15413 ((1 2 3 4) 5 6 7 8)
15416 @node Several files recursively
15417 @section Recursively Count Words in Different Files
15419 Besides a @code{while} loop, you can work on each of a list of files
15420 with recursion. A recursive version of @code{lengths-list-many-files}
15421 is short and simple.
15423 The recursive function has the usual parts: the `do-again-test', the
15424 `next-step-expression', and the recursive call. The `do-again-test'
15425 determines whether the function should call itself again, which it
15426 will do if the @code{list-of-files} contains any remaining elements;
15427 the `next-step-expression' resets the @code{list-of-files} to the
15428 @sc{cdr} of itself, so eventually the list will be empty; and the
15429 recursive call calls itself on the shorter list. The complete
15430 function is shorter than this description!
15431 @findex recursive-lengths-list-many-files
15435 (defun recursive-lengths-list-many-files (list-of-files)
15436 "Return list of lengths of each defun in LIST-OF-FILES."
15437 (if list-of-files ; @r{do-again-test}
15440 (expand-file-name (car list-of-files)))
15441 (recursive-lengths-list-many-files
15442 (cdr list-of-files)))))
15447 In a sentence, the function returns the lengths' list for the first of
15448 the @code{list-of-files} appended to the result of calling itself on
15449 the rest of the @code{list-of-files}.
15451 Here is a test of @code{recursive-lengths-list-many-files}, along with
15452 the results of running @code{lengths-list-file} on each of the files
15455 Install @code{recursive-lengths-list-many-files} and
15456 @code{lengths-list-file}, if necessary, and then evaluate the
15457 following expressions. You may need to change the files' pathnames;
15458 those here work when this Info file and the Emacs sources are located
15459 in their customary places. To change the expressions, copy them to
15460 the @file{*scratch*} buffer, edit them, and then evaluate them.
15462 The results are shown after the @samp{@result{}}. (These results are
15463 for files from Emacs version 22.1.1; files from other versions of
15464 Emacs may produce different results.)
15466 @c !!! 22.1.1 lisp sources location here
15469 (cd "/usr/local/share/emacs/22.1.1/")
15471 (lengths-list-file "./lisp/macros.el")
15472 @result{} (283 263 480 90)
15476 (lengths-list-file "./lisp/mail/mailalias.el")
15477 @result{} (38 32 29 95 178 180 321 218 324)
15481 (lengths-list-file "./lisp/makesum.el")
15486 (recursive-lengths-list-many-files
15487 '("./lisp/macros.el"
15488 "./lisp/mail/mailalias.el"
15489 "./lisp/makesum.el"))
15490 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15494 The @code{recursive-lengths-list-many-files} function produces the
15497 The next step is to prepare the data in the list for display in a graph.
15499 @node Prepare the data
15500 @section Prepare the Data for Display in a Graph
15502 The @code{recursive-lengths-list-many-files} function returns a list
15503 of numbers. Each number records the length of a function definition.
15504 What we need to do now is transform this data into a list of numbers
15505 suitable for generating a graph. The new list will tell how many
15506 functions definitions contain less than 10 words and
15507 symbols, how many contain between 10 and 19 words and symbols, how
15508 many contain between 20 and 29 words and symbols, and so on.
15510 In brief, we need to go through the lengths' list produced by the
15511 @code{recursive-lengths-list-many-files} function and count the number
15512 of defuns within each range of lengths, and produce a list of those
15516 * Data for Display in Detail::
15517 * Sorting:: Sorting lists.
15518 * Files List:: Making a list of files.
15519 * Counting function definitions::
15523 @node Data for Display in Detail
15524 @unnumberedsubsec The Data for Display in Detail
15527 Based on what we have done before, we can readily foresee that it
15528 should not be too hard to write a function that `@sc{cdr}s' down the
15529 lengths' list, looks at each element, determines which length range it
15530 is in, and increments a counter for that range.
15532 However, before beginning to write such a function, we should consider
15533 the advantages of sorting the lengths' list first, so the numbers are
15534 ordered from smallest to largest. First, sorting will make it easier
15535 to count the numbers in each range, since two adjacent numbers will
15536 either be in the same length range or in adjacent ranges. Second, by
15537 inspecting a sorted list, we can discover the highest and lowest
15538 number, and thereby determine the largest and smallest length range
15542 @subsection Sorting Lists
15545 Emacs contains a function to sort lists, called (as you might guess)
15546 @code{sort}. The @code{sort} function takes two arguments, the list
15547 to be sorted, and a predicate that determines whether the first of
15548 two list elements is ``less'' than the second.
15550 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15551 Type Object as an Argument}), a predicate is a function that
15552 determines whether some property is true or false. The @code{sort}
15553 function will reorder a list according to whatever property the
15554 predicate uses; this means that @code{sort} can be used to sort
15555 non-numeric lists by non-numeric criteria---it can, for example,
15556 alphabetize a list.
15559 The @code{<} function is used when sorting a numeric list. For example,
15562 (sort '(4 8 21 17 33 7 21 7) '<)
15570 (4 7 7 8 17 21 21 33)
15574 (Note that in this example, both the arguments are quoted so that the
15575 symbols are not evaluated before being passed to @code{sort} as
15578 Sorting the list returned by the
15579 @code{recursive-lengths-list-many-files} function is straightforward;
15580 it uses the @code{<} function:
15584 In GNU Emacs 22, eval
15586 (cd "/usr/local/share/emacs/22.0.50/")
15588 (recursive-lengths-list-many-files
15589 '("./lisp/macros.el"
15590 "./lisp/mail/mailalias.el"
15591 "./lisp/makesum.el"))
15599 (recursive-lengths-list-many-files
15600 '("./lisp/macros.el"
15601 "./lisp/mailalias.el"
15602 "./lisp/makesum.el"))
15612 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15616 (Note that in this example, the first argument to @code{sort} is not
15617 quoted, since the expression must be evaluated so as to produce the
15618 list that is passed to @code{sort}.)
15621 @subsection Making a List of Files
15623 The @code{recursive-lengths-list-many-files} function requires a list
15624 of files as its argument. For our test examples, we constructed such
15625 a list by hand; but the Emacs Lisp source directory is too large for
15626 us to do for that. Instead, we will write a function to do the job
15627 for us. In this function, we will use both a @code{while} loop and a
15630 @findex directory-files
15631 We did not have to write a function like this for older versions of
15632 GNU Emacs, since they placed all the @samp{.el} files in one
15633 directory. Instead, we were able to use the @code{directory-files}
15634 function, which lists the names of files that match a specified
15635 pattern within a single directory.
15637 However, recent versions of Emacs place Emacs Lisp files in
15638 sub-directories of the top level @file{lisp} directory. This
15639 re-arrangement eases navigation. For example, all the mail related
15640 files are in a @file{lisp} sub-directory called @file{mail}. But at
15641 the same time, this arrangement forces us to create a file listing
15642 function that descends into the sub-directories.
15644 @findex files-in-below-directory
15645 We can create this function, called @code{files-in-below-directory},
15646 using familiar functions such as @code{car}, @code{nthcdr}, and
15647 @code{substring} in conjunction with an existing function called
15648 @code{directory-files-and-attributes}. This latter function not only
15649 lists all the filenames in a directory, including the names
15650 of sub-directories, but also their attributes.
15652 To restate our goal: to create a function that will enable us
15653 to feed filenames to @code{recursive-lengths-list-many-files}
15654 as a list that looks like this (but with more elements):
15658 ("./lisp/macros.el"
15659 "./lisp/mail/rmail.el"
15660 "./lisp/makesum.el")
15664 The @code{directory-files-and-attributes} function returns a list of
15665 lists. Each of the lists within the main list consists of 13
15666 elements. The first element is a string that contains the name of the
15667 file---which, in GNU/Linux, may be a `directory file', that is to
15668 say, a file with the special attributes of a directory. The second
15669 element of the list is @code{t} for a directory, a string
15670 for symbolic link (the string is the name linked to), or @code{nil}.
15672 For example, the first @samp{.el} file in the @file{lisp/} directory
15673 is @file{abbrev.el}. Its name is
15674 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15675 directory or a symbolic link.
15678 This is how @code{directory-files-and-attributes} lists that file and
15690 (20615 27034 579989 697000)
15692 (20615 26327 734791 805000)
15704 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15705 directory. The beginning of its listing looks like this:
15716 (To learn about the different attributes, look at the documentation of
15717 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15718 function does not list the filename, so its first element is
15719 @code{directory-files-and-attributes}'s second element.)
15721 We will want our new function, @code{files-in-below-directory}, to
15722 list the @samp{.el} files in the directory it is told to check, and in
15723 any directories below that directory.
15725 This gives us a hint on how to construct
15726 @code{files-in-below-directory}: within a directory, the function
15727 should add @samp{.el} filenames to a list; and if, within a directory,
15728 the function comes upon a sub-directory, it should go into that
15729 sub-directory and repeat its actions.
15731 However, we should note that every directory contains a name that
15732 refers to itself, called @file{.}, (``dot'') and a name that refers to
15733 its parent directory, called @file{..} (``double dot''). (In
15734 @file{/}, the root directory, @file{..} refers to itself, since
15735 @file{/} has no parent.) Clearly, we do not want our
15736 @code{files-in-below-directory} function to enter those directories,
15737 since they always lead us, directly or indirectly, to the current
15740 Consequently, our @code{files-in-below-directory} function must do
15745 Check to see whether it is looking at a filename that ends in
15746 @samp{.el}; and if so, add its name to a list.
15749 Check to see whether it is looking at a filename that is the name of a
15750 directory; and if so,
15754 Check to see whether it is looking at @file{.} or @file{..}; and if
15758 Or else, go into that directory and repeat the process.
15762 Let's write a function definition to do these tasks. We will use a
15763 @code{while} loop to move from one filename to another within a
15764 directory, checking what needs to be done; and we will use a recursive
15765 call to repeat the actions on each sub-directory. The recursive
15766 pattern is `accumulate'
15767 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15768 using @code{append} as the combiner.
15771 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15772 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15774 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15775 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15778 @c /usr/local/share/emacs/22.1.1/lisp/
15781 Here is the function:
15785 (defun files-in-below-directory (directory)
15786 "List the .el files in DIRECTORY and in its sub-directories."
15787 ;; Although the function will be used non-interactively,
15788 ;; it will be easier to test if we make it interactive.
15789 ;; The directory will have a name such as
15790 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15791 (interactive "DDirectory name: ")
15794 (let (el-files-list
15795 (current-directory-list
15796 (directory-files-and-attributes directory t)))
15797 ;; while we are in the current directory
15798 (while current-directory-list
15802 ;; check to see whether filename ends in `.el'
15803 ;; and if so, append its name to a list.
15804 ((equal ".el" (substring (car (car current-directory-list)) -3))
15805 (setq el-files-list
15806 (cons (car (car current-directory-list)) el-files-list)))
15809 ;; check whether filename is that of a directory
15810 ((eq t (car (cdr (car current-directory-list))))
15811 ;; decide whether to skip or recurse
15814 (substring (car (car current-directory-list)) -1))
15815 ;; then do nothing since filename is that of
15816 ;; current directory or parent, "." or ".."
15820 ;; else descend into the directory and repeat the process
15821 (setq el-files-list
15823 (files-in-below-directory
15824 (car (car current-directory-list)))
15826 ;; move to the next filename in the list; this also
15827 ;; shortens the list so the while loop eventually comes to an end
15828 (setq current-directory-list (cdr current-directory-list)))
15829 ;; return the filenames
15834 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15835 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15837 The @code{files-in-below-directory} @code{directory-files} function
15838 takes one argument, the name of a directory.
15841 Thus, on my system,
15843 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15845 @c !!! 22.1.1 lisp sources location here
15849 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15854 tells me that in and below my Lisp sources directory are 1031
15857 @code{files-in-below-directory} returns a list in reverse alphabetical
15858 order. An expression to sort the list in alphabetical order looks
15864 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15871 "Test how long it takes to find lengths of all sorted elisp defuns."
15872 (insert "\n" (current-time-string) "\n")
15875 (recursive-lengths-list-many-files
15876 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15878 (insert (format "%s" (current-time-string))))
15881 @node Counting function definitions
15882 @subsection Counting function definitions
15884 Our immediate goal is to generate a list that tells us how many
15885 function definitions contain fewer than 10 words and symbols, how many
15886 contain between 10 and 19 words and symbols, how many contain between
15887 20 and 29 words and symbols, and so on.
15889 With a sorted list of numbers, this is easy: count how many elements
15890 of the list are smaller than 10, then, after moving past the numbers
15891 just counted, count how many are smaller than 20, then, after moving
15892 past the numbers just counted, count how many are smaller than 30, and
15893 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
15894 larger than the top of that range. We can call the list of such
15895 numbers the @code{top-of-ranges} list.
15898 If we wished, we could generate this list automatically, but it is
15899 simpler to write a list manually. Here it is:
15900 @vindex top-of-ranges
15904 (defvar top-of-ranges
15907 110 120 130 140 150
15908 160 170 180 190 200
15909 210 220 230 240 250
15910 260 270 280 290 300)
15911 "List specifying ranges for `defuns-per-range'.")
15915 To change the ranges, we edit this list.
15917 Next, we need to write the function that creates the list of the
15918 number of definitions within each range. Clearly, this function must
15919 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
15922 The @code{defuns-per-range} function must do two things again and
15923 again: it must count the number of definitions within a range
15924 specified by the current top-of-range value; and it must shift to the
15925 next higher value in the @code{top-of-ranges} list after counting the
15926 number of definitions in the current range. Since each of these
15927 actions is repetitive, we can use @code{while} loops for the job.
15928 One loop counts the number of definitions in the range defined by the
15929 current top-of-range value, and the other loop selects each of the
15930 top-of-range values in turn.
15932 Several entries of the @code{sorted-lengths} list are counted for each
15933 range; this means that the loop for the @code{sorted-lengths} list
15934 will be inside the loop for the @code{top-of-ranges} list, like a
15935 small gear inside a big gear.
15937 The inner loop counts the number of definitions within the range. It
15938 is a simple counting loop of the type we have seen before.
15939 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
15940 The true-or-false test of the loop tests whether the value from the
15941 @code{sorted-lengths} list is smaller than the current value of the
15942 top of the range. If it is, the function increments the counter and
15943 tests the next value from the @code{sorted-lengths} list.
15946 The inner loop looks like this:
15950 (while @var{length-element-smaller-than-top-of-range}
15951 (setq number-within-range (1+ number-within-range))
15952 (setq sorted-lengths (cdr sorted-lengths)))
15956 The outer loop must start with the lowest value of the
15957 @code{top-of-ranges} list, and then be set to each of the succeeding
15958 higher values in turn. This can be done with a loop like this:
15962 (while top-of-ranges
15963 @var{body-of-loop}@dots{}
15964 (setq top-of-ranges (cdr top-of-ranges)))
15969 Put together, the two loops look like this:
15973 (while top-of-ranges
15975 ;; @r{Count the number of elements within the current range.}
15976 (while @var{length-element-smaller-than-top-of-range}
15977 (setq number-within-range (1+ number-within-range))
15978 (setq sorted-lengths (cdr sorted-lengths)))
15980 ;; @r{Move to next range.}
15981 (setq top-of-ranges (cdr top-of-ranges)))
15985 In addition, in each circuit of the outer loop, Emacs should record
15986 the number of definitions within that range (the value of
15987 @code{number-within-range}) in a list. We can use @code{cons} for
15988 this purpose. (@xref{cons, , @code{cons}}.)
15990 The @code{cons} function works fine, except that the list it
15991 constructs will contain the number of definitions for the highest
15992 range at its beginning and the number of definitions for the lowest
15993 range at its end. This is because @code{cons} attaches new elements
15994 of the list to the beginning of the list, and since the two loops are
15995 working their way through the lengths' list from the lower end first,
15996 the @code{defuns-per-range-list} will end up largest number first.
15997 But we will want to print our graph with smallest values first and the
15998 larger later. The solution is to reverse the order of the
15999 @code{defuns-per-range-list}. We can do this using the
16000 @code{nreverse} function, which reverses the order of a list.
16007 (nreverse '(1 2 3 4))
16018 Note that the @code{nreverse} function is ``destructive''---that is,
16019 it changes the list to which it is applied; this contrasts with the
16020 @code{car} and @code{cdr} functions, which are non-destructive. In
16021 this case, we do not want the original @code{defuns-per-range-list},
16022 so it does not matter that it is destroyed. (The @code{reverse}
16023 function provides a reversed copy of a list, leaving the original list
16028 Put all together, the @code{defuns-per-range} looks like this:
16032 (defun defuns-per-range (sorted-lengths top-of-ranges)
16033 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16034 (let ((top-of-range (car top-of-ranges))
16035 (number-within-range 0)
16036 defuns-per-range-list)
16041 (while top-of-ranges
16047 ;; @r{Need number for numeric test.}
16048 (car sorted-lengths)
16049 (< (car sorted-lengths) top-of-range))
16053 ;; @r{Count number of definitions within current range.}
16054 (setq number-within-range (1+ number-within-range))
16055 (setq sorted-lengths (cdr sorted-lengths)))
16057 ;; @r{Exit inner loop but remain within outer loop.}
16061 (setq defuns-per-range-list
16062 (cons number-within-range defuns-per-range-list))
16063 (setq number-within-range 0) ; @r{Reset count to zero.}
16067 ;; @r{Move to next range.}
16068 (setq top-of-ranges (cdr top-of-ranges))
16069 ;; @r{Specify next top of range value.}
16070 (setq top-of-range (car top-of-ranges)))
16074 ;; @r{Exit outer loop and count the number of defuns larger than}
16075 ;; @r{ the largest top-of-range value.}
16076 (setq defuns-per-range-list
16078 (length sorted-lengths)
16079 defuns-per-range-list))
16083 ;; @r{Return a list of the number of definitions within each range,}
16084 ;; @r{ smallest to largest.}
16085 (nreverse defuns-per-range-list)))
16091 The function is straightforward except for one subtle feature. The
16092 true-or-false test of the inner loop looks like this:
16096 (and (car sorted-lengths)
16097 (< (car sorted-lengths) top-of-range))
16103 instead of like this:
16106 (< (car sorted-lengths) top-of-range)
16109 The purpose of the test is to determine whether the first item in the
16110 @code{sorted-lengths} list is less than the value of the top of the
16113 The simple version of the test works fine unless the
16114 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16115 @code{(car sorted-lengths)} expression function returns
16116 @code{nil}. The @code{<} function cannot compare a number to
16117 @code{nil}, which is an empty list, so Emacs signals an error and
16118 stops the function from attempting to continue to execute.
16120 The @code{sorted-lengths} list always becomes @code{nil} when the
16121 counter reaches the end of the list. This means that any attempt to
16122 use the @code{defuns-per-range} function with the simple version of
16123 the test will fail.
16125 We solve the problem by using the @code{(car sorted-lengths)}
16126 expression in conjunction with the @code{and} expression. The
16127 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16128 value so long as the list has at least one number within it, but
16129 returns @code{nil} if the list is empty. The @code{and} expression
16130 first evaluates the @code{(car sorted-lengths)} expression, and
16131 if it is @code{nil}, returns false @emph{without} evaluating the
16132 @code{<} expression. But if the @code{(car sorted-lengths)}
16133 expression returns a non-@code{nil} value, the @code{and} expression
16134 evaluates the @code{<} expression, and returns that value as the value
16135 of the @code{and} expression.
16137 @c colon in printed section title causes problem in Info cross reference
16138 This way, we avoid an error.
16141 (For information about @code{and}, see
16142 @ref{kill-new function, , The @code{kill-new} function}.)
16146 (@xref{kill-new function, , The @code{kill-new} function}, for
16147 information about @code{and}.)
16150 Here is a short test of the @code{defuns-per-range} function. First,
16151 evaluate the expression that binds (a shortened)
16152 @code{top-of-ranges} list to the list of values, then evaluate the
16153 expression for binding the @code{sorted-lengths} list, and then
16154 evaluate the @code{defuns-per-range} function.
16158 ;; @r{(Shorter list than we will use later.)}
16159 (setq top-of-ranges
16160 '(110 120 130 140 150
16161 160 170 180 190 200))
16163 (setq sorted-lengths
16164 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16166 (defuns-per-range sorted-lengths top-of-ranges)
16172 The list returned looks like this:
16175 (2 2 2 0 0 1 0 2 0 0 4)
16179 Indeed, there are two elements of the @code{sorted-lengths} list
16180 smaller than 110, two elements between 110 and 119, two elements
16181 between 120 and 129, and so on. There are four elements with a value
16184 @c The next step is to turn this numbers' list into a graph.
16185 @node Readying a Graph
16186 @chapter Readying a Graph
16187 @cindex Readying a graph
16188 @cindex Graph prototype
16189 @cindex Prototype graph
16190 @cindex Body of graph
16192 Our goal is to construct a graph showing the numbers of function
16193 definitions of various lengths in the Emacs lisp sources.
16195 As a practical matter, if you were creating a graph, you would
16196 probably use a program such as @code{gnuplot} to do the job.
16197 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16198 however, we create one from scratch, and in the process we will
16199 re-acquaint ourselves with some of what we learned before and learn
16202 In this chapter, we will first write a simple graph printing function.
16203 This first definition will be a @dfn{prototype}, a rapidly written
16204 function that enables us to reconnoiter this unknown graph-making
16205 territory. We will discover dragons, or find that they are myth.
16206 After scouting the terrain, we will feel more confident and enhance
16207 the function to label the axes automatically.
16210 * Columns of a graph::
16211 * graph-body-print:: How to print the body of a graph.
16212 * recursive-graph-body-print::
16214 * Line Graph Exercise::
16218 @node Columns of a graph
16219 @unnumberedsec Printing the Columns of a Graph
16222 Since Emacs is designed to be flexible and work with all kinds of
16223 terminals, including character-only terminals, the graph will need to
16224 be made from one of the `typewriter' symbols. An asterisk will do; as
16225 we enhance the graph-printing function, we can make the choice of
16226 symbol a user option.
16228 We can call this function @code{graph-body-print}; it will take a
16229 @code{numbers-list} as its only argument. At this stage, we will not
16230 label the graph, but only print its body.
16232 The @code{graph-body-print} function inserts a vertical column of
16233 asterisks for each element in the @code{numbers-list}. The height of
16234 each line is determined by the value of that element of the
16235 @code{numbers-list}.
16237 Inserting columns is a repetitive act; that means that this function can
16238 be written either with a @code{while} loop or recursively.
16240 Our first challenge is to discover how to print a column of asterisks.
16241 Usually, in Emacs, we print characters onto a screen horizontally,
16242 line by line, by typing. We have two routes we can follow: write our
16243 own column-insertion function or discover whether one exists in Emacs.
16245 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16246 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16247 command, except that the latter finds only those functions that are
16248 commands. The @kbd{M-x apropos} command lists all symbols that match
16249 a regular expression, including functions that are not interactive.
16252 What we want to look for is some command that prints or inserts
16253 columns. Very likely, the name of the function will contain either
16254 the word `print' or the word `insert' or the word `column'.
16255 Therefore, we can simply type @kbd{M-x apropos RET
16256 print\|insert\|column RET} and look at the result. On my system, this
16257 command once too takes quite some time, and then produced a list of 79
16258 functions and variables. Now it does not take much time at all and
16259 produces a list of 211 functions and variables. Scanning down the
16260 list, the only function that looks as if it might do the job is
16261 @code{insert-rectangle}.
16264 Indeed, this is the function we want; its documentation says:
16269 Insert text of RECTANGLE with upper left corner at point.
16270 RECTANGLE's first line is inserted at point,
16271 its second line is inserted at a point vertically under point, etc.
16272 RECTANGLE should be a list of strings.
16273 After this command, the mark is at the upper left corner
16274 and point is at the lower right corner.
16278 We can run a quick test, to make sure it does what we expect of it.
16280 Here is the result of placing the cursor after the
16281 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16282 (@code{eval-last-sexp}). The function inserts the strings
16283 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16284 point. Also the function returns @code{nil}.
16288 (insert-rectangle '("first" "second" "third"))first
16295 Of course, we won't be inserting the text of the
16296 @code{insert-rectangle} expression itself into the buffer in which we
16297 are making the graph, but will call the function from our program. We
16298 shall, however, have to make sure that point is in the buffer at the
16299 place where the @code{insert-rectangle} function will insert its
16302 If you are reading this in Info, you can see how this works by
16303 switching to another buffer, such as the @file{*scratch*} buffer,
16304 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16305 @code{insert-rectangle} expression into the minibuffer at the prompt,
16306 and then typing @key{RET}. This causes Emacs to evaluate the
16307 expression in the minibuffer, but to use as the value of point the
16308 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16309 keybinding for @code{eval-expression}. Also, @code{nil} does not
16310 appear in the @file{*scratch*} buffer since the expression is
16311 evaluated in the minibuffer.)
16313 We find when we do this that point ends up at the end of the last
16314 inserted line---that is to say, this function moves point as a
16315 side-effect. If we were to repeat the command, with point at this
16316 position, the next insertion would be below and to the right of the
16317 previous insertion. We don't want this! If we are going to make a
16318 bar graph, the columns need to be beside each other.
16320 So we discover that each cycle of the column-inserting @code{while}
16321 loop must reposition point to the place we want it, and that place
16322 will be at the top, not the bottom, of the column. Moreover, we
16323 remember that when we print a graph, we do not expect all the columns
16324 to be the same height. This means that the top of each column may be
16325 at a different height from the previous one. We cannot simply
16326 reposition point to the same line each time, but moved over to the
16327 right---or perhaps we can@dots{}
16329 We are planning to make the columns of the bar graph out of asterisks.
16330 The number of asterisks in the column is the number specified by the
16331 current element of the @code{numbers-list}. We need to construct a
16332 list of asterisks of the right length for each call to
16333 @code{insert-rectangle}. If this list consists solely of the requisite
16334 number of asterisks, then we will have position point the right number
16335 of lines above the base for the graph to print correctly. This could
16338 Alternatively, if we can figure out some way to pass
16339 @code{insert-rectangle} a list of the same length each time, then we
16340 can place point on the same line each time, but move it over one
16341 column to the right for each new column. If we do this, however, some
16342 of the entries in the list passed to @code{insert-rectangle} must be
16343 blanks rather than asterisks. For example, if the maximum height of
16344 the graph is 5, but the height of the column is 3, then
16345 @code{insert-rectangle} requires an argument that looks like this:
16348 (" " " " "*" "*" "*")
16351 This last proposal is not so difficult, so long as we can determine
16352 the column height. There are two ways for us to specify the column
16353 height: we can arbitrarily state what it will be, which would work
16354 fine for graphs of that height; or we can search through the list of
16355 numbers and use the maximum height of the list as the maximum height
16356 of the graph. If the latter operation were difficult, then the former
16357 procedure would be easiest, but there is a function built into Emacs
16358 that determines the maximum of its arguments. We can use that
16359 function. The function is called @code{max} and it returns the
16360 largest of all its arguments, which must be numbers. Thus, for
16368 returns 7. (A corresponding function called @code{min} returns the
16369 smallest of all its arguments.)
16373 However, we cannot simply call @code{max} on the @code{numbers-list};
16374 the @code{max} function expects numbers as its argument, not a list of
16375 numbers. Thus, the following expression,
16378 (max '(3 4 6 5 7 3))
16383 produces the following error message;
16386 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16390 We need a function that passes a list of arguments to a function.
16391 This function is @code{apply}. This function `applies' its first
16392 argument (a function) to its remaining arguments, the last of which
16399 (apply 'max 3 4 7 3 '(4 8 5))
16405 (Incidentally, I don't know how you would learn of this function
16406 without a book such as this. It is possible to discover other
16407 functions, like @code{search-forward} or @code{insert-rectangle}, by
16408 guessing at a part of their names and then using @code{apropos}. Even
16409 though its base in metaphor is clear---`apply' its first argument to
16410 the rest---I doubt a novice would come up with that particular word
16411 when using @code{apropos} or other aid. Of course, I could be wrong;
16412 after all, the function was first named by someone who had to invent
16415 The second and subsequent arguments to @code{apply} are optional, so
16416 we can use @code{apply} to call a function and pass the elements of a
16417 list to it, like this, which also returns 8:
16420 (apply 'max '(4 8 5))
16423 This latter way is how we will use @code{apply}. The
16424 @code{recursive-lengths-list-many-files} function returns a numbers'
16425 list to which we can apply @code{max} (we could also apply @code{max} to
16426 the sorted numbers' list; it does not matter whether the list is
16430 Hence, the operation for finding the maximum height of the graph is this:
16433 (setq max-graph-height (apply 'max numbers-list))
16436 Now we can return to the question of how to create a list of strings
16437 for a column of the graph. Told the maximum height of the graph
16438 and the number of asterisks that should appear in the column, the
16439 function should return a list of strings for the
16440 @code{insert-rectangle} command to insert.
16442 Each column is made up of asterisks or blanks. Since the function is
16443 passed the value of the height of the column and the number of
16444 asterisks in the column, the number of blanks can be found by
16445 subtracting the number of asterisks from the height of the column.
16446 Given the number of blanks and the number of asterisks, two
16447 @code{while} loops can be used to construct the list:
16451 ;;; @r{First version.}
16452 (defun column-of-graph (max-graph-height actual-height)
16453 "Return list of strings that is one column of a graph."
16454 (let ((insert-list nil)
16455 (number-of-top-blanks
16456 (- max-graph-height actual-height)))
16460 ;; @r{Fill in asterisks.}
16461 (while (> actual-height 0)
16462 (setq insert-list (cons "*" insert-list))
16463 (setq actual-height (1- actual-height)))
16467 ;; @r{Fill in blanks.}
16468 (while (> number-of-top-blanks 0)
16469 (setq insert-list (cons " " insert-list))
16470 (setq number-of-top-blanks
16471 (1- number-of-top-blanks)))
16475 ;; @r{Return whole list.}
16480 If you install this function and then evaluate the following
16481 expression you will see that it returns the list as desired:
16484 (column-of-graph 5 3)
16492 (" " " " "*" "*" "*")
16495 As written, @code{column-of-graph} contains a major flaw: the symbols
16496 used for the blank and for the marked entries in the column are
16497 `hard-coded' as a space and asterisk. This is fine for a prototype,
16498 but you, or another user, may wish to use other symbols. For example,
16499 in testing the graph function, you many want to use a period in place
16500 of the space, to make sure the point is being repositioned properly
16501 each time the @code{insert-rectangle} function is called; or you might
16502 want to substitute a @samp{+} sign or other symbol for the asterisk.
16503 You might even want to make a graph-column that is more than one
16504 display column wide. The program should be more flexible. The way to
16505 do that is to replace the blank and the asterisk with two variables
16506 that we can call @code{graph-blank} and @code{graph-symbol} and define
16507 those variables separately.
16509 Also, the documentation is not well written. These considerations
16510 lead us to the second version of the function:
16514 (defvar graph-symbol "*"
16515 "String used as symbol in graph, usually an asterisk.")
16519 (defvar graph-blank " "
16520 "String used as blank in graph, usually a blank space.
16521 graph-blank must be the same number of columns wide
16527 (For an explanation of @code{defvar}, see
16528 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16532 ;;; @r{Second version.}
16533 (defun column-of-graph (max-graph-height actual-height)
16534 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16538 The graph-symbols are contiguous entries at the end
16540 The list will be inserted as one column of a graph.
16541 The strings are either graph-blank or graph-symbol."
16545 (let ((insert-list nil)
16546 (number-of-top-blanks
16547 (- max-graph-height actual-height)))
16551 ;; @r{Fill in @code{graph-symbols}.}
16552 (while (> actual-height 0)
16553 (setq insert-list (cons graph-symbol insert-list))
16554 (setq actual-height (1- actual-height)))
16558 ;; @r{Fill in @code{graph-blanks}.}
16559 (while (> number-of-top-blanks 0)
16560 (setq insert-list (cons graph-blank insert-list))
16561 (setq number-of-top-blanks
16562 (1- number-of-top-blanks)))
16564 ;; @r{Return whole list.}
16569 If we wished, we could rewrite @code{column-of-graph} a third time to
16570 provide optionally for a line graph as well as for a bar graph. This
16571 would not be hard to do. One way to think of a line graph is that it
16572 is no more than a bar graph in which the part of each bar that is
16573 below the top is blank. To construct a column for a line graph, the
16574 function first constructs a list of blanks that is one shorter than
16575 the value, then it uses @code{cons} to attach a graph symbol to the
16576 list; then it uses @code{cons} again to attach the `top blanks' to
16579 It is easy to see how to write such a function, but since we don't
16580 need it, we will not do it. But the job could be done, and if it were
16581 done, it would be done with @code{column-of-graph}. Even more
16582 important, it is worth noting that few changes would have to be made
16583 anywhere else. The enhancement, if we ever wish to make it, is
16586 Now, finally, we come to our first actual graph printing function.
16587 This prints the body of a graph, not the labels for the vertical and
16588 horizontal axes, so we can call this @code{graph-body-print}.
16590 @node graph-body-print
16591 @section The @code{graph-body-print} Function
16592 @findex graph-body-print
16594 After our preparation in the preceding section, the
16595 @code{graph-body-print} function is straightforward. The function
16596 will print column after column of asterisks and blanks, using the
16597 elements of a numbers' list to specify the number of asterisks in each
16598 column. This is a repetitive act, which means we can use a
16599 decrementing @code{while} loop or recursive function for the job. In
16600 this section, we will write the definition using a @code{while} loop.
16602 The @code{column-of-graph} function requires the height of the graph
16603 as an argument, so we should determine and record that as a local variable.
16605 This leads us to the following template for the @code{while} loop
16606 version of this function:
16610 (defun graph-body-print (numbers-list)
16611 "@var{documentation}@dots{}"
16612 (let ((height @dots{}
16617 (while numbers-list
16618 @var{insert-columns-and-reposition-point}
16619 (setq numbers-list (cdr numbers-list)))))
16624 We need to fill in the slots of the template.
16626 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16627 determine the height of the graph.
16629 The @code{while} loop will cycle through the @code{numbers-list} one
16630 element at a time. As it is shortened by the @code{(setq numbers-list
16631 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16632 list is the value of the argument for @code{column-of-graph}.
16634 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16635 function inserts the list returned by @code{column-of-graph}. Since
16636 the @code{insert-rectangle} function moves point to the lower right of
16637 the inserted rectangle, we need to save the location of point at the
16638 time the rectangle is inserted, move back to that position after the
16639 rectangle is inserted, and then move horizontally to the next place
16640 from which @code{insert-rectangle} is called.
16642 If the inserted columns are one character wide, as they will be if
16643 single blanks and asterisks are used, the repositioning command is
16644 simply @code{(forward-char 1)}; however, the width of a column may be
16645 greater than one. This means that the repositioning command should be
16646 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16647 itself is the length of a @code{graph-blank} and can be found using
16648 the expression @code{(length graph-blank)}. The best place to bind
16649 the @code{symbol-width} variable to the value of the width of graph
16650 column is in the varlist of the @code{let} expression.
16653 These considerations lead to the following function definition:
16657 (defun graph-body-print (numbers-list)
16658 "Print a bar graph of the NUMBERS-LIST.
16659 The numbers-list consists of the Y-axis values."
16661 (let ((height (apply 'max numbers-list))
16662 (symbol-width (length graph-blank))
16667 (while numbers-list
16668 (setq from-position (point))
16670 (column-of-graph height (car numbers-list)))
16671 (goto-char from-position)
16672 (forward-char symbol-width)
16675 ;; @r{Draw graph column by column.}
16677 (setq numbers-list (cdr numbers-list)))
16680 ;; @r{Place point for X axis labels.}
16681 (forward-line height)
16688 The one unexpected expression in this function is the
16689 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16690 expression makes the graph printing operation more interesting to
16691 watch than it would be otherwise. The expression causes Emacs to
16692 `sit' or do nothing for a zero length of time and then redraw the
16693 screen. Placed here, it causes Emacs to redraw the screen column by
16694 column. Without it, Emacs would not redraw the screen until the
16697 We can test @code{graph-body-print} with a short list of numbers.
16701 Install @code{graph-symbol}, @code{graph-blank},
16702 @code{column-of-graph}, which are in
16704 @ref{Readying a Graph, , Readying a Graph},
16707 @ref{Columns of a graph},
16709 and @code{graph-body-print}.
16713 Copy the following expression:
16716 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16720 Switch to the @file{*scratch*} buffer and place the cursor where you
16721 want the graph to start.
16724 Type @kbd{M-:} (@code{eval-expression}).
16727 Yank the @code{graph-body-print} expression into the minibuffer
16728 with @kbd{C-y} (@code{yank)}.
16731 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16735 Emacs will print a graph like this:
16749 @node recursive-graph-body-print
16750 @section The @code{recursive-graph-body-print} Function
16751 @findex recursive-graph-body-print
16753 The @code{graph-body-print} function may also be written recursively.
16754 The recursive solution is divided into two parts: an outside `wrapper'
16755 that uses a @code{let} expression to determine the values of several
16756 variables that need only be found once, such as the maximum height of
16757 the graph, and an inside function that is called recursively to print
16761 The `wrapper' is uncomplicated:
16765 (defun recursive-graph-body-print (numbers-list)
16766 "Print a bar graph of the NUMBERS-LIST.
16767 The numbers-list consists of the Y-axis values."
16768 (let ((height (apply 'max numbers-list))
16769 (symbol-width (length graph-blank))
16771 (recursive-graph-body-print-internal
16778 The recursive function is a little more difficult. It has four parts:
16779 the `do-again-test', the printing code, the recursive call, and the
16780 `next-step-expression'. The `do-again-test' is a @code{when}
16781 expression that determines whether the @code{numbers-list} contains
16782 any remaining elements; if it does, the function prints one column of
16783 the graph using the printing code and calls itself again. The
16784 function calls itself again according to the value produced by the
16785 `next-step-expression' which causes the call to act on a shorter
16786 version of the @code{numbers-list}.
16790 (defun recursive-graph-body-print-internal
16791 (numbers-list height symbol-width)
16792 "Print a bar graph.
16793 Used within recursive-graph-body-print function."
16798 (setq from-position (point))
16800 (column-of-graph height (car numbers-list)))
16803 (goto-char from-position)
16804 (forward-char symbol-width)
16805 (sit-for 0) ; @r{Draw graph column by column.}
16806 (recursive-graph-body-print-internal
16807 (cdr numbers-list) height symbol-width)))
16812 After installation, this expression can be tested; here is a sample:
16815 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16819 Here is what @code{recursive-graph-body-print} produces:
16833 Either of these two functions, @code{graph-body-print} or
16834 @code{recursive-graph-body-print}, create the body of a graph.
16837 @section Need for Printed Axes
16839 A graph needs printed axes, so you can orient yourself. For a do-once
16840 project, it may be reasonable to draw the axes by hand using Emacs's
16841 Picture mode; but a graph drawing function may be used more than once.
16843 For this reason, I have written enhancements to the basic
16844 @code{print-graph-body} function that automatically print labels for
16845 the horizontal and vertical axes. Since the label printing functions
16846 do not contain much new material, I have placed their description in
16847 an appendix. @xref{Full Graph, , A Graph with Labeled Axes}.
16849 @node Line Graph Exercise
16852 Write a line graph version of the graph printing functions.
16854 @node Emacs Initialization
16855 @chapter Your @file{.emacs} File
16856 @cindex @file{.emacs} file
16857 @cindex Customizing your @file{.emacs} file
16858 @cindex Initialization file
16860 ``You don't have to like Emacs to like it''---this seemingly
16861 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16862 the box' Emacs is a generic tool. Most people who use it, customize
16863 it to suit themselves.
16865 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16866 expressions in Emacs Lisp you can change or extend Emacs.
16869 * Default Configuration::
16870 * Site-wide Init:: You can write site-wide init files.
16871 * defcustom:: Emacs will write code for you.
16872 * Beginning a .emacs File:: How to write a @code{.emacs file}.
16873 * Text and Auto-fill:: Automatically wrap lines.
16874 * Mail Aliases:: Use abbreviations for email addresses.
16875 * Indent Tabs Mode:: Don't use tabs with @TeX{}
16876 * Keybindings:: Create some personal keybindings.
16877 * Keymaps:: More about key binding.
16878 * Loading Files:: Load (i.e., evaluate) files automatically.
16879 * Autoload:: Make functions available.
16880 * Simple Extension:: Define a function; bind it to a key.
16881 * X11 Colors:: Colors in X.
16883 * Mode Line:: How to customize your mode line.
16887 @node Default Configuration
16888 @unnumberedsec Emacs's Default Configuration
16891 There are those who appreciate Emacs's default configuration. After
16892 all, Emacs starts you in C mode when you edit a C file, starts you in
16893 Fortran mode when you edit a Fortran file, and starts you in
16894 Fundamental mode when you edit an unadorned file. This all makes
16895 sense, if you do not know who is going to use Emacs. Who knows what a
16896 person hopes to do with an unadorned file? Fundamental mode is the
16897 right default for such a file, just as C mode is the right default for
16898 editing C code. (Enough programming languages have syntaxes
16899 that enable them to share or nearly share features, so C mode is
16900 now provided by CC mode, the `C Collection'.)
16902 But when you do know who is going to use Emacs---you,
16903 yourself---then it makes sense to customize Emacs.
16905 For example, I seldom want Fundamental mode when I edit an
16906 otherwise undistinguished file; I want Text mode. This is why I
16907 customize Emacs: so it suits me.
16909 You can customize and extend Emacs by writing or adapting a
16910 @file{~/.emacs} file. This is your personal initialization file; its
16911 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
16912 may also add @file{.el} to @file{~/.emacs} and call it a
16913 @file{~/.emacs.el} file. In the past, you were forbidden to type the
16914 extra keystrokes that the name @file{~/.emacs.el} requires, but now
16915 you may. The new format is consistent with the Emacs Lisp file
16916 naming conventions; the old format saves typing.}
16918 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
16919 code yourself; or you can use Emacs's @code{customize} feature to write
16920 the code for you. You can combine your own expressions and
16921 auto-written Customize expressions in your @file{.emacs} file.
16923 (I myself prefer to write my own expressions, except for those,
16924 particularly fonts, that I find easier to manipulate using the
16925 @code{customize} command. I combine the two methods.)
16927 Most of this chapter is about writing expressions yourself. It
16928 describes a simple @file{.emacs} file; for more information, see
16929 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
16930 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
16933 @node Site-wide Init
16934 @section Site-wide Initialization Files
16936 @cindex @file{default.el} init file
16937 @cindex @file{site-init.el} init file
16938 @cindex @file{site-load.el} init file
16939 In addition to your personal initialization file, Emacs automatically
16940 loads various site-wide initialization files, if they exist. These
16941 have the same form as your @file{.emacs} file, but are loaded by
16944 Two site-wide initialization files, @file{site-load.el} and
16945 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
16946 `dumped' version of Emacs is created, as is most common. (Dumped
16947 copies of Emacs load more quickly. However, once a file is loaded and
16948 dumped, a change to it does not lead to a change in Emacs unless you
16949 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
16950 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
16951 @file{INSTALL} file.)
16953 Three other site-wide initialization files are loaded automatically
16954 each time you start Emacs, if they exist. These are
16955 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
16956 file, and @file{default.el}, and the terminal type file, which are both
16957 loaded @emph{after} your @file{.emacs} file.
16959 Settings and definitions in your @file{.emacs} file will overwrite
16960 conflicting settings and definitions in a @file{site-start.el} file,
16961 if it exists; but the settings and definitions in a @file{default.el}
16962 or terminal type file will overwrite those in your @file{.emacs} file.
16963 (You can prevent interference from a terminal type file by setting
16964 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
16965 Simple Extension}.)
16967 @c Rewritten to avoid overfull hbox.
16968 The @file{INSTALL} file that comes in the distribution contains
16969 descriptions of the @file{site-init.el} and @file{site-load.el} files.
16971 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
16972 control loading. These files are in the @file{lisp} directory of the
16973 Emacs distribution and are worth perusing.
16975 The @file{loaddefs.el} file contains a good many suggestions as to
16976 what to put into your own @file{.emacs} file, or into a site-wide
16977 initialization file.
16980 @section Specifying Variables using @code{defcustom}
16983 You can specify variables using @code{defcustom} so that you and
16984 others can then use Emacs's @code{customize} feature to set their
16985 values. (You cannot use @code{customize} to write function
16986 definitions; but you can write @code{defuns} in your @file{.emacs}
16987 file. Indeed, you can write any Lisp expression in your @file{.emacs}
16990 The @code{customize} feature depends on the @code{defcustom} macro.
16991 Although you can use @code{defvar} or @code{setq} for variables that
16992 users set, the @code{defcustom} macro is designed for the job.
16994 You can use your knowledge of @code{defvar} for writing the
16995 first three arguments for @code{defcustom}. The first argument to
16996 @code{defcustom} is the name of the variable. The second argument is
16997 the variable's initial value, if any; and this value is set only if
16998 the value has not already been set. The third argument is the
17001 The fourth and subsequent arguments to @code{defcustom} specify types
17002 and options; these are not featured in @code{defvar}. (These
17003 arguments are optional.)
17005 Each of these arguments consists of a keyword followed by a value.
17006 Each keyword starts with the colon character @samp{:}.
17009 For example, the customizable user option variable
17010 @code{text-mode-hook} looks like this:
17014 (defcustom text-mode-hook nil
17015 "Normal hook run when entering Text mode and many related modes."
17017 :options '(turn-on-auto-fill flyspell-mode)
17023 The name of the variable is @code{text-mode-hook}; it has no default
17024 value; and its documentation string tells you what it does.
17026 The @code{:type} keyword tells Emacs the kind of data to which
17027 @code{text-mode-hook} should be set and how to display the value in a
17028 Customization buffer.
17030 The @code{:options} keyword specifies a suggested list of values for
17031 the variable. Usually, @code{:options} applies to a hook.
17032 The list is only a suggestion; it is not exclusive; a person who sets
17033 the variable may set it to other values; the list shown following the
17034 @code{:options} keyword is intended to offer convenient choices to a
17037 Finally, the @code{:group} keyword tells the Emacs Customization
17038 command in which group the variable is located. This tells where to
17041 The @code{defcustom} macro recognizes more than a dozen keywords.
17042 For more information, see @ref{Customization, , Writing Customization
17043 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17045 Consider @code{text-mode-hook} as an example.
17047 There are two ways to customize this variable. You can use the
17048 customization command or write the appropriate expressions yourself.
17051 Using the customization command, you can type:
17058 and find that the group for editing files of data is called `data'.
17059 Enter that group. Text Mode Hook is the first member. You can click
17060 on its various options, such as @code{turn-on-auto-fill}, to set the
17061 values. After you click on the button to
17064 Save for Future Sessions
17068 Emacs will write an expression into your @file{.emacs} file.
17069 It will look like this:
17073 (custom-set-variables
17074 ;; custom-set-variables was added by Custom.
17075 ;; If you edit it by hand, you could mess it up, so be careful.
17076 ;; Your init file should contain only one such instance.
17077 ;; If there is more than one, they won't work right.
17078 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17083 (The @code{text-mode-hook-identify} function tells
17084 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17085 It comes on automatically.)
17087 The @code{custom-set-variables} function works somewhat differently
17088 than a @code{setq}. While I have never learned the differences, I
17089 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17090 file by hand: I make the changes in what appears to me to be a
17091 reasonable manner and have not had any problems. Others prefer to use
17092 the Customization command and let Emacs do the work for them.
17094 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17095 This function sets the various font faces. Over time, I have set a
17096 considerable number of faces. Some of the time, I re-set them using
17097 @code{customize}; other times, I simply edit the
17098 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17100 The second way to customize your @code{text-mode-hook} is to set it
17101 yourself in your @file{.emacs} file using code that has nothing to do
17102 with the @code{custom-set-@dots{}} functions.
17105 When you do this, and later use @code{customize}, you will see a
17109 CHANGED outside Customize; operating on it here may be unreliable.
17113 This message is only a warning. If you click on the button to
17116 Save for Future Sessions
17120 Emacs will write a @code{custom-set-@dots{}} expression near the end
17121 of your @file{.emacs} file that will be evaluated after your
17122 hand-written expression. It will, therefore, overrule your
17123 hand-written expression. No harm will be done. When you do this,
17124 however, be careful to remember which expression is active; if you
17125 forget, you may confuse yourself.
17127 So long as you remember where the values are set, you will have no
17128 trouble. In any event, the values are always set in your
17129 initialization file, which is usually called @file{.emacs}.
17131 I myself use @code{customize} for hardly anything. Mostly, I write
17132 expressions myself.
17136 Incidentally, to be more complete concerning defines: @code{defsubst}
17137 defines an inline function. The syntax is just like that of
17138 @code{defun}. @code{defconst} defines a symbol as a constant. The
17139 intent is that neither programs nor users should ever change a value
17140 set by @code{defconst}. (You can change it; the value set is a
17141 variable; but please do not.)
17143 @node Beginning a .emacs File
17144 @section Beginning a @file{.emacs} File
17145 @cindex @file{.emacs} file, beginning of
17147 When you start Emacs, it loads your @file{.emacs} file unless you tell
17148 it not to by specifying @samp{-q} on the command line. (The
17149 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17151 A @file{.emacs} file contains Lisp expressions. Often, these are no
17152 more than expressions to set values; sometimes they are function
17155 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17156 Manual}, for a short description of initialization files.
17158 This chapter goes over some of the same ground, but is a walk among
17159 extracts from a complete, long-used @file{.emacs} file---my own.
17161 The first part of the file consists of comments: reminders to myself.
17162 By now, of course, I remember these things, but when I started, I did
17168 ;;;; Bob's .emacs file
17169 ; Robert J. Chassell
17170 ; 26 September 1985
17175 Look at that date! I started this file a long time ago. I have been
17176 adding to it ever since.
17180 ; Each section in this file is introduced by a
17181 ; line beginning with four semicolons; and each
17182 ; entry is introduced by a line beginning with
17183 ; three semicolons.
17188 This describes the usual conventions for comments in Emacs Lisp.
17189 Everything on a line that follows a semicolon is a comment. Two,
17190 three, and four semicolons are used as subsection and section markers.
17191 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17192 more about comments.)
17197 ; Control-h is the help key;
17198 ; after typing control-h, type a letter to
17199 ; indicate the subject about which you want help.
17200 ; For an explanation of the help facility,
17201 ; type control-h two times in a row.
17206 Just remember: type @kbd{C-h} two times for help.
17210 ; To find out about any mode, type control-h m
17211 ; while in that mode. For example, to find out
17212 ; about mail mode, enter mail mode and then type
17218 `Mode help', as I call this, is very helpful. Usually, it tells you
17219 all you need to know.
17221 Of course, you don't need to include comments like these in your
17222 @file{.emacs} file. I included them in mine because I kept forgetting
17223 about Mode help or the conventions for comments---but I was able to
17224 remember to look here to remind myself.
17226 @node Text and Auto-fill
17227 @section Text and Auto Fill Mode
17229 Now we come to the part that `turns on' Text mode and
17234 ;;; Text mode and Auto Fill mode
17235 ;; The next two lines put Emacs into Text mode
17236 ;; and Auto Fill mode, and are for writers who
17237 ;; want to start writing prose rather than code.
17238 (setq-default major-mode 'text-mode)
17239 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17243 Here is the first part of this @file{.emacs} file that does something
17244 besides remind a forgetful human!
17246 The first of the two lines in parentheses tells Emacs to turn on Text
17247 mode when you find a file, @emph{unless} that file should go into some
17248 other mode, such as C mode.
17250 @cindex Per-buffer, local variables list
17251 @cindex Local variables list, per-buffer,
17252 @cindex Automatic mode selection
17253 @cindex Mode selection, automatic
17254 When Emacs reads a file, it looks at the extension to the file name,
17255 if any. (The extension is the part that comes after a @samp{.}.) If
17256 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17257 on C mode. Also, Emacs looks at first nonblank line of the file; if
17258 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17259 possesses a list of extensions and specifications that it uses
17260 automatically. In addition, Emacs looks near the last page for a
17261 per-buffer, ``local variables list'', if any.
17264 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17267 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17271 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17272 Files'' in @cite{The GNU Emacs Manual}.
17275 Now, back to the @file{.emacs} file.
17278 Here is the line again; how does it work?
17280 @cindex Text Mode turned on
17282 (setq major-mode 'text-mode)
17286 This line is a short, but complete Emacs Lisp expression.
17288 We are already familiar with @code{setq}. It sets the following variable,
17289 @code{major-mode}, to the subsequent value, which is @code{text-mode}.
17290 The single quote mark before @code{text-mode} tells Emacs to deal directly
17291 with the @code{text-mode} symbol, not with whatever it might stand for.
17292 @xref{set & setq, , Setting the Value of a Variable},
17293 for a reminder of how @code{setq} works.
17294 The main point is that there is no difference between the procedure you
17295 use to set a value in your @file{.emacs} file and the procedure you use
17296 anywhere else in Emacs.
17299 Here is the next line:
17301 @cindex Auto Fill mode turned on
17304 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17308 In this line, the @code{add-hook} command adds
17309 @code{turn-on-auto-fill} to the variable.
17311 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17312 it!, turns on Auto Fill mode.
17314 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17315 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17316 turns on Auto Fill mode.
17318 In brief, the first line causes Emacs to enter Text mode when you edit a
17319 file, unless the file name extension, a first non-blank line, or local
17320 variables to tell Emacs otherwise.
17322 Text mode among other actions, sets the syntax table to work
17323 conveniently for writers. In Text mode, Emacs considers an apostrophe
17324 as part of a word like a letter; but Emacs does not consider a period
17325 or a space as part of a word. Thus, @kbd{M-f} moves you over
17326 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17327 the @samp{t} of @samp{it's}.
17329 The second line causes Emacs to turn on Auto Fill mode when it turns
17330 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17331 that is too wide and brings the excessively wide part of the line down
17332 to the next line. Emacs breaks lines between words, not within them.
17334 When Auto Fill mode is turned off, lines continue to the right as you
17335 type them. Depending on how you set the value of
17336 @code{truncate-lines}, the words you type either disappear off the
17337 right side of the screen, or else are shown, in a rather ugly and
17338 unreadable manner, as a continuation line on the screen.
17341 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17342 fill commands to insert two spaces after a colon:
17345 (setq colon-double-space t)
17349 @section Mail Aliases
17351 Here is a @code{setq} that `turns on' mail aliases, along with more
17357 ; To enter mail mode, type `C-x m'
17358 ; To enter RMAIL (for reading mail),
17360 (setq mail-aliases t)
17364 @cindex Mail aliases
17366 This @code{setq} command sets the value of the variable
17367 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17368 says, in effect, ``Yes, use mail aliases.''
17370 Mail aliases are convenient short names for long email addresses or
17371 for lists of email addresses. The file where you keep your `aliases'
17372 is @file{~/.mailrc}. You write an alias like this:
17375 alias geo george@@foobar.wiz.edu
17379 When you write a message to George, address it to @samp{geo}; the
17380 mailer will automatically expand @samp{geo} to the full address.
17382 @node Indent Tabs Mode
17383 @section Indent Tabs Mode
17384 @cindex Tabs, preventing
17385 @findex indent-tabs-mode
17387 By default, Emacs inserts tabs in place of multiple spaces when it
17388 formats a region. (For example, you might indent many lines of text
17389 all at once with the @code{indent-region} command.) Tabs look fine on
17390 a terminal or with ordinary printing, but they produce badly indented
17391 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17394 The following turns off Indent Tabs mode:
17398 ;;; Prevent Extraneous Tabs
17399 (setq-default indent-tabs-mode nil)
17403 Note that this line uses @code{setq-default} rather than the
17404 @code{setq} command that we have seen before. The @code{setq-default}
17405 command sets values only in buffers that do not have their own local
17406 values for the variable.
17409 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17411 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17415 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17416 Files'' in @cite{The GNU Emacs Manual}.
17421 @section Some Keybindings
17423 Now for some personal keybindings:
17427 ;;; Compare windows
17428 (global-set-key "\C-cw" 'compare-windows)
17432 @findex compare-windows
17433 @code{compare-windows} is a nifty command that compares the text in
17434 your current window with text in the next window. It makes the
17435 comparison by starting at point in each window, moving over text in
17436 each window as far as they match. I use this command all the time.
17438 This also shows how to set a key globally, for all modes.
17440 @cindex Setting a key globally
17441 @cindex Global set key
17442 @cindex Key setting globally
17443 @findex global-set-key
17444 The command is @code{global-set-key}. It is followed by the
17445 keybinding. In a @file{.emacs} file, the keybinding is written as
17446 shown: @code{\C-c} stands for `control-c', which means `press the
17447 control key and the @key{c} key at the same time'. The @code{w} means
17448 `press the @key{w} key'. The keybinding is surrounded by double
17449 quotation marks. In documentation, you would write this as
17450 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17451 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17452 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17453 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17456 The command invoked by the keys is @code{compare-windows}. Note that
17457 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17458 would first try to evaluate the symbol to determine its value.
17460 These three things, the double quotation marks, the backslash before
17461 the @samp{C}, and the single quote mark are necessary parts of
17462 keybinding that I tend to forget. Fortunately, I have come to
17463 remember that I should look at my existing @file{.emacs} file, and
17464 adapt what is there.
17466 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17467 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17468 set of keys, @kbd{C-c} followed by a single character, is strictly
17469 reserved for individuals' own use. (I call these `own' keys, since
17470 these are for my own use.) You should always be able to create such a
17471 keybinding for your own use without stomping on someone else's
17472 keybinding. If you ever write an extension to Emacs, please avoid
17473 taking any of these keys for public use. Create a key like @kbd{C-c
17474 C-w} instead. Otherwise, we will run out of `own' keys.
17477 Here is another keybinding, with a comment:
17481 ;;; Keybinding for `occur'
17482 ; I use occur a lot, so let's bind it to a key:
17483 (global-set-key "\C-co" 'occur)
17488 The @code{occur} command shows all the lines in the current buffer
17489 that contain a match for a regular expression. Matching lines are
17490 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17491 to jump to occurrences.
17493 @findex global-unset-key
17494 @cindex Unbinding key
17495 @cindex Key unbinding
17497 Here is how to unbind a key, so it does not
17503 (global-unset-key "\C-xf")
17507 There is a reason for this unbinding: I found I inadvertently typed
17508 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17509 file, as I intended, I accidentally set the width for filled text,
17510 almost always to a width I did not want. Since I hardly ever reset my
17511 default width, I simply unbound the key.
17513 @findex list-buffers, @r{rebound}
17514 @findex buffer-menu, @r{bound to key}
17516 The following rebinds an existing key:
17520 ;;; Rebind `C-x C-b' for `buffer-menu'
17521 (global-set-key "\C-x\C-b" 'buffer-menu)
17525 By default, @kbd{C-x C-b} runs the
17526 @code{list-buffers} command. This command lists
17527 your buffers in @emph{another} window. Since I
17528 almost always want to do something in that
17529 window, I prefer the @code{buffer-menu}
17530 command, which not only lists the buffers,
17531 but moves point into that window.
17536 @cindex Rebinding keys
17538 Emacs uses @dfn{keymaps} to record which keys call which commands.
17539 When you use @code{global-set-key} to set the keybinding for a single
17540 command in all parts of Emacs, you are specifying the keybinding in
17541 @code{current-global-map}.
17543 Specific modes, such as C mode or Text mode, have their own keymaps;
17544 the mode-specific keymaps override the global map that is shared by
17547 The @code{global-set-key} function binds, or rebinds, the global
17548 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17549 function @code{buffer-menu}:
17552 (global-set-key "\C-x\C-b" 'buffer-menu)
17555 Mode-specific keymaps are bound using the @code{define-key} function,
17556 which takes a specific keymap as an argument, as well as the key and
17557 the command. For example, my @file{.emacs} file contains the
17558 following expression to bind the @code{texinfo-insert-@@group} command
17559 to @kbd{C-c C-c g}:
17563 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17568 The @code{texinfo-insert-@@group} function itself is a little extension
17569 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17570 use this command all the time and prefer to type the three strokes
17571 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17572 (@samp{@@group} and its matching @samp{@@end group} are commands that
17573 keep all enclosed text together on one page; many multi-line examples
17574 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17577 Here is the @code{texinfo-insert-@@group} function definition:
17581 (defun texinfo-insert-@@group ()
17582 "Insert the string @@group in a Texinfo buffer."
17584 (beginning-of-line)
17585 (insert "@@group\n"))
17589 (Of course, I could have used Abbrev mode to save typing, rather than
17590 write a function to insert a word; but I prefer key strokes consistent
17591 with other Texinfo mode key bindings.)
17593 You will see numerous @code{define-key} expressions in
17594 @file{loaddefs.el} as well as in the various mode libraries, such as
17595 @file{cc-mode.el} and @file{lisp-mode.el}.
17597 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17598 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17599 Reference Manual}, for more information about keymaps.
17601 @node Loading Files
17602 @section Loading Files
17603 @cindex Loading files
17606 Many people in the GNU Emacs community have written extensions to
17607 Emacs. As time goes by, these extensions are often included in new
17608 releases. For example, the Calendar and Diary packages are now part
17609 of the standard GNU Emacs, as is Calc.
17611 You can use a @code{load} command to evaluate a complete file and
17612 thereby install all the functions and variables in the file into Emacs.
17615 @c (auto-compression-mode t)
17618 (load "~/emacs/slowsplit")
17621 This evaluates, i.e., loads, the @file{slowsplit.el} file or if it
17622 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17623 @file{emacs} sub-directory of your home directory. The file contains
17624 the function @code{split-window-quietly}, which John Robinson wrote in
17627 The @code{split-window-quietly} function splits a window with the
17628 minimum of redisplay. I installed it in 1989 because it worked well
17629 with the slow 1200 baud terminals I was then using. Nowadays, I only
17630 occasionally come across such a slow connection, but I continue to use
17631 the function because I like the way it leaves the bottom half of a
17632 buffer in the lower of the new windows and the top half in the upper
17636 To replace the key binding for the default
17637 @code{split-window-vertically}, you must also unset that key and bind
17638 the keys to @code{split-window-quietly}, like this:
17642 (global-unset-key "\C-x2")
17643 (global-set-key "\C-x2" 'split-window-quietly)
17648 If you load many extensions, as I do, then instead of specifying the
17649 exact location of the extension file, as shown above, you can specify
17650 that directory as part of Emacs's @code{load-path}. Then, when Emacs
17651 loads a file, it will search that directory as well as its default
17652 list of directories. (The default list is specified in @file{paths.h}
17653 when Emacs is built.)
17656 The following command adds your @file{~/emacs} directory to the
17657 existing load path:
17661 ;;; Emacs Load Path
17662 (setq load-path (cons "~/emacs" load-path))
17666 Incidentally, @code{load-library} is an interactive interface to the
17667 @code{load} function. The complete function looks like this:
17669 @findex load-library
17672 (defun load-library (library)
17673 "Load the library named LIBRARY.
17674 This is an interface to the function `load'."
17676 (list (completing-read "Load library: "
17677 (apply-partially 'locate-file-completion-table
17679 (get-load-suffixes)))))
17684 The name of the function, @code{load-library}, comes from the use of
17685 `library' as a conventional synonym for `file'. The source for the
17686 @code{load-library} command is in the @file{files.el} library.
17688 Another interactive command that does a slightly different job is
17689 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17690 Emacs, emacs, The GNU Emacs Manual}, for information on the
17691 distinction between @code{load-library} and this command.
17694 @section Autoloading
17697 Instead of installing a function by loading the file that contains it,
17698 or by evaluating the function definition, you can make the function
17699 available but not actually install it until it is first called. This
17700 is called @dfn{autoloading}.
17702 When you execute an autoloaded function, Emacs automatically evaluates
17703 the file that contains the definition, and then calls the function.
17705 Emacs starts quicker with autoloaded functions, since their libraries
17706 are not loaded right away; but you need to wait a moment when you
17707 first use such a function, while its containing file is evaluated.
17709 Rarely used functions are frequently autoloaded. The
17710 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17711 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17712 come to use a `rare' function frequently. When you do, you should
17713 load that function's file with a @code{load} expression in your
17714 @file{.emacs} file.
17716 In my @file{.emacs} file, I load 14 libraries that contain functions
17717 that would otherwise be autoloaded. (Actually, it would have been
17718 better to include these files in my `dumped' Emacs, but I forgot.
17719 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17720 Reference Manual}, and the @file{INSTALL} file for more about
17723 You may also want to include autoloaded expressions in your @file{.emacs}
17724 file. @code{autoload} is a built-in function that takes up to five
17725 arguments, the final three of which are optional. The first argument
17726 is the name of the function to be autoloaded; the second is the name
17727 of the file to be loaded. The third argument is documentation for the
17728 function, and the fourth tells whether the function can be called
17729 interactively. The fifth argument tells what type of
17730 object---@code{autoload} can handle a keymap or macro as well as a
17731 function (the default is a function).
17734 Here is a typical example:
17738 (autoload 'html-helper-mode
17739 "html-helper-mode" "Edit HTML documents" t)
17744 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17745 which is a standard part of the distribution.)
17748 This expression autoloads the @code{html-helper-mode} function. It
17749 takes it from the @file{html-helper-mode.el} file (or from the byte
17750 compiled version @file{html-helper-mode.elc}, if that exists.) The
17751 file must be located in a directory specified by @code{load-path}.
17752 The documentation says that this is a mode to help you edit documents
17753 written in the HyperText Markup Language. You can call this mode
17754 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17755 duplicate the function's regular documentation in the autoload
17756 expression because the regular function is not yet loaded, so its
17757 documentation is not available.)
17759 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17760 Manual}, for more information.
17762 @node Simple Extension
17763 @section A Simple Extension: @code{line-to-top-of-window}
17764 @findex line-to-top-of-window
17765 @cindex Simple extension in @file{.emacs} file
17767 Here is a simple extension to Emacs that moves the line point is on to
17768 the top of the window. I use this all the time, to make text easier
17771 You can put the following code into a separate file and then load it
17772 from your @file{.emacs} file, or you can include it within your
17773 @file{.emacs} file.
17776 Here is the definition:
17780 ;;; Line to top of window;
17781 ;;; replace three keystroke sequence C-u 0 C-l
17782 (defun line-to-top-of-window ()
17783 "Move the line point is on to top of window."
17790 Now for the keybinding.
17792 Nowadays, function keys as well as mouse button events and
17793 non-@sc{ascii} characters are written within square brackets, without
17794 quotation marks. (In Emacs version 18 and before, you had to write
17795 different function key bindings for each different make of terminal.)
17797 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17801 (global-set-key [f6] 'line-to-top-of-window)
17804 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17805 Your Init File, emacs, The GNU Emacs Manual}.
17807 @cindex Conditional 'twixt two versions of Emacs
17808 @cindex Version of Emacs, choosing
17809 @cindex Emacs version, choosing
17810 If you run two versions of GNU Emacs, such as versions 22 and 23, and
17811 use one @file{.emacs} file, you can select which code to evaluate with
17812 the following conditional:
17817 ((= 22 emacs-major-version)
17818 ;; evaluate version 22 code
17820 ((= 23 emacs-major-version)
17821 ;; evaluate version 23 code
17826 For example, recent versions blink
17827 their cursors by default. I hate such blinking, as well as other
17828 features, so I placed the following in my @file{.emacs}
17829 file@footnote{When I start instances of Emacs that do not load my
17830 @file{.emacs} file or any site file, I also turn off blinking:
17833 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17835 @exdent Or nowadays, using an even more sophisticated set of options,
17843 (when (>= emacs-major-version 21)
17844 (blink-cursor-mode 0)
17845 ;; Insert newline when you press `C-n' (next-line)
17846 ;; at the end of the buffer
17847 (setq next-line-add-newlines t)
17850 ;; Turn on image viewing
17851 (auto-image-file-mode t)
17854 ;; Turn on menu bar (this bar has text)
17855 ;; (Use numeric argument to turn on)
17859 ;; Turn off tool bar (this bar has icons)
17860 ;; (Use numeric argument to turn on)
17861 (tool-bar-mode nil)
17864 ;; Turn off tooltip mode for tool bar
17865 ;; (This mode causes icon explanations to pop up)
17866 ;; (Use numeric argument to turn on)
17868 ;; If tooltips turned on, make tips appear promptly
17869 (setq tooltip-delay 0.1) ; default is 0.7 second
17875 @section X11 Colors
17877 You can specify colors when you use Emacs with the MIT X Windowing
17880 I dislike the default colors and specify my own.
17883 Here are the expressions in my @file{.emacs}
17884 file that set values:
17888 ;; Set cursor color
17889 (set-cursor-color "white")
17892 (set-mouse-color "white")
17894 ;; Set foreground and background
17895 (set-foreground-color "white")
17896 (set-background-color "darkblue")
17900 ;;; Set highlighting colors for isearch and drag
17901 (set-face-foreground 'highlight "white")
17902 (set-face-background 'highlight "blue")
17906 (set-face-foreground 'region "cyan")
17907 (set-face-background 'region "blue")
17911 (set-face-foreground 'secondary-selection "skyblue")
17912 (set-face-background 'secondary-selection "darkblue")
17916 ;; Set calendar highlighting colors
17917 (setq calendar-load-hook
17919 (set-face-foreground 'diary-face "skyblue")
17920 (set-face-background 'holiday-face "slate blue")
17921 (set-face-foreground 'holiday-face "white")))
17925 The various shades of blue soothe my eye and prevent me from seeing
17926 the screen flicker.
17928 Alternatively, I could have set my specifications in various X
17929 initialization files. For example, I could set the foreground,
17930 background, cursor, and pointer (i.e., mouse) colors in my
17931 @file{~/.Xresources} file like this:
17935 Emacs*foreground: white
17936 Emacs*background: darkblue
17937 Emacs*cursorColor: white
17938 Emacs*pointerColor: white
17942 In any event, since it is not part of Emacs, I set the root color of
17943 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
17944 run more modern window managers, such as Enlightenment, Gnome, or KDE;
17945 in those cases, I often specify an image rather than a plain color.}:
17948 xsetroot -solid Navy -fg white &
17952 @node Miscellaneous
17953 @section Miscellaneous Settings for a @file{.emacs} File
17956 Here are a few miscellaneous settings:
17961 Set the shape and color of the mouse cursor:
17965 ; Cursor shapes are defined in
17966 ; `/usr/include/X11/cursorfont.h';
17967 ; for example, the `target' cursor is number 128;
17968 ; the `top_left_arrow' cursor is number 132.
17972 (let ((mpointer (x-get-resource "*mpointer"
17973 "*emacs*mpointer")))
17974 ;; If you have not set your mouse pointer
17975 ;; then set it, otherwise leave as is:
17976 (if (eq mpointer nil)
17977 (setq mpointer "132")) ; top_left_arrow
17980 (setq x-pointer-shape (string-to-int mpointer))
17981 (set-mouse-color "white"))
17986 Or you can set the values of a variety of features in an alist, like
17992 default-frame-alist
17993 '((cursor-color . "white")
17994 (mouse-color . "white")
17995 (foreground-color . "white")
17996 (background-color . "DodgerBlue4")
17997 ;; (cursor-type . bar)
17998 (cursor-type . box)
18001 (tool-bar-lines . 0)
18002 (menu-bar-lines . 1)
18006 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
18012 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
18013 into @kbd{@key{CTRL}-h}.@*
18014 (Some older keyboards needed this, although I have not seen the
18019 ;; Translate `C-h' to <DEL>.
18020 ; (keyboard-translate ?\C-h ?\C-?)
18022 ;; Translate <DEL> to `C-h'.
18023 (keyboard-translate ?\C-? ?\C-h)
18027 @item Turn off a blinking cursor!
18031 (if (fboundp 'blink-cursor-mode)
18032 (blink-cursor-mode -1))
18037 or start GNU Emacs with the command @code{emacs -nbc}.
18040 @item When using `grep'@*
18041 @samp{-i}@w{ } Ignore case distinctions@*
18042 @samp{-n}@w{ } Prefix each line of output with line number@*
18043 @samp{-H}@w{ } Print the filename for each match.@*
18044 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18047 (setq grep-command "grep -i -nH -e ")
18051 @c Evidently, no longer needed in GNU Emacs 22
18053 item Automatically uncompress compressed files when visiting them
18056 (load "uncompress")
18061 @item Find an existing buffer, even if it has a different name@*
18062 This avoids problems with symbolic links.
18065 (setq find-file-existing-other-name t)
18068 @item Set your language environment and default input method
18072 (set-language-environment "latin-1")
18073 ;; Remember you can enable or disable multilingual text input
18074 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18075 (setq default-input-method "latin-1-prefix")
18079 If you want to write with Chinese `GB' characters, set this instead:
18083 (set-language-environment "Chinese-GB")
18084 (setq default-input-method "chinese-tonepy")
18089 @subsubheading Fixing Unpleasant Key Bindings
18090 @cindex Key bindings, fixing
18091 @cindex Bindings, key, fixing unpleasant
18093 Some systems bind keys unpleasantly. Sometimes, for example, the
18094 @key{CTRL} key appears in an awkward spot rather than at the far left
18097 Usually, when people fix these sorts of keybindings, they do not
18098 change their @file{~/.emacs} file. Instead, they bind the proper keys
18099 on their consoles with the @code{loadkeys} or @code{install-keymap}
18100 commands in their boot script and then include @code{xmodmap} commands
18101 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18109 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18111 install-keymap emacs2
18117 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18118 Lock} key is at the far left of the home row:
18122 # Bind the key labeled `Caps Lock' to `Control'
18123 # (Such a broken user interface suggests that keyboard manufacturers
18124 # think that computers are typewriters from 1885.)
18126 xmodmap -e "clear Lock"
18127 xmodmap -e "add Control = Caps_Lock"
18133 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18134 key to a @key{META} key:
18138 # Some ill designed keyboards have a key labeled ALT and no Meta
18139 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18145 @section A Modified Mode Line
18146 @vindex mode-line-format
18147 @cindex Mode line format
18149 Finally, a feature I really like: a modified mode line.
18151 When I work over a network, I forget which machine I am using. Also,
18152 I tend to I lose track of where I am, and which line point is on.
18154 So I reset my mode line to look like this:
18157 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18160 I am visiting a file called @file{foo.texi}, on my machine
18161 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18162 Texinfo mode, and am at the top of the buffer.
18165 My @file{.emacs} file has a section that looks like this:
18169 ;; Set a Mode Line that tells me which machine, which directory,
18170 ;; and which line I am on, plus the other customary information.
18171 (setq-default mode-line-format
18175 "mouse-1: select window, mouse-2: delete others ..."))
18176 mode-line-mule-info
18178 mode-line-frame-identification
18182 mode-line-buffer-identification
18185 (system-name) 0 (string-match "\\..+" (system-name))))
18190 "mouse-1: select window, mouse-2: delete others ..."))
18191 (line-number-mode " Line %l ")
18197 "mouse-1: select window, mouse-2: delete others ..."))
18198 (:eval (mode-line-mode-name))
18201 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18210 Here, I redefine the default mode line. Most of the parts are from
18211 the original; but I make a few changes. I set the @emph{default} mode
18212 line format so as to permit various modes, such as Info, to override
18215 Many elements in the list are self-explanatory:
18216 @code{mode-line-modified} is a variable that tells whether the buffer
18217 has been modified, @code{mode-name} tells the name of the mode, and so
18218 on. However, the format looks complicated because of two features we
18219 have not discussed.
18221 @cindex Properties, in mode line example
18222 The first string in the mode line is a dash or hyphen, @samp{-}. In
18223 the old days, it would have been specified simply as @code{"-"}. But
18224 nowadays, Emacs can add properties to a string, such as highlighting
18225 or, as in this case, a help feature. If you place your mouse cursor
18226 over the hyphen, some help information appears (By default, you must
18227 wait seven-tenths of a second before the information appears. You can
18228 change that timing by changing the value of @code{tooltip-delay}.)
18231 The new string format has a special syntax:
18234 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18238 The @code{#(} begins a list. The first element of the list is the
18239 string itself, just one @samp{-}. The second and third
18240 elements specify the range over which the fourth element applies. A
18241 range starts @emph{after} a character, so a zero means the range
18242 starts just before the first character; a 1 means that the range ends
18243 just after the first character. The third element is the property for
18244 the range. It consists of a property list, a
18245 property name, in this case, @samp{help-echo}, followed by a value, in this
18246 case, a string. The second, third, and fourth elements of this new
18247 string format can be repeated.
18249 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18250 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18251 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18253 @code{mode-line-buffer-identification}
18254 displays the current buffer name. It is a list
18255 beginning @code{(#("%12b" 0 4 @dots{}}.
18256 The @code{#(} begins the list.
18258 The @samp{"%12b"} displays the current buffer name, using the
18259 @code{buffer-name} function with which we are familiar; the `12'
18260 specifies the maximum number of characters that will be displayed.
18261 When a name has fewer characters, whitespace is added to fill out to
18262 this number. (Buffer names can and often should be longer than 12
18263 characters; this length works well in a typical 80 column wide
18266 @code{:eval} says to evaluate the following form and use the result as
18267 a string to display. In this case, the expression displays the first
18268 component of the full system name. The end of the first component is
18269 a @samp{.} (`period'), so I use the @code{string-match} function to
18270 tell me the length of the first component. The substring from the
18271 zeroth character to that length is the name of the machine.
18274 This is the expression:
18279 (system-name) 0 (string-match "\\..+" (system-name))))
18283 @samp{%[} and @samp{%]} cause a pair of square brackets
18284 to appear for each recursive editing level. @samp{%n} says `Narrow'
18285 when narrowing is in effect. @samp{%P} tells you the percentage of
18286 the buffer that is above the bottom of the window, or `Top', `Bottom',
18287 or `All'. (A lower case @samp{p} tell you the percentage above the
18288 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18291 Remember, ``You don't have to like Emacs to like it''---your own
18292 Emacs can have different colors, different commands, and different
18293 keys than a default Emacs.
18295 On the other hand, if you want to bring up a plain `out of the box'
18296 Emacs, with no customization, type:
18303 This will start an Emacs that does @emph{not} load your
18304 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18311 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18312 first is built into the internals of Emacs and is always with you;
18313 the second requires that you instrument a function before you can use it.
18315 Both debuggers are described extensively in @ref{Debugging, ,
18316 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18317 In this chapter, I will walk through a short example of each.
18320 * debug:: How to use the built-in debugger.
18321 * debug-on-entry:: Start debugging when you call a function.
18322 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18323 * edebug:: How to use Edebug, a source level debugger.
18324 * Debugging Exercises::
18328 @section @code{debug}
18331 Suppose you have written a function definition that is intended to
18332 return the sum of the numbers 1 through a given number. (This is the
18333 @code{triangle} function discussed earlier. @xref{Decrementing
18334 Example, , Example with Decrementing Counter}, for a discussion.)
18335 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18337 However, your function definition has a bug. You have mistyped
18338 @samp{1=} for @samp{1-}. Here is the broken definition:
18340 @findex triangle-bugged
18343 (defun triangle-bugged (number)
18344 "Return sum of numbers 1 through NUMBER inclusive."
18346 (while (> number 0)
18347 (setq total (+ total number))
18348 (setq number (1= number))) ; @r{Error here.}
18353 If you are reading this in Info, you can evaluate this definition in
18354 the normal fashion. You will see @code{triangle-bugged} appear in the
18358 Now evaluate the @code{triangle-bugged} function with an
18362 (triangle-bugged 4)
18366 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18372 ---------- Buffer: *Backtrace* ----------
18373 Debugger entered--Lisp error: (void-function 1=)
18375 (setq number (1= number))
18376 (while (> number 0) (setq total (+ total number))
18377 (setq number (1= number)))
18378 (let ((total 0)) (while (> number 0) (setq total ...)
18379 (setq number ...)) total)
18383 eval((triangle-bugged 4))
18384 eval-last-sexp-1(nil)
18385 eval-last-sexp(nil)
18386 call-interactively(eval-last-sexp)
18387 ---------- Buffer: *Backtrace* ----------
18392 (I have reformatted this example slightly; the debugger does not fold
18393 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18394 the @file{*Backtrace*} buffer.)
18396 In practice, for a bug as simple as this, the `Lisp error' line will
18397 tell you what you need to know to correct the definition. The
18398 function @code{1=} is `void'.
18402 In GNU Emacs 20 and before, you will see:
18405 Symbol's function definition is void:@: 1=
18409 which has the same meaning as the @file{*Backtrace*} buffer line in
18413 However, suppose you are not quite certain what is going on?
18414 You can read the complete backtrace.
18416 In this case, you need to run a recent GNU Emacs, which automatically
18417 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18418 else, you need to start the debugger manually as described below.
18420 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18421 what Emacs did that led to the error. Emacs made an interactive call
18422 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18423 of the @code{triangle-bugged} expression. Each line above tells you
18424 what the Lisp interpreter evaluated next.
18427 The third line from the top of the buffer is
18430 (setq number (1= number))
18434 Emacs tried to evaluate this expression; in order to do so, it tried
18435 to evaluate the inner expression shown on the second line from the
18444 This is where the error occurred; as the top line says:
18447 Debugger entered--Lisp error: (void-function 1=)
18451 You can correct the mistake, re-evaluate the function definition, and
18452 then run your test again.
18454 @node debug-on-entry
18455 @section @code{debug-on-entry}
18456 @findex debug-on-entry
18458 A recent GNU Emacs starts the debugger automatically when your
18459 function has an error.
18462 GNU Emacs version 20 and before did not; it simply
18463 presented you with an error message. You had to start the debugger
18467 Incidentally, you can start the debugger manually for all versions of
18468 Emacs; the advantage is that the debugger runs even if you do not have
18469 a bug in your code. Sometimes your code will be free of bugs!
18471 You can enter the debugger when you call the function by calling
18472 @code{debug-on-entry}.
18479 M-x debug-on-entry RET triangle-bugged RET
18484 Now, evaluate the following:
18487 (triangle-bugged 5)
18491 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18492 you that it is beginning to evaluate the @code{triangle-bugged}
18497 ---------- Buffer: *Backtrace* ----------
18498 Debugger entered--entering a function:
18499 * triangle-bugged(5)
18500 eval((triangle-bugged 5))
18503 eval-last-sexp-1(nil)
18504 eval-last-sexp(nil)
18505 call-interactively(eval-last-sexp)
18506 ---------- Buffer: *Backtrace* ----------
18510 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18511 the first expression in @code{triangle-bugged}; the buffer will look
18516 ---------- Buffer: *Backtrace* ----------
18517 Debugger entered--beginning evaluation of function call form:
18518 * (let ((total 0)) (while (> number 0) (setq total ...)
18519 (setq number ...)) total)
18520 * triangle-bugged(5)
18521 eval((triangle-bugged 5))
18524 eval-last-sexp-1(nil)
18525 eval-last-sexp(nil)
18526 call-interactively(eval-last-sexp)
18527 ---------- Buffer: *Backtrace* ----------
18532 Now, type @kbd{d} again, eight times, slowly. Each time you type
18533 @kbd{d}, Emacs will evaluate another expression in the function
18537 Eventually, the buffer will look like this:
18541 ---------- Buffer: *Backtrace* ----------
18542 Debugger entered--beginning evaluation of function call form:
18543 * (setq number (1= number))
18544 * (while (> number 0) (setq total (+ total number))
18545 (setq number (1= number)))
18548 * (let ((total 0)) (while (> number 0) (setq total ...)
18549 (setq number ...)) total)
18550 * triangle-bugged(5)
18551 eval((triangle-bugged 5))
18554 eval-last-sexp-1(nil)
18555 eval-last-sexp(nil)
18556 call-interactively(eval-last-sexp)
18557 ---------- Buffer: *Backtrace* ----------
18563 Finally, after you type @kbd{d} two more times, Emacs will reach the
18564 error, and the top two lines of the @file{*Backtrace*} buffer will look
18569 ---------- Buffer: *Backtrace* ----------
18570 Debugger entered--Lisp error: (void-function 1=)
18573 ---------- Buffer: *Backtrace* ----------
18577 By typing @kbd{d}, you were able to step through the function.
18579 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18580 quits the trace, but does not cancel @code{debug-on-entry}.
18582 @findex cancel-debug-on-entry
18583 To cancel the effect of @code{debug-on-entry}, call
18584 @code{cancel-debug-on-entry} and the name of the function, like this:
18587 M-x cancel-debug-on-entry RET triangle-bugged RET
18591 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18593 @node debug-on-quit
18594 @section @code{debug-on-quit} and @code{(debug)}
18596 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18597 there are two other ways to start @code{debug}.
18599 @findex debug-on-quit
18600 You can start @code{debug} whenever you type @kbd{C-g}
18601 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18602 @code{t}. This is useful for debugging infinite loops.
18605 @cindex @code{(debug)} in code
18606 Or, you can insert a line that says @code{(debug)} into your code
18607 where you want the debugger to start, like this:
18611 (defun triangle-bugged (number)
18612 "Return sum of numbers 1 through NUMBER inclusive."
18614 (while (> number 0)
18615 (setq total (+ total number))
18616 (debug) ; @r{Start debugger.}
18617 (setq number (1= number))) ; @r{Error here.}
18622 The @code{debug} function is described in detail in @ref{Debugger, ,
18623 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18626 @section The @code{edebug} Source Level Debugger
18627 @cindex Source level debugger
18630 Edebug is a source level debugger. Edebug normally displays the
18631 source of the code you are debugging, with an arrow at the left that
18632 shows which line you are currently executing.
18634 You can walk through the execution of a function, line by line, or run
18635 quickly until reaching a @dfn{breakpoint} where execution stops.
18637 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18638 Lisp Reference Manual}.
18641 Here is a bugged function definition for @code{triangle-recursively}.
18642 @xref{Recursive triangle function, , Recursion in place of a counter},
18643 for a review of it.
18647 (defun triangle-recursively-bugged (number)
18648 "Return sum of numbers 1 through NUMBER inclusive.
18653 (triangle-recursively-bugged
18654 (1= number))))) ; @r{Error here.}
18659 Normally, you would install this definition by positioning your cursor
18660 after the function's closing parenthesis and typing @kbd{C-x C-e}
18661 (@code{eval-last-sexp}) or else by positioning your cursor within the
18662 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18663 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18667 However, to prepare this function definition for Edebug, you must
18668 first @dfn{instrument} the code using a different command. You can do
18669 this by positioning your cursor within or just after the definition
18673 M-x edebug-defun RET
18677 This will cause Emacs to load Edebug automatically if it is not
18678 already loaded, and properly instrument the function.
18680 After instrumenting the function, place your cursor after the
18681 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18684 (triangle-recursively-bugged 3)
18688 You will be jumped back to the source for
18689 @code{triangle-recursively-bugged} and the cursor positioned at the
18690 beginning of the @code{if} line of the function. Also, you will see
18691 an arrowhead at the left hand side of that line. The arrowhead marks
18692 the line where the function is executing. (In the following examples,
18693 we show the arrowhead with @samp{=>}; in a windowing system, you may
18694 see the arrowhead as a solid triangle in the window `fringe'.)
18697 =>@point{}(if (= number 1)
18702 In the example, the location of point is displayed with a star,
18703 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18706 In the example, the location of point is displayed as @samp{@point{}}
18707 (in a printed book, it is displayed with a five pointed star).
18710 If you now press @key{SPC}, point will move to the next expression to
18711 be executed; the line will look like this:
18714 =>(if @point{}(= number 1)
18718 As you continue to press @key{SPC}, point will move from expression to
18719 expression. At the same time, whenever an expression returns a value,
18720 that value will be displayed in the echo area. For example, after you
18721 move point past @code{number}, you will see the following:
18724 Result: 3 (#o3, #x3, ?\C-c)
18728 This means the value of @code{number} is 3, which is octal three,
18729 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18730 alphabet, in case you need to know this information).
18732 You can continue moving through the code until you reach the line with
18733 the error. Before evaluation, that line looks like this:
18736 => @point{}(1= number))))) ; @r{Error here.}
18741 When you press @key{SPC} once again, you will produce an error message
18745 Symbol's function definition is void:@: 1=
18751 Press @kbd{q} to quit Edebug.
18753 To remove instrumentation from a function definition, simply
18754 re-evaluate it with a command that does not instrument it.
18755 For example, you could place your cursor after the definition's
18756 closing parenthesis and type @kbd{C-x C-e}.
18758 Edebug does a great deal more than walk with you through a function.
18759 You can set it so it races through on its own, stopping only at an
18760 error or at specified stopping points; you can cause it to display the
18761 changing values of various expressions; you can find out how many
18762 times a function is called, and more.
18764 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18765 Lisp Reference Manual}.
18768 @node Debugging Exercises
18769 @section Debugging Exercises
18773 Install the @code{@value{COUNT-WORDS}} function and then cause it to
18774 enter the built-in debugger when you call it. Run the command on a
18775 region containing two words. You will need to press @kbd{d} a
18776 remarkable number of times. On your system, is a `hook' called after
18777 the command finishes? (For information on hooks, see @ref{Command
18778 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18782 Copy @code{@value{COUNT-WORDS}} into the @file{*scratch*} buffer,
18783 instrument the function for Edebug, and walk through its execution.
18784 The function does not need to have a bug, although you can introduce
18785 one if you wish. If the function lacks a bug, the walk-through
18786 completes without problems.
18789 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18790 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.,
18791 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18792 for commands made outside of the Edebug debugging buffer.)
18795 In the Edebug debugging buffer, use the @kbd{p}
18796 (@code{edebug-bounce-point}) command to see where in the region the
18797 @code{@value{COUNT-WORDS}} is working.
18800 Move point to some spot further down the function and then type the
18801 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18804 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18805 walk through the function on its own; use an upper case @kbd{T} for
18806 @code{edebug-Trace-fast-mode}.
18809 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18814 @chapter Conclusion
18816 We have now reached the end of this Introduction. You have now
18817 learned enough about programming in Emacs Lisp to set values, to write
18818 simple @file{.emacs} files for yourself and your friends, and write
18819 simple customizations and extensions to Emacs.
18821 This is a place to stop. Or, if you wish, you can now go onward, and
18824 You have learned some of the basic nuts and bolts of programming. But
18825 only some. There are a great many more brackets and hinges that are
18826 easy to use that we have not touched.
18828 A path you can follow right now lies among the sources to GNU Emacs
18831 @cite{The GNU Emacs Lisp Reference Manual}.
18834 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18835 Emacs Lisp Reference Manual}.
18838 The Emacs Lisp sources are an adventure. When you read the sources and
18839 come across a function or expression that is unfamiliar, you need to
18840 figure out or find out what it does.
18842 Go to the Reference Manual. It is a thorough, complete, and fairly
18843 easy-to-read description of Emacs Lisp. It is written not only for
18844 experts, but for people who know what you know. (The @cite{Reference
18845 Manual} comes with the standard GNU Emacs distribution. Like this
18846 introduction, it comes as a Texinfo source file, so you can read it
18847 on-line and as a typeset, printed book.)
18849 Go to the other on-line help that is part of GNU Emacs: the on-line
18850 documentation for all functions and variables, and @code{find-tag},
18851 the program that takes you to sources.
18853 Here is an example of how I explore the sources. Because of its name,
18854 @file{simple.el} is the file I looked at first, a long time ago. As
18855 it happens some of the functions in @file{simple.el} are complicated,
18856 or at least look complicated at first sight. The @code{open-line}
18857 function, for example, looks complicated.
18859 You may want to walk through this function slowly, as we did with the
18860 @code{forward-sentence} function. (@xref{forward-sentence, The
18861 @code{forward-sentence} function}.) Or you may want to skip that
18862 function and look at another, such as @code{split-line}. You don't
18863 need to read all the functions. According to
18864 @code{count-words-in-defun}, the @code{split-line} function contains
18865 102 words and symbols.
18867 Even though it is short, @code{split-line} contains expressions
18868 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18869 @code{current-column} and @code{insert-and-inherit}.
18871 Consider the @code{skip-chars-forward} function. (It is part of the
18872 function definition for @code{back-to-indentation}, which is shown in
18873 @ref{Review, , Review}.)
18875 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18876 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18877 function. This gives you the function documentation.
18879 You may be able to guess what is done by a well named function such as
18880 @code{indent-to}; or you can look it up, too. Incidentally, the
18881 @code{describe-function} function itself is in @file{help.el}; it is
18882 one of those long, but decipherable functions. You can look up
18883 @code{describe-function} using the @kbd{C-h f} command!
18885 In this instance, since the code is Lisp, the @file{*Help*} buffer
18886 contains the name of the library containing the function's source.
18887 You can put point over the name of the library and press the RET key,
18888 which in this situation is bound to @code{help-follow}, and be taken
18889 directly to the source, in the same way as @kbd{M-.}
18892 The definition for @code{describe-function} illustrates how to
18893 customize the @code{interactive} expression without using the standard
18894 character codes; and it shows how to create a temporary buffer.
18896 (The @code{indent-to} function is written in C rather than Emacs Lisp;
18897 it is a `built-in' function. @code{help-follow} takes you to its
18898 source as does @code{find-tag}, when properly set up.)
18900 You can look at a function's source using @code{find-tag}, which is
18901 bound to @kbd{M-.} Finally, you can find out what the Reference
18902 Manual has to say by visiting the manual in Info, and typing @kbd{i}
18903 (@code{Info-index}) and the name of the function, or by looking up the
18904 function in the index to a printed copy of the manual.
18906 Similarly, you can find out what is meant by
18907 @code{insert-and-inherit}.
18909 Other interesting source files include @file{paragraphs.el},
18910 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
18911 file includes short, easily understood functions as well as longer
18912 ones. The @file{loaddefs.el} file contains the many standard
18913 autoloads and many keymaps. I have never looked at it all; only at
18914 parts. @file{loadup.el} is the file that loads the standard parts of
18915 Emacs; it tells you a great deal about how Emacs is built.
18916 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
18917 Reference Manual}, for more about building.)
18919 As I said, you have learned some nuts and bolts; however, and very
18920 importantly, we have hardly touched major aspects of programming; I
18921 have said nothing about how to sort information, except to use the
18922 predefined @code{sort} function; I have said nothing about how to store
18923 information, except to use variables and lists; I have said nothing
18924 about how to write programs that write programs. These are topics for
18925 another, and different kind of book, a different kind of learning.
18927 What you have done is learn enough for much practical work with GNU
18928 Emacs. What you have done is get started. This is the end of a
18931 @c ================ Appendix ================
18934 @appendix The @code{the-the} Function
18936 @cindex Duplicated words function
18937 @cindex Words, duplicated
18939 Sometimes when you you write text, you duplicate words---as with ``you
18940 you'' near the beginning of this sentence. I find that most
18941 frequently, I duplicate ``the''; hence, I call the function for
18942 detecting duplicated words, @code{the-the}.
18945 As a first step, you could use the following regular expression to
18946 search for duplicates:
18949 \\(\\w+[ \t\n]+\\)\\1
18953 This regexp matches one or more word-constituent characters followed
18954 by one or more spaces, tabs, or newlines. However, it does not detect
18955 duplicated words on different lines, since the ending of the first
18956 word, the end of the line, is different from the ending of the second
18957 word, a space. (For more information about regular expressions, see
18958 @ref{Regexp Search, , Regular Expression Searches}, as well as
18959 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
18960 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
18961 The GNU Emacs Lisp Reference Manual}.)
18963 You might try searching just for duplicated word-constituent
18964 characters but that does not work since the pattern detects doubles
18965 such as the two occurrences of `th' in `with the'.
18967 Another possible regexp searches for word-constituent characters
18968 followed by non-word-constituent characters, reduplicated. Here,
18969 @w{@samp{\\w+}} matches one or more word-constituent characters and
18970 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
18973 \\(\\(\\w+\\)\\W*\\)\\1
18979 Here is the pattern that I use. It is not perfect, but good enough.
18980 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
18981 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
18982 any characters that are @emph{not} an @@-sign, space, newline, or tab.
18985 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
18988 One can write more complicated expressions, but I found that this
18989 expression is good enough, so I use it.
18991 Here is the @code{the-the} function, as I include it in my
18992 @file{.emacs} file, along with a handy global key binding:
18997 "Search forward for for a duplicated word."
18999 (message "Searching for for duplicated words ...")
19003 ;; This regexp is not perfect
19004 ;; but is fairly good over all:
19005 (if (re-search-forward
19006 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
19007 (message "Found duplicated word.")
19008 (message "End of buffer")))
19012 ;; Bind `the-the' to C-c \
19013 (global-set-key "\C-c\\" 'the-the)
19022 one two two three four five
19027 You can substitute the other regular expressions shown above in the
19028 function definition and try each of them on this list.
19031 @appendix Handling the Kill Ring
19032 @cindex Kill ring handling
19033 @cindex Handling the kill ring
19034 @cindex Ring, making a list like a
19036 The kill ring is a list that is transformed into a ring by the
19037 workings of the @code{current-kill} function. The @code{yank} and
19038 @code{yank-pop} commands use the @code{current-kill} function.
19040 This appendix describes the @code{current-kill} function as well as
19041 both the @code{yank} and the @code{yank-pop} commands, but first,
19042 consider the workings of the kill ring.
19045 * What the Kill Ring Does::
19047 * yank:: Paste a copy of a clipped element.
19048 * yank-pop:: Insert element pointed to.
19053 @node What the Kill Ring Does
19054 @unnumberedsec What the Kill Ring Does
19058 The kill ring has a default maximum length of sixty items; this number
19059 is too large for an explanation. Instead, set it to four. Please
19060 evaluate the following:
19064 (setq old-kill-ring-max kill-ring-max)
19065 (setq kill-ring-max 4)
19070 Then, please copy each line of the following indented example into the
19071 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19075 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19076 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19077 merely copy it to the kill ring. However, your machine may beep at
19078 you. Alternatively, for silence, you may copy the region of each line
19079 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19080 each line for this command to succeed, but it does not matter at which
19081 end you put point or mark.)
19085 Please invoke the calls in order, so that five elements attempt to
19086 fill the kill ring:
19091 second piece of text
19093 fourth line of text
19100 Then find the value of @code{kill-ring} by evaluating
19112 ("fifth bit of text" "fourth line of text"
19113 "third line" "second piece of text")
19118 The first element, @samp{first some text}, was dropped.
19121 To return to the old value for the length of the kill ring, evaluate:
19124 (setq kill-ring-max old-kill-ring-max)
19128 @appendixsec The @code{current-kill} Function
19129 @findex current-kill
19131 The @code{current-kill} function changes the element in the kill ring
19132 to which @code{kill-ring-yank-pointer} points. (Also, the
19133 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19134 to the latest element of the kill ring. The @code{kill-new}
19135 function is used directly or indirectly by @code{kill-append},
19136 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19137 and @code{kill-region}.)
19140 * Code for current-kill::
19141 * Understanding current-kill::
19145 @node Code for current-kill
19146 @unnumberedsubsec The code for @code{current-kill}
19151 The @code{current-kill} function is used by @code{yank} and by
19152 @code{yank-pop}. Here is the code for @code{current-kill}:
19156 (defun current-kill (n &optional do-not-move)
19157 "Rotate the yanking point by N places, and then return that kill.
19158 If N is zero, `interprogram-paste-function' is set, and calling it
19159 returns a string, then that string is added to the front of the
19160 kill ring and returned as the latest kill.
19163 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19164 yanking point; just return the Nth kill forward."
19165 (let ((interprogram-paste (and (= n 0)
19166 interprogram-paste-function
19167 (funcall interprogram-paste-function))))
19170 (if interprogram-paste
19172 ;; Disable the interprogram cut function when we add the new
19173 ;; text to the kill ring, so Emacs doesn't try to own the
19174 ;; selection, with identical text.
19175 (let ((interprogram-cut-function nil))
19176 (kill-new interprogram-paste))
19177 interprogram-paste)
19180 (or kill-ring (error "Kill ring is empty"))
19181 (let ((ARGth-kill-element
19182 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19183 (length kill-ring))
19186 (setq kill-ring-yank-pointer ARGth-kill-element))
19187 (car ARGth-kill-element)))))
19191 Remember also that the @code{kill-new} function sets
19192 @code{kill-ring-yank-pointer} to the latest element of the kill
19193 ring, which means that all the functions that call it set the value
19194 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19195 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19198 Here is the line in @code{kill-new}, which is explained in
19199 @ref{kill-new function, , The @code{kill-new} function}.
19202 (setq kill-ring-yank-pointer kill-ring)
19206 @node Understanding current-kill
19207 @unnumberedsubsec @code{current-kill} in Outline
19210 The @code{current-kill} function looks complex, but as usual, it can
19211 be understood by taking it apart piece by piece. First look at it in
19216 (defun current-kill (n &optional do-not-move)
19217 "Rotate the yanking point by N places, and then return that kill."
19223 This function takes two arguments, one of which is optional. It has a
19224 documentation string. It is @emph{not} interactive.
19227 * Body of current-kill::
19228 * Digression concerning error:: How to mislead humans, but not computers.
19229 * Determining the Element::
19233 @node Body of current-kill
19234 @unnumberedsubsubsec The Body of @code{current-kill}
19237 The body of the function definition is a @code{let} expression, which
19238 itself has a body as well as a @var{varlist}.
19240 The @code{let} expression declares a variable that will be only usable
19241 within the bounds of this function. This variable is called
19242 @code{interprogram-paste} and is for copying to another program. It
19243 is not for copying within this instance of GNU Emacs. Most window
19244 systems provide a facility for interprogram pasting. Sadly, that
19245 facility usually provides only for the last element. Most windowing
19246 systems have not adopted a ring of many possibilities, even though
19247 Emacs has provided it for decades.
19249 The @code{if} expression has two parts, one if there exists
19250 @code{interprogram-paste} and one if not.
19253 Let us consider the `if not' or else-part of the @code{current-kill}
19254 function. (The then-part uses the @code{kill-new} function, which
19255 we have already described. @xref{kill-new function, , The
19256 @code{kill-new} function}.)
19260 (or kill-ring (error "Kill ring is empty"))
19261 (let ((ARGth-kill-element
19262 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19263 (length kill-ring))
19266 (setq kill-ring-yank-pointer ARGth-kill-element))
19267 (car ARGth-kill-element))
19272 The code first checks whether the kill ring has content; otherwise it
19276 Note that the @code{or} expression is very similar to testing length
19283 (if (zerop (length kill-ring)) ; @r{if-part}
19284 (error "Kill ring is empty")) ; @r{then-part}
19290 If there is not anything in the kill ring, its length must be zero and
19291 an error message sent to the user: @samp{Kill ring is empty}. The
19292 @code{current-kill} function uses an @code{or} expression which is
19293 simpler. But an @code{if} expression reminds us what goes on.
19295 This @code{if} expression uses the function @code{zerop} which returns
19296 true if the value it is testing is zero. When @code{zerop} tests
19297 true, the then-part of the @code{if} is evaluated. The then-part is a
19298 list starting with the function @code{error}, which is a function that
19299 is similar to the @code{message} function
19300 (@pxref{message, , The @code{message} Function}) in that
19301 it prints a one-line message in the echo area. However, in addition
19302 to printing a message, @code{error} also stops evaluation of the
19303 function within which it is embedded. This means that the rest of the
19304 function will not be evaluated if the length of the kill ring is zero.
19306 Then the @code{current-kill} function selects the element to return.
19307 The selection depends on the number of places that @code{current-kill}
19308 rotates and on where @code{kill-ring-yank-pointer} points.
19310 Next, either the optional @code{do-not-move} argument is true or the
19311 current value of @code{kill-ring-yank-pointer} is set to point to the
19312 list. Finally, another expression returns the first element of the
19313 list even if the @code{do-not-move} argument is true.
19316 @node Digression concerning error
19317 @unnumberedsubsubsec Digression about the word `error'
19320 In my opinion, it is slightly misleading, at least to humans, to use
19321 the term `error' as the name of the @code{error} function. A better
19322 term would be `cancel'. Strictly speaking, of course, you cannot
19323 point to, much less rotate a pointer to a list that has no length, so
19324 from the point of view of the computer, the word `error' is correct.
19325 But a human expects to attempt this sort of thing, if only to find out
19326 whether the kill ring is full or empty. This is an act of
19329 From the human point of view, the act of exploration and discovery is
19330 not necessarily an error, and therefore should not be labeled as one,
19331 even in the bowels of a computer. As it is, the code in Emacs implies
19332 that a human who is acting virtuously, by exploring his or her
19333 environment, is making an error. This is bad. Even though the computer
19334 takes the same steps as it does when there is an `error', a term such as
19335 `cancel' would have a clearer connotation.
19338 @node Determining the Element
19339 @unnumberedsubsubsec Determining the Element
19342 Among other actions, the else-part of the @code{if} expression sets
19343 the value of @code{kill-ring-yank-pointer} to
19344 @code{ARGth-kill-element} when the kill ring has something in it and
19345 the value of @code{do-not-move} is @code{nil}.
19348 The code looks like this:
19352 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19353 (length kill-ring))
19358 This needs some examination. Unless it is not supposed to move the
19359 pointer, the @code{current-kill} function changes where
19360 @code{kill-ring-yank-pointer} points.
19362 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19363 expression does. Also, clearly, @code{ARGth-kill-element} is being
19364 set to be equal to some @sc{cdr} of the kill ring, using the
19365 @code{nthcdr} function that is described in an earlier section.
19366 (@xref{copy-region-as-kill}.) How does it do this?
19368 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19369 works by repeatedly taking the @sc{cdr} of a list---it takes the
19370 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19373 The two following expressions produce the same result:
19377 (setq kill-ring-yank-pointer (cdr kill-ring))
19379 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19383 However, the @code{nthcdr} expression is more complicated. It uses
19384 the @code{mod} function to determine which @sc{cdr} to select.
19386 (You will remember to look at inner functions first; indeed, we will
19387 have to go inside the @code{mod}.)
19389 The @code{mod} function returns the value of its first argument modulo
19390 the second; that is to say, it returns the remainder after dividing
19391 the first argument by the second. The value returned has the same
19392 sign as the second argument.
19400 @result{} 0 ;; @r{because there is no remainder}
19407 In this case, the first argument is often smaller than the second.
19419 We can guess what the @code{-} function does. It is like @code{+} but
19420 subtracts instead of adds; the @code{-} function subtracts its second
19421 argument from its first. Also, we already know what the @code{length}
19422 function does (@pxref{length}). It returns the length of a list.
19424 And @code{n} is the name of the required argument to the
19425 @code{current-kill} function.
19428 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19429 expression returns the whole list, as you can see by evaluating the
19434 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19435 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19436 (nthcdr (mod (- 0 4) 4)
19437 '("fourth line of text"
19439 "second piece of text"
19440 "first some text"))
19445 When the first argument to the @code{current-kill} function is one,
19446 the @code{nthcdr} expression returns the list without its first
19451 (nthcdr (mod (- 1 4) 4)
19452 '("fourth line of text"
19454 "second piece of text"
19455 "first some text"))
19459 @cindex @samp{global variable} defined
19460 @cindex @samp{variable, global}, defined
19461 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19462 are @dfn{global variables}. That means that any expression in Emacs
19463 Lisp can access them. They are not like the local variables set by
19464 @code{let} or like the symbols in an argument list.
19465 Local variables can only be accessed
19466 within the @code{let} that defines them or the function that specifies
19467 them in an argument list (and within expressions called by them).
19470 @c texi2dvi fails when the name of the section is within ifnottex ...
19471 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19472 @ref{defun, , The @code{defun} Macro}.)
19476 @appendixsec @code{yank}
19479 After learning about @code{current-kill}, the code for the
19480 @code{yank} function is almost easy.
19482 The @code{yank} function does not use the
19483 @code{kill-ring-yank-pointer} variable directly. It calls
19484 @code{insert-for-yank} which calls @code{current-kill} which sets the
19485 @code{kill-ring-yank-pointer} variable.
19488 The code looks like this:
19493 (defun yank (&optional arg)
19494 "Reinsert (\"paste\") the last stretch of killed text.
19495 More precisely, reinsert the stretch of killed text most recently
19496 killed OR yanked. Put point at end, and set mark at beginning.
19497 With just \\[universal-argument] as argument, same but put point at
19498 beginning (and mark at end). With argument N, reinsert the Nth most
19499 recently killed stretch of killed text.
19501 When this command inserts killed text into the buffer, it honors
19502 `yank-excluded-properties' and `yank-handler' as described in the
19503 doc string for `insert-for-yank-1', which see.
19505 See also the command \\[yank-pop]."
19509 (setq yank-window-start (window-start))
19510 ;; If we don't get all the way thru, make last-command indicate that
19511 ;; for the following command.
19512 (setq this-command t)
19513 (push-mark (point))
19516 (insert-for-yank (current-kill (cond
19521 ;; This is like exchange-point-and-mark,
19522 ;; but doesn't activate the mark.
19523 ;; It is cleaner to avoid activation, even though the command
19524 ;; loop would deactivate the mark because we inserted text.
19525 (goto-char (prog1 (mark t)
19526 (set-marker (mark-marker) (point) (current-buffer)))))
19529 ;; If we do get all the way thru, make this-command indicate that.
19530 (if (eq this-command t)
19531 (setq this-command 'yank))
19536 The key expression is @code{insert-for-yank}, which inserts the string
19537 returned by @code{current-kill}, but removes some text properties from
19540 However, before getting to that expression, the function sets the value
19541 of @code{yank-window-start} to the position returned by the
19542 @code{(window-start)} expression, the position at which the display
19543 currently starts. The @code{yank} function also sets
19544 @code{this-command} and pushes the mark.
19546 After it yanks the appropriate element, if the optional argument is a
19547 @sc{cons} rather than a number or nothing, it puts point at beginning
19548 of the yanked text and mark at its end.
19550 (The @code{prog1} function is like @code{progn} but returns the value
19551 of its first argument rather than the value of its last argument. Its
19552 first argument is forced to return the buffer's mark as an integer.
19553 You can see the documentation for these functions by placing point
19554 over them in this buffer and then typing @kbd{C-h f}
19555 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19558 The last part of the function tells what to do when it succeeds.
19561 @appendixsec @code{yank-pop}
19564 After understanding @code{yank} and @code{current-kill}, you know how
19565 to approach the @code{yank-pop} function. Leaving out the
19566 documentation to save space, it looks like this:
19571 (defun yank-pop (&optional arg)
19574 (if (not (eq last-command 'yank))
19575 (error "Previous command was not a yank"))
19578 (setq this-command 'yank)
19579 (unless arg (setq arg 1))
19580 (let ((inhibit-read-only t)
19581 (before (< (point) (mark t))))
19585 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19586 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19587 (setq yank-undo-function nil)
19590 (set-marker (mark-marker) (point) (current-buffer))
19591 (insert-for-yank (current-kill arg))
19592 ;; Set the window start back where it was in the yank command,
19594 (set-window-start (selected-window) yank-window-start t)
19598 ;; This is like exchange-point-and-mark,
19599 ;; but doesn't activate the mark.
19600 ;; It is cleaner to avoid activation, even though the command
19601 ;; loop would deactivate the mark because we inserted text.
19602 (goto-char (prog1 (mark t)
19603 (set-marker (mark-marker)
19605 (current-buffer))))))
19610 The function is interactive with a small @samp{p} so the prefix
19611 argument is processed and passed to the function. The command can
19612 only be used after a previous yank; otherwise an error message is
19613 sent. This check uses the variable @code{last-command} which is set
19614 by @code{yank} and is discussed elsewhere.
19615 (@xref{copy-region-as-kill}.)
19617 The @code{let} clause sets the variable @code{before} to true or false
19618 depending whether point is before or after mark and then the region
19619 between point and mark is deleted. This is the region that was just
19620 inserted by the previous yank and it is this text that will be
19623 @code{funcall} calls its first argument as a function, passing
19624 remaining arguments to it. The first argument is whatever the
19625 @code{or} expression returns. The two remaining arguments are the
19626 positions of point and mark set by the preceding @code{yank} command.
19628 There is more, but that is the hardest part.
19631 @appendixsec The @file{ring.el} File
19632 @cindex @file{ring.el} file
19634 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19635 provides many of the features we just discussed. But functions such
19636 as @code{kill-ring-yank-pointer} do not use this library, possibly
19637 because they were written earlier.
19640 @appendix A Graph with Labeled Axes
19642 Printed axes help you understand a graph. They convey scale. In an
19643 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19644 wrote the code to print the body of a graph. Here we write the code
19645 for printing and labeling vertical and horizontal axes, along with the
19649 * Labeled Example::
19650 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19651 * print-Y-axis:: Print a label for the vertical axis.
19652 * print-X-axis:: Print a horizontal label.
19653 * Print Whole Graph:: The function to print a complete graph.
19657 @node Labeled Example
19658 @unnumberedsec Labeled Example Graph
19661 Since insertions fill a buffer to the right and below point, the new
19662 graph printing function should first print the Y or vertical axis,
19663 then the body of the graph, and finally the X or horizontal axis.
19664 This sequence lays out for us the contents of the function:
19674 Print body of graph.
19681 Here is an example of how a finished graph should look:
19694 1 - ****************
19701 In this graph, both the vertical and the horizontal axes are labeled
19702 with numbers. However, in some graphs, the horizontal axis is time
19703 and would be better labeled with months, like this:
19717 Indeed, with a little thought, we can easily come up with a variety of
19718 vertical and horizontal labeling schemes. Our task could become
19719 complicated. But complications breed confusion. Rather than permit
19720 this, it is better choose a simple labeling scheme for our first
19721 effort, and to modify or replace it later.
19724 These considerations suggest the following outline for the
19725 @code{print-graph} function:
19729 (defun print-graph (numbers-list)
19730 "@var{documentation}@dots{}"
19731 (let ((height @dots{}
19735 (print-Y-axis height @dots{} )
19736 (graph-body-print numbers-list)
19737 (print-X-axis @dots{} )))
19741 We can work on each part of the @code{print-graph} function definition
19744 @node print-graph Varlist
19745 @appendixsec The @code{print-graph} Varlist
19746 @cindex @code{print-graph} varlist
19748 In writing the @code{print-graph} function, the first task is to write
19749 the varlist in the @code{let} expression. (We will leave aside for the
19750 moment any thoughts about making the function interactive or about the
19751 contents of its documentation string.)
19753 The varlist should set several values. Clearly, the top of the label
19754 for the vertical axis must be at least the height of the graph, which
19755 means that we must obtain this information here. Note that the
19756 @code{print-graph-body} function also requires this information. There
19757 is no reason to calculate the height of the graph in two different
19758 places, so we should change @code{print-graph-body} from the way we
19759 defined it earlier to take advantage of the calculation.
19761 Similarly, both the function for printing the X axis labels and the
19762 @code{print-graph-body} function need to learn the value of the width of
19763 each symbol. We can perform the calculation here and change the
19764 definition for @code{print-graph-body} from the way we defined it in the
19767 The length of the label for the horizontal axis must be at least as long
19768 as the graph. However, this information is used only in the function
19769 that prints the horizontal axis, so it does not need to be calculated here.
19771 These thoughts lead us directly to the following form for the varlist
19772 in the @code{let} for @code{print-graph}:
19776 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19777 (symbol-width (length graph-blank)))
19782 As we shall see, this expression is not quite right.
19786 @appendixsec The @code{print-Y-axis} Function
19787 @cindex Axis, print vertical
19788 @cindex Y axis printing
19789 @cindex Vertical axis printing
19790 @cindex Print vertical axis
19792 The job of the @code{print-Y-axis} function is to print a label for
19793 the vertical axis that looks like this:
19811 The function should be passed the height of the graph, and then should
19812 construct and insert the appropriate numbers and marks.
19815 * print-Y-axis in Detail::
19816 * Height of label:: What height for the Y axis?
19817 * Compute a Remainder:: How to compute the remainder of a division.
19818 * Y Axis Element:: Construct a line for the Y axis.
19819 * Y-axis-column:: Generate a list of Y axis labels.
19820 * print-Y-axis Penultimate:: A not quite final version.
19824 @node print-Y-axis in Detail
19825 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19828 It is easy enough to see in the figure what the Y axis label should
19829 look like; but to say in words, and then to write a function
19830 definition to do the job is another matter. It is not quite true to
19831 say that we want a number and a tic every five lines: there are only
19832 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19833 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19834 and 9). It is better to say that we want a number and a tic mark on
19835 the base line (number 1) and then that we want a number and a tic on
19836 the fifth line from the bottom and on every line that is a multiple of
19840 @node Height of label
19841 @unnumberedsubsec What height should the label be?
19844 The next issue is what height the label should be? Suppose the maximum
19845 height of tallest column of the graph is seven. Should the highest
19846 label on the Y axis be @samp{5 -}, and should the graph stick up above
19847 the label? Or should the highest label be @samp{7 -}, and mark the peak
19848 of the graph? Or should the highest label be @code{10 -}, which is a
19849 multiple of five, and be higher than the topmost value of the graph?
19851 The latter form is preferred. Most graphs are drawn within rectangles
19852 whose sides are an integral number of steps long---5, 10, 15, and so
19853 on for a step distance of five. But as soon as we decide to use a
19854 step height for the vertical axis, we discover that the simple
19855 expression in the varlist for computing the height is wrong. The
19856 expression is @code{(apply 'max numbers-list)}. This returns the
19857 precise height, not the maximum height plus whatever is necessary to
19858 round up to the nearest multiple of five. A more complex expression
19861 As usual in cases like this, a complex problem becomes simpler if it is
19862 divided into several smaller problems.
19864 First, consider the case when the highest value of the graph is an
19865 integral multiple of five---when it is 5, 10, 15, or some higher
19866 multiple of five. We can use this value as the Y axis height.
19868 A fairly simply way to determine whether a number is a multiple of
19869 five is to divide it by five and see if the division results in a
19870 remainder. If there is no remainder, the number is a multiple of
19871 five. Thus, seven divided by five has a remainder of two, and seven
19872 is not an integral multiple of five. Put in slightly different
19873 language, more reminiscent of the classroom, five goes into seven
19874 once, with a remainder of two. However, five goes into ten twice,
19875 with no remainder: ten is an integral multiple of five.
19877 @node Compute a Remainder
19878 @appendixsubsec Side Trip: Compute a Remainder
19880 @findex % @r{(remainder function)}
19881 @cindex Remainder function, @code{%}
19882 In Lisp, the function for computing a remainder is @code{%}. The
19883 function returns the remainder of its first argument divided by its
19884 second argument. As it happens, @code{%} is a function in Emacs Lisp
19885 that you cannot discover using @code{apropos}: you find nothing if you
19886 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
19887 learn of the existence of @code{%} is to read about it in a book such
19888 as this or in the Emacs Lisp sources.
19890 You can try the @code{%} function by evaluating the following two
19902 The first expression returns 2 and the second expression returns 0.
19904 To test whether the returned value is zero or some other number, we
19905 can use the @code{zerop} function. This function returns @code{t} if
19906 its argument, which must be a number, is zero.
19918 Thus, the following expression will return @code{t} if the height
19919 of the graph is evenly divisible by five:
19922 (zerop (% height 5))
19926 (The value of @code{height}, of course, can be found from @code{(apply
19927 'max numbers-list)}.)
19929 On the other hand, if the value of @code{height} is not a multiple of
19930 five, we want to reset the value to the next higher multiple of five.
19931 This is straightforward arithmetic using functions with which we are
19932 already familiar. First, we divide the value of @code{height} by five
19933 to determine how many times five goes into the number. Thus, five
19934 goes into twelve twice. If we add one to this quotient and multiply by
19935 five, we will obtain the value of the next multiple of five that is
19936 larger than the height. Five goes into twelve twice. Add one to two,
19937 and multiply by five; the result is fifteen, which is the next multiple
19938 of five that is higher than twelve. The Lisp expression for this is:
19941 (* (1+ (/ height 5)) 5)
19945 For example, if you evaluate the following, the result is 15:
19948 (* (1+ (/ 12 5)) 5)
19951 All through this discussion, we have been using `five' as the value
19952 for spacing labels on the Y axis; but we may want to use some other
19953 value. For generality, we should replace `five' with a variable to
19954 which we can assign a value. The best name I can think of for this
19955 variable is @code{Y-axis-label-spacing}.
19958 Using this term, and an @code{if} expression, we produce the
19963 (if (zerop (% height Y-axis-label-spacing))
19966 (* (1+ (/ height Y-axis-label-spacing))
19967 Y-axis-label-spacing))
19972 This expression returns the value of @code{height} itself if the height
19973 is an even multiple of the value of the @code{Y-axis-label-spacing} or
19974 else it computes and returns a value of @code{height} that is equal to
19975 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
19977 We can now include this expression in the @code{let} expression of the
19978 @code{print-graph} function (after first setting the value of
19979 @code{Y-axis-label-spacing}):
19980 @vindex Y-axis-label-spacing
19984 (defvar Y-axis-label-spacing 5
19985 "Number of lines from one Y axis label to next.")
19990 (let* ((height (apply 'max numbers-list))
19991 (height-of-top-line
19992 (if (zerop (% height Y-axis-label-spacing))
19997 (* (1+ (/ height Y-axis-label-spacing))
19998 Y-axis-label-spacing)))
19999 (symbol-width (length graph-blank))))
20005 (Note use of the @code{let*} function: the initial value of height is
20006 computed once by the @code{(apply 'max numbers-list)} expression and
20007 then the resulting value of @code{height} is used to compute its
20008 final value. @xref{fwd-para let, , The @code{let*} expression}, for
20009 more about @code{let*}.)
20011 @node Y Axis Element
20012 @appendixsubsec Construct a Y Axis Element
20014 When we print the vertical axis, we want to insert strings such as
20015 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20016 Moreover, we want the numbers and dashes to line up, so shorter
20017 numbers must be padded with leading spaces. If some of the strings
20018 use two digit numbers, the strings with single digit numbers must
20019 include a leading blank space before the number.
20021 @findex number-to-string
20022 To figure out the length of the number, the @code{length} function is
20023 used. But the @code{length} function works only with a string, not with
20024 a number. So the number has to be converted from being a number to
20025 being a string. This is done with the @code{number-to-string} function.
20030 (length (number-to-string 35))
20033 (length (number-to-string 100))
20039 (@code{number-to-string} is also called @code{int-to-string}; you will
20040 see this alternative name in various sources.)
20042 In addition, in each label, each number is followed by a string such
20043 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20044 This variable is defined with @code{defvar}:
20049 (defvar Y-axis-tic " - "
20050 "String that follows number in a Y axis label.")
20054 The length of the Y label is the sum of the length of the Y axis tic
20055 mark and the length of the number of the top of the graph.
20058 (length (concat (number-to-string height) Y-axis-tic)))
20061 This value will be calculated by the @code{print-graph} function in
20062 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20063 did not think to include this in the varlist when we first proposed it.)
20065 To make a complete vertical axis label, a tic mark is concatenated
20066 with a number; and the two together may be preceded by one or more
20067 spaces depending on how long the number is. The label consists of
20068 three parts: the (optional) leading spaces, the number, and the tic
20069 mark. The function is passed the value of the number for the specific
20070 row, and the value of the width of the top line, which is calculated
20071 (just once) by @code{print-graph}.
20075 (defun Y-axis-element (number full-Y-label-width)
20076 "Construct a NUMBERed label element.
20077 A numbered element looks like this ` 5 - ',
20078 and is padded as needed so all line up with
20079 the element for the largest number."
20082 (let* ((leading-spaces
20083 (- full-Y-label-width
20085 (concat (number-to-string number)
20090 (make-string leading-spaces ? )
20091 (number-to-string number)
20096 The @code{Y-axis-element} function concatenates together the leading
20097 spaces, if any; the number, as a string; and the tic mark.
20099 To figure out how many leading spaces the label will need, the
20100 function subtracts the actual length of the label---the length of the
20101 number plus the length of the tic mark---from the desired label width.
20103 @findex make-string
20104 Blank spaces are inserted using the @code{make-string} function. This
20105 function takes two arguments: the first tells it how long the string
20106 will be and the second is a symbol for the character to insert, in a
20107 special format. The format is a question mark followed by a blank
20108 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20109 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20110 syntax for characters. (Of course, you might want to replace the
20111 blank space by some other character @dots{} You know what to do.)
20113 The @code{number-to-string} function is used in the concatenation
20114 expression, to convert the number to a string that is concatenated
20115 with the leading spaces and the tic mark.
20117 @node Y-axis-column
20118 @appendixsubsec Create a Y Axis Column
20120 The preceding functions provide all the tools needed to construct a
20121 function that generates a list of numbered and blank strings to insert
20122 as the label for the vertical axis:
20124 @findex Y-axis-column
20127 (defun Y-axis-column (height width-of-label)
20128 "Construct list of Y axis labels and blank strings.
20129 For HEIGHT of line above base and WIDTH-OF-LABEL."
20133 (while (> height 1)
20134 (if (zerop (% height Y-axis-label-spacing))
20135 ;; @r{Insert label.}
20138 (Y-axis-element height width-of-label)
20142 ;; @r{Else, insert blanks.}
20145 (make-string width-of-label ? )
20147 (setq height (1- height)))
20148 ;; @r{Insert base line.}
20150 (cons (Y-axis-element 1 width-of-label) Y-axis))
20151 (nreverse Y-axis)))
20155 In this function, we start with the value of @code{height} and
20156 repetitively subtract one from its value. After each subtraction, we
20157 test to see whether the value is an integral multiple of the
20158 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20159 using the @code{Y-axis-element} function; if not, we construct a
20160 blank label using the @code{make-string} function. The base line
20161 consists of the number one followed by a tic mark.
20164 @node print-Y-axis Penultimate
20165 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20167 The list constructed by the @code{Y-axis-column} function is passed to
20168 the @code{print-Y-axis} function, which inserts the list as a column.
20170 @findex print-Y-axis
20173 (defun print-Y-axis (height full-Y-label-width)
20174 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20175 Height must be the maximum height of the graph.
20176 Full width is the width of the highest label element."
20177 ;; Value of height and full-Y-label-width
20178 ;; are passed by `print-graph'.
20181 (let ((start (point)))
20183 (Y-axis-column height full-Y-label-width))
20184 ;; @r{Place point ready for inserting graph.}
20186 ;; @r{Move point forward by value of} full-Y-label-width
20187 (forward-char full-Y-label-width)))
20191 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20192 insert the Y axis labels created by the @code{Y-axis-column} function.
20193 In addition, it places point at the correct position for printing the body of
20196 You can test @code{print-Y-axis}:
20204 Y-axis-label-spacing
20213 Copy the following expression:
20216 (print-Y-axis 12 5)
20220 Switch to the @file{*scratch*} buffer and place the cursor where you
20221 want the axis labels to start.
20224 Type @kbd{M-:} (@code{eval-expression}).
20227 Yank the @code{graph-body-print} expression into the minibuffer
20228 with @kbd{C-y} (@code{yank)}.
20231 Press @key{RET} to evaluate the expression.
20234 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20235 }}}. (The @code{print-graph} function will pass the value of
20236 @code{height-of-top-line}, which in this case will end up as 15,
20237 thereby getting rid of what might appear as a bug.)
20241 @appendixsec The @code{print-X-axis} Function
20242 @cindex Axis, print horizontal
20243 @cindex X axis printing
20244 @cindex Print horizontal axis
20245 @cindex Horizontal axis printing
20247 X axis labels are much like Y axis labels, except that the ticks are on a
20248 line above the numbers. Labels should look like this:
20257 The first tic is under the first column of the graph and is preceded by
20258 several blank spaces. These spaces provide room in rows above for the Y
20259 axis labels. The second, third, fourth, and subsequent ticks are all
20260 spaced equally, according to the value of @code{X-axis-label-spacing}.
20262 The second row of the X axis consists of numbers, preceded by several
20263 blank spaces and also separated according to the value of the variable
20264 @code{X-axis-label-spacing}.
20266 The value of the variable @code{X-axis-label-spacing} should itself be
20267 measured in units of @code{symbol-width}, since you may want to change
20268 the width of the symbols that you are using to print the body of the
20269 graph without changing the ways the graph is labeled.
20272 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
20273 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20277 @node Similarities differences
20278 @unnumberedsubsec Similarities and differences
20281 The @code{print-X-axis} function is constructed in more or less the
20282 same fashion as the @code{print-Y-axis} function except that it has
20283 two lines: the line of tic marks and the numbers. We will write a
20284 separate function to print each line and then combine them within the
20285 @code{print-X-axis} function.
20287 This is a three step process:
20291 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20294 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20297 Write a function to print both lines, the @code{print-X-axis} function,
20298 using @code{print-X-axis-tic-line} and
20299 @code{print-X-axis-numbered-line}.
20302 @node X Axis Tic Marks
20303 @appendixsubsec X Axis Tic Marks
20305 The first function should print the X axis tic marks. We must specify
20306 the tic marks themselves and their spacing:
20310 (defvar X-axis-label-spacing
20311 (if (boundp 'graph-blank)
20312 (* 5 (length graph-blank)) 5)
20313 "Number of units from one X axis label to next.")
20318 (Note that the value of @code{graph-blank} is set by another
20319 @code{defvar}. The @code{boundp} predicate checks whether it has
20320 already been set; @code{boundp} returns @code{nil} if it has not. If
20321 @code{graph-blank} were unbound and we did not use this conditional
20322 construction, in a recent GNU Emacs, we would enter the debugger and
20323 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20324 @w{(void-variable graph-blank)}}.)
20327 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20331 (defvar X-axis-tic-symbol "|"
20332 "String to insert to point to a column in X axis.")
20337 The goal is to make a line that looks like this:
20343 The first tic is indented so that it is under the first column, which is
20344 indented to provide space for the Y axis labels.
20346 A tic element consists of the blank spaces that stretch from one tic to
20347 the next plus a tic symbol. The number of blanks is determined by the
20348 width of the tic symbol and the @code{X-axis-label-spacing}.
20351 The code looks like this:
20355 ;;; X-axis-tic-element
20359 ;; @r{Make a string of blanks.}
20360 (- (* symbol-width X-axis-label-spacing)
20361 (length X-axis-tic-symbol))
20363 ;; @r{Concatenate blanks with tic symbol.}
20369 Next, we determine how many blanks are needed to indent the first tic
20370 mark to the first column of the graph. This uses the value of
20371 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20374 The code to make @code{X-axis-leading-spaces}
20379 ;; X-axis-leading-spaces
20381 (make-string full-Y-label-width ? )
20386 We also need to determine the length of the horizontal axis, which is
20387 the length of the numbers list, and the number of ticks in the horizontal
20394 (length numbers-list)
20400 (* symbol-width X-axis-label-spacing)
20404 ;; number-of-X-ticks
20405 (if (zerop (% (X-length tic-width)))
20406 (/ (X-length tic-width))
20407 (1+ (/ (X-length tic-width))))
20412 All this leads us directly to the function for printing the X axis tic line:
20414 @findex print-X-axis-tic-line
20417 (defun print-X-axis-tic-line
20418 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20419 "Print ticks for X axis."
20420 (insert X-axis-leading-spaces)
20421 (insert X-axis-tic-symbol) ; @r{Under first column.}
20424 ;; @r{Insert second tic in the right spot.}
20427 (- (* symbol-width X-axis-label-spacing)
20428 ;; @r{Insert white space up to second tic symbol.}
20429 (* 2 (length X-axis-tic-symbol)))
20431 X-axis-tic-symbol))
20434 ;; @r{Insert remaining ticks.}
20435 (while (> number-of-X-tics 1)
20436 (insert X-axis-tic-element)
20437 (setq number-of-X-tics (1- number-of-X-tics))))
20441 The line of numbers is equally straightforward:
20444 First, we create a numbered element with blank spaces before each number:
20446 @findex X-axis-element
20449 (defun X-axis-element (number)
20450 "Construct a numbered X axis element."
20451 (let ((leading-spaces
20452 (- (* symbol-width X-axis-label-spacing)
20453 (length (number-to-string number)))))
20454 (concat (make-string leading-spaces ? )
20455 (number-to-string number))))
20459 Next, we create the function to print the numbered line, starting with
20460 the number ``1'' under the first column:
20462 @findex print-X-axis-numbered-line
20465 (defun print-X-axis-numbered-line
20466 (number-of-X-tics X-axis-leading-spaces)
20467 "Print line of X-axis numbers"
20468 (let ((number X-axis-label-spacing))
20469 (insert X-axis-leading-spaces)
20475 ;; @r{Insert white space up to next number.}
20476 (- (* symbol-width X-axis-label-spacing) 2)
20478 (number-to-string number)))
20481 ;; @r{Insert remaining numbers.}
20482 (setq number (+ number X-axis-label-spacing))
20483 (while (> number-of-X-tics 1)
20484 (insert (X-axis-element number))
20485 (setq number (+ number X-axis-label-spacing))
20486 (setq number-of-X-tics (1- number-of-X-tics)))))
20490 Finally, we need to write the @code{print-X-axis} that uses
20491 @code{print-X-axis-tic-line} and
20492 @code{print-X-axis-numbered-line}.
20494 The function must determine the local values of the variables used by both
20495 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20496 then it must call them. Also, it must print the carriage return that
20497 separates the two lines.
20499 The function consists of a varlist that specifies five local variables,
20500 and calls to each of the two line printing functions:
20502 @findex print-X-axis
20505 (defun print-X-axis (numbers-list)
20506 "Print X axis labels to length of NUMBERS-LIST."
20507 (let* ((leading-spaces
20508 (make-string full-Y-label-width ? ))
20511 ;; symbol-width @r{is provided by} graph-body-print
20512 (tic-width (* symbol-width X-axis-label-spacing))
20513 (X-length (length numbers-list))
20521 ;; @r{Make a string of blanks.}
20522 (- (* symbol-width X-axis-label-spacing)
20523 (length X-axis-tic-symbol))
20527 ;; @r{Concatenate blanks with tic symbol.}
20528 X-axis-tic-symbol))
20532 (if (zerop (% X-length tic-width))
20533 (/ X-length tic-width)
20534 (1+ (/ X-length tic-width)))))
20537 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20539 (print-X-axis-numbered-line tic-number leading-spaces)))
20544 You can test @code{print-X-axis}:
20548 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20549 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20550 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20553 Copy the following expression:
20558 (let ((full-Y-label-width 5)
20561 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20566 Switch to the @file{*scratch*} buffer and place the cursor where you
20567 want the axis labels to start.
20570 Type @kbd{M-:} (@code{eval-expression}).
20573 Yank the test expression into the minibuffer
20574 with @kbd{C-y} (@code{yank)}.
20577 Press @key{RET} to evaluate the expression.
20581 Emacs will print the horizontal axis like this:
20591 @node Print Whole Graph
20592 @appendixsec Printing the Whole Graph
20593 @cindex Printing the whole graph
20594 @cindex Whole graph printing
20595 @cindex Graph, printing all
20597 Now we are nearly ready to print the whole graph.
20599 The function to print the graph with the proper labels follows the
20600 outline we created earlier (@pxref{Full Graph, , A Graph with Labeled
20601 Axes}), but with additions.
20604 Here is the outline:
20608 (defun print-graph (numbers-list)
20609 "@var{documentation}@dots{}"
20610 (let ((height @dots{}
20614 (print-Y-axis height @dots{} )
20615 (graph-body-print numbers-list)
20616 (print-X-axis @dots{} )))
20621 * The final version:: A few changes.
20622 * Test print-graph:: Run a short test.
20623 * Graphing words in defuns:: Executing the final code.
20624 * lambda:: How to write an anonymous function.
20625 * mapcar:: Apply a function to elements of a list.
20626 * Another Bug:: Yet another bug @dots{} most insidious.
20627 * Final printed graph:: The graph itself!
20631 @node The final version
20632 @unnumberedsubsec Changes for the Final Version
20635 The final version is different from what we planned in two ways:
20636 first, it contains additional values calculated once in the varlist;
20637 second, it carries an option to specify the labels' increment per row.
20638 This latter feature turns out to be essential; otherwise, a graph may
20639 have more rows than fit on a display or on a sheet of paper.
20642 This new feature requires a change to the @code{Y-axis-column}
20643 function, to add @code{vertical-step} to it. The function looks like
20646 @findex Y-axis-column @r{Final version.}
20649 ;;; @r{Final version.}
20650 (defun Y-axis-column
20651 (height width-of-label &optional vertical-step)
20652 "Construct list of labels for Y axis.
20653 HEIGHT is maximum height of graph.
20654 WIDTH-OF-LABEL is maximum width of label.
20655 VERTICAL-STEP, an option, is a positive integer
20656 that specifies how much a Y axis label increments
20657 for each line. For example, a step of 5 means
20658 that each line is five units of the graph."
20662 (number-per-line (or vertical-step 1)))
20663 (while (> height 1)
20664 (if (zerop (% height Y-axis-label-spacing))
20667 ;; @r{Insert label.}
20671 (* height number-per-line)
20676 ;; @r{Else, insert blanks.}
20679 (make-string width-of-label ? )
20681 (setq height (1- height)))
20684 ;; @r{Insert base line.}
20685 (setq Y-axis (cons (Y-axis-element
20686 (or vertical-step 1)
20689 (nreverse Y-axis)))
20693 The values for the maximum height of graph and the width of a symbol
20694 are computed by @code{print-graph} in its @code{let} expression; so
20695 @code{graph-body-print} must be changed to accept them.
20697 @findex graph-body-print @r{Final version.}
20700 ;;; @r{Final version.}
20701 (defun graph-body-print (numbers-list height symbol-width)
20702 "Print a bar graph of the NUMBERS-LIST.
20703 The numbers-list consists of the Y-axis values.
20704 HEIGHT is maximum height of graph.
20705 SYMBOL-WIDTH is number of each column."
20708 (let (from-position)
20709 (while numbers-list
20710 (setq from-position (point))
20712 (column-of-graph height (car numbers-list)))
20713 (goto-char from-position)
20714 (forward-char symbol-width)
20717 ;; @r{Draw graph column by column.}
20719 (setq numbers-list (cdr numbers-list)))
20720 ;; @r{Place point for X axis labels.}
20721 (forward-line height)
20727 Finally, the code for the @code{print-graph} function:
20729 @findex print-graph @r{Final version.}
20732 ;;; @r{Final version.}
20734 (numbers-list &optional vertical-step)
20735 "Print labeled bar graph of the NUMBERS-LIST.
20736 The numbers-list consists of the Y-axis values.
20740 Optionally, VERTICAL-STEP, a positive integer,
20741 specifies how much a Y axis label increments for
20742 each line. For example, a step of 5 means that
20743 each row is five units."
20746 (let* ((symbol-width (length graph-blank))
20747 ;; @code{height} @r{is both the largest number}
20748 ;; @r{and the number with the most digits.}
20749 (height (apply 'max numbers-list))
20752 (height-of-top-line
20753 (if (zerop (% height Y-axis-label-spacing))
20756 (* (1+ (/ height Y-axis-label-spacing))
20757 Y-axis-label-spacing)))
20760 (vertical-step (or vertical-step 1))
20761 (full-Y-label-width
20767 (* height-of-top-line vertical-step))
20773 height-of-top-line full-Y-label-width vertical-step)
20777 numbers-list height-of-top-line symbol-width)
20778 (print-X-axis numbers-list)))
20782 @node Test print-graph
20783 @appendixsubsec Testing @code{print-graph}
20786 We can test the @code{print-graph} function with a short list of numbers:
20790 Install the final versions of @code{Y-axis-column},
20791 @code{graph-body-print}, and @code{print-graph} (in addition to the
20795 Copy the following expression:
20798 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20802 Switch to the @file{*scratch*} buffer and place the cursor where you
20803 want the axis labels to start.
20806 Type @kbd{M-:} (@code{eval-expression}).
20809 Yank the test expression into the minibuffer
20810 with @kbd{C-y} (@code{yank)}.
20813 Press @key{RET} to evaluate the expression.
20817 Emacs will print a graph that looks like this:
20838 On the other hand, if you pass @code{print-graph} a
20839 @code{vertical-step} value of 2, by evaluating this expression:
20842 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20847 The graph looks like this:
20868 (A question: is the `2' on the bottom of the vertical axis a bug or a
20869 feature? If you think it is a bug, and should be a `1' instead, (or
20870 even a `0'), you can modify the sources.)
20872 @node Graphing words in defuns
20873 @appendixsubsec Graphing Numbers of Words and Symbols
20875 Now for the graph for which all this code was written: a graph that
20876 shows how many function definitions contain fewer than 10 words and
20877 symbols, how many contain between 10 and 19 words and symbols, how
20878 many contain between 20 and 29 words and symbols, and so on.
20880 This is a multi-step process. First make sure you have loaded all the
20884 It is a good idea to reset the value of @code{top-of-ranges} in case
20885 you have set it to some different value. You can evaluate the
20890 (setq top-of-ranges
20893 110 120 130 140 150
20894 160 170 180 190 200
20895 210 220 230 240 250
20896 260 270 280 290 300)
20901 Next create a list of the number of words and symbols in each range.
20905 Evaluate the following:
20909 (setq list-for-graph
20912 (recursive-lengths-list-many-files
20913 (directory-files "/usr/local/emacs/lisp"
20921 On my old machine, this took about an hour. It looked though 303 Lisp
20922 files in my copy of Emacs version 19.23. After all that computing,
20923 the @code{list-for-graph} had this value:
20927 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
20928 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
20933 This means that my copy of Emacs had 537 function definitions with
20934 fewer than 10 words or symbols in them, 1,027 function definitions
20935 with 10 to 19 words or symbols in them, 955 function definitions with
20936 20 to 29 words or symbols in them, and so on.
20938 Clearly, just by looking at this list we can see that most function
20939 definitions contain ten to thirty words and symbols.
20941 Now for printing. We do @emph{not} want to print a graph that is
20942 1,030 lines high @dots{} Instead, we should print a graph that is
20943 fewer than twenty-five lines high. A graph that height can be
20944 displayed on almost any monitor, and easily printed on a sheet of paper.
20946 This means that each value in @code{list-for-graph} must be reduced to
20947 one-fiftieth its present value.
20949 Here is a short function to do just that, using two functions we have
20950 not yet seen, @code{mapcar} and @code{lambda}.
20954 (defun one-fiftieth (full-range)
20955 "Return list, each number one-fiftieth of previous."
20956 (mapcar (lambda (arg) (/ arg 50)) full-range))
20961 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
20962 @cindex Anonymous function
20965 @code{lambda} is the symbol for an anonymous function, a function
20966 without a name. Every time you use an anonymous function, you need to
20967 include its whole body.
20974 (lambda (arg) (/ arg 50))
20978 is a function definition that says `return the value resulting from
20979 dividing whatever is passed to me as @code{arg} by 50'.
20982 Earlier, for example, we had a function @code{multiply-by-seven}; it
20983 multiplied its argument by 7. This function is similar, except it
20984 divides its argument by 50; and, it has no name. The anonymous
20985 equivalent of @code{multiply-by-seven} is:
20988 (lambda (number) (* 7 number))
20992 (@xref{defun, , The @code{defun} Macro}.)
20996 If we want to multiply 3 by 7, we can write:
20998 @c !!! Clear print-postscript-figures if the computer formatting this
20999 @c document is too small and cannot handle all the diagrams and figures.
21000 @c clear print-postscript-figures
21001 @c set print-postscript-figures
21002 @c lambda example diagram #1
21006 (multiply-by-seven 3)
21007 \_______________/ ^
21013 @ifset print-postscript-figures
21016 @center @image{lambda-1}
21017 %%%% old method of including an image
21018 % \input /usr/local/lib/tex/inputs/psfig.tex
21019 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21024 @ifclear print-postscript-figures
21028 (multiply-by-seven 3)
21029 \_______________/ ^
21038 This expression returns 21.
21042 Similarly, we can write:
21044 @c lambda example diagram #2
21048 ((lambda (number) (* 7 number)) 3)
21049 \____________________________/ ^
21051 anonymous function argument
21055 @ifset print-postscript-figures
21058 @center @image{lambda-2}
21059 %%%% old method of including an image
21060 % \input /usr/local/lib/tex/inputs/psfig.tex
21061 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21066 @ifclear print-postscript-figures
21070 ((lambda (number) (* 7 number)) 3)
21071 \____________________________/ ^
21073 anonymous function argument
21081 If we want to divide 100 by 50, we can write:
21083 @c lambda example diagram #3
21087 ((lambda (arg) (/ arg 50)) 100)
21088 \______________________/ \_/
21090 anonymous function argument
21094 @ifset print-postscript-figures
21097 @center @image{lambda-3}
21098 %%%% old method of including an image
21099 % \input /usr/local/lib/tex/inputs/psfig.tex
21100 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21105 @ifclear print-postscript-figures
21109 ((lambda (arg) (/ arg 50)) 100)
21110 \______________________/ \_/
21112 anonymous function argument
21119 This expression returns 2. The 100 is passed to the function, which
21120 divides that number by 50.
21122 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21123 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21124 expressions derive from the Lambda Calculus.
21127 @appendixsubsec The @code{mapcar} Function
21130 @code{mapcar} is a function that calls its first argument with each
21131 element of its second argument, in turn. The second argument must be
21134 The @samp{map} part of the name comes from the mathematical phrase,
21135 `mapping over a domain', meaning to apply a function to each of the
21136 elements in a domain. The mathematical phrase is based on the
21137 metaphor of a surveyor walking, one step at a time, over an area he is
21138 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21147 (mapcar '1+ '(2 4 6))
21153 The function @code{1+} which adds one to its argument, is executed on
21154 @emph{each} element of the list, and a new list is returned.
21156 Contrast this with @code{apply}, which applies its first argument to
21158 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21162 In the definition of @code{one-fiftieth}, the first argument is the
21163 anonymous function:
21166 (lambda (arg) (/ arg 50))
21170 and the second argument is @code{full-range}, which will be bound to
21171 @code{list-for-graph}.
21174 The whole expression looks like this:
21177 (mapcar (lambda (arg) (/ arg 50)) full-range))
21180 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21181 Lisp Reference Manual}, for more about @code{mapcar}.
21183 Using the @code{one-fiftieth} function, we can generate a list in
21184 which each element is one-fiftieth the size of the corresponding
21185 element in @code{list-for-graph}.
21189 (setq fiftieth-list-for-graph
21190 (one-fiftieth list-for-graph))
21195 The resulting list looks like this:
21199 (10 20 19 15 11 9 6 5 4 3 3 2 2
21200 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21205 This, we are almost ready to print! (We also notice the loss of
21206 information: many of the higher ranges are 0, meaning that fewer than
21207 50 defuns had that many words or symbols---but not necessarily meaning
21208 that none had that many words or symbols.)
21211 @appendixsubsec Another Bug @dots{} Most Insidious
21212 @cindex Bug, most insidious type
21213 @cindex Insidious type of bug
21215 I said `almost ready to print'! Of course, there is a bug in the
21216 @code{print-graph} function @dots{} It has a @code{vertical-step}
21217 option, but not a @code{horizontal-step} option. The
21218 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21219 @code{print-graph} function will print only by ones.
21221 This is a classic example of what some consider the most insidious
21222 type of bug, the bug of omission. This is not the kind of bug you can
21223 find by studying the code, for it is not in the code; it is an omitted
21224 feature. Your best actions are to try your program early and often;
21225 and try to arrange, as much as you can, to write code that is easy to
21226 understand and easy to change. Try to be aware, whenever you can,
21227 that whatever you have written, @emph{will} be rewritten, if not soon,
21228 eventually. A hard maxim to follow.
21230 It is the @code{print-X-axis-numbered-line} function that needs the
21231 work; and then the @code{print-X-axis} and the @code{print-graph}
21232 functions need to be adapted. Not much needs to be done; there is one
21233 nicety: the numbers ought to line up under the tic marks. This takes
21237 Here is the corrected @code{print-X-axis-numbered-line}:
21241 (defun print-X-axis-numbered-line
21242 (number-of-X-tics X-axis-leading-spaces
21243 &optional horizontal-step)
21244 "Print line of X-axis numbers"
21245 (let ((number X-axis-label-spacing)
21246 (horizontal-step (or horizontal-step 1)))
21249 (insert X-axis-leading-spaces)
21250 ;; @r{Delete extra leading spaces.}
21253 (length (number-to-string horizontal-step)))))
21258 ;; @r{Insert white space.}
21260 X-axis-label-spacing)
21263 (number-to-string horizontal-step)))
21267 (* number horizontal-step))))
21270 ;; @r{Insert remaining numbers.}
21271 (setq number (+ number X-axis-label-spacing))
21272 (while (> number-of-X-tics 1)
21273 (insert (X-axis-element
21274 (* number horizontal-step)))
21275 (setq number (+ number X-axis-label-spacing))
21276 (setq number-of-X-tics (1- number-of-X-tics)))))
21281 If you are reading this in Info, you can see the new versions of
21282 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21283 reading this in a printed book, you can see the changed lines here
21284 (the full text is too much to print).
21289 (defun print-X-axis (numbers-list horizontal-step)
21291 (print-X-axis-numbered-line
21292 tic-number leading-spaces horizontal-step))
21300 &optional vertical-step horizontal-step)
21302 (print-X-axis numbers-list horizontal-step))
21310 (defun print-X-axis (numbers-list horizontal-step)
21311 "Print X axis labels to length of NUMBERS-LIST.
21312 Optionally, HORIZONTAL-STEP, a positive integer,
21313 specifies how much an X axis label increments for
21317 ;; Value of symbol-width and full-Y-label-width
21318 ;; are passed by `print-graph'.
21319 (let* ((leading-spaces
21320 (make-string full-Y-label-width ? ))
21321 ;; symbol-width @r{is provided by} graph-body-print
21322 (tic-width (* symbol-width X-axis-label-spacing))
21323 (X-length (length numbers-list))
21329 ;; @r{Make a string of blanks.}
21330 (- (* symbol-width X-axis-label-spacing)
21331 (length X-axis-tic-symbol))
21335 ;; @r{Concatenate blanks with tic symbol.}
21336 X-axis-tic-symbol))
21338 (if (zerop (% X-length tic-width))
21339 (/ X-length tic-width)
21340 (1+ (/ X-length tic-width)))))
21344 (print-X-axis-tic-line
21345 tic-number leading-spaces X-tic)
21347 (print-X-axis-numbered-line
21348 tic-number leading-spaces horizontal-step)))
21355 (numbers-list &optional vertical-step horizontal-step)
21356 "Print labeled bar graph of the NUMBERS-LIST.
21357 The numbers-list consists of the Y-axis values.
21361 Optionally, VERTICAL-STEP, a positive integer,
21362 specifies how much a Y axis label increments for
21363 each line. For example, a step of 5 means that
21364 each row is five units.
21368 Optionally, HORIZONTAL-STEP, a positive integer,
21369 specifies how much an X axis label increments for
21371 (let* ((symbol-width (length graph-blank))
21372 ;; @code{height} @r{is both the largest number}
21373 ;; @r{and the number with the most digits.}
21374 (height (apply 'max numbers-list))
21377 (height-of-top-line
21378 (if (zerop (% height Y-axis-label-spacing))
21381 (* (1+ (/ height Y-axis-label-spacing))
21382 Y-axis-label-spacing)))
21385 (vertical-step (or vertical-step 1))
21386 (full-Y-label-width
21390 (* height-of-top-line vertical-step))
21395 height-of-top-line full-Y-label-width vertical-step)
21397 numbers-list height-of-top-line symbol-width)
21398 (print-X-axis numbers-list horizontal-step)))
21405 Graphing Definitions Re-listed
21408 Here are all the graphing definitions in their final form:
21412 (defvar top-of-ranges
21415 110 120 130 140 150
21416 160 170 180 190 200
21417 210 220 230 240 250)
21418 "List specifying ranges for `defuns-per-range'.")
21422 (defvar graph-symbol "*"
21423 "String used as symbol in graph, usually an asterisk.")
21427 (defvar graph-blank " "
21428 "String used as blank in graph, usually a blank space.
21429 graph-blank must be the same number of columns wide
21434 (defvar Y-axis-tic " - "
21435 "String that follows number in a Y axis label.")
21439 (defvar Y-axis-label-spacing 5
21440 "Number of lines from one Y axis label to next.")
21444 (defvar X-axis-tic-symbol "|"
21445 "String to insert to point to a column in X axis.")
21449 (defvar X-axis-label-spacing
21450 (if (boundp 'graph-blank)
21451 (* 5 (length graph-blank)) 5)
21452 "Number of units from one X axis label to next.")
21458 (defun count-words-in-defun ()
21459 "Return the number of words and symbols in a defun."
21460 (beginning-of-defun)
21462 (end (save-excursion (end-of-defun) (point))))
21467 (and (< (point) end)
21469 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21471 (setq count (1+ count)))
21478 (defun lengths-list-file (filename)
21479 "Return list of definitions' lengths within FILE.
21480 The returned list is a list of numbers.
21481 Each number is the number of words or
21482 symbols in one function definition."
21486 (message "Working on `%s' ... " filename)
21488 (let ((buffer (find-file-noselect filename))
21490 (set-buffer buffer)
21491 (setq buffer-read-only t)
21493 (goto-char (point-min))
21497 (while (re-search-forward "^(defun" nil t)
21499 (cons (count-words-in-defun) lengths-list)))
21500 (kill-buffer buffer)
21507 (defun lengths-list-many-files (list-of-files)
21508 "Return list of lengths of defuns in LIST-OF-FILES."
21509 (let (lengths-list)
21510 ;;; @r{true-or-false-test}
21511 (while list-of-files
21517 ;;; @r{Generate a lengths' list.}
21519 (expand-file-name (car list-of-files)))))
21520 ;;; @r{Make files' list shorter.}
21521 (setq list-of-files (cdr list-of-files)))
21522 ;;; @r{Return final value of lengths' list.}
21529 (defun defuns-per-range (sorted-lengths top-of-ranges)
21530 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21531 (let ((top-of-range (car top-of-ranges))
21532 (number-within-range 0)
21533 defuns-per-range-list)
21538 (while top-of-ranges
21542 ;; @r{Need number for numeric test.}
21543 (car sorted-lengths)
21544 (< (car sorted-lengths) top-of-range))
21546 ;; @r{Count number of definitions within current range.}
21547 (setq number-within-range (1+ number-within-range))
21548 (setq sorted-lengths (cdr sorted-lengths)))
21552 ;; @r{Exit inner loop but remain within outer loop.}
21554 (setq defuns-per-range-list
21555 (cons number-within-range defuns-per-range-list))
21556 (setq number-within-range 0) ; @r{Reset count to zero.}
21558 ;; @r{Move to next range.}
21559 (setq top-of-ranges (cdr top-of-ranges))
21560 ;; @r{Specify next top of range value.}
21561 (setq top-of-range (car top-of-ranges)))
21565 ;; @r{Exit outer loop and count the number of defuns larger than}
21566 ;; @r{ the largest top-of-range value.}
21567 (setq defuns-per-range-list
21569 (length sorted-lengths)
21570 defuns-per-range-list))
21572 ;; @r{Return a list of the number of definitions within each range,}
21573 ;; @r{ smallest to largest.}
21574 (nreverse defuns-per-range-list)))
21580 (defun column-of-graph (max-graph-height actual-height)
21581 "Return list of MAX-GRAPH-HEIGHT strings;
21582 ACTUAL-HEIGHT are graph-symbols.
21583 The graph-symbols are contiguous entries at the end
21585 The list will be inserted as one column of a graph.
21586 The strings are either graph-blank or graph-symbol."
21590 (let ((insert-list nil)
21591 (number-of-top-blanks
21592 (- max-graph-height actual-height)))
21594 ;; @r{Fill in @code{graph-symbols}.}
21595 (while (> actual-height 0)
21596 (setq insert-list (cons graph-symbol insert-list))
21597 (setq actual-height (1- actual-height)))
21601 ;; @r{Fill in @code{graph-blanks}.}
21602 (while (> number-of-top-blanks 0)
21603 (setq insert-list (cons graph-blank insert-list))
21604 (setq number-of-top-blanks
21605 (1- number-of-top-blanks)))
21607 ;; @r{Return whole list.}
21614 (defun Y-axis-element (number full-Y-label-width)
21615 "Construct a NUMBERed label element.
21616 A numbered element looks like this ` 5 - ',
21617 and is padded as needed so all line up with
21618 the element for the largest number."
21621 (let* ((leading-spaces
21622 (- full-Y-label-width
21624 (concat (number-to-string number)
21629 (make-string leading-spaces ? )
21630 (number-to-string number)
21637 (defun print-Y-axis
21638 (height full-Y-label-width &optional vertical-step)
21639 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21640 Height must be the maximum height of the graph.
21641 Full width is the width of the highest label element.
21642 Optionally, print according to VERTICAL-STEP."
21645 ;; Value of height and full-Y-label-width
21646 ;; are passed by `print-graph'.
21647 (let ((start (point)))
21649 (Y-axis-column height full-Y-label-width vertical-step))
21652 ;; @r{Place point ready for inserting graph.}
21654 ;; @r{Move point forward by value of} full-Y-label-width
21655 (forward-char full-Y-label-width)))
21661 (defun print-X-axis-tic-line
21662 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21663 "Print ticks for X axis."
21664 (insert X-axis-leading-spaces)
21665 (insert X-axis-tic-symbol) ; @r{Under first column.}
21668 ;; @r{Insert second tic in the right spot.}
21671 (- (* symbol-width X-axis-label-spacing)
21672 ;; @r{Insert white space up to second tic symbol.}
21673 (* 2 (length X-axis-tic-symbol)))
21675 X-axis-tic-symbol))
21678 ;; @r{Insert remaining ticks.}
21679 (while (> number-of-X-tics 1)
21680 (insert X-axis-tic-element)
21681 (setq number-of-X-tics (1- number-of-X-tics))))
21687 (defun X-axis-element (number)
21688 "Construct a numbered X axis element."
21689 (let ((leading-spaces
21690 (- (* symbol-width X-axis-label-spacing)
21691 (length (number-to-string number)))))
21692 (concat (make-string leading-spaces ? )
21693 (number-to-string number))))
21699 (defun graph-body-print (numbers-list height symbol-width)
21700 "Print a bar graph of the NUMBERS-LIST.
21701 The numbers-list consists of the Y-axis values.
21702 HEIGHT is maximum height of graph.
21703 SYMBOL-WIDTH is number of each column."
21706 (let (from-position)
21707 (while numbers-list
21708 (setq from-position (point))
21710 (column-of-graph height (car numbers-list)))
21711 (goto-char from-position)
21712 (forward-char symbol-width)
21715 ;; @r{Draw graph column by column.}
21717 (setq numbers-list (cdr numbers-list)))
21718 ;; @r{Place point for X axis labels.}
21719 (forward-line height)
21726 (defun Y-axis-column
21727 (height width-of-label &optional vertical-step)
21728 "Construct list of labels for Y axis.
21729 HEIGHT is maximum height of graph.
21730 WIDTH-OF-LABEL is maximum width of label.
21733 VERTICAL-STEP, an option, is a positive integer
21734 that specifies how much a Y axis label increments
21735 for each line. For example, a step of 5 means
21736 that each line is five units of the graph."
21738 (number-per-line (or vertical-step 1)))
21741 (while (> height 1)
21742 (if (zerop (% height Y-axis-label-spacing))
21743 ;; @r{Insert label.}
21747 (* height number-per-line)
21752 ;; @r{Else, insert blanks.}
21755 (make-string width-of-label ? )
21757 (setq height (1- height)))
21760 ;; @r{Insert base line.}
21761 (setq Y-axis (cons (Y-axis-element
21762 (or vertical-step 1)
21765 (nreverse Y-axis)))
21771 (defun print-X-axis-numbered-line
21772 (number-of-X-tics X-axis-leading-spaces
21773 &optional horizontal-step)
21774 "Print line of X-axis numbers"
21775 (let ((number X-axis-label-spacing)
21776 (horizontal-step (or horizontal-step 1)))
21779 (insert X-axis-leading-spaces)
21781 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21784 ;; @r{Insert white space up to next number.}
21785 (- (* symbol-width X-axis-label-spacing)
21786 (1- (length (number-to-string horizontal-step)))
21789 (number-to-string (* number horizontal-step))))
21792 ;; @r{Insert remaining numbers.}
21793 (setq number (+ number X-axis-label-spacing))
21794 (while (> number-of-X-tics 1)
21795 (insert (X-axis-element (* number horizontal-step)))
21796 (setq number (+ number X-axis-label-spacing))
21797 (setq number-of-X-tics (1- number-of-X-tics)))))
21803 (defun print-X-axis (numbers-list horizontal-step)
21804 "Print X axis labels to length of NUMBERS-LIST.
21805 Optionally, HORIZONTAL-STEP, a positive integer,
21806 specifies how much an X axis label increments for
21810 ;; Value of symbol-width and full-Y-label-width
21811 ;; are passed by `print-graph'.
21812 (let* ((leading-spaces
21813 (make-string full-Y-label-width ? ))
21814 ;; symbol-width @r{is provided by} graph-body-print
21815 (tic-width (* symbol-width X-axis-label-spacing))
21816 (X-length (length numbers-list))
21822 ;; @r{Make a string of blanks.}
21823 (- (* symbol-width X-axis-label-spacing)
21824 (length X-axis-tic-symbol))
21828 ;; @r{Concatenate blanks with tic symbol.}
21829 X-axis-tic-symbol))
21831 (if (zerop (% X-length tic-width))
21832 (/ X-length tic-width)
21833 (1+ (/ X-length tic-width)))))
21837 (print-X-axis-tic-line
21838 tic-number leading-spaces X-tic)
21840 (print-X-axis-numbered-line
21841 tic-number leading-spaces horizontal-step)))
21847 (defun one-fiftieth (full-range)
21848 "Return list, each number of which is 1/50th previous."
21849 (mapcar (lambda (arg) (/ arg 50)) full-range))
21856 (numbers-list &optional vertical-step horizontal-step)
21857 "Print labeled bar graph of the NUMBERS-LIST.
21858 The numbers-list consists of the Y-axis values.
21862 Optionally, VERTICAL-STEP, a positive integer,
21863 specifies how much a Y axis label increments for
21864 each line. For example, a step of 5 means that
21865 each row is five units.
21869 Optionally, HORIZONTAL-STEP, a positive integer,
21870 specifies how much an X axis label increments for
21872 (let* ((symbol-width (length graph-blank))
21873 ;; @code{height} @r{is both the largest number}
21874 ;; @r{and the number with the most digits.}
21875 (height (apply 'max numbers-list))
21878 (height-of-top-line
21879 (if (zerop (% height Y-axis-label-spacing))
21882 (* (1+ (/ height Y-axis-label-spacing))
21883 Y-axis-label-spacing)))
21886 (vertical-step (or vertical-step 1))
21887 (full-Y-label-width
21891 (* height-of-top-line vertical-step))
21897 height-of-top-line full-Y-label-width vertical-step)
21899 numbers-list height-of-top-line symbol-width)
21900 (print-X-axis numbers-list horizontal-step)))
21907 @node Final printed graph
21908 @appendixsubsec The Printed Graph
21910 When made and installed, you can call the @code{print-graph} command
21916 (print-graph fiftieth-list-for-graph 50 10)
21946 50 - ***************** * *
21948 10 50 100 150 200 250 300 350
21955 The largest group of functions contain 10--19 words and symbols each.
21957 @node Free Software and Free Manuals
21958 @appendix Free Software and Free Manuals
21960 @strong{by Richard M. Stallman}
21963 The biggest deficiency in free operating systems is not in the
21964 software---it is the lack of good free manuals that we can include in
21965 these systems. Many of our most important programs do not come with
21966 full manuals. Documentation is an essential part of any software
21967 package; when an important free software package does not come with a
21968 free manual, that is a major gap. We have many such gaps today.
21970 Once upon a time, many years ago, I thought I would learn Perl. I got
21971 a copy of a free manual, but I found it hard to read. When I asked
21972 Perl users about alternatives, they told me that there were better
21973 introductory manuals---but those were not free.
21975 Why was this? The authors of the good manuals had written them for
21976 O'Reilly Associates, which published them with restrictive terms---no
21977 copying, no modification, source files not available---which exclude
21978 them from the free software community.
21980 That wasn't the first time this sort of thing has happened, and (to
21981 our community's great loss) it was far from the last. Proprietary
21982 manual publishers have enticed a great many authors to restrict their
21983 manuals since then. Many times I have heard a GNU user eagerly tell me
21984 about a manual that he is writing, with which he expects to help the
21985 GNU project---and then had my hopes dashed, as he proceeded to explain
21986 that he had signed a contract with a publisher that would restrict it
21987 so that we cannot use it.
21989 Given that writing good English is a rare skill among programmers, we
21990 can ill afford to lose manuals this way.
21992 Free documentation, like free software, is a matter of freedom, not
21993 price. The problem with these manuals was not that O'Reilly Associates
21994 charged a price for printed copies---that in itself is fine. The Free
21995 Software Foundation @uref{http://shop.fsf.org, sells printed copies} of
21996 free @uref{http://www.gnu.org/doc/doc.html, GNU manuals}, too.
21997 But GNU manuals are available in source code form, while these manuals
21998 are available only on paper. GNU manuals come with permission to copy
21999 and modify; the Perl manuals do not. These restrictions are the
22002 The criterion for a free manual is pretty much the same as for free
22003 software: it is a matter of giving all users certain
22004 freedoms. Redistribution (including commercial redistribution) must be
22005 permitted, so that the manual can accompany every copy of the program,
22006 on-line or on paper. Permission for modification is crucial too.
22008 As a general rule, I don't believe that it is essential for people to
22009 have permission to modify all sorts of articles and books. The issues
22010 for writings are not necessarily the same as those for software. For
22011 example, I don't think you or I are obliged to give permission to
22012 modify articles like this one, which describe our actions and our
22015 But there is a particular reason why the freedom to modify is crucial
22016 for documentation for free software. When people exercise their right
22017 to modify the software, and add or change its features, if they are
22018 conscientious they will change the manual too---so they can provide
22019 accurate and usable documentation with the modified program. A manual
22020 which forbids programmers to be conscientious and finish the job, or
22021 more precisely requires them to write a new manual from scratch if
22022 they change the program, does not fill our community's needs.
22024 While a blanket prohibition on modification is unacceptable, some
22025 kinds of limits on the method of modification pose no problem. For
22026 example, requirements to preserve the original author's copyright
22027 notice, the distribution terms, or the list of authors, are ok. It is
22028 also no problem to require modified versions to include notice that
22029 they were modified, even to have entire sections that may not be
22030 deleted or changed, as long as these sections deal with nontechnical
22031 topics. (Some GNU manuals have them.)
22033 These kinds of restrictions are not a problem because, as a practical
22034 matter, they don't stop the conscientious programmer from adapting the
22035 manual to fit the modified program. In other words, they don't block
22036 the free software community from making full use of the manual.
22038 However, it must be possible to modify all the technical content of
22039 the manual, and then distribute the result in all the usual media,
22040 through all the usual channels; otherwise, the restrictions do block
22041 the community, the manual is not free, and so we need another manual.
22043 Unfortunately, it is often hard to find someone to write another
22044 manual when a proprietary manual exists. The obstacle is that many
22045 users think that a proprietary manual is good enough---so they don't
22046 see the need to write a free manual. They do not see that the free
22047 operating system has a gap that needs filling.
22049 Why do users think that proprietary manuals are good enough? Some have
22050 not considered the issue. I hope this article will do something to
22053 Other users consider proprietary manuals acceptable for the same
22054 reason so many people consider proprietary software acceptable: they
22055 judge in purely practical terms, not using freedom as a
22056 criterion. These people are entitled to their opinions, but since
22057 those opinions spring from values which do not include freedom, they
22058 are no guide for those of us who do value freedom.
22060 Please spread the word about this issue. We continue to lose manuals
22061 to proprietary publishing. If we spread the word that proprietary
22062 manuals are not sufficient, perhaps the next person who wants to help
22063 GNU by writing documentation will realize, before it is too late, that
22064 he must above all make it free.
22066 We can also encourage commercial publishers to sell free, copylefted
22067 manuals instead of proprietary ones. One way you can help this is to
22068 check the distribution terms of a manual before you buy it, and prefer
22069 copylefted manuals to non-copylefted ones.
22073 Note: The Free Software Foundation maintains a page on its Web site
22074 that lists free books available from other publishers:@*
22075 @uref{http://www.gnu.org/doc/other-free-books.html}
22077 @node GNU Free Documentation License
22078 @appendix GNU Free Documentation License
22080 @cindex FDL, GNU Free Documentation License
22081 @include doclicense.texi
22087 MENU ENTRY: NODE NAME.
22093 @c Place biographical information on right-hand (verso) page
22096 \par\vfill\supereject
22098 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22099 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22102 % \par\vfill\supereject
22103 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22104 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22105 %\page\hbox{}%\page
22106 %\page\hbox{}%\page
22113 @c ================ Biographical information ================
22117 @center About the Author
22122 @node About the Author
22123 @unnumbered About the Author
22127 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22128 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22129 world on software freedom. Chassell was a founding Director and
22130 Treasurer of the Free Software Foundation, Inc. He is co-author of
22131 the @cite{Texinfo} manual, and has edited more than a dozen other
22132 books. He graduated from Cambridge University, in England. He has an
22133 abiding interest in social and economic history and flies his own
22140 @c @c Prevent page number on blank verso, so eject it first.
22142 @c \par\vfill\supereject
22147 @c @evenheading @thispage @| @| @thistitle
22148 @c @oddheading @| @| @thispage