1 \input texinfo @c -*-texinfo-*-
2 @comment %**start of header
3 @setfilename ../../info/eintr
4 @c setfilename emacs-lisp-intro.info
5 @c sethtmlfilename emacs-lisp-intro.html
6 @settitle Programming in Emacs Lisp
12 @c <<<< For hard copy printing, this file is now
13 @c set for smallbook, which works for all sizes
14 @c of paper, and with PostScript figures >>>>
20 @set print-postscript-figures
22 @c clear print-postscript-figures
25 @comment %**end of header
27 @c per rms and peterb, use 10pt fonts for the main text, mostly to
28 @c save on paper cost.
29 @c Do this inside @tex for now, so current makeinfo does not complain.
35 \global\hbadness=6666 % don't worry about not-too-underfull boxes
38 @set edition-number 3.10
39 @set update-date 28 October 2009
40 @c FIXME can this be updated? -- xfq
43 ## Summary of shell commands to create various output formats:
45 pushd /usr/local/src/emacs/lispintro/
49 makeinfo --paragraph-indent=0 --verbose emacs-lisp-intro.texi
51 ## ;; (progn (when (bufferp (get-buffer "*info*")) (kill-buffer "*info*")) (info "/usr/local/src/emacs/info/eintr"))
54 texi2dvi emacs-lisp-intro.texi
56 ## xdvi -margins 24pt -topmargin 4pt -offsets 24pt -geometry 760x1140 -s 5 -useTeXpages -mousemode 1 emacs-lisp-intro.dvi &
59 makeinfo --html --no-split --verbose emacs-lisp-intro.texi
61 ## galeon emacs-lisp-intro.html
64 makeinfo --fill-column=70 --no-split --paragraph-indent=0 \
65 --verbose --no-headers --output=emacs-lisp-intro.txt emacs-lisp-intro.texi
71 mtusb # mount -v -t ext3 /dev/sda /mnt
72 cp -v /u/intro/emacs-lisp-intro.texi /mnt/backup/intro/emacs-lisp-intro.texi
73 umtusb # umount -v /mnt
76 ## Other shell commands
78 pushd /usr/local/src/emacs/lispintro/
82 texi2dvi --pdf emacs-lisp-intro.texi
83 # xpdf emacs-lisp-intro.pdf &
85 ## DocBook -- note file extension
86 makeinfo --docbook --no-split --paragraph-indent=0 \
87 --verbose --output=emacs-lisp-intro.docbook emacs-lisp-intro.texi
89 ## XML with a Texinfo DTD -- note file extension
90 makeinfo --xml --no-split --paragraph-indent=0 \
91 --verbose --output=emacs-lisp-intro.texinfoxml emacs-lisp-intro.texi
93 ## PostScript (needs DVI)
94 # gv emacs-lisp-intro.ps &
95 # Create DVI if we lack it
96 # texi2dvi emacs-lisp-intro.texi
97 dvips emacs-lisp-intro.dvi -o emacs-lisp-intro.ps
100 # Use OpenOffice to view RTF
101 # Create HTML if we lack it
102 # makeinfo --no-split --html emacs-lisp-intro.texi
103 /usr/local/src/html2rtf.pl emacs-lisp-intro.html
106 /usr/bin/rtf2latex emacs-lisp-intro.rtf
112 @c ================ Included Figures ================
114 @c Set print-postscript-figures if you print PostScript figures.
115 @c If you clear this, the ten figures will be printed as ASCII diagrams.
116 @c (This is not relevant to Info, since Info only handles ASCII.)
117 @c Your site may require editing changes to print PostScript; in this
118 @c case, search for `print-postscript-figures' and make appropriate changes.
120 @c ================ How to Create an Info file ================
122 @c If you have `makeinfo' installed, run the following command
124 @c makeinfo emacs-lisp-intro.texi
126 @c or, if you want a single, large Info file, and no paragraph indents:
127 @c makeinfo --no-split --paragraph-indent=0 --verbose emacs-lisp-intro.texi
129 @c After creating the Info file, edit your Info `dir' file, if the
130 @c `dircategory' section below does not enable your system to
131 @c install the manual automatically.
132 @c (The `dir' file is often in the `/usr/local/share/info/' directory.)
134 @c ================ How to Create an HTML file ================
136 @c To convert to HTML format
137 @c makeinfo --html --no-split --verbose emacs-lisp-intro.texi
139 @c ================ How to Print a Book in Various Sizes ================
141 @c This book can be printed in any of three different sizes.
142 @c In the above header, set @-commands appropriately.
152 @c European A4 size paper:
157 @c ================ How to Typeset and Print ================
159 @c If you do not include PostScript figures, run either of the
160 @c following command sequences, or similar commands suited to your
163 @c texi2dvi emacs-lisp-intro.texi
164 @c lpr -d emacs-lisp-intro.dvi
168 @c tex emacs-lisp-intro.texi
169 @c texindex emacs-lisp-intro.??
170 @c tex emacs-lisp-intro.texi
171 @c lpr -d emacs-lisp-intro.dvi
173 @c If you include the PostScript figures, and you have old software,
174 @c you may need to convert the .dvi file to a .ps file before
175 @c printing. Run either of the following command sequences, or one
178 @c dvips -f < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
182 @c postscript -p < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
185 @c (Note: if you edit the book so as to change the length of the
186 @c table of contents, you may have to change the value of `pageno' below.)
188 @c ================ End of Formatting Sections ================
190 @c For next or subsequent edition:
191 @c create function using with-output-to-temp-buffer
192 @c create a major mode, with keymaps
193 @c run an asynchronous process, like grep or diff
195 @c For 8.5 by 11 inch format: do not use such a small amount of
196 @c whitespace between paragraphs as smallbook format
199 \global\parskip 6pt plus 1pt
203 @c For all sized formats: print within-book cross
204 @c reference with ``...'' rather than [...]
206 @c This works with the texinfo.tex file, version 2003-05-04.08,
207 @c in the Texinfo version 4.6 of the 2003 Jun 13 distribution.
210 \if \xrefprintnodename
211 \global\def\xrefprintnodename#1{\unskip, ``#1''}
213 \global\def\xrefprintnodename#1{ ``#1''}
215 % \global\def\xrefprintnodename#1{, ``#1''}
218 @c ----------------------------------------------------
220 @dircategory GNU Emacs Lisp
222 * Emacs Lisp Intro: (eintr).
223 A simple introduction to Emacs Lisp programming.
227 This is an @cite{Introduction to Programming in Emacs Lisp}, for
228 people who are not programmers.
230 Edition @value{edition-number}, @value{update-date}
233 <p>The homepage for GNU Emacs is at
234 <a href="http://www.gnu.org/software/emacs/">http://www.gnu.org/software/emacs/</a>.
235 <br>To view this manual in other formats, click
236 <a href="/software/emacs/emacs-lisp-intro/emacs-lisp-intro.html">here</a>.
240 Copyright @copyright{} 1990--1995, 1997, 2001--2013 Free Software
247 GNU Press, @hfill @uref{http://www.fsf.org/campaigns/gnu-press/}@*
248 a division of the @hfill email: @email{sales@@fsf.org}@*
249 Free Software Foundation, Inc. @hfill Tel: +1 (617) 542-5942@*
250 51 Franklin Street, Fifth Floor @hfill Fax: +1 (617) 542-2652@*
251 Boston, MA 02110-1301 USA
258 GNU Press, http://www.fsf.org/campaigns/gnu-press/
259 a division of the email: sales@@fsf.org
260 Free Software Foundation, Inc. Tel: +1 (617) 542-5942
261 51 Franklin Street, Fifth Floor Fax: +1 (617) 542-2652
262 Boston, MA 02110-1301 USA
267 @c Printed copies are available from @uref{http://shop.fsf.org/} for $35 each.@*
270 Permission is granted to copy, distribute and/or modify this document
271 under the terms of the GNU Free Documentation License, Version 1.3 or
272 any later version published by the Free Software Foundation; there
273 being no Invariant Section, with the Front-Cover Texts being ``A GNU
274 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
275 the license is included in the section entitled ``GNU Free
276 Documentation License''.
278 (a) The FSF's Back-Cover Text is: ``You have the freedom to
279 copy and modify this GNU manual. Buying copies from the FSF
280 supports it in developing GNU and promoting software freedom.''
283 @c half title; two lines here, so do not use `shorttitlepage'
286 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
288 {\begingroup\hbox{}\vskip 0.25in \chaprm%
289 \centerline{Programming in Emacs Lisp}%
290 \endgroup\page\hbox{}\page}
295 @center @titlefont{An Introduction to}
297 @center @titlefont{Programming in Emacs Lisp}
299 @center Revised Third Edition
301 @center by Robert J. Chassell
304 @vskip 0pt plus 1filll
310 @evenheading @thispage @| @| @thischapter
311 @oddheading @thissection @| @| @thispage
315 @c Keep T.O.C. short by tightening up for largebook
318 \global\parskip 2pt plus 1pt
319 \global\advance\baselineskip by -1pt
329 @top An Introduction to Programming in Emacs Lisp
333 This master menu first lists each chapter and index; then it lists
334 every node in every chapter.
337 @c >>>> Set pageno appropriately <<<<
339 @c The first page of the Preface is a roman numeral; it is the first
340 @c right handed page after the Table of Contents; hence the following
341 @c setting must be for an odd negative number.
344 @c global@pageno = -11
347 @set COUNT-WORDS count-words-example
348 @c Length of variable name chosen so that things still line up when expanded.
351 * Preface:: What to look for.
352 * List Processing:: What is Lisp?
353 * Practicing Evaluation:: Running several programs.
354 * Writing Defuns:: How to write function definitions.
355 * Buffer Walk Through:: Exploring a few buffer-related functions.
356 * More Complex:: A few, even more complex functions.
357 * Narrowing & Widening:: Restricting your and Emacs attention to
359 * car cdr & cons:: Fundamental functions in Lisp.
360 * Cutting & Storing Text:: Removing text and saving it.
361 * List Implementation:: How lists are implemented in the computer.
362 * Yanking:: Pasting stored text.
363 * Loops & Recursion:: How to repeat a process.
364 * Regexp Search:: Regular expression searches.
365 * Counting Words:: A review of repetition and regexps.
366 * Words in a defun:: Counting words in a @code{defun}.
367 * Readying a Graph:: A prototype graph printing function.
368 * Emacs Initialization:: How to write a @file{.emacs} file.
369 * Debugging:: How to run the Emacs Lisp debuggers.
370 * Conclusion:: Now you have the basics.
371 * the-the:: An appendix: how to find reduplicated words.
372 * Kill Ring:: An appendix: how the kill ring works.
373 * Full Graph:: How to create a graph with labeled axes.
374 * Free Software and Free Manuals::
375 * GNU Free Documentation License::
380 --- The Detailed Node Listing ---
384 * Why:: Why learn Emacs Lisp?
385 * On Reading this Text:: Read, gain familiarity, pick up habits....
386 * Who You Are:: For whom this is written.
388 * Note for Novices:: You can read this as a novice.
393 * Lisp Lists:: What are lists?
394 * Run a Program:: Any list in Lisp is a program ready to run.
395 * Making Errors:: Generating an error message.
396 * Names & Definitions:: Names of symbols and function definitions.
397 * Lisp Interpreter:: What the Lisp interpreter does.
398 * Evaluation:: Running a program.
399 * Variables:: Returning a value from a variable.
400 * Arguments:: Passing information to a function.
401 * set & setq:: Setting the value of a variable.
402 * Summary:: The major points.
403 * Error Message Exercises::
407 * Numbers Lists:: List have numbers, other lists, in them.
408 * Lisp Atoms:: Elemental entities.
409 * Whitespace in Lists:: Formatting lists to be readable.
410 * Typing Lists:: How GNU Emacs helps you type lists.
414 * Complications:: Variables, Special forms, Lists within.
415 * Byte Compiling:: Specially processing code for speed.
419 * How the Interpreter Acts:: Returns and Side Effects...
420 * Evaluating Inner Lists:: Lists within lists...
424 * fill-column Example::
425 * Void Function:: The error message for a symbol
427 * Void Variable:: The error message for a symbol without a value.
431 * Data types:: Types of data passed to a function.
432 * Args as Variable or List:: An argument can be the value
433 of a variable or list.
434 * Variable Number of Arguments:: Some functions may take a
435 variable number of arguments.
436 * Wrong Type of Argument:: Passing an argument of the wrong type
438 * message:: A useful function for sending messages.
440 Setting the Value of a Variable
442 * Using set:: Setting values.
443 * Using setq:: Setting a quoted value.
444 * Counting:: Using @code{setq} to count.
446 Practicing Evaluation
448 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
450 * Buffer Names:: Buffers and files are different.
451 * Getting Buffers:: Getting a buffer itself, not merely its name.
452 * Switching Buffers:: How to change to another buffer.
453 * Buffer Size & Locations:: Where point is located and the size of
455 * Evaluation Exercise::
457 How To Write Function Definitions
459 * Primitive Functions::
460 * defun:: The @code{defun} macro.
461 * Install:: Install a function definition.
462 * Interactive:: Making a function interactive.
463 * Interactive Options:: Different options for @code{interactive}.
464 * Permanent Installation:: Installing code permanently.
465 * let:: Creating and initializing local variables.
467 * else:: If--then--else expressions.
468 * Truth & Falsehood:: What Lisp considers false and true.
469 * save-excursion:: Keeping track of point, mark, and buffer.
473 Install a Function Definition
475 * Effect of installation::
476 * Change a defun:: How to change a function definition.
478 Make a Function Interactive
480 * Interactive multiply-by-seven:: An overview.
481 * multiply-by-seven in detail:: The interactive version.
485 * Prevent confusion::
486 * Parts of let Expression::
487 * Sample let Expression::
488 * Uninitialized let Variables::
490 The @code{if} Special Form
492 * if in more detail::
493 * type-of-animal in detail:: An example of an @code{if} expression.
495 Truth and Falsehood in Emacs Lisp
497 * nil explained:: @code{nil} has two meanings.
499 @code{save-excursion}
501 * Point and mark:: A review of various locations.
502 * Template for save-excursion::
504 A Few Buffer--Related Functions
506 * Finding More:: How to find more information.
507 * simplified-beginning-of-buffer:: Shows @code{goto-char},
508 @code{point-min}, and @code{push-mark}.
509 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
510 * append-to-buffer:: Uses @code{save-excursion} and
511 @code{insert-buffer-substring}.
512 * Buffer Related Review:: Review.
515 The Definition of @code{mark-whole-buffer}
517 * mark-whole-buffer overview::
518 * Body of mark-whole-buffer:: Only three lines of code.
520 The Definition of @code{append-to-buffer}
522 * append-to-buffer overview::
523 * append interactive:: A two part interactive expression.
524 * append-to-buffer body:: Incorporates a @code{let} expression.
525 * append save-excursion:: How the @code{save-excursion} works.
527 A Few More Complex Functions
529 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
530 * insert-buffer:: Read-only, and with @code{or}.
531 * beginning-of-buffer:: Shows @code{goto-char},
532 @code{point-min}, and @code{push-mark}.
533 * Second Buffer Related Review::
534 * optional Exercise::
536 The Definition of @code{insert-buffer}
538 * insert-buffer code::
539 * insert-buffer interactive:: When you can read, but not write.
540 * insert-buffer body:: The body has an @code{or} and a @code{let}.
541 * if & or:: Using an @code{if} instead of an @code{or}.
542 * Insert or:: How the @code{or} expression works.
543 * Insert let:: Two @code{save-excursion} expressions.
544 * New insert-buffer::
546 The Interactive Expression in @code{insert-buffer}
548 * Read-only buffer:: When a buffer cannot be modified.
549 * b for interactive:: An existing buffer or else its name.
551 Complete Definition of @code{beginning-of-buffer}
553 * Optional Arguments::
554 * beginning-of-buffer opt arg:: Example with optional argument.
555 * beginning-of-buffer complete::
557 @code{beginning-of-buffer} with an Argument
559 * Disentangle beginning-of-buffer::
560 * Large buffer case::
561 * Small buffer case::
563 Narrowing and Widening
565 * Narrowing advantages:: The advantages of narrowing
566 * save-restriction:: The @code{save-restriction} special form.
567 * what-line:: The number of the line that point is on.
570 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
572 * Strange Names:: An historical aside: why the strange names?
573 * car & cdr:: Functions for extracting part of a list.
574 * cons:: Constructing a list.
575 * nthcdr:: Calling @code{cdr} repeatedly.
577 * setcar:: Changing the first element of a list.
578 * setcdr:: Changing the rest of a list.
584 * length:: How to find the length of a list.
586 Cutting and Storing Text
588 * Storing Text:: Text is stored in a list.
589 * zap-to-char:: Cutting out text up to a character.
590 * kill-region:: Cutting text out of a region.
591 * copy-region-as-kill:: A definition for copying text.
592 * Digression into C:: Minor note on C programming language macros.
593 * defvar:: How to give a variable an initial value.
594 * cons & search-fwd Review::
599 * Complete zap-to-char:: The complete implementation.
600 * zap-to-char interactive:: A three part interactive expression.
601 * zap-to-char body:: A short overview.
602 * search-forward:: How to search for a string.
603 * progn:: The @code{progn} special form.
604 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
608 * Complete kill-region:: The function definition.
609 * condition-case:: Dealing with a problem.
612 @code{copy-region-as-kill}
614 * Complete copy-region-as-kill:: The complete function definition.
615 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
617 The Body of @code{copy-region-as-kill}
619 * last-command & this-command::
620 * kill-append function::
621 * kill-new function::
623 Initializing a Variable with @code{defvar}
625 * See variable current value::
626 * defvar and asterisk::
628 How Lists are Implemented
631 * Symbols as Chest:: Exploring a powerful metaphor.
636 * Kill Ring Overview::
637 * kill-ring-yank-pointer:: The kill ring is a list.
638 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
642 * while:: Causing a stretch of code to repeat.
644 * Recursion:: Causing a function to call itself.
649 * Looping with while:: Repeat so long as test returns true.
650 * Loop Example:: A @code{while} loop that uses a list.
651 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
652 * Incrementing Loop:: A loop with an incrementing counter.
653 * Incrementing Loop Details::
654 * Decrementing Loop:: A loop with a decrementing counter.
656 Details of an Incrementing Loop
658 * Incrementing Example:: Counting pebbles in a triangle.
659 * Inc Example parts:: The parts of the function definition.
660 * Inc Example altogether:: Putting the function definition together.
662 Loop with a Decrementing Counter
664 * Decrementing Example:: More pebbles on the beach.
665 * Dec Example parts:: The parts of the function definition.
666 * Dec Example altogether:: Putting the function definition together.
668 Save your time: @code{dolist} and @code{dotimes}
675 * Building Robots:: Same model, different serial number ...
676 * Recursive Definition Parts:: Walk until you stop ...
677 * Recursion with list:: Using a list as the test whether to recurse.
678 * Recursive triangle function::
679 * Recursion with cond::
680 * Recursive Patterns:: Often used templates.
681 * No Deferment:: Don't store up work ...
682 * No deferment solution::
684 Recursion in Place of a Counter
686 * Recursive Example arg of 1 or 2::
687 * Recursive Example arg of 3 or 4::
695 Regular Expression Searches
697 * sentence-end:: The regular expression for @code{sentence-end}.
698 * re-search-forward:: Very similar to @code{search-forward}.
699 * forward-sentence:: A straightforward example of regexp search.
700 * forward-paragraph:: A somewhat complex example.
701 * etags:: How to create your own @file{TAGS} table.
703 * re-search Exercises::
705 @code{forward-sentence}
707 * Complete forward-sentence::
708 * fwd-sentence while loops:: Two @code{while} loops.
709 * fwd-sentence re-search:: A regular expression search.
711 @code{forward-paragraph}: a Goldmine of Functions
713 * forward-paragraph in brief:: Key parts of the function definition.
714 * fwd-para let:: The @code{let*} expression.
715 * fwd-para while:: The forward motion @code{while} loop.
717 Counting: Repetition and Regexps
720 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
721 * recursive-count-words:: Start with case of no words in region.
722 * Counting Exercise::
724 The @code{@value{COUNT-WORDS}} Function
726 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
727 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
729 Counting Words in a @code{defun}
731 * Divide and Conquer::
732 * Words and Symbols:: What to count?
733 * Syntax:: What constitutes a word or symbol?
734 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
735 * Several defuns:: Counting several defuns in a file.
736 * Find a File:: Do you want to look at a file?
737 * lengths-list-file:: A list of the lengths of many definitions.
738 * Several files:: Counting in definitions in different files.
739 * Several files recursively:: Recursively counting in different files.
740 * Prepare the data:: Prepare the data for display in a graph.
742 Count Words in @code{defuns} in Different Files
744 * lengths-list-many-files:: Return a list of the lengths of defuns.
745 * append:: Attach one list to another.
747 Prepare the Data for Display in a Graph
749 * Data for Display in Detail::
750 * Sorting:: Sorting lists.
751 * Files List:: Making a list of files.
752 * Counting function definitions::
756 * Columns of a graph::
757 * graph-body-print:: How to print the body of a graph.
758 * recursive-graph-body-print::
760 * Line Graph Exercise::
762 Your @file{.emacs} File
764 * Default Configuration::
765 * Site-wide Init:: You can write site-wide init files.
766 * defcustom:: Emacs will write code for you.
767 * Beginning a .emacs File:: How to write a @code{.emacs file}.
768 * Text and Auto-fill:: Automatically wrap lines.
769 * Mail Aliases:: Use abbreviations for email addresses.
770 * Indent Tabs Mode:: Don't use tabs with @TeX{}
771 * Keybindings:: Create some personal keybindings.
772 * Keymaps:: More about key binding.
773 * Loading Files:: Load (i.e., evaluate) files automatically.
774 * Autoload:: Make functions available.
775 * Simple Extension:: Define a function; bind it to a key.
776 * X11 Colors:: Colors in X.
778 * Mode Line:: How to customize your mode line.
782 * debug:: How to use the built-in debugger.
783 * debug-on-entry:: Start debugging when you call a function.
784 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
785 * edebug:: How to use Edebug, a source level debugger.
786 * Debugging Exercises::
788 Handling the Kill Ring
790 * What the Kill Ring Does::
792 * yank:: Paste a copy of a clipped element.
793 * yank-pop:: Insert element pointed to.
796 The @code{current-kill} Function
798 * Code for current-kill::
799 * Understanding current-kill::
801 @code{current-kill} in Outline
803 * Body of current-kill::
804 * Digression concerning error:: How to mislead humans, but not computers.
805 * Determining the Element::
807 A Graph with Labeled Axes
810 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
811 * print-Y-axis:: Print a label for the vertical axis.
812 * print-X-axis:: Print a horizontal label.
813 * Print Whole Graph:: The function to print a complete graph.
815 The @code{print-Y-axis} Function
817 * print-Y-axis in Detail::
818 * Height of label:: What height for the Y axis?
819 * Compute a Remainder:: How to compute the remainder of a division.
820 * Y Axis Element:: Construct a line for the Y axis.
821 * Y-axis-column:: Generate a list of Y axis labels.
822 * print-Y-axis Penultimate:: A not quite final version.
824 The @code{print-X-axis} Function
826 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
827 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
829 Printing the Whole Graph
831 * The final version:: A few changes.
832 * Test print-graph:: Run a short test.
833 * Graphing words in defuns:: Executing the final code.
834 * lambda:: How to write an anonymous function.
835 * mapcar:: Apply a function to elements of a list.
836 * Another Bug:: Yet another bug @dots{} most insidious.
837 * Final printed graph:: The graph itself!
845 Most of the GNU Emacs integrated environment is written in the programming
846 language called Emacs Lisp. The code written in this programming
847 language is the software---the sets of instructions---that tell the
848 computer what to do when you give it commands. Emacs is designed so
849 that you can write new code in Emacs Lisp and easily install it as an
850 extension to the editor.
852 (GNU Emacs is sometimes called an ``extensible editor'', but it does
853 much more than provide editing capabilities. It is better to refer to
854 Emacs as an ``extensible computing environment''. However, that
855 phrase is quite a mouthful. It is easier to refer to Emacs simply as
856 an editor. Moreover, everything you do in Emacs---find the Mayan date
857 and phases of the moon, simplify polynomials, debug code, manage
858 files, read letters, write books---all these activities are kinds of
859 editing in the most general sense of the word.)
862 * Why:: Why learn Emacs Lisp?
863 * On Reading this Text:: Read, gain familiarity, pick up habits....
864 * Who You Are:: For whom this is written.
866 * Note for Novices:: You can read this as a novice.
872 @unnumberedsec Why Study Emacs Lisp?
875 Although Emacs Lisp is usually thought of in association only with Emacs,
876 it is a full computer programming language. You can use Emacs Lisp as
877 you would any other programming language.
879 Perhaps you want to understand programming; perhaps you want to extend
880 Emacs; or perhaps you want to become a programmer. This introduction to
881 Emacs Lisp is designed to get you started: to guide you in learning the
882 fundamentals of programming, and more importantly, to show you how you
883 can teach yourself to go further.
885 @node On Reading this Text
886 @unnumberedsec On Reading this Text
888 All through this document, you will see little sample programs you can
889 run inside of Emacs. If you read this document in Info inside of GNU
890 Emacs, you can run the programs as they appear. (This is easy to do and
891 is explained when the examples are presented.) Alternatively, you can
892 read this introduction as a printed book while sitting beside a computer
893 running Emacs. (This is what I like to do; I like printed books.) If
894 you don't have a running Emacs beside you, you can still read this book,
895 but in this case, it is best to treat it as a novel or as a travel guide
896 to a country not yet visited: interesting, but not the same as being
899 Much of this introduction is dedicated to walkthroughs or guided tours
900 of code used in GNU Emacs. These tours are designed for two purposes:
901 first, to give you familiarity with real, working code (code you use
902 every day); and, second, to give you familiarity with the way Emacs
903 works. It is interesting to see how a working environment is
906 hope that you will pick up the habit of browsing through source code.
907 You can learn from it and mine it for ideas. Having GNU Emacs is like
908 having a dragon's cave of treasures.
910 In addition to learning about Emacs as an editor and Emacs Lisp as a
911 programming language, the examples and guided tours will give you an
912 opportunity to get acquainted with Emacs as a Lisp programming
913 environment. GNU Emacs supports programming and provides tools that
914 you will want to become comfortable using, such as @kbd{M-.} (the key
915 which invokes the @code{find-tag} command). You will also learn about
916 buffers and other objects that are part of the environment.
917 Learning about these features of Emacs is like learning new routes
918 around your home town.
921 In addition, I have written several programs as extended examples.
922 Although these are examples, the programs are real. I use them.
923 Other people use them. You may use them. Beyond the fragments of
924 programs used for illustrations, there is very little in here that is
925 `just for teaching purposes'; what you see is used. This is a great
926 advantage of Emacs Lisp: it is easy to learn to use it for work.
929 Finally, I hope to convey some of the skills for using Emacs to
930 learn aspects of programming that you don't know. You can often use
931 Emacs to help you understand what puzzles you or to find out how to do
932 something new. This self-reliance is not only a pleasure, but an
936 @unnumberedsec For Whom This is Written
938 This text is written as an elementary introduction for people who are
939 not programmers. If you are a programmer, you may not be satisfied with
940 this primer. The reason is that you may have become expert at reading
941 reference manuals and be put off by the way this text is organized.
943 An expert programmer who reviewed this text said to me:
946 @i{I prefer to learn from reference manuals. I ``dive into'' each
947 paragraph, and ``come up for air'' between paragraphs.}
949 @i{When I get to the end of a paragraph, I assume that that subject is
950 done, finished, that I know everything I need (with the
951 possible exception of the case when the next paragraph starts talking
952 about it in more detail). I expect that a well written reference manual
953 will not have a lot of redundancy, and that it will have excellent
954 pointers to the (one) place where the information I want is.}
957 This introduction is not written for this person!
959 Firstly, I try to say everything at least three times: first, to
960 introduce it; second, to show it in context; and third, to show it in a
961 different context, or to review it.
963 Secondly, I hardly ever put all the information about a subject in one
964 place, much less in one paragraph. To my way of thinking, that imposes
965 too heavy a burden on the reader. Instead I try to explain only what
966 you need to know at the time. (Sometimes I include a little extra
967 information so you won't be surprised later when the additional
968 information is formally introduced.)
970 When you read this text, you are not expected to learn everything the
971 first time. Frequently, you need only make, as it were, a `nodding
972 acquaintance' with some of the items mentioned. My hope is that I have
973 structured the text and given you enough hints that you will be alert to
974 what is important, and concentrate on it.
976 You will need to ``dive into'' some paragraphs; there is no other way
977 to read them. But I have tried to keep down the number of such
978 paragraphs. This book is intended as an approachable hill, rather than
979 as a daunting mountain.
981 This introduction to @cite{Programming in Emacs Lisp} has a companion
984 @cite{The GNU Emacs Lisp Reference Manual}.
987 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
988 Emacs Lisp Reference Manual}.
990 The reference manual has more detail than this introduction. In the
991 reference manual, all the information about one topic is concentrated
992 in one place. You should turn to it if you are like the programmer
993 quoted above. And, of course, after you have read this
994 @cite{Introduction}, you will find the @cite{Reference Manual} useful
995 when you are writing your own programs.
998 @unnumberedsec Lisp History
1001 Lisp was first developed in the late 1950s at the Massachusetts
1002 Institute of Technology for research in artificial intelligence. The
1003 great power of the Lisp language makes it superior for other purposes as
1004 well, such as writing editor commands and integrated environments.
1008 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
1009 in the 1960s. It is somewhat inspired by Common Lisp, which became a
1010 standard in the 1980s. However, Emacs Lisp is much simpler than Common
1011 Lisp. (The standard Emacs distribution contains an optional extensions
1012 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
1014 @node Note for Novices
1015 @unnumberedsec A Note for Novices
1017 If you don't know GNU Emacs, you can still read this document
1018 profitably. However, I recommend you learn Emacs, if only to learn to
1019 move around your computer screen. You can teach yourself how to use
1020 Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This
1021 means you press and release the @key{CTRL} key and the @kbd{h} at the
1022 same time, and then press and release @kbd{t}.)
1024 Also, I often refer to one of Emacs's standard commands by listing the
1025 keys which you press to invoke the command and then giving the name of
1026 the command in parentheses, like this: @kbd{M-C-\}
1027 (@code{indent-region}). What this means is that the
1028 @code{indent-region} command is customarily invoked by typing
1029 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
1030 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
1031 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
1032 @key{META} key, @key{CTRL} key and @key{\} key all at the same time.
1033 (On many modern keyboards the @key{META} key is labeled
1035 Sometimes a combination like this is called a keychord, since it is
1036 similar to the way you play a chord on a piano. If your keyboard does
1037 not have a @key{META} key, the @key{ESC} key prefix is used in place
1038 of it. In this case, @kbd{M-C-\} means that you press and release your
1039 @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at
1040 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
1041 along with the key that is labeled @key{ALT} and, at the same time,
1042 press the @key{\} key.
1044 In addition to typing a lone keychord, you can prefix what you type
1045 with @kbd{C-u}, which is called the `universal argument'. The
1046 @kbd{C-u} keychord passes an argument to the subsequent command.
1047 Thus, to indent a region of plain text by 6 spaces, mark the region,
1048 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
1049 Emacs either passes the number 4 to the command or otherwise runs the
1050 command differently than it would otherwise.) @xref{Arguments, ,
1051 Numeric Arguments, emacs, The GNU Emacs Manual}.
1053 If you are reading this in Info using GNU Emacs, you can read through
1054 this whole document just by pressing the space bar, @key{SPC}.
1055 (To learn about Info, type @kbd{C-h i} and then select Info.)
1057 A note on terminology: when I use the word Lisp alone, I often am
1058 referring to the various dialects of Lisp in general, but when I speak
1059 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
1062 @unnumberedsec Thank You
1064 My thanks to all who helped me with this book. My especial thanks to
1065 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
1066 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.
1067 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
1068 @w{Philip Johnson} and @w{David Stampe} for their patient
1069 encouragement. My mistakes are my own.
1073 @email{bob@@gnu.org}
1076 @c ================ Beginning of main text ================
1078 @c Start main text on right-hand (verso) page
1081 \par\vfill\supereject
1084 \par\vfill\supereject
1086 \par\vfill\supereject
1088 \par\vfill\supereject
1092 @c Note: this resetting of the page number back to 1 causes TeX to gripe
1093 @c about already having seen page numbers 1-4 before (in the preface):
1094 @c pdfTeX warning (ext4): destination with the same identifier (name{1})
1095 @c has been already used, duplicate ignored
1096 @c I guess that is harmless (what happens if a later part of the text
1097 @c makes a link to something in the first 4 pages though?).
1098 @c E.g., note that the Emacs manual has a preface, but does not bother
1099 @c resetting the page numbers back to 1 after that.
1102 @evenheading @thispage @| @| @thischapter
1103 @oddheading @thissection @| @| @thispage
1107 @node List Processing
1108 @chapter List Processing
1110 To the untutored eye, Lisp is a strange programming language. In Lisp
1111 code there are parentheses everywhere. Some people even claim that
1112 the name stands for `Lots of Isolated Silly Parentheses'. But the
1113 claim is unwarranted. Lisp stands for LISt Processing, and the
1114 programming language handles @emph{lists} (and lists of lists) by
1115 putting them between parentheses. The parentheses mark the boundaries
1116 of the list. Sometimes a list is preceded by a single apostrophe or
1117 quotation mark, @samp{'}@footnote{The single apostrophe or quotation
1118 mark is an abbreviation for the function @code{quote}; you need not
1119 think about functions now; functions are defined in @ref{Making
1120 Errors, , Generate an Error Message}.} Lists are the basis of Lisp.
1123 * Lisp Lists:: What are lists?
1124 * Run a Program:: Any list in Lisp is a program ready to run.
1125 * Making Errors:: Generating an error message.
1126 * Names & Definitions:: Names of symbols and function definitions.
1127 * Lisp Interpreter:: What the Lisp interpreter does.
1128 * Evaluation:: Running a program.
1129 * Variables:: Returning a value from a variable.
1130 * Arguments:: Passing information to a function.
1131 * set & setq:: Setting the value of a variable.
1132 * Summary:: The major points.
1133 * Error Message Exercises::
1140 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1141 This list is preceded by a single apostrophe. It could just as well be
1142 written as follows, which looks more like the kind of list you are likely
1143 to be familiar with:
1155 The elements of this list are the names of the four different flowers,
1156 separated from each other by whitespace and surrounded by parentheses,
1157 like flowers in a field with a stone wall around them.
1158 @cindex Flowers in a field
1161 * Numbers Lists:: List have numbers, other lists, in them.
1162 * Lisp Atoms:: Elemental entities.
1163 * Whitespace in Lists:: Formatting lists to be readable.
1164 * Typing Lists:: How GNU Emacs helps you type lists.
1169 @unnumberedsubsec Numbers, Lists inside of Lists
1172 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1173 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1174 separated by whitespace.
1176 In Lisp, both data and programs are represented the same way; that is,
1177 they are both lists of words, numbers, or other lists, separated by
1178 whitespace and surrounded by parentheses. (Since a program looks like
1179 data, one program may easily serve as data for another; this is a very
1180 powerful feature of Lisp.) (Incidentally, these two parenthetical
1181 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1182 @samp{.} as punctuation marks.)
1185 Here is another list, this time with a list inside of it:
1188 '(this list has (a list inside of it))
1191 The components of this list are the words @samp{this}, @samp{list},
1192 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1193 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1194 @samp{of}, @samp{it}.
1197 @subsection Lisp Atoms
1200 In Lisp, what we have been calling words are called @dfn{atoms}. This
1201 term comes from the historical meaning of the word atom, which means
1202 `indivisible'. As far as Lisp is concerned, the words we have been
1203 using in the lists cannot be divided into any smaller parts and still
1204 mean the same thing as part of a program; likewise with numbers and
1205 single character symbols like @samp{+}. On the other hand, unlike an
1206 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1207 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1209 In a list, atoms are separated from each other by whitespace. They can be
1210 right next to a parenthesis.
1212 @cindex @samp{empty list} defined
1213 Technically speaking, a list in Lisp consists of parentheses surrounding
1214 atoms separated by whitespace or surrounding other lists or surrounding
1215 both atoms and other lists. A list can have just one atom in it or
1216 have nothing in it at all. A list with nothing in it looks like this:
1217 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1218 empty list is considered both an atom and a list at the same time.
1220 @cindex Symbolic expressions, introduced
1221 @cindex @samp{expression} defined
1222 @cindex @samp{form} defined
1223 The printed representation of both atoms and lists are called
1224 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1225 The word @dfn{expression} by itself can refer to either the printed
1226 representation, or to the atom or list as it is held internally in the
1227 computer. Often, people use the term @dfn{expression}
1228 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1229 as a synonym for expression.)
1231 Incidentally, the atoms that make up our universe were named such when
1232 they were thought to be indivisible; but it has been found that physical
1233 atoms are not indivisible. Parts can split off an atom or it can
1234 fission into two parts of roughly equal size. Physical atoms were named
1235 prematurely, before their truer nature was found. In Lisp, certain
1236 kinds of atom, such as an array, can be separated into parts; but the
1237 mechanism for doing this is different from the mechanism for splitting a
1238 list. As far as list operations are concerned, the atoms of a list are
1241 As in English, the meanings of the component letters of a Lisp atom
1242 are different from the meaning the letters make as a word. For
1243 example, the word for the South American sloth, the @samp{ai}, is
1244 completely different from the two words, @samp{a}, and @samp{i}.
1246 There are many kinds of atom in nature but only a few in Lisp: for
1247 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1248 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1249 listed in the examples above are all symbols. In everyday Lisp
1250 conversation, the word ``atom'' is not often used, because programmers
1251 usually try to be more specific about what kind of atom they are dealing
1252 with. Lisp programming is mostly about symbols (and sometimes numbers)
1253 within lists. (Incidentally, the preceding three word parenthetical
1254 remark is a proper list in Lisp, since it consists of atoms, which in
1255 this case are symbols, separated by whitespace and enclosed by
1256 parentheses, without any non-Lisp punctuation.)
1259 Text between double quotation marks---even sentences or
1260 paragraphs---is also an atom. Here is an example:
1261 @cindex Text between double quotation marks
1264 '(this list includes "text between quotation marks.")
1267 @cindex @samp{string} defined
1269 In Lisp, all of the quoted text including the punctuation mark and the
1270 blank spaces is a single atom. This kind of atom is called a
1271 @dfn{string} (for `string of characters') and is the sort of thing that
1272 is used for messages that a computer can print for a human to read.
1273 Strings are a different kind of atom than numbers or symbols and are
1276 @node Whitespace in Lists
1277 @subsection Whitespace in Lists
1278 @cindex Whitespace in lists
1281 The amount of whitespace in a list does not matter. From the point of view
1282 of the Lisp language,
1293 is exactly the same as this:
1296 '(this list looks like this)
1299 Both examples show what to Lisp is the same list, the list made up of
1300 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1301 @samp{this} in that order.
1303 Extra whitespace and newlines are designed to make a list more readable
1304 by humans. When Lisp reads the expression, it gets rid of all the extra
1305 whitespace (but it needs to have at least one space between atoms in
1306 order to tell them apart.)
1308 Odd as it seems, the examples we have seen cover almost all of what Lisp
1309 lists look like! Every other list in Lisp looks more or less like one
1310 of these examples, except that the list may be longer and more complex.
1311 In brief, a list is between parentheses, a string is between quotation
1312 marks, a symbol looks like a word, and a number looks like a number.
1313 (For certain situations, square brackets, dots and a few other special
1314 characters may be used; however, we will go quite far without them.)
1317 @subsection GNU Emacs Helps You Type Lists
1318 @cindex Help typing lists
1319 @cindex Formatting help
1321 When you type a Lisp expression in GNU Emacs using either Lisp
1322 Interaction mode or Emacs Lisp mode, you have available to you several
1323 commands to format the Lisp expression so it is easy to read. For
1324 example, pressing the @key{TAB} key automatically indents the line the
1325 cursor is on by the right amount. A command to properly indent the
1326 code in a region is customarily bound to @kbd{M-C-\}. Indentation is
1327 designed so that you can see which elements of a list belong to which
1328 list---elements of a sub-list are indented more than the elements of
1331 In addition, when you type a closing parenthesis, Emacs momentarily
1332 jumps the cursor back to the matching opening parenthesis, so you can
1333 see which one it is. This is very useful, since every list you type
1334 in Lisp must have its closing parenthesis match its opening
1335 parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs
1336 Manual}, for more information about Emacs's modes.)
1339 @section Run a Program
1340 @cindex Run a program
1341 @cindex Program, running one
1343 @cindex @samp{evaluate} defined
1344 A list in Lisp---any list---is a program ready to run. If you run it
1345 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1346 of three things: do nothing except return to you the list itself; send
1347 you an error message; or, treat the first symbol in the list as a
1348 command to do something. (Usually, of course, it is the last of these
1349 three things that you really want!)
1351 @c use code for the single apostrophe, not samp.
1352 The single apostrophe, @code{'}, that I put in front of some of the
1353 example lists in preceding sections is called a @dfn{quote}; when it
1354 precedes a list, it tells Lisp to do nothing with the list, other than
1355 take it as it is written. But if there is no quote preceding a list,
1356 the first item of the list is special: it is a command for the computer
1357 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1358 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1359 understands that the @code{+} is an instruction to do something with the
1360 rest of the list: add the numbers that follow.
1363 If you are reading this inside of GNU Emacs in Info, here is how you can
1364 evaluate such a list: place your cursor immediately after the right
1365 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1371 @c use code for the number four, not samp.
1373 You will see the number @code{4} appear in the echo area. (In the
1374 jargon, what you have just done is ``evaluate the list.'' The echo area
1375 is the line at the bottom of the screen that displays or ``echoes''
1376 text.) Now try the same thing with a quoted list: place the cursor
1377 right after the following list and type @kbd{C-x C-e}:
1380 '(this is a quoted list)
1384 You will see @code{(this is a quoted list)} appear in the echo area.
1386 @cindex Lisp interpreter, explained
1387 @cindex Interpreter, Lisp, explained
1388 In both cases, what you are doing is giving a command to the program
1389 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1390 interpreter a command to evaluate the expression. The name of the Lisp
1391 interpreter comes from the word for the task done by a human who comes
1392 up with the meaning of an expression---who ``interprets'' it.
1394 You can also evaluate an atom that is not part of a list---one that is
1395 not surrounded by parentheses; again, the Lisp interpreter translates
1396 from the humanly readable expression to the language of the computer.
1397 But before discussing this (@pxref{Variables}), we will discuss what the
1398 Lisp interpreter does when you make an error.
1401 @section Generate an Error Message
1402 @cindex Generate an error message
1403 @cindex Error message generation
1405 Partly so you won't worry if you do it accidentally, we will now give
1406 a command to the Lisp interpreter that generates an error message.
1407 This is a harmless activity; and indeed, we will often try to generate
1408 error messages intentionally. Once you understand the jargon, error
1409 messages can be informative. Instead of being called ``error''
1410 messages, they should be called ``help'' messages. They are like
1411 signposts to a traveler in a strange country; deciphering them can be
1412 hard, but once understood, they can point the way.
1414 The error message is generated by a built-in GNU Emacs debugger. We
1415 will `enter the debugger'. You get out of the debugger by typing @code{q}.
1417 What we will do is evaluate a list that is not quoted and does not
1418 have a meaningful command as its first element. Here is a list almost
1419 exactly the same as the one we just used, but without the single-quote
1420 in front of it. Position the cursor right after it and type @kbd{C-x
1424 (this is an unquoted list)
1429 What you see depends on which version of Emacs you are running. GNU
1430 Emacs version 22 provides more information than version 20 and before.
1431 First, the more recent result of generating an error; then the
1432 earlier, version 20 result.
1436 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1437 you will see the following in it:
1440 A @file{*Backtrace*} window will open up and you should see the
1445 ---------- Buffer: *Backtrace* ----------
1446 Debugger entered--Lisp error: (void-function this)
1447 (this is an unquoted list)
1448 eval((this is an unquoted list))
1449 eval-last-sexp-1(nil)
1451 call-interactively(eval-last-sexp)
1452 ---------- Buffer: *Backtrace* ----------
1458 Your cursor will be in this window (you may have to wait a few seconds
1459 before it becomes visible). To quit the debugger and make the
1460 debugger window go away, type:
1467 Please type @kbd{q} right now, so you become confident that you can
1468 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1471 @cindex @samp{function} defined
1472 Based on what we already know, we can almost read this error message.
1474 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1475 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1476 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1477 an abbreviation for `evaluate' and @code{sexp} is an abbreviation for
1478 `symbolic expression'. The command means `evaluate last symbolic
1479 expression', which is the expression just before your cursor.
1481 Each line above tells you what the Lisp interpreter evaluated next.
1482 The most recent action is at the top. The buffer is called the
1483 @file{*Backtrace*} buffer because it enables you to track Emacs
1487 At the top of the @file{*Backtrace*} buffer, you see the line:
1490 Debugger entered--Lisp error: (void-function this)
1494 The Lisp interpreter tried to evaluate the first atom of the list, the
1495 word @samp{this}. It is this action that generated the error message
1496 @samp{void-function this}.
1498 The message contains the words @samp{void-function} and @samp{this}.
1500 @cindex @samp{function} defined
1501 The word @samp{function} was mentioned once before. It is a very
1502 important word. For our purposes, we can define it by saying that a
1503 @dfn{function} is a set of instructions to the computer that tell the
1504 computer to do something.
1506 Now we can begin to understand the error message: @samp{void-function
1507 this}. The function (that is, the word @samp{this}) does not have a
1508 definition of any set of instructions for the computer to carry out.
1510 The slightly odd word, @samp{void-function}, is designed to cover the
1511 way Emacs Lisp is implemented, which is that when a symbol does not
1512 have a function definition attached to it, the place that should
1513 contain the instructions is `void'.
1515 On the other hand, since we were able to add 2 plus 2 successfully, by
1516 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1517 have a set of instructions for the computer to obey and those
1518 instructions must be to add the numbers that follow the @code{+}.
1520 It is possible to prevent Emacs entering the debugger in cases like
1521 this. We do not explain how to do that here, but we will mention what
1522 the result looks like, because you may encounter a similar situation
1523 if there is a bug in some Emacs code that you are using. In such
1524 cases, you will see only one line of error message; it will appear in
1525 the echo area and look like this:
1528 Symbol's function definition is void:@: this
1533 (Also, your terminal may beep at you---some do, some don't; and others
1534 blink. This is just a device to get your attention.)
1536 The message goes away as soon as you type a key, even just to
1539 We know the meaning of the word @samp{Symbol}. It refers to the first
1540 atom of the list, the word @samp{this}. The word @samp{function}
1541 refers to the instructions that tell the computer what to do.
1542 (Technically, the symbol tells the computer where to find the
1543 instructions, but this is a complication we can ignore for the
1546 The error message can be understood: @samp{Symbol's function
1547 definition is void:@: this}. The symbol (that is, the word
1548 @samp{this}) lacks instructions for the computer to carry out.
1550 @node Names & Definitions
1551 @section Symbol Names and Function Definitions
1552 @cindex Symbol names
1554 We can articulate another characteristic of Lisp based on what we have
1555 discussed so far---an important characteristic: a symbol, like
1556 @code{+}, is not itself the set of instructions for the computer to
1557 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1558 of locating the definition or set of instructions. What we see is the
1559 name through which the instructions can be found. Names of people
1560 work the same way. I can be referred to as @samp{Bob}; however, I am
1561 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1562 consciousness consistently associated with a particular life-form.
1563 The name is not me, but it can be used to refer to me.
1565 In Lisp, one set of instructions can be attached to several names.
1566 For example, the computer instructions for adding numbers can be
1567 linked to the symbol @code{plus} as well as to the symbol @code{+}
1568 (and are in some dialects of Lisp). Among humans, I can be referred
1569 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1571 On the other hand, a symbol can have only one function definition
1572 attached to it at a time. Otherwise, the computer would be confused as
1573 to which definition to use. If this were the case among people, only
1574 one person in the world could be named @samp{Bob}. However, the function
1575 definition to which the name refers can be changed readily.
1576 (@xref{Install, , Install a Function Definition}.)
1578 Since Emacs Lisp is large, it is customary to name symbols in a way
1579 that identifies the part of Emacs to which the function belongs.
1580 Thus, all the names for functions that deal with Texinfo start with
1581 @samp{texinfo-} and those for functions that deal with reading mail
1582 start with @samp{rmail-}.
1584 @node Lisp Interpreter
1585 @section The Lisp Interpreter
1586 @cindex Lisp interpreter, what it does
1587 @cindex Interpreter, what it does
1589 Based on what we have seen, we can now start to figure out what the
1590 Lisp interpreter does when we command it to evaluate a list.
1591 First, it looks to see whether there is a quote before the list; if
1592 there is, the interpreter just gives us the list. On the other
1593 hand, if there is no quote, the interpreter looks at the first element
1594 in the list and sees whether it has a function definition. If it does,
1595 the interpreter carries out the instructions in the function definition.
1596 Otherwise, the interpreter prints an error message.
1598 This is how Lisp works. Simple. There are added complications which we
1599 will get to in a minute, but these are the fundamentals. Of course, to
1600 write Lisp programs, you need to know how to write function definitions
1601 and attach them to names, and how to do this without confusing either
1602 yourself or the computer.
1605 * Complications:: Variables, Special forms, Lists within.
1606 * Byte Compiling:: Specially processing code for speed.
1611 @unnumberedsubsec Complications
1614 Now, for the first complication. In addition to lists, the Lisp
1615 interpreter can evaluate a symbol that is not quoted and does not have
1616 parentheses around it. The Lisp interpreter will attempt to determine
1617 the symbol's value as a @dfn{variable}. This situation is described
1618 in the section on variables. (@xref{Variables}.)
1620 @cindex Special form
1621 The second complication occurs because some functions are unusual and
1622 do not work in the usual manner. Those that don't are called
1623 @dfn{special forms}. They are used for special jobs, like defining a
1624 function, and there are not many of them. In the next few chapters,
1625 you will be introduced to several of the more important special forms.
1626 And there are also @dfn{macros}. Macro is a construct defined in
1627 Lisp, which differs from a function in that it translates a Lisp
1628 expression into another expression which is to be evaluated instead of
1629 the original expression. (@xref{Lisp macro}.)
1632 The third and final complication is this: if the function that the
1633 Lisp interpreter is looking at is not a special form, and if it is part
1634 of a list, the Lisp interpreter looks to see whether the list has a list
1635 inside of it. If there is an inner list, the Lisp interpreter first
1636 figures out what it should do with the inside list, and then it works on
1637 the outside list. If there is yet another list embedded inside the
1638 inner list, it works on that one first, and so on. It always works on
1639 the innermost list first. The interpreter works on the innermost list
1640 first, to evaluate the result of that list. The result may be
1641 used by the enclosing expression.
1643 Otherwise, the interpreter works left to right, from one expression to
1646 @node Byte Compiling
1647 @subsection Byte Compiling
1648 @cindex Byte compiling
1650 One other aspect of interpreting: the Lisp interpreter is able to
1651 interpret two kinds of entity: humanly readable code, on which we will
1652 focus exclusively, and specially processed code, called @dfn{byte
1653 compiled} code, which is not humanly readable. Byte compiled code
1654 runs faster than humanly readable code.
1656 You can transform humanly readable code into byte compiled code by
1657 running one of the compile commands such as @code{byte-compile-file}.
1658 Byte compiled code is usually stored in a file that ends with a
1659 @file{.elc} extension rather than a @file{.el} extension. You will
1660 see both kinds of file in the @file{emacs/lisp} directory; the files
1661 to read are those with @file{.el} extensions.
1663 As a practical matter, for most things you might do to customize or
1664 extend Emacs, you do not need to byte compile; and I will not discuss
1665 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1666 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1673 When the Lisp interpreter works on an expression, the term for the
1674 activity is called @dfn{evaluation}. We say that the interpreter
1675 `evaluates the expression'. I've used this term several times before.
1676 The word comes from its use in everyday language, `to ascertain the
1677 value or amount of; to appraise', according to @cite{Webster's New
1678 Collegiate Dictionary}.
1681 * How the Interpreter Acts:: Returns and Side Effects...
1682 * Evaluating Inner Lists:: Lists within lists...
1686 @node How the Interpreter Acts
1687 @unnumberedsubsec How the Lisp Interpreter Acts
1690 @cindex @samp{returned value} explained
1691 After evaluating an expression, the Lisp interpreter will most likely
1692 @dfn{return} the value that the computer produces by carrying out the
1693 instructions it found in the function definition, or perhaps it will
1694 give up on that function and produce an error message. (The interpreter
1695 may also find itself tossed, so to speak, to a different function or it
1696 may attempt to repeat continually what it is doing for ever and ever in
1697 what is called an `infinite loop'. These actions are less common; and
1698 we can ignore them.) Most frequently, the interpreter returns a value.
1700 @cindex @samp{side effect} defined
1701 At the same time the interpreter returns a value, it may do something
1702 else as well, such as move a cursor or copy a file; this other kind of
1703 action is called a @dfn{side effect}. Actions that we humans think are
1704 important, such as printing results, are often ``side effects'' to the
1705 Lisp interpreter. The jargon can sound peculiar, but it turns out that
1706 it is fairly easy to learn to use side effects.
1708 In summary, evaluating a symbolic expression most commonly causes the
1709 Lisp interpreter to return a value and perhaps carry out a side effect;
1710 or else produce an error.
1712 @node Evaluating Inner Lists
1713 @subsection Evaluating Inner Lists
1714 @cindex Inner list evaluation
1715 @cindex Evaluating inner lists
1717 If evaluation applies to a list that is inside another list, the outer
1718 list may use the value returned by the first evaluation as information
1719 when the outer list is evaluated. This explains why inner expressions
1720 are evaluated first: the values they return are used by the outer
1724 We can investigate this process by evaluating another addition example.
1725 Place your cursor after the following expression and type @kbd{C-x C-e}:
1732 The number 8 will appear in the echo area.
1734 What happens is that the Lisp interpreter first evaluates the inner
1735 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1736 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1737 returns the value 8. Since there are no more enclosing expressions to
1738 evaluate, the interpreter prints that value in the echo area.
1740 Now it is easy to understand the name of the command invoked by the
1741 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1742 letters @code{sexp} are an abbreviation for `symbolic expression', and
1743 @code{eval} is an abbreviation for `evaluate'. The command means
1744 `evaluate last symbolic expression'.
1746 As an experiment, you can try evaluating the expression by putting the
1747 cursor at the beginning of the next line immediately following the
1748 expression, or inside the expression.
1751 Here is another copy of the expression:
1758 If you place the cursor at the beginning of the blank line that
1759 immediately follows the expression and type @kbd{C-x C-e}, you will
1760 still get the value 8 printed in the echo area. Now try putting the
1761 cursor inside the expression. If you put it right after the next to
1762 last parenthesis (so it appears to sit on top of the last parenthesis),
1763 you will get a 6 printed in the echo area! This is because the command
1764 evaluates the expression @code{(+ 3 3)}.
1766 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1767 you will get the number itself. In Lisp, if you evaluate a number, you
1768 get the number itself---this is how numbers differ from symbols. If you
1769 evaluate a list starting with a symbol like @code{+}, you will get a
1770 value returned that is the result of the computer carrying out the
1771 instructions in the function definition attached to that name. If a
1772 symbol by itself is evaluated, something different happens, as we will
1773 see in the next section.
1779 In Emacs Lisp, a symbol can have a value attached to it just as it can
1780 have a function definition attached to it. The two are different.
1781 The function definition is a set of instructions that a computer will
1782 obey. A value, on the other hand, is something, such as number or a
1783 name, that can vary (which is why such a symbol is called a variable).
1784 The value of a symbol can be any expression in Lisp, such as a symbol,
1785 number, list, or string. A symbol that has a value is often called a
1788 A symbol can have both a function definition and a value attached to
1789 it at the same time. Or it can have just one or the other.
1790 The two are separate. This is somewhat similar
1791 to the way the name Cambridge can refer to the city in Massachusetts
1792 and have some information attached to the name as well, such as
1793 ``great programming center''.
1796 (Incidentally, in Emacs Lisp, a symbol can have two
1797 other things attached to it, too: a property list and a documentation
1798 string; these are discussed later.)
1801 Another way to think about this is to imagine a symbol as being a chest
1802 of drawers. The function definition is put in one drawer, the value in
1803 another, and so on. What is put in the drawer holding the value can be
1804 changed without affecting the contents of the drawer holding the
1805 function definition, and vice-verse.
1808 * fill-column Example::
1809 * Void Function:: The error message for a symbol
1811 * Void Variable:: The error message for a symbol without a value.
1815 @node fill-column Example
1816 @unnumberedsubsec @code{fill-column}, an Example Variable
1819 @findex fill-column, @r{an example variable}
1820 @cindex Example variable, @code{fill-column}
1821 @cindex Variable, example of, @code{fill-column}
1822 The variable @code{fill-column} illustrates a symbol with a value
1823 attached to it: in every GNU Emacs buffer, this symbol is set to some
1824 value, usually 72 or 70, but sometimes to some other value. To find the
1825 value of this symbol, evaluate it by itself. If you are reading this in
1826 Info inside of GNU Emacs, you can do this by putting the cursor after
1827 the symbol and typing @kbd{C-x C-e}:
1834 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1835 area. This is the value for which @code{fill-column} is set for me as I
1836 write this. It may be different for you in your Info buffer. Notice
1837 that the value returned as a variable is printed in exactly the same way
1838 as the value returned by a function carrying out its instructions. From
1839 the point of view of the Lisp interpreter, a value returned is a value
1840 returned. What kind of expression it came from ceases to matter once
1843 A symbol can have any value attached to it or, to use the jargon, we can
1844 @dfn{bind} the variable to a value: to a number, such as 72; to a
1845 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1846 oak)}; we can even bind a variable to a function definition.
1848 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1849 Setting the Value of a Variable}, for information about one way to do
1853 @subsection Error Message for a Symbol Without a Function
1854 @cindex Symbol without function error
1855 @cindex Error for symbol without function
1857 When we evaluated @code{fill-column} to find its value as a variable,
1858 we did not place parentheses around the word. This is because we did
1859 not intend to use it as a function name.
1861 If @code{fill-column} were the first or only element of a list, the
1862 Lisp interpreter would attempt to find the function definition
1863 attached to it. But @code{fill-column} has no function definition.
1864 Try evaluating this:
1872 You will create a @file{*Backtrace*} buffer that says:
1876 ---------- Buffer: *Backtrace* ----------
1877 Debugger entered--Lisp error: (void-function fill-column)
1880 eval-last-sexp-1(nil)
1882 call-interactively(eval-last-sexp)
1883 ---------- Buffer: *Backtrace* ----------
1888 (Remember, to quit the debugger and make the debugger window go away,
1889 type @kbd{q} in the @file{*Backtrace*} buffer.)
1893 In GNU Emacs 20 and before, you will produce an error message that says:
1896 Symbol's function definition is void:@: fill-column
1900 (The message will go away as soon as you move the cursor or type
1905 @subsection Error Message for a Symbol Without a Value
1906 @cindex Symbol without value error
1907 @cindex Error for symbol without value
1909 If you attempt to evaluate a symbol that does not have a value bound to
1910 it, you will receive an error message. You can see this by
1911 experimenting with our 2 plus 2 addition. In the following expression,
1912 put your cursor right after the @code{+}, before the first number 2,
1921 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1926 ---------- Buffer: *Backtrace* ----------
1927 Debugger entered--Lisp error: (void-variable +)
1929 eval-last-sexp-1(nil)
1931 call-interactively(eval-last-sexp)
1932 ---------- Buffer: *Backtrace* ----------
1937 (Again, you can quit the debugger by
1938 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1940 This backtrace is different from the very first error message we saw,
1941 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1942 In this case, the function does not have a value as a variable; while
1943 in the other error message, the function (the word `this') did not
1946 In this experiment with the @code{+}, what we did was cause the Lisp
1947 interpreter to evaluate the @code{+} and look for the value of the
1948 variable instead of the function definition. We did this by placing the
1949 cursor right after the symbol rather than after the parenthesis of the
1950 enclosing list as we did before. As a consequence, the Lisp interpreter
1951 evaluated the preceding s-expression, which in this case was
1954 Since @code{+} does not have a value bound to it, just the function
1955 definition, the error message reported that the symbol's value as a
1960 In GNU Emacs version 20 and before, your error message will say:
1963 Symbol's value as variable is void:@: +
1967 The meaning is the same as in GNU Emacs 22.
1973 @cindex Passing information to functions
1975 To see how information is passed to functions, let's look again at
1976 our old standby, the addition of two plus two. In Lisp, this is written
1983 If you evaluate this expression, the number 4 will appear in your echo
1984 area. What the Lisp interpreter does is add the numbers that follow
1987 @cindex @samp{argument} defined
1988 The numbers added by @code{+} are called the @dfn{arguments} of the
1989 function @code{+}. These numbers are the information that is given to
1990 or @dfn{passed} to the function.
1992 The word `argument' comes from the way it is used in mathematics and
1993 does not refer to a disputation between two people; instead it refers to
1994 the information presented to the function, in this case, to the
1995 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1996 that follow the function. The values returned by the evaluation of
1997 these atoms or lists are passed to the function. Different functions
1998 require different numbers of arguments; some functions require none at
1999 all.@footnote{It is curious to track the path by which the word `argument'
2000 came to have two different meanings, one in mathematics and the other in
2001 everyday English. According to the @cite{Oxford English Dictionary},
2002 the word derives from the Latin for @samp{to make clear, prove}; thus it
2003 came to mean, by one thread of derivation, `the evidence offered as
2004 proof', which is to say, `the information offered', which led to its
2005 meaning in Lisp. But in the other thread of derivation, it came to mean
2006 `to assert in a manner against which others may make counter
2007 assertions', which led to the meaning of the word as a disputation.
2008 (Note here that the English word has two different definitions attached
2009 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
2010 have two different function definitions at the same time.)}
2013 * Data types:: Types of data passed to a function.
2014 * Args as Variable or List:: An argument can be the value
2015 of a variable or list.
2016 * Variable Number of Arguments:: Some functions may take a
2017 variable number of arguments.
2018 * Wrong Type of Argument:: Passing an argument of the wrong type
2020 * message:: A useful function for sending messages.
2024 @subsection Arguments' Data Types
2026 @cindex Types of data
2027 @cindex Arguments' data types
2029 The type of data that should be passed to a function depends on what
2030 kind of information it uses. The arguments to a function such as
2031 @code{+} must have values that are numbers, since @code{+} adds numbers.
2032 Other functions use different kinds of data for their arguments.
2036 For example, the @code{concat} function links together or unites two or
2037 more strings of text to produce a string. The arguments are strings.
2038 Concatenating the two character strings @code{abc}, @code{def} produces
2039 the single string @code{abcdef}. This can be seen by evaluating the
2043 (concat "abc" "def")
2047 The value produced by evaluating this expression is @code{"abcdef"}.
2049 A function such as @code{substring} uses both a string and numbers as
2050 arguments. The function returns a part of the string, a substring of
2051 the first argument. This function takes three arguments. Its first
2052 argument is the string of characters, the second and third arguments are
2053 numbers that indicate the beginning and end of the substring. The
2054 numbers are a count of the number of characters (including spaces and
2055 punctuation) from the beginning of the string.
2058 For example, if you evaluate the following:
2061 (substring "The quick brown fox jumped." 16 19)
2065 you will see @code{"fox"} appear in the echo area. The arguments are the
2066 string and the two numbers.
2068 Note that the string passed to @code{substring} is a single atom even
2069 though it is made up of several words separated by spaces. Lisp counts
2070 everything between the two quotation marks as part of the string,
2071 including the spaces. You can think of the @code{substring} function as
2072 a kind of `atom smasher' since it takes an otherwise indivisible atom
2073 and extracts a part. However, @code{substring} is only able to extract
2074 a substring from an argument that is a string, not from another type of
2075 atom such as a number or symbol.
2077 @node Args as Variable or List
2078 @subsection An Argument as the Value of a Variable or List
2080 An argument can be a symbol that returns a value when it is evaluated.
2081 For example, when the symbol @code{fill-column} by itself is evaluated,
2082 it returns a number. This number can be used in an addition.
2085 Position the cursor after the following expression and type @kbd{C-x
2093 The value will be a number two more than what you get by evaluating
2094 @code{fill-column} alone. For me, this is 74, because my value of
2095 @code{fill-column} is 72.
2097 As we have just seen, an argument can be a symbol that returns a value
2098 when evaluated. In addition, an argument can be a list that returns a
2099 value when it is evaluated. For example, in the following expression,
2100 the arguments to the function @code{concat} are the strings
2101 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2102 @code{(number-to-string (+ 2 fill-column))}.
2104 @c For GNU Emacs 22, need number-to-string
2106 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2110 If you evaluate this expression---and if, as with my Emacs,
2111 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2112 appear in the echo area. (Note that you must put spaces after the
2113 word @samp{The} and before the word @samp{red} so they will appear in
2114 the final string. The function @code{number-to-string} converts the
2115 integer that the addition function returns to a string.
2116 @code{number-to-string} is also known as @code{int-to-string}.)
2118 @node Variable Number of Arguments
2119 @subsection Variable Number of Arguments
2120 @cindex Variable number of arguments
2121 @cindex Arguments, variable number of
2123 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2124 number of arguments. (The @code{*} is the symbol for multiplication.)
2125 This can be seen by evaluating each of the following expressions in
2126 the usual way. What you will see in the echo area is printed in this
2127 text after @samp{@result{}}, which you may read as `evaluates to'.
2130 In the first set, the functions have no arguments:
2141 In this set, the functions have one argument each:
2152 In this set, the functions have three arguments each:
2156 (+ 3 4 5) @result{} 12
2158 (* 3 4 5) @result{} 60
2162 @node Wrong Type of Argument
2163 @subsection Using the Wrong Type Object as an Argument
2164 @cindex Wrong type of argument
2165 @cindex Argument, wrong type of
2167 When a function is passed an argument of the wrong type, the Lisp
2168 interpreter produces an error message. For example, the @code{+}
2169 function expects the values of its arguments to be numbers. As an
2170 experiment we can pass it the quoted symbol @code{hello} instead of a
2171 number. Position the cursor after the following expression and type
2179 When you do this you will generate an error message. What has happened
2180 is that @code{+} has tried to add the 2 to the value returned by
2181 @code{'hello}, but the value returned by @code{'hello} is the symbol
2182 @code{hello}, not a number. Only numbers can be added. So @code{+}
2183 could not carry out its addition.
2186 You will create and enter a @file{*Backtrace*} buffer that says:
2191 ---------- Buffer: *Backtrace* ----------
2192 Debugger entered--Lisp error:
2193 (wrong-type-argument number-or-marker-p hello)
2195 eval((+ 2 (quote hello)))
2196 eval-last-sexp-1(nil)
2198 call-interactively(eval-last-sexp)
2199 ---------- Buffer: *Backtrace* ----------
2204 As usual, the error message tries to be helpful and makes sense after you
2205 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2206 the abbreviation @code{'hello}.}
2208 The first part of the error message is straightforward; it says
2209 @samp{wrong type argument}. Next comes the mysterious jargon word
2210 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2211 kind of argument the @code{+} expected.
2213 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2214 trying to determine whether the information presented it (the value of
2215 the argument) is a number or a marker (a special object representing a
2216 buffer position). What it does is test to see whether the @code{+} is
2217 being given numbers to add. It also tests to see whether the
2218 argument is something called a marker, which is a specific feature of
2219 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2220 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2221 its position is kept as a marker. The mark can be considered a
2222 number---the number of characters the location is from the beginning
2223 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2224 numeric value of marker positions as numbers.
2226 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2227 practice started in the early days of Lisp programming. The @samp{p}
2228 stands for `predicate'. In the jargon used by the early Lisp
2229 researchers, a predicate refers to a function to determine whether some
2230 property is true or false. So the @samp{p} tells us that
2231 @code{number-or-marker-p} is the name of a function that determines
2232 whether it is true or false that the argument supplied is a number or
2233 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2234 a function that tests whether its argument has the value of zero, and
2235 @code{listp}, a function that tests whether its argument is a list.
2237 Finally, the last part of the error message is the symbol @code{hello}.
2238 This is the value of the argument that was passed to @code{+}. If the
2239 addition had been passed the correct type of object, the value passed
2240 would have been a number, such as 37, rather than a symbol like
2241 @code{hello}. But then you would not have got the error message.
2245 In GNU Emacs version 20 and before, the echo area displays an error
2249 Wrong type argument:@: number-or-marker-p, hello
2252 This says, in different words, the same as the top line of the
2253 @file{*Backtrace*} buffer.
2257 @subsection The @code{message} Function
2260 Like @code{+}, the @code{message} function takes a variable number of
2261 arguments. It is used to send messages to the user and is so useful
2262 that we will describe it here.
2265 A message is printed in the echo area. For example, you can print a
2266 message in your echo area by evaluating the following list:
2269 (message "This message appears in the echo area!")
2272 The whole string between double quotation marks is a single argument
2273 and is printed @i{in toto}. (Note that in this example, the message
2274 itself will appear in the echo area within double quotes; that is
2275 because you see the value returned by the @code{message} function. In
2276 most uses of @code{message} in programs that you write, the text will
2277 be printed in the echo area as a side-effect, without the quotes.
2278 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2279 detail}, for an example of this.)
2281 However, if there is a @samp{%s} in the quoted string of characters, the
2282 @code{message} function does not print the @samp{%s} as such, but looks
2283 to the argument that follows the string. It evaluates the second
2284 argument and prints the value at the location in the string where the
2288 You can see this by positioning the cursor after the following
2289 expression and typing @kbd{C-x C-e}:
2292 (message "The name of this buffer is: %s." (buffer-name))
2296 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2297 echo area. The function @code{buffer-name} returns the name of the
2298 buffer as a string, which the @code{message} function inserts in place
2301 To print a value as an integer, use @samp{%d} in the same way as
2302 @samp{%s}. For example, to print a message in the echo area that
2303 states the value of the @code{fill-column}, evaluate the following:
2306 (message "The value of fill-column is %d." fill-column)
2310 On my system, when I evaluate this list, @code{"The value of
2311 fill-column is 72."} appears in my echo area@footnote{Actually, you
2312 can use @code{%s} to print a number. It is non-specific. @code{%d}
2313 prints only the part of a number left of a decimal point, and not
2314 anything that is not a number.}.
2316 If there is more than one @samp{%s} in the quoted string, the value of
2317 the first argument following the quoted string is printed at the
2318 location of the first @samp{%s} and the value of the second argument is
2319 printed at the location of the second @samp{%s}, and so on.
2322 For example, if you evaluate the following,
2326 (message "There are %d %s in the office!"
2327 (- fill-column 14) "pink elephants")
2332 a rather whimsical message will appear in your echo area. On my system
2333 it says, @code{"There are 58 pink elephants in the office!"}.
2335 The expression @code{(- fill-column 14)} is evaluated and the resulting
2336 number is inserted in place of the @samp{%d}; and the string in double
2337 quotes, @code{"pink elephants"}, is treated as a single argument and
2338 inserted in place of the @samp{%s}. (That is to say, a string between
2339 double quotes evaluates to itself, like a number.)
2341 Finally, here is a somewhat complex example that not only illustrates
2342 the computation of a number, but also shows how you can use an
2343 expression within an expression to generate the text that is substituted
2348 (message "He saw %d %s"
2352 "The quick brown foxes jumped." 16 21)
2357 In this example, @code{message} has three arguments: the string,
2358 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2359 the expression beginning with the function @code{concat}. The value
2360 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2361 in place of the @samp{%d}; and the value returned by the expression
2362 beginning with @code{concat} is inserted in place of the @samp{%s}.
2364 When your fill column is 70 and you evaluate the expression, the
2365 message @code{"He saw 38 red foxes leaping."} appears in your echo
2369 @section Setting the Value of a Variable
2370 @cindex Variable, setting value
2371 @cindex Setting value of variable
2373 @cindex @samp{bind} defined
2374 There are several ways by which a variable can be given a value. One of
2375 the ways is to use either the function @code{set} or the function
2376 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2377 jargon for this process is to @dfn{bind} a variable to a value.)
2379 The following sections not only describe how @code{set} and @code{setq}
2380 work but also illustrate how arguments are passed.
2383 * Using set:: Setting values.
2384 * Using setq:: Setting a quoted value.
2385 * Counting:: Using @code{setq} to count.
2389 @subsection Using @code{set}
2392 To set the value of the symbol @code{flowers} to the list @code{'(rose
2393 violet daisy buttercup)}, evaluate the following expression by
2394 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2397 (set 'flowers '(rose violet daisy buttercup))
2401 The list @code{(rose violet daisy buttercup)} will appear in the echo
2402 area. This is what is @emph{returned} by the @code{set} function. As a
2403 side effect, the symbol @code{flowers} is bound to the list; that is,
2404 the symbol @code{flowers}, which can be viewed as a variable, is given
2405 the list as its value. (This process, by the way, illustrates how a
2406 side effect to the Lisp interpreter, setting the value, can be the
2407 primary effect that we humans are interested in. This is because every
2408 Lisp function must return a value if it does not get an error, but it
2409 will only have a side effect if it is designed to have one.)
2411 After evaluating the @code{set} expression, you can evaluate the symbol
2412 @code{flowers} and it will return the value you just set. Here is the
2413 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2420 When you evaluate @code{flowers}, the list
2421 @code{(rose violet daisy buttercup)} appears in the echo area.
2423 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2424 in front of it, what you will see in the echo area is the symbol itself,
2425 @code{flowers}. Here is the quoted symbol, so you can try this:
2431 Note also, that when you use @code{set}, you need to quote both
2432 arguments to @code{set}, unless you want them evaluated. Since we do
2433 not want either argument evaluated, neither the variable
2434 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2435 are quoted. (When you use @code{set} without quoting its first
2436 argument, the first argument is evaluated before anything else is
2437 done. If you did this and @code{flowers} did not have a value
2438 already, you would get an error message that the @samp{Symbol's value
2439 as variable is void}; on the other hand, if @code{flowers} did return
2440 a value after it was evaluated, the @code{set} would attempt to set
2441 the value that was returned. There are situations where this is the
2442 right thing for the function to do; but such situations are rare.)
2445 @subsection Using @code{setq}
2448 As a practical matter, you almost always quote the first argument to
2449 @code{set}. The combination of @code{set} and a quoted first argument
2450 is so common that it has its own name: the special form @code{setq}.
2451 This special form is just like @code{set} except that the first argument
2452 is quoted automatically, so you don't need to type the quote mark
2453 yourself. Also, as an added convenience, @code{setq} permits you to set
2454 several different variables to different values, all in one expression.
2456 To set the value of the variable @code{carnivores} to the list
2457 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2461 (setq carnivores '(lion tiger leopard))
2465 This is exactly the same as using @code{set} except the first argument
2466 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2467 means @code{quote}.)
2470 With @code{set}, the expression would look like this:
2473 (set 'carnivores '(lion tiger leopard))
2476 Also, @code{setq} can be used to assign different values to
2477 different variables. The first argument is bound to the value
2478 of the second argument, the third argument is bound to the value of the
2479 fourth argument, and so on. For example, you could use the following to
2480 assign a list of trees to the symbol @code{trees} and a list of herbivores
2481 to the symbol @code{herbivores}:
2485 (setq trees '(pine fir oak maple)
2486 herbivores '(gazelle antelope zebra))
2491 (The expression could just as well have been on one line, but it might
2492 not have fit on a page; and humans find it easier to read nicely
2495 Although I have been using the term `assign', there is another way of
2496 thinking about the workings of @code{set} and @code{setq}; and that is to
2497 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2498 list. This latter way of thinking is very common and in forthcoming
2499 chapters we shall come upon at least one symbol that has `pointer' as
2500 part of its name. The name is chosen because the symbol has a value,
2501 specifically a list, attached to it; or, expressed another way,
2502 the symbol is set to ``point'' to the list.
2505 @subsection Counting
2508 Here is an example that shows how to use @code{setq} in a counter. You
2509 might use this to count how many times a part of your program repeats
2510 itself. First set a variable to zero; then add one to the number each
2511 time the program repeats itself. To do this, you need a variable that
2512 serves as a counter, and two expressions: an initial @code{setq}
2513 expression that sets the counter variable to zero; and a second
2514 @code{setq} expression that increments the counter each time it is
2519 (setq counter 0) ; @r{Let's call this the initializer.}
2521 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2523 counter ; @r{This is the counter.}
2528 (The text following the @samp{;} are comments. @xref{Change a
2529 defun, , Change a Function Definition}.)
2531 If you evaluate the first of these expressions, the initializer,
2532 @code{(setq counter 0)}, and then evaluate the third expression,
2533 @code{counter}, the number @code{0} will appear in the echo area. If
2534 you then evaluate the second expression, the incrementer, @code{(setq
2535 counter (+ counter 1))}, the counter will get the value 1. So if you
2536 again evaluate @code{counter}, the number @code{1} will appear in the
2537 echo area. Each time you evaluate the second expression, the value of
2538 the counter will be incremented.
2540 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2541 the Lisp interpreter first evaluates the innermost list; this is the
2542 addition. In order to evaluate this list, it must evaluate the variable
2543 @code{counter} and the number @code{1}. When it evaluates the variable
2544 @code{counter}, it receives its current value. It passes this value and
2545 the number @code{1} to the @code{+} which adds them together. The sum
2546 is then returned as the value of the inner list and passed to the
2547 @code{setq} which sets the variable @code{counter} to this new value.
2548 Thus, the value of the variable, @code{counter}, is changed.
2553 Learning Lisp is like climbing a hill in which the first part is the
2554 steepest. You have now climbed the most difficult part; what remains
2555 becomes easier as you progress onwards.
2563 Lisp programs are made up of expressions, which are lists or single atoms.
2566 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2567 surrounded by parentheses. A list can be empty.
2570 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2571 character symbols like @code{+}, strings of characters between double
2572 quotation marks, or numbers.
2575 A number evaluates to itself.
2578 A string between double quotes also evaluates to itself.
2581 When you evaluate a symbol by itself, its value is returned.
2584 When you evaluate a list, the Lisp interpreter looks at the first symbol
2585 in the list and then at the function definition bound to that symbol.
2586 Then the instructions in the function definition are carried out.
2589 A single quotation mark,
2596 , tells the Lisp interpreter that it should
2597 return the following expression as written, and not evaluate it as it
2598 would if the quote were not there.
2601 Arguments are the information passed to a function. The arguments to a
2602 function are computed by evaluating the rest of the elements of the list
2603 of which the function is the first element.
2606 A function always returns a value when it is evaluated (unless it gets
2607 an error); in addition, it may also carry out some action called a
2608 ``side effect''. In many cases, a function's primary purpose is to
2609 create a side effect.
2612 @node Error Message Exercises
2615 A few simple exercises:
2619 Generate an error message by evaluating an appropriate symbol that is
2620 not within parentheses.
2623 Generate an error message by evaluating an appropriate symbol that is
2624 between parentheses.
2627 Create a counter that increments by two rather than one.
2630 Write an expression that prints a message in the echo area when
2634 @node Practicing Evaluation
2635 @chapter Practicing Evaluation
2636 @cindex Practicing evaluation
2637 @cindex Evaluation practice
2639 Before learning how to write a function definition in Emacs Lisp, it is
2640 useful to spend a little time evaluating various expressions that have
2641 already been written. These expressions will be lists with the
2642 functions as their first (and often only) element. Since some of the
2643 functions associated with buffers are both simple and interesting, we
2644 will start with those. In this section, we will evaluate a few of
2645 these. In another section, we will study the code of several other
2646 buffer-related functions, to see how they were written.
2649 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2651 * Buffer Names:: Buffers and files are different.
2652 * Getting Buffers:: Getting a buffer itself, not merely its name.
2653 * Switching Buffers:: How to change to another buffer.
2654 * Buffer Size & Locations:: Where point is located and the size of
2656 * Evaluation Exercise::
2660 @node How to Evaluate
2661 @unnumberedsec How to Evaluate
2664 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2665 command to move the cursor or to scroll the screen, @i{you are evaluating
2666 an expression,} the first element of which is a function. @i{This is
2669 @cindex @samp{interactive function} defined
2670 @cindex @samp{command} defined
2671 When you type keys, you cause the Lisp interpreter to evaluate an
2672 expression and that is how you get your results. Even typing plain text
2673 involves evaluating an Emacs Lisp function, in this case, one that uses
2674 @code{self-insert-command}, which simply inserts the character you
2675 typed. The functions you evaluate by typing keystrokes are called
2676 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2677 interactive will be illustrated in the chapter on how to write function
2678 definitions. @xref{Interactive, , Making a Function Interactive}.
2680 In addition to typing keyboard commands, we have seen a second way to
2681 evaluate an expression: by positioning the cursor after a list and
2682 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2683 section. There are other ways to evaluate an expression as well; these
2684 will be described as we come to them.
2686 Besides being used for practicing evaluation, the functions shown in the
2687 next few sections are important in their own right. A study of these
2688 functions makes clear the distinction between buffers and files, how to
2689 switch to a buffer, and how to determine a location within it.
2692 @section Buffer Names
2694 @findex buffer-file-name
2696 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2697 the difference between a file and a buffer. When you evaluate the
2698 following expression, @code{(buffer-name)}, the name of the buffer
2699 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2700 the name of the file to which the buffer refers appears in the echo
2701 area. Usually, the name returned by @code{(buffer-name)} is the same as
2702 the name of the file to which it refers, and the name returned by
2703 @code{(buffer-file-name)} is the full path-name of the file.
2705 A file and a buffer are two different entities. A file is information
2706 recorded permanently in the computer (unless you delete it). A buffer,
2707 on the other hand, is information inside of Emacs that will vanish at
2708 the end of the editing session (or when you kill the buffer). Usually,
2709 a buffer contains information that you have copied from a file; we say
2710 the buffer is @dfn{visiting} that file. This copy is what you work on
2711 and modify. Changes to the buffer do not change the file, until you
2712 save the buffer. When you save the buffer, the buffer is copied to the file
2713 and is thus saved permanently.
2716 If you are reading this in Info inside of GNU Emacs, you can evaluate
2717 each of the following expressions by positioning the cursor after it and
2718 typing @kbd{C-x C-e}.
2729 When I do this in Info, the value returned by evaluating
2730 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2731 evaluating @code{(buffer-file-name)} is @file{nil}.
2733 On the other hand, while I am writing this document, the value
2734 returned by evaluating @code{(buffer-name)} is
2735 @file{"introduction.texinfo"}, and the value returned by evaluating
2736 @code{(buffer-file-name)} is
2737 @file{"/gnu/work/intro/introduction.texinfo"}.
2739 @cindex @code{nil}, history of word
2740 The former is the name of the buffer and the latter is the name of the
2741 file. In Info, the buffer name is @file{"*info*"}. Info does not
2742 point to any file, so the result of evaluating
2743 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2744 from the Latin word for `nothing'; in this case, it means that the
2745 buffer is not associated with any file. (In Lisp, @code{nil} is also
2746 used to mean `false' and is a synonym for the empty list, @code{()}.)
2748 When I am writing, the name of my buffer is
2749 @file{"introduction.texinfo"}. The name of the file to which it
2750 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2752 (In the expressions, the parentheses tell the Lisp interpreter to
2753 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2754 functions; without the parentheses, the interpreter would attempt to
2755 evaluate the symbols as variables. @xref{Variables}.)
2757 In spite of the distinction between files and buffers, you will often
2758 find that people refer to a file when they mean a buffer and vice-verse.
2759 Indeed, most people say, ``I am editing a file,'' rather than saying,
2760 ``I am editing a buffer which I will soon save to a file.'' It is
2761 almost always clear from context what people mean. When dealing with
2762 computer programs, however, it is important to keep the distinction in mind,
2763 since the computer is not as smart as a person.
2765 @cindex Buffer, history of word
2766 The word `buffer', by the way, comes from the meaning of the word as a
2767 cushion that deadens the force of a collision. In early computers, a
2768 buffer cushioned the interaction between files and the computer's
2769 central processing unit. The drums or tapes that held a file and the
2770 central processing unit were pieces of equipment that were very
2771 different from each other, working at their own speeds, in spurts. The
2772 buffer made it possible for them to work together effectively.
2773 Eventually, the buffer grew from being an intermediary, a temporary
2774 holding place, to being the place where work is done. This
2775 transformation is rather like that of a small seaport that grew into a
2776 great city: once it was merely the place where cargo was warehoused
2777 temporarily before being loaded onto ships; then it became a business
2778 and cultural center in its own right.
2780 Not all buffers are associated with files. For example, a
2781 @file{*scratch*} buffer does not visit any file. Similarly, a
2782 @file{*Help*} buffer is not associated with any file.
2784 In the old days, when you lacked a @file{~/.emacs} file and started an
2785 Emacs session by typing the command @code{emacs} alone, without naming
2786 any files, Emacs started with the @file{*scratch*} buffer visible.
2787 Nowadays, you will see a splash screen. You can follow one of the
2788 commands suggested on the splash screen, visit a file, or press the
2789 spacebar to reach the @file{*scratch*} buffer.
2791 If you switch to the @file{*scratch*} buffer, type
2792 @code{(buffer-name)}, position the cursor after it, and then type
2793 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2794 will be returned and will appear in the echo area. @code{"*scratch*"}
2795 is the name of the buffer. When you type @code{(buffer-file-name)} in
2796 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2797 in the echo area, just as it does when you evaluate
2798 @code{(buffer-file-name)} in Info.
2800 Incidentally, if you are in the @file{*scratch*} buffer and want the
2801 value returned by an expression to appear in the @file{*scratch*}
2802 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2803 instead of @kbd{C-x C-e}. This causes the value returned to appear
2804 after the expression. The buffer will look like this:
2807 (buffer-name)"*scratch*"
2811 You cannot do this in Info since Info is read-only and it will not allow
2812 you to change the contents of the buffer. But you can do this in any
2813 buffer you can edit; and when you write code or documentation (such as
2814 this book), this feature is very useful.
2816 @node Getting Buffers
2817 @section Getting Buffers
2818 @findex current-buffer
2819 @findex other-buffer
2820 @cindex Getting a buffer
2822 The @code{buffer-name} function returns the @emph{name} of the buffer;
2823 to get the buffer @emph{itself}, a different function is needed: the
2824 @code{current-buffer} function. If you use this function in code, what
2825 you get is the buffer itself.
2827 A name and the object or entity to which the name refers are different
2828 from each other. You are not your name. You are a person to whom
2829 others refer by name. If you ask to speak to George and someone hands you
2830 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2831 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2832 not be satisfied. You do not want to speak to the name, but to the
2833 person to whom the name refers. A buffer is similar: the name of the
2834 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2835 get a buffer itself, you need to use a function such as
2836 @code{current-buffer}.
2838 However, there is a slight complication: if you evaluate
2839 @code{current-buffer} in an expression on its own, as we will do here,
2840 what you see is a printed representation of the name of the buffer
2841 without the contents of the buffer. Emacs works this way for two
2842 reasons: the buffer may be thousands of lines long---too long to be
2843 conveniently displayed; and, another buffer may have the same contents
2844 but a different name, and it is important to distinguish between them.
2847 Here is an expression containing the function:
2854 If you evaluate this expression in Info in Emacs in the usual way,
2855 @file{#<buffer *info*>} will appear in the echo area. The special
2856 format indicates that the buffer itself is being returned, rather than
2859 Incidentally, while you can type a number or symbol into a program, you
2860 cannot do that with the printed representation of a buffer: the only way
2861 to get a buffer itself is with a function such as @code{current-buffer}.
2863 A related function is @code{other-buffer}. This returns the most
2864 recently selected buffer other than the one you are in currently, not
2865 a printed representation of its name. If you have recently switched
2866 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2867 will return that buffer.
2870 You can see this by evaluating the expression:
2877 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2878 the name of whatever other buffer you switched back from most
2879 recently@footnote{Actually, by default, if the buffer from which you
2880 just switched is visible to you in another window, @code{other-buffer}
2881 will choose the most recent buffer that you cannot see; this is a
2882 subtlety that I often forget.}.
2884 @node Switching Buffers
2885 @section Switching Buffers
2886 @findex switch-to-buffer
2888 @cindex Switching to a buffer
2890 The @code{other-buffer} function actually provides a buffer when it is
2891 used as an argument to a function that requires one. We can see this
2892 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2895 But first, a brief introduction to the @code{switch-to-buffer}
2896 function. When you switched back and forth from Info to the
2897 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2898 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2899 rather, to save typing, you probably only typed @kbd{RET} if the
2900 default buffer was @file{*scratch*}, or if it was different, then you
2901 typed just part of the name, such as @code{*sc}, pressed your
2902 @kbd{TAB} key to cause it to expand to the full name, and then typed
2903 @kbd{RET}.} when prompted in the minibuffer for the name of
2904 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2905 b}, cause the Lisp interpreter to evaluate the interactive function
2906 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2907 different keystrokes call or run different functions. For example,
2908 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2909 @code{forward-sentence}, and so on.
2911 By writing @code{switch-to-buffer} in an expression, and giving it a
2912 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2916 (switch-to-buffer (other-buffer))
2920 The symbol @code{switch-to-buffer} is the first element of the list,
2921 so the Lisp interpreter will treat it as a function and carry out the
2922 instructions that are attached to it. But before doing that, the
2923 interpreter will note that @code{other-buffer} is inside parentheses
2924 and work on that symbol first. @code{other-buffer} is the first (and
2925 in this case, the only) element of this list, so the Lisp interpreter
2926 calls or runs the function. It returns another buffer. Next, the
2927 interpreter runs @code{switch-to-buffer}, passing to it, as an
2928 argument, the other buffer, which is what Emacs will switch to. If
2929 you are reading this in Info, try this now. Evaluate the expression.
2930 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2931 expression will move you to your most recent other buffer that you
2932 cannot see. If you really want to go to your most recently selected
2933 buffer, even if you can still see it, you need to evaluate the
2934 following more complex expression:
2937 (switch-to-buffer (other-buffer (current-buffer) t))
2941 In this case, the first argument to @code{other-buffer} tells it which
2942 buffer to skip---the current one---and the second argument tells
2943 @code{other-buffer} it is OK to switch to a visible buffer.
2944 In regular use, @code{switch-to-buffer} takes you to an invisible
2945 window since you would most likely use @kbd{C-x o} (@code{other-window})
2946 to go to another visible buffer.}
2948 In the programming examples in later sections of this document, you will
2949 see the function @code{set-buffer} more often than
2950 @code{switch-to-buffer}. This is because of a difference between
2951 computer programs and humans: humans have eyes and expect to see the
2952 buffer on which they are working on their computer terminals. This is
2953 so obvious, it almost goes without saying. However, programs do not
2954 have eyes. When a computer program works on a buffer, that buffer does
2955 not need to be visible on the screen.
2957 @code{switch-to-buffer} is designed for humans and does two different
2958 things: it switches the buffer to which Emacs's attention is directed; and
2959 it switches the buffer displayed in the window to the new buffer.
2960 @code{set-buffer}, on the other hand, does only one thing: it switches
2961 the attention of the computer program to a different buffer. The buffer
2962 on the screen remains unchanged (of course, normally nothing happens
2963 there until the command finishes running).
2965 @cindex @samp{call} defined
2966 Also, we have just introduced another jargon term, the word @dfn{call}.
2967 When you evaluate a list in which the first symbol is a function, you
2968 are calling that function. The use of the term comes from the notion of
2969 the function as an entity that can do something for you if you `call'
2970 it---just as a plumber is an entity who can fix a leak if you call him
2973 @node Buffer Size & Locations
2974 @section Buffer Size and the Location of Point
2975 @cindex Size of buffer
2977 @cindex Point location
2978 @cindex Location of point
2980 Finally, let's look at several rather simple functions,
2981 @code{buffer-size}, @code{point}, @code{point-min}, and
2982 @code{point-max}. These give information about the size of a buffer and
2983 the location of point within it.
2985 The function @code{buffer-size} tells you the size of the current
2986 buffer; that is, the function returns a count of the number of
2987 characters in the buffer.
2994 You can evaluate this in the usual way, by positioning the
2995 cursor after the expression and typing @kbd{C-x C-e}.
2997 @cindex @samp{point} defined
2998 In Emacs, the current position of the cursor is called @dfn{point}.
2999 The expression @code{(point)} returns a number that tells you where the
3000 cursor is located as a count of the number of characters from the
3001 beginning of the buffer up to point.
3004 You can see the character count for point in this buffer by evaluating
3005 the following expression in the usual way:
3012 As I write this, the value of @code{point} is 65724. The @code{point}
3013 function is frequently used in some of the examples later in this
3017 The value of point depends, of course, on its location within the
3018 buffer. If you evaluate point in this spot, the number will be larger:
3025 For me, the value of point in this location is 66043, which means that
3026 there are 319 characters (including spaces) between the two
3027 expressions. (Doubtless, you will see different numbers, since I will
3028 have edited this since I first evaluated point.)
3030 @cindex @samp{narrowing} defined
3031 The function @code{point-min} is somewhat similar to @code{point}, but
3032 it returns the value of the minimum permissible value of point in the
3033 current buffer. This is the number 1 unless @dfn{narrowing} is in
3034 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3035 or a program, to operations on just a part of a buffer.
3036 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3037 function @code{point-max} returns the value of the maximum permissible
3038 value of point in the current buffer.
3040 @node Evaluation Exercise
3043 Find a file with which you are working and move towards its middle.
3044 Find its buffer name, file name, length, and your position in the file.
3046 @node Writing Defuns
3047 @chapter How To Write Function Definitions
3048 @cindex Definition writing
3049 @cindex Function definition writing
3050 @cindex Writing a function definition
3052 When the Lisp interpreter evaluates a list, it looks to see whether the
3053 first symbol on the list has a function definition attached to it; or,
3054 put another way, whether the symbol points to a function definition. If
3055 it does, the computer carries out the instructions in the definition. A
3056 symbol that has a function definition is called, simply, a function
3057 (although, properly speaking, the definition is the function and the
3058 symbol refers to it.)
3061 * Primitive Functions::
3062 * defun:: The @code{defun} macro.
3063 * Install:: Install a function definition.
3064 * Interactive:: Making a function interactive.
3065 * Interactive Options:: Different options for @code{interactive}.
3066 * Permanent Installation:: Installing code permanently.
3067 * let:: Creating and initializing local variables.
3069 * else:: If--then--else expressions.
3070 * Truth & Falsehood:: What Lisp considers false and true.
3071 * save-excursion:: Keeping track of point, mark, and buffer.
3077 @node Primitive Functions
3078 @unnumberedsec An Aside about Primitive Functions
3080 @cindex Primitive functions
3081 @cindex Functions, primitive
3083 @cindex C language primitives
3084 @cindex Primitives written in C
3085 All functions are defined in terms of other functions, except for a few
3086 @dfn{primitive} functions that are written in the C programming
3087 language. When you write functions' definitions, you will write them in
3088 Emacs Lisp and use other functions as your building blocks. Some of the
3089 functions you will use will themselves be written in Emacs Lisp (perhaps
3090 by you) and some will be primitives written in C@. The primitive
3091 functions are used exactly like those written in Emacs Lisp and behave
3092 like them. They are written in C so we can easily run GNU Emacs on any
3093 computer that has sufficient power and can run C.
3095 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3096 distinguish between the use of functions written in C and the use of
3097 functions written in Emacs Lisp. The difference is irrelevant. I
3098 mention the distinction only because it is interesting to know. Indeed,
3099 unless you investigate, you won't know whether an already-written
3100 function is written in Emacs Lisp or C.
3103 @section The @code{defun} Macro
3106 @cindex @samp{function definition} defined
3107 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3108 it that tells the computer what to do when the function is called.
3109 This code is called the @dfn{function definition} and is created by
3110 evaluating a Lisp expression that starts with the symbol @code{defun}
3111 (which is an abbreviation for @emph{define function}).
3113 In subsequent sections, we will look at function definitions from the
3114 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3115 we will describe a simple function definition so you can see how it
3116 looks. This function definition uses arithmetic because it makes for a
3117 simple example. Some people dislike examples using arithmetic; however,
3118 if you are such a person, do not despair. Hardly any of the code we
3119 will study in the remainder of this introduction involves arithmetic or
3120 mathematics. The examples mostly involve text in one way or another.
3122 A function definition has up to five parts following the word
3127 The name of the symbol to which the function definition should be
3131 A list of the arguments that will be passed to the function. If no
3132 arguments will be passed to the function, this is an empty list,
3136 Documentation describing the function. (Technically optional, but
3137 strongly recommended.)
3140 Optionally, an expression to make the function interactive so you can
3141 use it by typing @kbd{M-x} and then the name of the function; or by
3142 typing an appropriate key or keychord.
3144 @cindex @samp{body} defined
3146 The code that instructs the computer what to do: the @dfn{body} of the
3147 function definition.
3150 It is helpful to think of the five parts of a function definition as
3151 being organized in a template, with slots for each part:
3155 (defun @var{function-name} (@var{arguments}@dots{})
3156 "@var{optional-documentation}@dots{}"
3157 (interactive @var{argument-passing-info}) ; @r{optional}
3162 As an example, here is the code for a function that multiplies its
3163 argument by 7. (This example is not interactive. @xref{Interactive,
3164 , Making a Function Interactive}, for that information.)
3168 (defun multiply-by-seven (number)
3169 "Multiply NUMBER by seven."
3174 This definition begins with a parenthesis and the symbol @code{defun},
3175 followed by the name of the function.
3177 @cindex @samp{argument list} defined
3178 The name of the function is followed by a list that contains the
3179 arguments that will be passed to the function. This list is called
3180 the @dfn{argument list}. In this example, the list has only one
3181 element, the symbol, @code{number}. When the function is used, the
3182 symbol will be bound to the value that is used as the argument to the
3185 Instead of choosing the word @code{number} for the name of the argument,
3186 I could have picked any other name. For example, I could have chosen
3187 the word @code{multiplicand}. I picked the word `number' because it
3188 tells what kind of value is intended for this slot; but I could just as
3189 well have chosen the word `multiplicand' to indicate the role that the
3190 value placed in this slot will play in the workings of the function. I
3191 could have called it @code{foogle}, but that would have been a bad
3192 choice because it would not tell humans what it means. The choice of
3193 name is up to the programmer and should be chosen to make the meaning of
3196 Indeed, you can choose any name you wish for a symbol in an argument
3197 list, even the name of a symbol used in some other function: the name
3198 you use in an argument list is private to that particular definition.
3199 In that definition, the name refers to a different entity than any use
3200 of the same name outside the function definition. Suppose you have a
3201 nick-name `Shorty' in your family; when your family members refer to
3202 `Shorty', they mean you. But outside your family, in a movie, for
3203 example, the name `Shorty' refers to someone else. Because a name in an
3204 argument list is private to the function definition, you can change the
3205 value of such a symbol inside the body of a function without changing
3206 its value outside the function. The effect is similar to that produced
3207 by a @code{let} expression. (@xref{let, , @code{let}}.)
3210 Note also that we discuss the word `number' in two different ways: as a
3211 symbol that appears in the code, and as the name of something that will
3212 be replaced by a something else during the evaluation of the function.
3213 In the first case, @code{number} is a symbol, not a number; it happens
3214 that within the function, it is a variable who value is the number in
3215 question, but our primary interest in it is as a symbol. On the other
3216 hand, when we are talking about the function, our interest is that we
3217 will substitute a number for the word @var{number}. To keep this
3218 distinction clear, we use different typography for the two
3219 circumstances. When we talk about this function, or about how it works,
3220 we refer to this number by writing @var{number}. In the function
3221 itself, we refer to it by writing @code{number}.
3224 The argument list is followed by the documentation string that
3225 describes the function. This is what you see when you type
3226 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3227 write a documentation string like this, you should make the first line
3228 a complete sentence since some commands, such as @code{apropos}, print
3229 only the first line of a multi-line documentation string. Also, you
3230 should not indent the second line of a documentation string, if you
3231 have one, because that looks odd when you use @kbd{C-h f}
3232 (@code{describe-function}). The documentation string is optional, but
3233 it is so useful, it should be included in almost every function you
3236 @findex * @r{(multiplication)}
3237 The third line of the example consists of the body of the function
3238 definition. (Most functions' definitions, of course, are longer than
3239 this.) In this function, the body is the list, @code{(* 7 number)}, which
3240 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3241 @code{*} is the function for multiplication, just as @code{+} is the
3242 function for addition.)
3244 When you use the @code{multiply-by-seven} function, the argument
3245 @code{number} evaluates to the actual number you want used. Here is an
3246 example that shows how @code{multiply-by-seven} is used; but don't try
3247 to evaluate this yet!
3250 (multiply-by-seven 3)
3254 The symbol @code{number}, specified in the function definition in the
3255 next section, is given or ``bound to'' the value 3 in the actual use of
3256 the function. Note that although @code{number} was inside parentheses
3257 in the function definition, the argument passed to the
3258 @code{multiply-by-seven} function is not in parentheses. The
3259 parentheses are written in the function definition so the computer can
3260 figure out where the argument list ends and the rest of the function
3263 If you evaluate this example, you are likely to get an error message.
3264 (Go ahead, try it!) This is because we have written the function
3265 definition, but not yet told the computer about the definition---we have
3266 not yet installed (or `loaded') the function definition in Emacs.
3267 Installing a function is the process that tells the Lisp interpreter the
3268 definition of the function. Installation is described in the next
3272 @section Install a Function Definition
3273 @cindex Install a Function Definition
3274 @cindex Definition installation
3275 @cindex Function definition installation
3277 If you are reading this inside of Info in Emacs, you can try out the
3278 @code{multiply-by-seven} function by first evaluating the function
3279 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3280 the function definition follows. Place the cursor after the last
3281 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3282 do this, @code{multiply-by-seven} will appear in the echo area. (What
3283 this means is that when a function definition is evaluated, the value it
3284 returns is the name of the defined function.) At the same time, this
3285 action installs the function definition.
3289 (defun multiply-by-seven (number)
3290 "Multiply NUMBER by seven."
3296 By evaluating this @code{defun}, you have just installed
3297 @code{multiply-by-seven} in Emacs. The function is now just as much a
3298 part of Emacs as @code{forward-word} or any other editing function you
3299 use. (@code{multiply-by-seven} will stay installed until you quit
3300 Emacs. To reload code automatically whenever you start Emacs, see
3301 @ref{Permanent Installation, , Installing Code Permanently}.)
3304 * Effect of installation::
3305 * Change a defun:: How to change a function definition.
3309 @node Effect of installation
3310 @unnumberedsubsec The effect of installation
3313 You can see the effect of installing @code{multiply-by-seven} by
3314 evaluating the following sample. Place the cursor after the following
3315 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3319 (multiply-by-seven 3)
3322 If you wish, you can read the documentation for the function by typing
3323 @kbd{C-h f} (@code{describe-function}) and then the name of the
3324 function, @code{multiply-by-seven}. When you do this, a
3325 @file{*Help*} window will appear on your screen that says:
3329 multiply-by-seven is a Lisp function.
3330 (multiply-by-seven NUMBER)
3332 Multiply NUMBER by seven.
3337 (To return to a single window on your screen, type @kbd{C-x 1}.)
3339 @node Change a defun
3340 @subsection Change a Function Definition
3341 @cindex Changing a function definition
3342 @cindex Function definition, how to change
3343 @cindex Definition, how to change
3345 If you want to change the code in @code{multiply-by-seven}, just rewrite
3346 it. To install the new version in place of the old one, evaluate the
3347 function definition again. This is how you modify code in Emacs. It is
3350 As an example, you can change the @code{multiply-by-seven} function to
3351 add the number to itself seven times instead of multiplying the number
3352 by seven. It produces the same answer, but by a different path. At
3353 the same time, we will add a comment to the code; a comment is text
3354 that the Lisp interpreter ignores, but that a human reader may find
3355 useful or enlightening. The comment is that this is the ``second
3360 (defun multiply-by-seven (number) ; @r{Second version.}
3361 "Multiply NUMBER by seven."
3362 (+ number number number number number number number))
3366 @cindex Comments in Lisp code
3367 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3368 line that follows a semicolon is a comment. The end of the line is the
3369 end of the comment. To stretch a comment over two or more lines, begin
3370 each line with a semicolon.
3372 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3373 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3374 Reference Manual}, for more about comments.
3376 You can install this version of the @code{multiply-by-seven} function by
3377 evaluating it in the same way you evaluated the first function: place
3378 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3380 In summary, this is how you write code in Emacs Lisp: you write a
3381 function; install it; test it; and then make fixes or enhancements and
3385 @section Make a Function Interactive
3386 @cindex Interactive functions
3389 You make a function interactive by placing a list that begins with
3390 the special form @code{interactive} immediately after the
3391 documentation. A user can invoke an interactive function by typing
3392 @kbd{M-x} and then the name of the function; or by typing the keys to
3393 which it is bound, for example, by typing @kbd{C-n} for
3394 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3396 Interestingly, when you call an interactive function interactively,
3397 the value returned is not automatically displayed in the echo area.
3398 This is because you often call an interactive function for its side
3399 effects, such as moving forward by a word or line, and not for the
3400 value returned. If the returned value were displayed in the echo area
3401 each time you typed a key, it would be very distracting.
3404 * Interactive multiply-by-seven:: An overview.
3405 * multiply-by-seven in detail:: The interactive version.
3409 @node Interactive multiply-by-seven
3410 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3413 Both the use of the special form @code{interactive} and one way to
3414 display a value in the echo area can be illustrated by creating an
3415 interactive version of @code{multiply-by-seven}.
3422 (defun multiply-by-seven (number) ; @r{Interactive version.}
3423 "Multiply NUMBER by seven."
3425 (message "The result is %d" (* 7 number)))
3430 You can install this code by placing your cursor after it and typing
3431 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3432 Then, you can use this code by typing @kbd{C-u} and a number and then
3433 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3434 @samp{The result is @dots{}} followed by the product will appear in the
3437 Speaking more generally, you invoke a function like this in either of two
3442 By typing a prefix argument that contains the number to be passed, and
3443 then typing @kbd{M-x} and the name of the function, as with
3444 @kbd{C-u 3 M-x forward-sentence}; or,
3447 By typing whatever key or keychord the function is bound to, as with
3452 Both the examples just mentioned work identically to move point forward
3453 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3454 it could not be used as an example of key binding.)
3456 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3459 A prefix argument is passed to an interactive function by typing the
3460 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3461 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3462 type @kbd{C-u} without a number, it defaults to 4).
3464 @node multiply-by-seven in detail
3465 @subsection An Interactive @code{multiply-by-seven}
3467 Let's look at the use of the special form @code{interactive} and then at
3468 the function @code{message} in the interactive version of
3469 @code{multiply-by-seven}. You will recall that the function definition
3474 (defun multiply-by-seven (number) ; @r{Interactive version.}
3475 "Multiply NUMBER by seven."
3477 (message "The result is %d" (* 7 number)))
3481 In this function, the expression, @code{(interactive "p")}, is a list of
3482 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3483 the function and use its value for the argument of the function.
3486 The argument will be a number. This means that the symbol
3487 @code{number} will be bound to a number in the line:
3490 (message "The result is %d" (* 7 number))
3495 For example, if your prefix argument is 5, the Lisp interpreter will
3496 evaluate the line as if it were:
3499 (message "The result is %d" (* 7 5))
3503 (If you are reading this in GNU Emacs, you can evaluate this expression
3504 yourself.) First, the interpreter will evaluate the inner list, which
3505 is @code{(* 7 5)}. This returns a value of 35. Next, it
3506 will evaluate the outer list, passing the values of the second and
3507 subsequent elements of the list to the function @code{message}.
3509 As we have seen, @code{message} is an Emacs Lisp function especially
3510 designed for sending a one line message to a user. (@xref{message, ,
3511 The @code{message} function}.) In summary, the @code{message}
3512 function prints its first argument in the echo area as is, except for
3513 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3514 which we have not mentioned). When it sees a control sequence, the
3515 function looks to the second or subsequent arguments and prints the
3516 value of the argument in the location in the string where the control
3517 sequence is located.
3519 In the interactive @code{multiply-by-seven} function, the control string
3520 is @samp{%d}, which requires a number, and the value returned by
3521 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3522 is printed in place of the @samp{%d} and the message is @samp{The result
3525 (Note that when you call the function @code{multiply-by-seven}, the
3526 message is printed without quotes, but when you call @code{message}, the
3527 text is printed in double quotes. This is because the value returned by
3528 @code{message} is what appears in the echo area when you evaluate an
3529 expression whose first element is @code{message}; but when embedded in a
3530 function, @code{message} prints the text as a side effect without
3533 @node Interactive Options
3534 @section Different Options for @code{interactive}
3535 @cindex Options for @code{interactive}
3536 @cindex Interactive options
3538 In the example, @code{multiply-by-seven} used @code{"p"} as the
3539 argument to @code{interactive}. This argument told Emacs to interpret
3540 your typing either @kbd{C-u} followed by a number or @key{META}
3541 followed by a number as a command to pass that number to the function
3542 as its argument. Emacs has more than twenty characters predefined for
3543 use with @code{interactive}. In almost every case, one of these
3544 options will enable you to pass the right information interactively to
3545 a function. (@xref{Interactive Codes, , Code Characters for
3546 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3549 Consider the function @code{zap-to-char}. Its interactive expression
3553 (interactive "p\ncZap to char: ")
3556 The first part of the argument to @code{interactive} is @samp{p}, with
3557 which you are already familiar. This argument tells Emacs to
3558 interpret a `prefix', as a number to be passed to the function. You
3559 can specify a prefix either by typing @kbd{C-u} followed by a number
3560 or by typing @key{META} followed by a number. The prefix is the
3561 number of specified characters. Thus, if your prefix is three and the
3562 specified character is @samp{x}, then you will delete all the text up
3563 to and including the third next @samp{x}. If you do not set a prefix,
3564 then you delete all the text up to and including the specified
3565 character, but no more.
3567 The @samp{c} tells the function the name of the character to which to delete.
3569 More formally, a function with two or more arguments can have
3570 information passed to each argument by adding parts to the string that
3571 follows @code{interactive}. When you do this, the information is
3572 passed to each argument in the same order it is specified in the
3573 @code{interactive} list. In the string, each part is separated from
3574 the next part by a @samp{\n}, which is a newline. For example, you
3575 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3576 This causes Emacs to pass the value of the prefix argument (if there
3577 is one) and the character.
3579 In this case, the function definition looks like the following, where
3580 @code{arg} and @code{char} are the symbols to which @code{interactive}
3581 binds the prefix argument and the specified character:
3585 (defun @var{name-of-function} (arg char)
3586 "@var{documentation}@dots{}"
3587 (interactive "p\ncZap to char: ")
3588 @var{body-of-function}@dots{})
3593 (The space after the colon in the prompt makes it look better when you
3594 are prompted. @xref{copy-to-buffer, , The Definition of
3595 @code{copy-to-buffer}}, for an example.)
3597 When a function does not take arguments, @code{interactive} does not
3598 require any. Such a function contains the simple expression
3599 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3602 Alternatively, if the special letter-codes are not right for your
3603 application, you can pass your own arguments to @code{interactive} as
3606 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3607 for an example. @xref{Using Interactive, , Using @code{Interactive},
3608 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3609 explanation about this technique.
3611 @node Permanent Installation
3612 @section Install Code Permanently
3613 @cindex Install code permanently
3614 @cindex Permanent code installation
3615 @cindex Code installation
3617 When you install a function definition by evaluating it, it will stay
3618 installed until you quit Emacs. The next time you start a new session
3619 of Emacs, the function will not be installed unless you evaluate the
3620 function definition again.
3622 At some point, you may want to have code installed automatically
3623 whenever you start a new session of Emacs. There are several ways of
3628 If you have code that is just for yourself, you can put the code for the
3629 function definition in your @file{.emacs} initialization file. When you
3630 start Emacs, your @file{.emacs} file is automatically evaluated and all
3631 the function definitions within it are installed.
3632 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3635 Alternatively, you can put the function definitions that you want
3636 installed in one or more files of their own and use the @code{load}
3637 function to cause Emacs to evaluate and thereby install each of the
3638 functions in the files.
3639 @xref{Loading Files, , Loading Files}.
3642 Thirdly, if you have code that your whole site will use, it is usual
3643 to put it in a file called @file{site-init.el} that is loaded when
3644 Emacs is built. This makes the code available to everyone who uses
3645 your machine. (See the @file{INSTALL} file that is part of the Emacs
3649 Finally, if you have code that everyone who uses Emacs may want, you
3650 can post it on a computer network or send a copy to the Free Software
3651 Foundation. (When you do this, please license the code and its
3652 documentation under a license that permits other people to run, copy,
3653 study, modify, and redistribute the code and which protects you from
3654 having your work taken from you.) If you send a copy of your code to
3655 the Free Software Foundation, and properly protect yourself and
3656 others, it may be included in the next release of Emacs. In large
3657 part, this is how Emacs has grown over the past years, by donations.
3663 The @code{let} expression is a special form in Lisp that you will need
3664 to use in most function definitions.
3666 @code{let} is used to attach or bind a symbol to a value in such a way
3667 that the Lisp interpreter will not confuse the variable with a
3668 variable of the same name that is not part of the function.
3670 To understand why the @code{let} special form is necessary, consider
3671 the situation in which you own a home that you generally refer to as
3672 `the house', as in the sentence, ``The house needs painting.'' If you
3673 are visiting a friend and your host refers to `the house', he is
3674 likely to be referring to @emph{his} house, not yours, that is, to a
3677 If your friend is referring to his house and you think he is referring
3678 to your house, you may be in for some confusion. The same thing could
3679 happen in Lisp if a variable that is used inside of one function has
3680 the same name as a variable that is used inside of another function,
3681 and the two are not intended to refer to the same value. The
3682 @code{let} special form prevents this kind of confusion.
3685 * Prevent confusion::
3686 * Parts of let Expression::
3687 * Sample let Expression::
3688 * Uninitialized let Variables::
3692 @node Prevent confusion
3693 @unnumberedsubsec @code{let} Prevents Confusion
3696 @cindex @samp{local variable} defined
3697 @cindex @samp{variable, local}, defined
3698 The @code{let} special form prevents confusion. @code{let} creates a
3699 name for a @dfn{local variable} that overshadows any use of the same
3700 name outside the @code{let} expression. This is like understanding
3701 that whenever your host refers to `the house', he means his house, not
3702 yours. (Symbols used in argument lists work the same way.
3703 @xref{defun, , The @code{defun} Macro}.)
3705 Local variables created by a @code{let} expression retain their value
3706 @emph{only} within the @code{let} expression itself (and within
3707 expressions called within the @code{let} expression); the local
3708 variables have no effect outside the @code{let} expression.
3710 Another way to think about @code{let} is that it is like a @code{setq}
3711 that is temporary and local. The values set by @code{let} are
3712 automatically undone when the @code{let} is finished. The setting
3713 only affects expressions that are inside the bounds of the @code{let}
3714 expression. In computer science jargon, we would say ``the binding of
3715 a symbol is visible only in functions called in the @code{let} form;
3716 in Emacs Lisp, scoping is dynamic, not lexical.''
3718 @code{let} can create more than one variable at once. Also,
3719 @code{let} gives each variable it creates an initial value, either a
3720 value specified by you, or @code{nil}. (In the jargon, this is called
3721 `binding the variable to the value'.) After @code{let} has created
3722 and bound the variables, it executes the code in the body of the
3723 @code{let}, and returns the value of the last expression in the body,
3724 as the value of the whole @code{let} expression. (`Execute' is a jargon
3725 term that means to evaluate a list; it comes from the use of the word
3726 meaning `to give practical effect to' (@cite{Oxford English
3727 Dictionary}). Since you evaluate an expression to perform an action,
3728 `execute' has evolved as a synonym to `evaluate'.)
3730 @node Parts of let Expression
3731 @subsection The Parts of a @code{let} Expression
3732 @cindex @code{let} expression, parts of
3733 @cindex Parts of @code{let} expression
3735 @cindex @samp{varlist} defined
3736 A @code{let} expression is a list of three parts. The first part is
3737 the symbol @code{let}. The second part is a list, called a
3738 @dfn{varlist}, each element of which is either a symbol by itself or a
3739 two-element list, the first element of which is a symbol. The third
3740 part of the @code{let} expression is the body of the @code{let}. The
3741 body usually consists of one or more lists.
3744 A template for a @code{let} expression looks like this:
3747 (let @var{varlist} @var{body}@dots{})
3751 The symbols in the varlist are the variables that are given initial
3752 values by the @code{let} special form. Symbols by themselves are given
3753 the initial value of @code{nil}; and each symbol that is the first
3754 element of a two-element list is bound to the value that is returned
3755 when the Lisp interpreter evaluates the second element.
3757 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3758 this case, in a @code{let} expression, Emacs binds the symbol
3759 @code{thread} to an initial value of @code{nil}, and binds the symbol
3760 @code{needles} to an initial value of 3.
3762 When you write a @code{let} expression, what you do is put the
3763 appropriate expressions in the slots of the @code{let} expression
3766 If the varlist is composed of two-element lists, as is often the case,
3767 the template for the @code{let} expression looks like this:
3771 (let ((@var{variable} @var{value})
3772 (@var{variable} @var{value})
3778 @node Sample let Expression
3779 @subsection Sample @code{let} Expression
3780 @cindex Sample @code{let} expression
3781 @cindex @code{let} expression sample
3783 The following expression creates and gives initial values
3784 to the two variables @code{zebra} and @code{tiger}. The body of the
3785 @code{let} expression is a list which calls the @code{message} function.
3789 (let ((zebra 'stripes)
3791 (message "One kind of animal has %s and another is %s."
3796 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3798 The two variables are @code{zebra} and @code{tiger}. Each variable is
3799 the first element of a two-element list and each value is the second
3800 element of its two-element list. In the varlist, Emacs binds the
3801 variable @code{zebra} to the value @code{stripes}@footnote{According
3802 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3803 become impossibly dangerous as they grow older'' but the claim here is
3804 that they do not become fierce like a tiger. (1997, W. W. Norton and
3805 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3806 variable @code{tiger} to the value @code{fierce}. In this example,
3807 both values are symbols preceded by a quote. The values could just as
3808 well have been another list or a string. The body of the @code{let}
3809 follows after the list holding the variables. In this example, the
3810 body is a list that uses the @code{message} function to print a string
3814 You may evaluate the example in the usual fashion, by placing the
3815 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3816 this, the following will appear in the echo area:
3819 "One kind of animal has stripes and another is fierce."
3822 As we have seen before, the @code{message} function prints its first
3823 argument, except for @samp{%s}. In this example, the value of the variable
3824 @code{zebra} is printed at the location of the first @samp{%s} and the
3825 value of the variable @code{tiger} is printed at the location of the
3828 @node Uninitialized let Variables
3829 @subsection Uninitialized Variables in a @code{let} Statement
3830 @cindex Uninitialized @code{let} variables
3831 @cindex @code{let} variables uninitialized
3833 If you do not bind the variables in a @code{let} statement to specific
3834 initial values, they will automatically be bound to an initial value of
3835 @code{nil}, as in the following expression:
3844 "Here are %d variables with %s, %s, and %s value."
3845 birch pine fir oak))
3850 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3853 If you evaluate this expression in the usual way, the following will
3854 appear in your echo area:
3857 "Here are 3 variables with nil, nil, and some value."
3861 In this example, Emacs binds the symbol @code{birch} to the number 3,
3862 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3863 the symbol @code{oak} to the value @code{some}.
3865 Note that in the first part of the @code{let}, the variables @code{pine}
3866 and @code{fir} stand alone as atoms that are not surrounded by
3867 parentheses; this is because they are being bound to @code{nil}, the
3868 empty list. But @code{oak} is bound to @code{some} and so is a part of
3869 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3870 number 3 and so is in a list with that number. (Since a number
3871 evaluates to itself, the number does not need to be quoted. Also, the
3872 number is printed in the message using a @samp{%d} rather than a
3873 @samp{%s}.) The four variables as a group are put into a list to
3874 delimit them from the body of the @code{let}.
3877 @section The @code{if} Special Form
3879 @cindex Conditional with @code{if}
3881 A third special form, in addition to @code{defun} and @code{let}, is the
3882 conditional @code{if}. This form is used to instruct the computer to
3883 make decisions. You can write function definitions without using
3884 @code{if}, but it is used often enough, and is important enough, to be
3885 included here. It is used, for example, in the code for the
3886 function @code{beginning-of-buffer}.
3888 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3889 @emph{then} an expression is evaluated.'' If the test is not true, the
3890 expression is not evaluated. For example, you might make a decision
3891 such as, ``if it is warm and sunny, then go to the beach!''
3894 * if in more detail::
3895 * type-of-animal in detail:: An example of an @code{if} expression.
3899 @node if in more detail
3900 @unnumberedsubsec @code{if} in more detail
3903 @cindex @samp{if-part} defined
3904 @cindex @samp{then-part} defined
3905 An @code{if} expression written in Lisp does not use the word `then';
3906 the test and the action are the second and third elements of the list
3907 whose first element is @code{if}. Nonetheless, the test part of an
3908 @code{if} expression is often called the @dfn{if-part} and the second
3909 argument is often called the @dfn{then-part}.
3911 Also, when an @code{if} expression is written, the true-or-false-test
3912 is usually written on the same line as the symbol @code{if}, but the
3913 action to carry out if the test is true, the ``then-part'', is written
3914 on the second and subsequent lines. This makes the @code{if}
3915 expression easier to read.
3919 (if @var{true-or-false-test}
3920 @var{action-to-carry-out-if-test-is-true})
3925 The true-or-false-test will be an expression that
3926 is evaluated by the Lisp interpreter.
3928 Here is an example that you can evaluate in the usual manner. The test
3929 is whether the number 5 is greater than the number 4. Since it is, the
3930 message @samp{5 is greater than 4!} will be printed.
3934 (if (> 5 4) ; @r{if-part}
3935 (message "5 is greater than 4!")) ; @r{then-part}
3940 (The function @code{>} tests whether its first argument is greater than
3941 its second argument and returns true if it is.)
3942 @findex > (greater than)
3944 Of course, in actual use, the test in an @code{if} expression will not
3945 be fixed for all time as it is by the expression @code{(> 5 4)}.
3946 Instead, at least one of the variables used in the test will be bound to
3947 a value that is not known ahead of time. (If the value were known ahead
3948 of time, we would not need to run the test!)
3950 For example, the value may be bound to an argument of a function
3951 definition. In the following function definition, the character of the
3952 animal is a value that is passed to the function. If the value bound to
3953 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3954 tiger!} will be printed; otherwise, @code{nil} will be returned.
3958 (defun type-of-animal (characteristic)
3959 "Print message in echo area depending on CHARACTERISTIC.
3960 If the CHARACTERISTIC is the symbol `fierce',
3961 then warn of a tiger."
3962 (if (equal characteristic 'fierce)
3963 (message "It's a tiger!")))
3969 If you are reading this inside of GNU Emacs, you can evaluate the
3970 function definition in the usual way to install it in Emacs, and then you
3971 can evaluate the following two expressions to see the results:
3975 (type-of-animal 'fierce)
3977 (type-of-animal 'zebra)
3982 @c Following sentences rewritten to prevent overfull hbox.
3984 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
3985 following message printed in the echo area: @code{"It's a tiger!"}; and
3986 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
3987 printed in the echo area.
3989 @node type-of-animal in detail
3990 @subsection The @code{type-of-animal} Function in Detail
3992 Let's look at the @code{type-of-animal} function in detail.
3994 The function definition for @code{type-of-animal} was written by filling
3995 the slots of two templates, one for a function definition as a whole, and
3996 a second for an @code{if} expression.
3999 The template for every function that is not interactive is:
4003 (defun @var{name-of-function} (@var{argument-list})
4004 "@var{documentation}@dots{}"
4010 The parts of the function that match this template look like this:
4014 (defun type-of-animal (characteristic)
4015 "Print message in echo area depending on CHARACTERISTIC.
4016 If the CHARACTERISTIC is the symbol `fierce',
4017 then warn of a tiger."
4018 @var{body: the} @code{if} @var{expression})
4022 The name of function is @code{type-of-animal}; it is passed the value
4023 of one argument. The argument list is followed by a multi-line
4024 documentation string. The documentation string is included in the
4025 example because it is a good habit to write documentation string for
4026 every function definition. The body of the function definition
4027 consists of the @code{if} expression.
4030 The template for an @code{if} expression looks like this:
4034 (if @var{true-or-false-test}
4035 @var{action-to-carry-out-if-the-test-returns-true})
4040 In the @code{type-of-animal} function, the code for the @code{if}
4045 (if (equal characteristic 'fierce)
4046 (message "It's a tiger!")))
4051 Here, the true-or-false-test is the expression:
4054 (equal characteristic 'fierce)
4058 In Lisp, @code{equal} is a function that determines whether its first
4059 argument is equal to its second argument. The second argument is the
4060 quoted symbol @code{'fierce} and the first argument is the value of the
4061 symbol @code{characteristic}---in other words, the argument passed to
4064 In the first exercise of @code{type-of-animal}, the argument
4065 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4066 is equal to @code{fierce}, the expression, @code{(equal characteristic
4067 'fierce)}, returns a value of true. When this happens, the @code{if}
4068 evaluates the second argument or then-part of the @code{if}:
4069 @code{(message "It's tiger!")}.
4071 On the other hand, in the second exercise of @code{type-of-animal}, the
4072 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4073 is not equal to @code{fierce}, so the then-part is not evaluated and
4074 @code{nil} is returned by the @code{if} expression.
4077 @section If--then--else Expressions
4080 An @code{if} expression may have an optional third argument, called
4081 the @dfn{else-part}, for the case when the true-or-false-test returns
4082 false. When this happens, the second argument or then-part of the
4083 overall @code{if} expression is @emph{not} evaluated, but the third or
4084 else-part @emph{is} evaluated. You might think of this as the cloudy
4085 day alternative for the decision ``if it is warm and sunny, then go to
4086 the beach, else read a book!''.
4088 The word ``else'' is not written in the Lisp code; the else-part of an
4089 @code{if} expression comes after the then-part. In the written Lisp, the
4090 else-part is usually written to start on a line of its own and is
4091 indented less than the then-part:
4095 (if @var{true-or-false-test}
4096 @var{action-to-carry-out-if-the-test-returns-true}
4097 @var{action-to-carry-out-if-the-test-returns-false})
4101 For example, the following @code{if} expression prints the message @samp{4
4102 is not greater than 5!} when you evaluate it in the usual way:
4106 (if (> 4 5) ; @r{if-part}
4107 (message "4 falsely greater than 5!") ; @r{then-part}
4108 (message "4 is not greater than 5!")) ; @r{else-part}
4113 Note that the different levels of indentation make it easy to
4114 distinguish the then-part from the else-part. (GNU Emacs has several
4115 commands that automatically indent @code{if} expressions correctly.
4116 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4118 We can extend the @code{type-of-animal} function to include an
4119 else-part by simply incorporating an additional part to the @code{if}
4123 You can see the consequences of doing this if you evaluate the following
4124 version of the @code{type-of-animal} function definition to install it
4125 and then evaluate the two subsequent expressions to pass different
4126 arguments to the function.
4130 (defun type-of-animal (characteristic) ; @r{Second version.}
4131 "Print message in echo area depending on CHARACTERISTIC.
4132 If the CHARACTERISTIC is the symbol `fierce',
4133 then warn of a tiger;
4134 else say it's not fierce."
4135 (if (equal characteristic 'fierce)
4136 (message "It's a tiger!")
4137 (message "It's not fierce!")))
4144 (type-of-animal 'fierce)
4146 (type-of-animal 'zebra)
4151 @c Following sentence rewritten to prevent overfull hbox.
4153 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4154 following message printed in the echo area: @code{"It's a tiger!"}; but
4155 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4156 @code{"It's not fierce!"}.
4158 (Of course, if the @var{characteristic} were @code{ferocious}, the
4159 message @code{"It's not fierce!"} would be printed; and it would be
4160 misleading! When you write code, you need to take into account the
4161 possibility that some such argument will be tested by the @code{if}
4162 and write your program accordingly.)
4164 @node Truth & Falsehood
4165 @section Truth and Falsehood in Emacs Lisp
4166 @cindex Truth and falsehood in Emacs Lisp
4167 @cindex Falsehood and truth in Emacs Lisp
4170 There is an important aspect to the truth test in an @code{if}
4171 expression. So far, we have spoken of `true' and `false' as values of
4172 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4173 `false' is just our old friend @code{nil}. Anything else---anything
4176 The expression that tests for truth is interpreted as @dfn{true}
4177 if the result of evaluating it is a value that is not @code{nil}. In
4178 other words, the result of the test is considered true if the value
4179 returned is a number such as 47, a string such as @code{"hello"}, or a
4180 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4181 long as it is not empty), or even a buffer!
4184 * nil explained:: @code{nil} has two meanings.
4189 @unnumberedsubsec An explanation of @code{nil}
4192 Before illustrating a test for truth, we need an explanation of @code{nil}.
4194 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4195 empty list. Second, it means false and is the value returned when a
4196 true-or-false-test tests false. @code{nil} can be written as an empty
4197 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4198 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4199 to use @code{nil} for false and @code{()} for the empty list.
4201 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4202 list---is considered true. This means that if an evaluation returns
4203 something that is not an empty list, an @code{if} expression will test
4204 true. For example, if a number is put in the slot for the test, it
4205 will be evaluated and will return itself, since that is what numbers
4206 do when evaluated. In this conditional, the @code{if} expression will
4207 test true. The expression tests false only when @code{nil}, an empty
4208 list, is returned by evaluating the expression.
4210 You can see this by evaluating the two expressions in the following examples.
4212 In the first example, the number 4 is evaluated as the test in the
4213 @code{if} expression and returns itself; consequently, the then-part
4214 of the expression is evaluated and returned: @samp{true} appears in
4215 the echo area. In the second example, the @code{nil} indicates false;
4216 consequently, the else-part of the expression is evaluated and
4217 returned: @samp{false} appears in the echo area.
4234 Incidentally, if some other useful value is not available for a test that
4235 returns true, then the Lisp interpreter will return the symbol @code{t}
4236 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4237 when evaluated, as you can see by evaluating it in the usual way:
4245 On the other hand, this function returns @code{nil} if the test is false.
4251 @node save-excursion
4252 @section @code{save-excursion}
4253 @findex save-excursion
4254 @cindex Region, what it is
4255 @cindex Preserving point, mark, and buffer
4256 @cindex Point, mark, buffer preservation
4260 The @code{save-excursion} function is the third and final special form
4261 that we will discuss in this chapter.
4263 In Emacs Lisp programs used for editing, the @code{save-excursion}
4264 function is very common. It saves the location of point and mark,
4265 executes the body of the function, and then restores point and mark to
4266 their previous positions if their locations were changed. Its primary
4267 purpose is to keep the user from being surprised and disturbed by
4268 unexpected movement of point or mark.
4271 * Point and mark:: A review of various locations.
4272 * Template for save-excursion::
4276 @node Point and mark
4277 @unnumberedsubsec Point and Mark
4280 Before discussing @code{save-excursion}, however, it may be useful
4281 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4282 the current location of the cursor. Wherever the cursor
4283 is, that is point. More precisely, on terminals where the cursor
4284 appears to be on top of a character, point is immediately before the
4285 character. In Emacs Lisp, point is an integer. The first character in
4286 a buffer is number one, the second is number two, and so on. The
4287 function @code{point} returns the current position of the cursor as a
4288 number. Each buffer has its own value for point.
4290 The @dfn{mark} is another position in the buffer; its value can be set
4291 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4292 a mark has been set, you can use the command @kbd{C-x C-x}
4293 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4294 and set the mark to be the previous position of point. In addition, if
4295 you set another mark, the position of the previous mark is saved in the
4296 mark ring. Many mark positions can be saved this way. You can jump the
4297 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4300 The part of the buffer between point and mark is called @dfn{the
4301 region}. Numerous commands work on the region, including
4302 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4303 @code{print-region}.
4305 The @code{save-excursion} special form saves the locations of point and
4306 mark and restores those positions after the code within the body of the
4307 special form is evaluated by the Lisp interpreter. Thus, if point were
4308 in the beginning of a piece of text and some code moved point to the end
4309 of the buffer, the @code{save-excursion} would put point back to where
4310 it was before, after the expressions in the body of the function were
4313 In Emacs, a function frequently moves point as part of its internal
4314 workings even though a user would not expect this. For example,
4315 @code{count-lines-region} moves point. To prevent the user from being
4316 bothered by jumps that are both unexpected and (from the user's point of
4317 view) unnecessary, @code{save-excursion} is often used to keep point and
4318 mark in the location expected by the user. The use of
4319 @code{save-excursion} is good housekeeping.
4321 To make sure the house stays clean, @code{save-excursion} restores the
4322 values of point and mark even if something goes wrong in the code inside
4323 of it (or, to be more precise and to use the proper jargon, ``in case of
4324 abnormal exit''). This feature is very helpful.
4326 In addition to recording the values of point and mark,
4327 @code{save-excursion} keeps track of the current buffer, and restores
4328 it, too. This means you can write code that will change the buffer and
4329 have @code{save-excursion} switch you back to the original buffer.
4330 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4331 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4333 @node Template for save-excursion
4334 @subsection Template for a @code{save-excursion} Expression
4337 The template for code using @code{save-excursion} is simple:
4347 The body of the function is one or more expressions that will be
4348 evaluated in sequence by the Lisp interpreter. If there is more than
4349 one expression in the body, the value of the last one will be returned
4350 as the value of the @code{save-excursion} function. The other
4351 expressions in the body are evaluated only for their side effects; and
4352 @code{save-excursion} itself is used only for its side effect (which
4353 is restoring the positions of point and mark).
4356 In more detail, the template for a @code{save-excursion} expression
4362 @var{first-expression-in-body}
4363 @var{second-expression-in-body}
4364 @var{third-expression-in-body}
4366 @var{last-expression-in-body})
4371 An expression, of course, may be a symbol on its own or a list.
4373 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4374 within the body of a @code{let} expression. It looks like this:
4387 In the last few chapters we have introduced a macro and a fair number
4388 of functions and special forms. Here they are described in brief,
4389 along with a few similar functions that have not been mentioned yet.
4392 @item eval-last-sexp
4393 Evaluate the last symbolic expression before the current location of
4394 point. The value is printed in the echo area unless the function is
4395 invoked with an argument; in that case, the output is printed in the
4396 current buffer. This command is normally bound to @kbd{C-x C-e}.
4399 Define function. This macro has up to five parts: the name, a
4400 template for the arguments that will be passed to the function,
4401 documentation, an optional interactive declaration, and the body of
4405 For example, in an early version of Emacs, the function definition was
4406 as follows. (It is slightly more complex now that it seeks the first
4407 non-whitespace character rather than the first visible character.)
4411 (defun back-to-indentation ()
4412 "Move point to first visible character on line."
4414 (beginning-of-line 1)
4415 (skip-chars-forward " \t"))
4422 (defun backward-to-indentation (&optional arg)
4423 "Move backward ARG lines and position at first nonblank character."
4425 (forward-line (- (or arg 1)))
4426 (skip-chars-forward " \t"))
4428 (defun back-to-indentation ()
4429 "Move point to the first non-whitespace character on this line."
4431 (beginning-of-line 1)
4432 (skip-syntax-forward " " (line-end-position))
4433 ;; Move back over chars that have whitespace syntax but have the p flag.
4434 (backward-prefix-chars))
4438 Declare to the interpreter that the function can be used
4439 interactively. This special form may be followed by a string with one
4440 or more parts that pass the information to the arguments of the
4441 function, in sequence. These parts may also tell the interpreter to
4442 prompt for information. Parts of the string are separated by
4443 newlines, @samp{\n}.
4446 Common code characters are:
4450 The name of an existing buffer.
4453 The name of an existing file.
4456 The numeric prefix argument. (Note that this `p' is lower case.)
4459 Point and the mark, as two numeric arguments, smallest first. This
4460 is the only code letter that specifies two successive arguments
4464 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4465 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4469 Declare that a list of variables is for use within the body of the
4470 @code{let} and give them an initial value, either @code{nil} or a
4471 specified value; then evaluate the rest of the expressions in the body
4472 of the @code{let} and return the value of the last one. Inside the
4473 body of the @code{let}, the Lisp interpreter does not see the values of
4474 the variables of the same names that are bound outside of the
4482 (let ((foo (buffer-name))
4483 (bar (buffer-size)))
4485 "This buffer is %s and has %d characters."
4490 @item save-excursion
4491 Record the values of point and mark and the current buffer before
4492 evaluating the body of this special form. Restore the values of point
4493 and mark and buffer afterward.
4500 (message "We are %d characters into this buffer."
4503 (goto-char (point-min)) (point))))
4508 Evaluate the first argument to the function; if it is true, evaluate
4509 the second argument; else evaluate the third argument, if there is one.
4511 The @code{if} special form is called a @dfn{conditional}. There are
4512 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4520 (if (= 22 emacs-major-version)
4521 (message "This is version 22 Emacs")
4522 (message "This is not version 22 Emacs"))
4531 The @code{<} function tests whether its first argument is smaller than
4532 its second argument. A corresponding function, @code{>}, tests whether
4533 the first argument is greater than the second. Likewise, @code{<=}
4534 tests whether the first argument is less than or equal to the second and
4535 @code{>=} tests whether the first argument is greater than or equal to
4536 the second. In all cases, both arguments must be numbers or markers
4537 (markers indicate positions in buffers).
4541 The @code{=} function tests whether two arguments, both numbers or
4547 Test whether two objects are the same. @code{equal} uses one meaning
4548 of the word `same' and @code{eq} uses another: @code{equal} returns
4549 true if the two objects have a similar structure and contents, such as
4550 two copies of the same book. On the other hand, @code{eq}, returns
4551 true if both arguments are actually the same object.
4560 The @code{string-lessp} function tests whether its first argument is
4561 smaller than the second argument. A shorter, alternative name for the
4562 same function (a @code{defalias}) is @code{string<}.
4564 The arguments to @code{string-lessp} must be strings or symbols; the
4565 ordering is lexicographic, so case is significant. The print names of
4566 symbols are used instead of the symbols themselves.
4568 @cindex @samp{empty string} defined
4569 An empty string, @samp{""}, a string with no characters in it, is
4570 smaller than any string of characters.
4572 @code{string-equal} provides the corresponding test for equality. Its
4573 shorter, alternative name is @code{string=}. There are no string test
4574 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4577 Print a message in the echo area. The first argument is a string that
4578 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4579 arguments that follow the string. The argument used by @samp{%s} must
4580 be a string or a symbol; the argument used by @samp{%d} must be a
4581 number. The argument used by @samp{%c} must be an @sc{ascii} code
4582 number; it will be printed as the character with that @sc{ascii} code.
4583 (Various other %-sequences have not been mentioned.)
4587 The @code{setq} function sets the value of its first argument to the
4588 value of the second argument. The first argument is automatically
4589 quoted by @code{setq}. It does the same for succeeding pairs of
4590 arguments. Another function, @code{set}, takes only two arguments and
4591 evaluates both of them before setting the value returned by its first
4592 argument to the value returned by its second argument.
4595 Without an argument, return the name of the buffer, as a string.
4597 @item buffer-file-name
4598 Without an argument, return the name of the file the buffer is
4601 @item current-buffer
4602 Return the buffer in which Emacs is active; it may not be
4603 the buffer that is visible on the screen.
4606 Return the most recently selected buffer (other than the buffer passed
4607 to @code{other-buffer} as an argument and other than the current
4610 @item switch-to-buffer
4611 Select a buffer for Emacs to be active in and display it in the current
4612 window so users can look at it. Usually bound to @kbd{C-x b}.
4615 Switch Emacs's attention to a buffer on which programs will run. Don't
4616 alter what the window is showing.
4619 Return the number of characters in the current buffer.
4622 Return the value of the current position of the cursor, as an
4623 integer counting the number of characters from the beginning of the
4627 Return the minimum permissible value of point in
4628 the current buffer. This is 1, unless narrowing is in effect.
4631 Return the value of the maximum permissible value of point in the
4632 current buffer. This is the end of the buffer, unless narrowing is in
4637 @node defun Exercises
4642 Write a non-interactive function that doubles the value of its
4643 argument, a number. Make that function interactive.
4646 Write a function that tests whether the current value of
4647 @code{fill-column} is greater than the argument passed to the function,
4648 and if so, prints an appropriate message.
4651 @node Buffer Walk Through
4652 @chapter A Few Buffer--Related Functions
4654 In this chapter we study in detail several of the functions used in GNU
4655 Emacs. This is called a ``walk-through''. These functions are used as
4656 examples of Lisp code, but are not imaginary examples; with the
4657 exception of the first, simplified function definition, these functions
4658 show the actual code used in GNU Emacs. You can learn a great deal from
4659 these definitions. The functions described here are all related to
4660 buffers. Later, we will study other functions.
4663 * Finding More:: How to find more information.
4664 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4665 @code{point-min}, and @code{push-mark}.
4666 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4667 * append-to-buffer:: Uses @code{save-excursion} and
4668 @code{insert-buffer-substring}.
4669 * Buffer Related Review:: Review.
4670 * Buffer Exercises::
4674 @section Finding More Information
4676 @findex describe-function, @r{introduced}
4677 @cindex Find function documentation
4678 In this walk-through, I will describe each new function as we come to
4679 it, sometimes in detail and sometimes briefly. If you are interested,
4680 you can get the full documentation of any Emacs Lisp function at any
4681 time by typing @kbd{C-h f} and then the name of the function (and then
4682 @key{RET}). Similarly, you can get the full documentation for a
4683 variable by typing @kbd{C-h v} and then the name of the variable (and
4686 @cindex Find source of function
4687 @c In version 22, tells location both of C and of Emacs Lisp
4688 Also, @code{describe-function} will tell you the location of the
4689 function definition.
4691 Put point into the name of the file that contains the function and
4692 press the @key{RET} key. In this case, @key{RET} means
4693 @code{push-button} rather than `return' or `enter'. Emacs will take
4694 you directly to the function definition.
4699 If you move point over the file name and press
4700 the @key{RET} key, which in this case means @code{help-follow} rather
4701 than `return' or `enter', Emacs will take you directly to the function
4705 More generally, if you want to see a function in its original source
4706 file, you can use the @code{find-tag} function to jump to it.
4707 @code{find-tag} works with a wide variety of languages, not just
4708 Lisp, and C, and it works with non-programming text as well. For
4709 example, @code{find-tag} will jump to the various nodes in the
4710 Texinfo source file of this document.
4711 The @code{find-tag} function depends on `tags tables' that record
4712 the locations of the functions, variables, and other items to which
4713 @code{find-tag} jumps.
4715 To use the @code{find-tag} command, type @kbd{M-.} (i.e., press the
4716 period key while holding down the @key{META} key, or else type the
4717 @key{ESC} key and then type the period key), and then, at the prompt,
4718 type in the name of the function whose source code you want to see,
4719 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4720 switch buffers and display the source code for the function on your
4721 screen. To switch back to your current buffer, type @kbd{C-x b
4722 @key{RET}}. (On some keyboards, the @key{META} key is labeled
4725 @c !!! 22.1.1 tags table location in this paragraph
4726 @cindex TAGS table, specifying
4728 Depending on how the initial default values of your copy of Emacs are
4729 set, you may also need to specify the location of your `tags table',
4730 which is a file called @file{TAGS}. For example, if you are
4731 interested in Emacs sources, the tags table you will most likely want,
4732 if it has already been created for you, will be in a subdirectory of
4733 the @file{/usr/local/share/emacs/} directory; thus you would use the
4734 @code{M-x visit-tags-table} command and specify a pathname such as
4735 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4736 has not already been created, you will have to create it yourself. It
4737 will be in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4740 To create a @file{TAGS} file in a specific directory, switch to that
4741 directory in Emacs using @kbd{M-x cd} command, or list the directory
4742 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4743 @w{@code{etags *.el}} as the command to execute:
4746 M-x compile RET etags *.el RET
4749 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4751 After you become more familiar with Emacs Lisp, you will find that you will
4752 frequently use @code{find-tag} to navigate your way around source code;
4753 and you will create your own @file{TAGS} tables.
4755 @cindex Library, as term for `file'
4756 Incidentally, the files that contain Lisp code are conventionally
4757 called @dfn{libraries}. The metaphor is derived from that of a
4758 specialized library, such as a law library or an engineering library,
4759 rather than a general library. Each library, or file, contains
4760 functions that relate to a particular topic or activity, such as
4761 @file{abbrev.el} for handling abbreviations and other typing
4762 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4763 libraries provide code for a single activity, as the various
4764 @file{rmail@dots{}} files provide code for reading electronic mail.)
4765 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4766 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4767 by topic keywords.''
4769 @node simplified-beginning-of-buffer
4770 @section A Simplified @code{beginning-of-buffer} Definition
4771 @findex simplified-beginning-of-buffer
4773 The @code{beginning-of-buffer} command is a good function to start with
4774 since you are likely to be familiar with it and it is easy to
4775 understand. Used as an interactive command, @code{beginning-of-buffer}
4776 moves the cursor to the beginning of the buffer, leaving the mark at the
4777 previous position. It is generally bound to @kbd{M-<}.
4779 In this section, we will discuss a shortened version of the function
4780 that shows how it is most frequently used. This shortened function
4781 works as written, but it does not contain the code for a complex option.
4782 In another section, we will describe the entire function.
4783 (@xref{beginning-of-buffer, , Complete Definition of
4784 @code{beginning-of-buffer}}.)
4786 Before looking at the code, let's consider what the function
4787 definition has to contain: it must include an expression that makes
4788 the function interactive so it can be called by typing @kbd{M-x
4789 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4790 must include code to leave a mark at the original position in the
4791 buffer; and it must include code to move the cursor to the beginning
4795 Here is the complete text of the shortened version of the function:
4799 (defun simplified-beginning-of-buffer ()
4800 "Move point to the beginning of the buffer;
4801 leave mark at previous position."
4804 (goto-char (point-min)))
4808 Like all function definitions, this definition has five parts following
4809 the macro @code{defun}:
4813 The name: in this example, @code{simplified-beginning-of-buffer}.
4816 A list of the arguments: in this example, an empty list, @code{()},
4819 The documentation string.
4822 The interactive expression.
4829 In this function definition, the argument list is empty; this means that
4830 this function does not require any arguments. (When we look at the
4831 definition for the complete function, we will see that it may be passed
4832 an optional argument.)
4834 The interactive expression tells Emacs that the function is intended to
4835 be used interactively. In this example, @code{interactive} does not have
4836 an argument because @code{simplified-beginning-of-buffer} does not
4840 The body of the function consists of the two lines:
4845 (goto-char (point-min))
4849 The first of these lines is the expression, @code{(push-mark)}. When
4850 this expression is evaluated by the Lisp interpreter, it sets a mark at
4851 the current position of the cursor, wherever that may be. The position
4852 of this mark is saved in the mark ring.
4854 The next line is @code{(goto-char (point-min))}. This expression
4855 jumps the cursor to the minimum point in the buffer, that is, to the
4856 beginning of the buffer (or to the beginning of the accessible portion
4857 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4858 Narrowing and Widening}.)
4860 The @code{push-mark} command sets a mark at the place where the cursor
4861 was located before it was moved to the beginning of the buffer by the
4862 @code{(goto-char (point-min))} expression. Consequently, you can, if
4863 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4865 That is all there is to the function definition!
4867 @findex describe-function
4868 When you are reading code such as this and come upon an unfamiliar
4869 function, such as @code{goto-char}, you can find out what it does by
4870 using the @code{describe-function} command. To use this command, type
4871 @kbd{C-h f} and then type in the name of the function and press
4872 @key{RET}. The @code{describe-function} command will print the
4873 function's documentation string in a @file{*Help*} window. For
4874 example, the documentation for @code{goto-char} is:
4878 Set point to POSITION, a number or marker.
4879 Beginning of buffer is position (point-min), end is (point-max).
4884 The function's one argument is the desired position.
4887 (The prompt for @code{describe-function} will offer you the symbol
4888 under or preceding the cursor, so you can save typing by positioning
4889 the cursor right over or after the function and then typing @kbd{C-h f
4892 The @code{end-of-buffer} function definition is written in the same way as
4893 the @code{beginning-of-buffer} definition except that the body of the
4894 function contains the expression @code{(goto-char (point-max))} in place
4895 of @code{(goto-char (point-min))}.
4897 @node mark-whole-buffer
4898 @section The Definition of @code{mark-whole-buffer}
4899 @findex mark-whole-buffer
4901 The @code{mark-whole-buffer} function is no harder to understand than the
4902 @code{simplified-beginning-of-buffer} function. In this case, however,
4903 we will look at the complete function, not a shortened version.
4905 The @code{mark-whole-buffer} function is not as commonly used as the
4906 @code{beginning-of-buffer} function, but is useful nonetheless: it
4907 marks a whole buffer as a region by putting point at the beginning and
4908 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4912 * mark-whole-buffer overview::
4913 * Body of mark-whole-buffer:: Only three lines of code.
4917 @node mark-whole-buffer overview
4918 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4922 In GNU Emacs 22, the code for the complete function looks like this:
4926 (defun mark-whole-buffer ()
4927 "Put point at beginning and mark at end of buffer.
4928 You probably should not use this function in Lisp programs;
4929 it is usually a mistake for a Lisp function to use any subroutine
4930 that uses or sets the mark."
4933 (push-mark (point-max) nil t)
4934 (goto-char (point-min)))
4939 Like all other functions, the @code{mark-whole-buffer} function fits
4940 into the template for a function definition. The template looks like
4945 (defun @var{name-of-function} (@var{argument-list})
4946 "@var{documentation}@dots{}"
4947 (@var{interactive-expression}@dots{})
4952 Here is how the function works: the name of the function is
4953 @code{mark-whole-buffer}; it is followed by an empty argument list,
4954 @samp{()}, which means that the function does not require arguments.
4955 The documentation comes next.
4957 The next line is an @code{(interactive)} expression that tells Emacs
4958 that the function will be used interactively. These details are similar
4959 to the @code{simplified-beginning-of-buffer} function described in the
4963 @node Body of mark-whole-buffer
4964 @subsection Body of @code{mark-whole-buffer}
4966 The body of the @code{mark-whole-buffer} function consists of three
4973 (push-mark (point-max) nil t)
4974 (goto-char (point-min))
4978 The first of these lines is the expression, @code{(push-mark (point))}.
4980 This line does exactly the same job as the first line of the body of
4981 the @code{simplified-beginning-of-buffer} function, which is written
4982 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
4983 at the current position of the cursor.
4985 I don't know why the expression in @code{mark-whole-buffer} is written
4986 @code{(push-mark (point))} and the expression in
4987 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
4988 whoever wrote the code did not know that the arguments for
4989 @code{push-mark} are optional and that if @code{push-mark} is not
4990 passed an argument, the function automatically sets mark at the
4991 location of point by default. Or perhaps the expression was written
4992 so as to parallel the structure of the next line. In any case, the
4993 line causes Emacs to determine the position of point and set a mark
4996 In earlier versions of GNU Emacs, the next line of
4997 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
4998 expression sets a mark at the point in the buffer that has the highest
4999 number. This will be the end of the buffer (or, if the buffer is
5000 narrowed, the end of the accessible portion of the buffer.
5001 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
5002 narrowing.) After this mark has been set, the previous mark, the one
5003 set at point, is no longer set, but Emacs remembers its position, just
5004 as all other recent marks are always remembered. This means that you
5005 can, if you wish, go back to that position by typing @kbd{C-u
5009 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5013 (push-mark (point-max) nil t)
5017 The expression works nearly the same as before. It sets a mark at the
5018 highest numbered place in the buffer that it can. However, in this
5019 version, @code{push-mark} has two additional arguments. The second
5020 argument to @code{push-mark} is @code{nil}. This tells the function
5021 it @emph{should} display a message that says `Mark set' when it pushes
5022 the mark. The third argument is @code{t}. This tells
5023 @code{push-mark} to activate the mark when Transient Mark mode is
5024 turned on. Transient Mark mode highlights the currently active
5025 region. It is often turned off.
5027 Finally, the last line of the function is @code{(goto-char
5028 (point-min)))}. This is written exactly the same way as it is written
5029 in @code{beginning-of-buffer}. The expression moves the cursor to
5030 the minimum point in the buffer, that is, to the beginning of the buffer
5031 (or to the beginning of the accessible portion of the buffer). As a
5032 result of this, point is placed at the beginning of the buffer and mark
5033 is set at the end of the buffer. The whole buffer is, therefore, the
5036 @node append-to-buffer
5037 @section The Definition of @code{append-to-buffer}
5038 @findex append-to-buffer
5040 The @code{append-to-buffer} command is more complex than the
5041 @code{mark-whole-buffer} command. What it does is copy the region
5042 (that is, the part of the buffer between point and mark) from the
5043 current buffer to a specified buffer.
5046 * append-to-buffer overview::
5047 * append interactive:: A two part interactive expression.
5048 * append-to-buffer body:: Incorporates a @code{let} expression.
5049 * append save-excursion:: How the @code{save-excursion} works.
5053 @node append-to-buffer overview
5054 @unnumberedsubsec An Overview of @code{append-to-buffer}
5057 @findex insert-buffer-substring
5058 The @code{append-to-buffer} command uses the
5059 @code{insert-buffer-substring} function to copy the region.
5060 @code{insert-buffer-substring} is described by its name: it takes a
5061 string of characters from part of a buffer, a ``substring'', and
5062 inserts them into another buffer.
5064 Most of @code{append-to-buffer} is
5065 concerned with setting up the conditions for
5066 @code{insert-buffer-substring} to work: the code must specify both the
5067 buffer to which the text will go, the window it comes from and goes
5068 to, and the region that will be copied.
5071 Here is the complete text of the function:
5075 (defun append-to-buffer (buffer start end)
5076 "Append to specified buffer the text of the region.
5077 It is inserted into that buffer before its point.
5081 When calling from a program, give three arguments:
5082 BUFFER (or buffer name), START and END.
5083 START and END specify the portion of the current buffer to be copied."
5085 (list (read-buffer "Append to buffer: " (other-buffer
5086 (current-buffer) t))
5087 (region-beginning) (region-end)))
5090 (let ((oldbuf (current-buffer)))
5092 (let* ((append-to (get-buffer-create buffer))
5093 (windows (get-buffer-window-list append-to t t))
5095 (set-buffer append-to)
5096 (setq point (point))
5097 (barf-if-buffer-read-only)
5098 (insert-buffer-substring oldbuf start end)
5099 (dolist (window windows)
5100 (when (= (window-point window) point)
5101 (set-window-point window (point))))))))
5105 The function can be understood by looking at it as a series of
5106 filled-in templates.
5108 The outermost template is for the function definition. In this
5109 function, it looks like this (with several slots filled in):
5113 (defun append-to-buffer (buffer start end)
5114 "@var{documentation}@dots{}"
5115 (interactive @dots{})
5120 The first line of the function includes its name and three arguments.
5121 The arguments are the @code{buffer} to which the text will be copied, and
5122 the @code{start} and @code{end} of the region in the current buffer that
5125 The next part of the function is the documentation, which is clear and
5126 complete. As is conventional, the three arguments are written in
5127 upper case so you will notice them easily. Even better, they are
5128 described in the same order as in the argument list.
5130 Note that the documentation distinguishes between a buffer and its
5131 name. (The function can handle either.)
5133 @node append interactive
5134 @subsection The @code{append-to-buffer} Interactive Expression
5136 Since the @code{append-to-buffer} function will be used interactively,
5137 the function must have an @code{interactive} expression. (For a
5138 review of @code{interactive}, see @ref{Interactive, , Making a
5139 Function Interactive}.) The expression reads as follows:
5145 "Append to buffer: "
5146 (other-buffer (current-buffer) t))
5153 This expression is not one with letters standing for parts, as
5154 described earlier. Instead, it starts a list with these parts:
5156 The first part of the list is an expression to read the name of a
5157 buffer and return it as a string. That is @code{read-buffer}. The
5158 function requires a prompt as its first argument, @samp{"Append to
5159 buffer: "}. Its second argument tells the command what value to
5160 provide if you don't specify anything.
5162 In this case that second argument is an expression containing the
5163 function @code{other-buffer}, an exception, and a @samp{t}, standing
5166 The first argument to @code{other-buffer}, the exception, is yet
5167 another function, @code{current-buffer}. That is not going to be
5168 returned. The second argument is the symbol for true, @code{t}. that
5169 tells @code{other-buffer} that it may show visible buffers (except in
5170 this case, it will not show the current buffer, which makes sense).
5173 The expression looks like this:
5176 (other-buffer (current-buffer) t)
5179 The second and third arguments to the @code{list} expression are
5180 @code{(region-beginning)} and @code{(region-end)}. These two
5181 functions specify the beginning and end of the text to be appended.
5184 Originally, the command used the letters @samp{B} and @samp{r}.
5185 The whole @code{interactive} expression looked like this:
5188 (interactive "BAppend to buffer:@: \nr")
5192 But when that was done, the default value of the buffer switched to
5193 was invisible. That was not wanted.
5195 (The prompt was separated from the second argument with a newline,
5196 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5197 two arguments that follow the symbol @code{buffer} in the function's
5198 argument list (that is, @code{start} and @code{end}) to the values of
5199 point and mark. That argument worked fine.)
5201 @node append-to-buffer body
5202 @subsection The Body of @code{append-to-buffer}
5205 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5207 (defun append-to-buffer (buffer start end)
5208 "Append to specified buffer the text of the region.
5209 It is inserted into that buffer before its point.
5211 When calling from a program, give three arguments:
5212 BUFFER (or buffer name), START and END.
5213 START and END specify the portion of the current buffer to be copied."
5215 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5216 (region-beginning) (region-end)))
5217 (let ((oldbuf (current-buffer)))
5219 (let* ((append-to (get-buffer-create buffer))
5220 (windows (get-buffer-window-list append-to t t))
5222 (set-buffer append-to)
5223 (setq point (point))
5224 (barf-if-buffer-read-only)
5225 (insert-buffer-substring oldbuf start end)
5226 (dolist (window windows)
5227 (when (= (window-point window) point)
5228 (set-window-point window (point))))))))
5231 The body of the @code{append-to-buffer} function begins with @code{let}.
5233 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5234 @code{let} expression is to create and give initial values to one or
5235 more variables that will only be used within the body of the
5236 @code{let}. This means that such a variable will not be confused with
5237 any variable of the same name outside the @code{let} expression.
5239 We can see how the @code{let} expression fits into the function as a
5240 whole by showing a template for @code{append-to-buffer} with the
5241 @code{let} expression in outline:
5245 (defun append-to-buffer (buffer start end)
5246 "@var{documentation}@dots{}"
5247 (interactive @dots{})
5248 (let ((@var{variable} @var{value}))
5253 The @code{let} expression has three elements:
5257 The symbol @code{let};
5260 A varlist containing, in this case, a single two-element list,
5261 @code{(@var{variable} @var{value})};
5264 The body of the @code{let} expression.
5268 In the @code{append-to-buffer} function, the varlist looks like this:
5271 (oldbuf (current-buffer))
5275 In this part of the @code{let} expression, the one variable,
5276 @code{oldbuf}, is bound to the value returned by the
5277 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5278 used to keep track of the buffer in which you are working and from
5279 which you will copy.
5281 The element or elements of a varlist are surrounded by a set of
5282 parentheses so the Lisp interpreter can distinguish the varlist from
5283 the body of the @code{let}. As a consequence, the two-element list
5284 within the varlist is surrounded by a circumscribing set of parentheses.
5285 The line looks like this:
5289 (let ((oldbuf (current-buffer)))
5295 The two parentheses before @code{oldbuf} might surprise you if you did
5296 not realize that the first parenthesis before @code{oldbuf} marks the
5297 boundary of the varlist and the second parenthesis marks the beginning
5298 of the two-element list, @code{(oldbuf (current-buffer))}.
5300 @node append save-excursion
5301 @subsection @code{save-excursion} in @code{append-to-buffer}
5303 The body of the @code{let} expression in @code{append-to-buffer}
5304 consists of a @code{save-excursion} expression.
5306 The @code{save-excursion} function saves the locations of point and
5307 mark, and restores them to those positions after the expressions in the
5308 body of the @code{save-excursion} complete execution. In addition,
5309 @code{save-excursion} keeps track of the original buffer, and
5310 restores it. This is how @code{save-excursion} is used in
5311 @code{append-to-buffer}.
5314 @cindex Indentation for formatting
5315 @cindex Formatting convention
5316 Incidentally, it is worth noting here that a Lisp function is normally
5317 formatted so that everything that is enclosed in a multi-line spread is
5318 indented more to the right than the first symbol. In this function
5319 definition, the @code{let} is indented more than the @code{defun}, and
5320 the @code{save-excursion} is indented more than the @code{let}, like
5336 This formatting convention makes it easy to see that the lines in
5337 the body of the @code{save-excursion} are enclosed by the parentheses
5338 associated with @code{save-excursion}, just as the
5339 @code{save-excursion} itself is enclosed by the parentheses associated
5340 with the @code{let}:
5344 (let ((oldbuf (current-buffer)))
5347 (set-buffer @dots{})
5348 (insert-buffer-substring oldbuf start end)
5354 The use of the @code{save-excursion} function can be viewed as a process
5355 of filling in the slots of a template:
5360 @var{first-expression-in-body}
5361 @var{second-expression-in-body}
5363 @var{last-expression-in-body})
5369 In this function, the body of the @code{save-excursion} contains only
5370 one expression, the @code{let*} expression. You know about a
5371 @code{let} function. The @code{let*} function is different. It has a
5372 @samp{*} in its name. It enables Emacs to set each variable in its
5373 varlist in sequence, one after another.
5375 Its critical feature is that variables later in the varlist can make
5376 use of the values to which Emacs set variables earlier in the varlist.
5377 @xref{fwd-para let, , The @code{let*} expression}.
5379 We will skip functions like @code{let*} and focus on two: the
5380 @code{set-buffer} function and the @code{insert-buffer-substring}
5384 In the old days, the @code{set-buffer} expression was simply
5387 (set-buffer (get-buffer-create buffer))
5395 (set-buffer append-to)
5399 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5400 on in the @code{let*} expression. That extra binding would not be
5401 necessary except for that @code{append-to} is used later in the
5402 varlist as an argument to @code{get-buffer-window-list}.
5407 (let ((oldbuf (current-buffer)))
5409 (let* ((append-to (get-buffer-create buffer))
5410 (windows (get-buffer-window-list append-to t t))
5412 (set-buffer append-to)
5413 (setq point (point))
5414 (barf-if-buffer-read-only)
5415 (insert-buffer-substring oldbuf start end)
5416 (dolist (window windows)
5417 (when (= (window-point window) point)
5418 (set-window-point window (point))))))))
5421 The @code{append-to-buffer} function definition inserts text from the
5422 buffer in which you are currently to a named buffer. It happens that
5423 @code{insert-buffer-substring} copies text from another buffer to the
5424 current buffer, just the reverse---that is why the
5425 @code{append-to-buffer} definition starts out with a @code{let} that
5426 binds the local symbol @code{oldbuf} to the value returned by
5427 @code{current-buffer}.
5430 The @code{insert-buffer-substring} expression looks like this:
5433 (insert-buffer-substring oldbuf start end)
5437 The @code{insert-buffer-substring} function copies a string
5438 @emph{from} the buffer specified as its first argument and inserts the
5439 string into the present buffer. In this case, the argument to
5440 @code{insert-buffer-substring} is the value of the variable created
5441 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5442 was the current buffer when you gave the @code{append-to-buffer}
5445 After @code{insert-buffer-substring} has done its work,
5446 @code{save-excursion} will restore the action to the original buffer
5447 and @code{append-to-buffer} will have done its job.
5450 Written in skeletal form, the workings of the body look like this:
5454 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5455 (save-excursion ; @r{Keep track of buffer.}
5457 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5459 @var{change-back-to-original-buffer-when-finished}
5460 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5464 In summary, @code{append-to-buffer} works as follows: it saves the
5465 value of the current buffer in the variable called @code{oldbuf}. It
5466 gets the new buffer (creating one if need be) and switches Emacs's
5467 attention to it. Using the value of @code{oldbuf}, it inserts the
5468 region of text from the old buffer into the new buffer; and then using
5469 @code{save-excursion}, it brings you back to your original buffer.
5471 In looking at @code{append-to-buffer}, you have explored a fairly
5472 complex function. It shows how to use @code{let} and
5473 @code{save-excursion}, and how to change to and come back from another
5474 buffer. Many function definitions use @code{let},
5475 @code{save-excursion}, and @code{set-buffer} this way.
5477 @node Buffer Related Review
5480 Here is a brief summary of the various functions discussed in this chapter.
5483 @item describe-function
5484 @itemx describe-variable
5485 Print the documentation for a function or variable.
5486 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5489 Find the file containing the source for a function or variable and
5490 switch buffers to it, positioning point at the beginning of the item.
5491 Conventionally bound to @kbd{M-.} (that's a period following the
5494 @item save-excursion
5495 Save the location of point and mark and restore their values after the
5496 arguments to @code{save-excursion} have been evaluated. Also, remember
5497 the current buffer and return to it.
5500 Set mark at a location and record the value of the previous mark on the
5501 mark ring. The mark is a location in the buffer that will keep its
5502 relative position even if text is added to or removed from the buffer.
5505 Set point to the location specified by the value of the argument, which
5506 can be a number, a marker, or an expression that returns the number of
5507 a position, such as @code{(point-min)}.
5509 @item insert-buffer-substring
5510 Copy a region of text from a buffer that is passed to the function as
5511 an argument and insert the region into the current buffer.
5513 @item mark-whole-buffer
5514 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5517 Switch the attention of Emacs to another buffer, but do not change the
5518 window being displayed. Used when the program rather than a human is
5519 to work on a different buffer.
5521 @item get-buffer-create
5523 Find a named buffer or create one if a buffer of that name does not
5524 exist. The @code{get-buffer} function returns @code{nil} if the named
5525 buffer does not exist.
5529 @node Buffer Exercises
5534 Write your own @code{simplified-end-of-buffer} function definition;
5535 then test it to see whether it works.
5538 Use @code{if} and @code{get-buffer} to write a function that prints a
5539 message telling you whether a buffer exists.
5542 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5547 @chapter A Few More Complex Functions
5549 In this chapter, we build on what we have learned in previous chapters
5550 by looking at more complex functions. The @code{copy-to-buffer}
5551 function illustrates use of two @code{save-excursion} expressions in
5552 one definition, while the @code{insert-buffer} function illustrates
5553 use of an asterisk in an @code{interactive} expression, use of
5554 @code{or}, and the important distinction between a name and the object
5555 to which the name refers.
5558 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5559 * insert-buffer:: Read-only, and with @code{or}.
5560 * beginning-of-buffer:: Shows @code{goto-char},
5561 @code{point-min}, and @code{push-mark}.
5562 * Second Buffer Related Review::
5563 * optional Exercise::
5566 @node copy-to-buffer
5567 @section The Definition of @code{copy-to-buffer}
5568 @findex copy-to-buffer
5570 After understanding how @code{append-to-buffer} works, it is easy to
5571 understand @code{copy-to-buffer}. This function copies text into a
5572 buffer, but instead of adding to the second buffer, it replaces all the
5573 previous text in the second buffer.
5576 The body of @code{copy-to-buffer} looks like this,
5581 (interactive "BCopy to buffer: \nr")
5582 (let ((oldbuf (current-buffer)))
5583 (with-current-buffer (get-buffer-create buffer)
5584 (barf-if-buffer-read-only)
5587 (insert-buffer-substring oldbuf start end)))))
5591 The @code{copy-to-buffer} function has a simpler @code{interactive}
5592 expression than @code{append-to-buffer}.
5595 The definition then says
5598 (with-current-buffer (get-buffer-create buffer) @dots{}
5601 First, look at the earliest inner expression; that is evaluated first.
5602 That expression starts with @code{get-buffer-create buffer}. The
5603 function tells the computer to use the buffer with the name specified
5604 as the one to which you are copying, or if such a buffer does not
5605 exist, to create it. Then, the @code{with-current-buffer} function
5606 evaluates its body with that buffer temporarily current.
5608 (This demonstrates another way to shift the computer's attention but
5609 not the user's. The @code{append-to-buffer} function showed how to do
5610 the same with @code{save-excursion} and @code{set-buffer}.
5611 @code{with-current-buffer} is a newer, and arguably easier,
5614 The @code{barf-if-buffer-read-only} function sends you an error
5615 message saying the buffer is read-only if you cannot modify it.
5617 The next line has the @code{erase-buffer} function as its sole
5618 contents. That function erases the buffer.
5620 Finally, the last two lines contain the @code{save-excursion}
5621 expression with @code{insert-buffer-substring} as its body.
5622 The @code{insert-buffer-substring} expression copies the text from
5623 the buffer you are in (and you have not seen the computer shift its
5624 attention, so you don't know that that buffer is now called
5627 Incidentally, this is what is meant by `replacement'. To replace text,
5628 Emacs erases the previous text and then inserts new text.
5631 In outline, the body of @code{copy-to-buffer} looks like this:
5635 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5636 (@var{with-the-buffer-you-are-copying-to}
5637 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5640 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5645 @section The Definition of @code{insert-buffer}
5646 @findex insert-buffer
5648 @code{insert-buffer} is yet another buffer-related function. This
5649 command copies another buffer @emph{into} the current buffer. It is the
5650 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5651 copy a region of text @emph{from} the current buffer to another buffer.
5653 Here is a discussion based on the original code. The code was
5654 simplified in 2003 and is harder to understand.
5656 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5657 a discussion of the new body.)
5659 In addition, this code illustrates the use of @code{interactive} with a
5660 buffer that might be @dfn{read-only} and the important distinction
5661 between the name of an object and the object actually referred to.
5664 * insert-buffer code::
5665 * insert-buffer interactive:: When you can read, but not write.
5666 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5667 * if & or:: Using an @code{if} instead of an @code{or}.
5668 * Insert or:: How the @code{or} expression works.
5669 * Insert let:: Two @code{save-excursion} expressions.
5670 * New insert-buffer::
5674 @node insert-buffer code
5675 @unnumberedsubsec The Code for @code{insert-buffer}
5679 Here is the earlier code:
5683 (defun insert-buffer (buffer)
5684 "Insert after point the contents of BUFFER.
5685 Puts mark after the inserted text.
5686 BUFFER may be a buffer or a buffer name."
5687 (interactive "*bInsert buffer:@: ")
5690 (or (bufferp buffer)
5691 (setq buffer (get-buffer buffer)))
5692 (let (start end newmark)
5696 (setq start (point-min) end (point-max)))
5699 (insert-buffer-substring buffer start end)
5700 (setq newmark (point)))
5701 (push-mark newmark)))
5706 As with other function definitions, you can use a template to see an
5707 outline of the function:
5711 (defun insert-buffer (buffer)
5712 "@var{documentation}@dots{}"
5713 (interactive "*bInsert buffer:@: ")
5718 @node insert-buffer interactive
5719 @subsection The Interactive Expression in @code{insert-buffer}
5720 @findex interactive, @r{example use of}
5722 In @code{insert-buffer}, the argument to the @code{interactive}
5723 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5727 * Read-only buffer:: When a buffer cannot be modified.
5728 * b for interactive:: An existing buffer or else its name.
5731 @node Read-only buffer
5732 @unnumberedsubsubsec A Read-only Buffer
5733 @cindex Read-only buffer
5734 @cindex Asterisk for read-only buffer
5735 @findex * @r{for read-only buffer}
5737 The asterisk is for the situation when the current buffer is a
5738 read-only buffer---a buffer that cannot be modified. If
5739 @code{insert-buffer} is called when the current buffer is read-only, a
5740 message to this effect is printed in the echo area and the terminal
5741 may beep or blink at you; you will not be permitted to insert anything
5742 into current buffer. The asterisk does not need to be followed by a
5743 newline to separate it from the next argument.
5745 @node b for interactive
5746 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5748 The next argument in the interactive expression starts with a lower
5749 case @samp{b}. (This is different from the code for
5750 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5751 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5752 The lower-case @samp{b} tells the Lisp interpreter that the argument
5753 for @code{insert-buffer} should be an existing buffer or else its
5754 name. (The upper-case @samp{B} option provides for the possibility
5755 that the buffer does not exist.) Emacs will prompt you for the name
5756 of the buffer, offering you a default buffer, with name completion
5757 enabled. If the buffer does not exist, you receive a message that
5758 says ``No match''; your terminal may beep at you as well.
5760 The new and simplified code generates a list for @code{interactive}.
5761 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5762 functions with which we are already familiar and the @code{progn}
5763 special form with which we are not. (It will be described later.)
5765 @node insert-buffer body
5766 @subsection The Body of the @code{insert-buffer} Function
5768 The body of the @code{insert-buffer} function has two major parts: an
5769 @code{or} expression and a @code{let} expression. The purpose of the
5770 @code{or} expression is to ensure that the argument @code{buffer} is
5771 bound to a buffer and not just the name of a buffer. The body of the
5772 @code{let} expression contains the code which copies the other buffer
5773 into the current buffer.
5776 In outline, the two expressions fit into the @code{insert-buffer}
5781 (defun insert-buffer (buffer)
5782 "@var{documentation}@dots{}"
5783 (interactive "*bInsert buffer:@: ")
5788 (let (@var{varlist})
5789 @var{body-of-}@code{let}@dots{} )
5793 To understand how the @code{or} expression ensures that the argument
5794 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5795 is first necessary to understand the @code{or} function.
5797 Before doing this, let me rewrite this part of the function using
5798 @code{if} so that you can see what is done in a manner that will be familiar.
5801 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5803 The job to be done is to make sure the value of @code{buffer} is a
5804 buffer itself and not the name of a buffer. If the value is the name,
5805 then the buffer itself must be got.
5807 You can imagine yourself at a conference where an usher is wandering
5808 around holding a list with your name on it and looking for you: the
5809 usher is ``bound'' to your name, not to you; but when the usher finds
5810 you and takes your arm, the usher becomes ``bound'' to you.
5813 In Lisp, you might describe this situation like this:
5817 (if (not (holding-on-to-guest))
5818 (find-and-take-arm-of-guest))
5822 We want to do the same thing with a buffer---if we do not have the
5823 buffer itself, we want to get it.
5826 Using a predicate called @code{bufferp} that tells us whether we have a
5827 buffer (rather than its name), we can write the code like this:
5831 (if (not (bufferp buffer)) ; @r{if-part}
5832 (setq buffer (get-buffer buffer))) ; @r{then-part}
5837 Here, the true-or-false-test of the @code{if} expression is
5838 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5839 @w{@code{(setq buffer (get-buffer buffer))}}.
5841 In the test, the function @code{bufferp} returns true if its argument is
5842 a buffer---but false if its argument is the name of the buffer. (The
5843 last character of the function name @code{bufferp} is the character
5844 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5845 indicates that the function is a predicate, which is a term that means
5846 that the function will determine whether some property is true or false.
5847 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5851 The function @code{not} precedes the expression @code{(bufferp buffer)},
5852 so the true-or-false-test looks like this:
5855 (not (bufferp buffer))
5859 @code{not} is a function that returns true if its argument is false
5860 and false if its argument is true. So if @code{(bufferp buffer)}
5861 returns true, the @code{not} expression returns false and vice-verse:
5862 what is ``not true'' is false and what is ``not false'' is true.
5864 Using this test, the @code{if} expression works as follows: when the
5865 value of the variable @code{buffer} is actually a buffer rather than
5866 its name, the true-or-false-test returns false and the @code{if}
5867 expression does not evaluate the then-part. This is fine, since we do
5868 not need to do anything to the variable @code{buffer} if it really is
5871 On the other hand, when the value of @code{buffer} is not a buffer
5872 itself, but the name of a buffer, the true-or-false-test returns true
5873 and the then-part of the expression is evaluated. In this case, the
5874 then-part is @code{(setq buffer (get-buffer buffer))}. This
5875 expression uses the @code{get-buffer} function to return an actual
5876 buffer itself, given its name. The @code{setq} then sets the variable
5877 @code{buffer} to the value of the buffer itself, replacing its previous
5878 value (which was the name of the buffer).
5881 @subsection The @code{or} in the Body
5883 The purpose of the @code{or} expression in the @code{insert-buffer}
5884 function is to ensure that the argument @code{buffer} is bound to a
5885 buffer and not just to the name of a buffer. The previous section shows
5886 how the job could have been done using an @code{if} expression.
5887 However, the @code{insert-buffer} function actually uses @code{or}.
5888 To understand this, it is necessary to understand how @code{or} works.
5891 An @code{or} function can have any number of arguments. It evaluates
5892 each argument in turn and returns the value of the first of its
5893 arguments that is not @code{nil}. Also, and this is a crucial feature
5894 of @code{or}, it does not evaluate any subsequent arguments after
5895 returning the first non-@code{nil} value.
5898 The @code{or} expression looks like this:
5902 (or (bufferp buffer)
5903 (setq buffer (get-buffer buffer)))
5908 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5909 This expression returns true (a non-@code{nil} value) if the buffer is
5910 actually a buffer, and not just the name of a buffer. In the @code{or}
5911 expression, if this is the case, the @code{or} expression returns this
5912 true value and does not evaluate the next expression---and this is fine
5913 with us, since we do not want to do anything to the value of
5914 @code{buffer} if it really is a buffer.
5916 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5917 which it will be if the value of @code{buffer} is the name of a buffer,
5918 the Lisp interpreter evaluates the next element of the @code{or}
5919 expression. This is the expression @code{(setq buffer (get-buffer
5920 buffer))}. This expression returns a non-@code{nil} value, which
5921 is the value to which it sets the variable @code{buffer}---and this
5922 value is a buffer itself, not the name of a buffer.
5924 The result of all this is that the symbol @code{buffer} is always
5925 bound to a buffer itself rather than to the name of a buffer. All
5926 this is necessary because the @code{set-buffer} function in a
5927 following line only works with a buffer itself, not with the name to a
5931 Incidentally, using @code{or}, the situation with the usher would be
5935 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5939 @subsection The @code{let} Expression in @code{insert-buffer}
5941 After ensuring that the variable @code{buffer} refers to a buffer itself
5942 and not just to the name of a buffer, the @code{insert-buffer function}
5943 continues with a @code{let} expression. This specifies three local
5944 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5945 to the initial value @code{nil}. These variables are used inside the
5946 remainder of the @code{let} and temporarily hide any other occurrence of
5947 variables of the same name in Emacs until the end of the @code{let}.
5950 The body of the @code{let} contains two @code{save-excursion}
5951 expressions. First, we will look at the inner @code{save-excursion}
5952 expression in detail. The expression looks like this:
5958 (setq start (point-min) end (point-max)))
5963 The expression @code{(set-buffer buffer)} changes Emacs's attention
5964 from the current buffer to the one from which the text will copied.
5965 In that buffer, the variables @code{start} and @code{end} are set to
5966 the beginning and end of the buffer, using the commands
5967 @code{point-min} and @code{point-max}. Note that we have here an
5968 illustration of how @code{setq} is able to set two variables in the
5969 same expression. The first argument of @code{setq} is set to the
5970 value of its second, and its third argument is set to the value of its
5973 After the body of the inner @code{save-excursion} is evaluated, the
5974 @code{save-excursion} restores the original buffer, but @code{start} and
5975 @code{end} remain set to the values of the beginning and end of the
5976 buffer from which the text will be copied.
5979 The outer @code{save-excursion} expression looks like this:
5984 (@var{inner-}@code{save-excursion}@var{-expression}
5985 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
5986 (insert-buffer-substring buffer start end)
5987 (setq newmark (point)))
5992 The @code{insert-buffer-substring} function copies the text
5993 @emph{into} the current buffer @emph{from} the region indicated by
5994 @code{start} and @code{end} in @code{buffer}. Since the whole of the
5995 second buffer lies between @code{start} and @code{end}, the whole of
5996 the second buffer is copied into the buffer you are editing. Next,
5997 the value of point, which will be at the end of the inserted text, is
5998 recorded in the variable @code{newmark}.
6000 After the body of the outer @code{save-excursion} is evaluated, point
6001 and mark are relocated to their original places.
6003 However, it is convenient to locate a mark at the end of the newly
6004 inserted text and locate point at its beginning. The @code{newmark}
6005 variable records the end of the inserted text. In the last line of
6006 the @code{let} expression, the @code{(push-mark newmark)} expression
6007 function sets a mark to this location. (The previous location of the
6008 mark is still accessible; it is recorded on the mark ring and you can
6009 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6010 located at the beginning of the inserted text, which is where it was
6011 before you called the insert function, the position of which was saved
6012 by the first @code{save-excursion}.
6015 The whole @code{let} expression looks like this:
6019 (let (start end newmark)
6023 (setq start (point-min) end (point-max)))
6024 (insert-buffer-substring buffer start end)
6025 (setq newmark (point)))
6026 (push-mark newmark))
6030 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6031 function uses @code{let}, @code{save-excursion}, and
6032 @code{set-buffer}. In addition, the function illustrates one way to
6033 use @code{or}. All these functions are building blocks that we will
6034 find and use again and again.
6036 @node New insert-buffer
6037 @subsection New Body for @code{insert-buffer}
6038 @findex insert-buffer, new version body
6039 @findex new version body for insert-buffer
6041 The body in the GNU Emacs 22 version is more confusing than the original.
6044 It consists of two expressions,
6050 (insert-buffer-substring (get-buffer buffer))
6058 except, and this is what confuses novices, very important work is done
6059 inside the @code{push-mark} expression.
6061 The @code{get-buffer} function returns a buffer with the name
6062 provided. You will note that the function is @emph{not} called
6063 @code{get-buffer-create}; it does not create a buffer if one does not
6064 already exist. The buffer returned by @code{get-buffer}, an existing
6065 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6066 whole of the buffer (since you did not specify anything else).
6068 The location into which the buffer is inserted is recorded by
6069 @code{push-mark}. Then the function returns @code{nil}, the value of
6070 its last command. Put another way, the @code{insert-buffer} function
6071 exists only to produce a side effect, inserting another buffer, not to
6074 @node beginning-of-buffer
6075 @section Complete Definition of @code{beginning-of-buffer}
6076 @findex beginning-of-buffer
6078 The basic structure of the @code{beginning-of-buffer} function has
6079 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6080 Simplified @code{beginning-of-buffer} Definition}.)
6081 This section describes the complex part of the definition.
6083 As previously described, when invoked without an argument,
6084 @code{beginning-of-buffer} moves the cursor to the beginning of the
6085 buffer (in truth, the beginning of the accessible portion of the
6086 buffer), leaving the mark at the previous position. However, when the
6087 command is invoked with a number between one and ten, the function
6088 considers that number to be a fraction of the length of the buffer,
6089 measured in tenths, and Emacs moves the cursor that fraction of the
6090 way from the beginning of the buffer. Thus, you can either call this
6091 function with the key command @kbd{M-<}, which will move the cursor to
6092 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6093 M-<} which will move the cursor to a point 70% of the way through the
6094 buffer. If a number bigger than ten is used for the argument, it
6095 moves to the end of the buffer.
6097 The @code{beginning-of-buffer} function can be called with or without an
6098 argument. The use of the argument is optional.
6101 * Optional Arguments::
6102 * beginning-of-buffer opt arg:: Example with optional argument.
6103 * beginning-of-buffer complete::
6106 @node Optional Arguments
6107 @subsection Optional Arguments
6109 Unless told otherwise, Lisp expects that a function with an argument in
6110 its function definition will be called with a value for that argument.
6111 If that does not happen, you get an error and a message that says
6112 @samp{Wrong number of arguments}.
6114 @cindex Optional arguments
6117 However, optional arguments are a feature of Lisp: a particular
6118 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6119 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6120 @samp{optional} is part of the keyword.) In a function definition, if
6121 an argument follows the keyword @code{&optional}, no value need be
6122 passed to that argument when the function is called.
6125 The first line of the function definition of @code{beginning-of-buffer}
6126 therefore looks like this:
6129 (defun beginning-of-buffer (&optional arg)
6133 In outline, the whole function looks like this:
6137 (defun beginning-of-buffer (&optional arg)
6138 "@var{documentation}@dots{}"
6140 (or (@var{is-the-argument-a-cons-cell} arg)
6141 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6143 (let (@var{determine-size-and-set-it})
6145 (@var{if-there-is-an-argument}
6146 @var{figure-out-where-to-go}
6153 The function is similar to the @code{simplified-beginning-of-buffer}
6154 function except that the @code{interactive} expression has @code{"P"}
6155 as an argument and the @code{goto-char} function is followed by an
6156 if-then-else expression that figures out where to put the cursor if
6157 there is an argument that is not a cons cell.
6159 (Since I do not explain a cons cell for many more chapters, please
6160 consider ignoring the function @code{consp}. @xref{List
6161 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6162 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6165 The @code{"P"} in the @code{interactive} expression tells Emacs to
6166 pass a prefix argument, if there is one, to the function in raw form.
6167 A prefix argument is made by typing the @key{META} key followed by a
6168 number, or by typing @kbd{C-u} and then a number. (If you don't type
6169 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6170 @code{"p"} in the @code{interactive} expression causes the function to
6171 convert a prefix arg to a number.)
6173 The true-or-false-test of the @code{if} expression looks complex, but
6174 it is not: it checks whether @code{arg} has a value that is not
6175 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6176 does; it checks whether its argument is a cons cell.) If @code{arg}
6177 has a value that is not @code{nil} (and is not a cons cell), which
6178 will be the case if @code{beginning-of-buffer} is called with a
6179 numeric argument, then this true-or-false-test will return true and
6180 the then-part of the @code{if} expression will be evaluated. On the
6181 other hand, if @code{beginning-of-buffer} is not called with an
6182 argument, the value of @code{arg} will be @code{nil} and the else-part
6183 of the @code{if} expression will be evaluated. The else-part is
6184 simply @code{point-min}, and when this is the outcome, the whole
6185 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6186 is how we saw the @code{beginning-of-buffer} function in its
6189 @node beginning-of-buffer opt arg
6190 @subsection @code{beginning-of-buffer} with an Argument
6192 When @code{beginning-of-buffer} is called with an argument, an
6193 expression is evaluated which calculates what value to pass to
6194 @code{goto-char}. This expression is rather complicated at first sight.
6195 It includes an inner @code{if} expression and much arithmetic. It looks
6200 (if (> (buffer-size) 10000)
6201 ;; @r{Avoid overflow for large buffer sizes!}
6202 (* (prefix-numeric-value arg)
6207 size (prefix-numeric-value arg))) 10)))
6212 * Disentangle beginning-of-buffer::
6213 * Large buffer case::
6214 * Small buffer case::
6218 @node Disentangle beginning-of-buffer
6219 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6222 Like other complex-looking expressions, the conditional expression
6223 within @code{beginning-of-buffer} can be disentangled by looking at it
6224 as parts of a template, in this case, the template for an if-then-else
6225 expression. In skeletal form, the expression looks like this:
6229 (if (@var{buffer-is-large}
6230 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6231 @var{else-use-alternate-calculation}
6235 The true-or-false-test of this inner @code{if} expression checks the
6236 size of the buffer. The reason for this is that the old version 18
6237 Emacs used numbers that are no bigger than eight million or so and in
6238 the computation that followed, the programmer feared that Emacs might
6239 try to use over-large numbers if the buffer were large. The term
6240 `overflow', mentioned in the comment, means numbers that are over
6241 large. More recent versions of Emacs use larger numbers, but this
6242 code has not been touched, if only because people now look at buffers
6243 that are far, far larger than ever before.
6245 There are two cases: if the buffer is large and if it is not.
6247 @node Large buffer case
6248 @unnumberedsubsubsec What happens in a large buffer
6250 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6251 whether the size of the buffer is greater than 10,000 characters. To do
6252 this, it uses the @code{>} function and the computation of @code{size}
6253 that comes from the let expression.
6255 In the old days, the function @code{buffer-size} was used. Not only
6256 was that function called several times, it gave the size of the whole
6257 buffer, not the accessible part. The computation makes much more
6258 sense when it handles just the accessible part. (@xref{Narrowing &
6259 Widening, , Narrowing and Widening}, for more information on focusing
6260 attention to an `accessible' part.)
6263 The line looks like this:
6271 When the buffer is large, the then-part of the @code{if} expression is
6272 evaluated. It reads like this (after formatting for easy reading):
6277 (prefix-numeric-value arg)
6283 This expression is a multiplication, with two arguments to the function
6286 The first argument is @code{(prefix-numeric-value arg)}. When
6287 @code{"P"} is used as the argument for @code{interactive}, the value
6288 passed to the function as its argument is passed a ``raw prefix
6289 argument'', and not a number. (It is a number in a list.) To perform
6290 the arithmetic, a conversion is necessary, and
6291 @code{prefix-numeric-value} does the job.
6293 @findex / @r{(division)}
6295 The second argument is @code{(/ size 10)}. This expression divides
6296 the numeric value by ten---the numeric value of the size of the
6297 accessible portion of the buffer. This produces a number that tells
6298 how many characters make up one tenth of the buffer size. (In Lisp,
6299 @code{/} is used for division, just as @code{*} is used for
6303 In the multiplication expression as a whole, this amount is multiplied
6304 by the value of the prefix argument---the multiplication looks like this:
6308 (* @var{numeric-value-of-prefix-arg}
6309 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6314 If, for example, the prefix argument is @samp{7}, the one-tenth value
6315 will be multiplied by 7 to give a position 70% of the way through.
6318 The result of all this is that if the accessible portion of the buffer
6319 is large, the @code{goto-char} expression reads like this:
6323 (goto-char (* (prefix-numeric-value arg)
6328 This puts the cursor where we want it.
6330 @node Small buffer case
6331 @unnumberedsubsubsec What happens in a small buffer
6333 If the buffer contains fewer than 10,000 characters, a slightly
6334 different computation is performed. You might think this is not
6335 necessary, since the first computation could do the job. However, in
6336 a small buffer, the first method may not put the cursor on exactly the
6337 desired line; the second method does a better job.
6340 The code looks like this:
6342 @c Keep this on one line.
6344 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6349 This is code in which you figure out what happens by discovering how the
6350 functions are embedded in parentheses. It is easier to read if you
6351 reformat it with each expression indented more deeply than its
6352 enclosing expression:
6360 (prefix-numeric-value arg)))
6367 Looking at parentheses, we see that the innermost operation is
6368 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6369 a number. In the following expression, this number is multiplied by
6370 the size of the accessible portion of the buffer:
6373 (* size (prefix-numeric-value arg))
6377 This multiplication creates a number that may be larger than the size of
6378 the buffer---seven times larger if the argument is 7, for example. Ten
6379 is then added to this number and finally the large number is divided by
6380 ten to provide a value that is one character larger than the percentage
6381 position in the buffer.
6383 The number that results from all this is passed to @code{goto-char} and
6384 the cursor is moved to that point.
6387 @node beginning-of-buffer complete
6388 @subsection The Complete @code{beginning-of-buffer}
6391 Here is the complete text of the @code{beginning-of-buffer} function:
6397 (defun beginning-of-buffer (&optional arg)
6398 "Move point to the beginning of the buffer;
6399 leave mark at previous position.
6400 With \\[universal-argument] prefix,
6401 do not set mark at previous position.
6403 put point N/10 of the way from the beginning.
6405 If the buffer is narrowed,
6406 this command uses the beginning and size
6407 of the accessible part of the buffer.
6411 Don't use this command in Lisp programs!
6412 \(goto-char (point-min)) is faster
6413 and avoids clobbering the mark."
6416 (and transient-mark-mode mark-active)
6420 (let ((size (- (point-max) (point-min))))
6421 (goto-char (if (and arg (not (consp arg)))
6424 ;; Avoid overflow for large buffer sizes!
6425 (* (prefix-numeric-value arg)
6427 (/ (+ 10 (* size (prefix-numeric-value arg)))
6430 (if arg (forward-line 1)))
6435 From before GNU Emacs 22
6438 (defun beginning-of-buffer (&optional arg)
6439 "Move point to the beginning of the buffer;
6440 leave mark at previous position.
6441 With arg N, put point N/10 of the way
6442 from the true beginning.
6445 Don't use this in Lisp programs!
6446 \(goto-char (point-min)) is faster
6447 and does not set the mark."
6454 (if (> (buffer-size) 10000)
6455 ;; @r{Avoid overflow for large buffer sizes!}
6456 (* (prefix-numeric-value arg)
6457 (/ (buffer-size) 10))
6460 (/ (+ 10 (* (buffer-size)
6461 (prefix-numeric-value arg)))
6464 (if arg (forward-line 1)))
6470 Except for two small points, the previous discussion shows how this
6471 function works. The first point deals with a detail in the
6472 documentation string, and the second point concerns the last line of
6476 In the documentation string, there is reference to an expression:
6479 \\[universal-argument]
6483 A @samp{\\} is used before the first square bracket of this
6484 expression. This @samp{\\} tells the Lisp interpreter to substitute
6485 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6486 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6487 be different. (@xref{Documentation Tips, , Tips for Documentation
6488 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6492 Finally, the last line of the @code{beginning-of-buffer} command says
6493 to move point to the beginning of the next line if the command is
6494 invoked with an argument:
6497 (if arg (forward-line 1)))
6501 This puts the cursor at the beginning of the first line after the
6502 appropriate tenths position in the buffer. This is a flourish that
6503 means that the cursor is always located @emph{at least} the requested
6504 tenths of the way through the buffer, which is a nicety that is,
6505 perhaps, not necessary, but which, if it did not occur, would be sure
6508 On the other hand, it also means that if you specify the command with
6509 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6510 argument' is simply a cons cell, then the command puts you at the
6511 beginning of the second line @dots{} I don't know whether this is
6512 intended or whether no one has dealt with the code to avoid this
6515 @node Second Buffer Related Review
6518 Here is a brief summary of some of the topics covered in this chapter.
6522 Evaluate each argument in sequence, and return the value of the first
6523 argument that is not @code{nil}; if none return a value that is not
6524 @code{nil}, return @code{nil}. In brief, return the first true value
6525 of the arguments; return a true value if one @emph{or} any of the
6529 Evaluate each argument in sequence, and if any are @code{nil}, return
6530 @code{nil}; if none are @code{nil}, return the value of the last
6531 argument. In brief, return a true value only if all the arguments are
6532 true; return a true value if one @emph{and} each of the others is
6536 A keyword used to indicate that an argument to a function definition
6537 is optional; this means that the function can be evaluated without the
6538 argument, if desired.
6540 @item prefix-numeric-value
6541 Convert the `raw prefix argument' produced by @code{(interactive
6542 "P")} to a numeric value.
6545 Move point forward to the beginning of the next line, or if the argument
6546 is greater than one, forward that many lines. If it can't move as far
6547 forward as it is supposed to, @code{forward-line} goes forward as far as
6548 it can and then returns a count of the number of additional lines it was
6549 supposed to move but couldn't.
6552 Delete the entire contents of the current buffer.
6555 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6558 @node optional Exercise
6559 @section @code{optional} Argument Exercise
6561 Write an interactive function with an optional argument that tests
6562 whether its argument, a number, is greater than or equal to, or else,
6563 less than the value of @code{fill-column}, and tells you which, in a
6564 message. However, if you do not pass an argument to the function, use
6565 56 as a default value.
6567 @node Narrowing & Widening
6568 @chapter Narrowing and Widening
6569 @cindex Focusing attention (narrowing)
6573 Narrowing is a feature of Emacs that makes it possible for you to focus
6574 on a specific part of a buffer, and work without accidentally changing
6575 other parts. Narrowing is normally disabled since it can confuse
6579 * Narrowing advantages:: The advantages of narrowing
6580 * save-restriction:: The @code{save-restriction} special form.
6581 * what-line:: The number of the line that point is on.
6586 @node Narrowing advantages
6587 @unnumberedsec The Advantages of Narrowing
6590 With narrowing, the rest of a buffer is made invisible, as if it weren't
6591 there. This is an advantage if, for example, you want to replace a word
6592 in one part of a buffer but not in another: you narrow to the part you want
6593 and the replacement is carried out only in that section, not in the rest
6594 of the buffer. Searches will only work within a narrowed region, not
6595 outside of one, so if you are fixing a part of a document, you can keep
6596 yourself from accidentally finding parts you do not need to fix by
6597 narrowing just to the region you want.
6598 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6600 However, narrowing does make the rest of the buffer invisible, which
6601 can scare people who inadvertently invoke narrowing and think they
6602 have deleted a part of their file. Moreover, the @code{undo} command
6603 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6604 (nor should it), so people can become quite desperate if they do not
6605 know that they can return the rest of a buffer to visibility with the
6606 @code{widen} command.
6607 (The key binding for @code{widen} is @kbd{C-x n w}.)
6609 Narrowing is just as useful to the Lisp interpreter as to a human.
6610 Often, an Emacs Lisp function is designed to work on just part of a
6611 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6612 buffer that has been narrowed. The @code{what-line} function, for
6613 example, removes the narrowing from a buffer, if it has any narrowing
6614 and when it has finished its job, restores the narrowing to what it was.
6615 On the other hand, the @code{count-lines} function
6616 uses narrowing to restrict itself to just that portion
6617 of the buffer in which it is interested and then restores the previous
6620 @node save-restriction
6621 @section The @code{save-restriction} Special Form
6622 @findex save-restriction
6624 In Emacs Lisp, you can use the @code{save-restriction} special form to
6625 keep track of whatever narrowing is in effect, if any. When the Lisp
6626 interpreter meets with @code{save-restriction}, it executes the code
6627 in the body of the @code{save-restriction} expression, and then undoes
6628 any changes to narrowing that the code caused. If, for example, the
6629 buffer is narrowed and the code that follows @code{save-restriction}
6630 gets rid of the narrowing, @code{save-restriction} returns the buffer
6631 to its narrowed region afterwards. In the @code{what-line} command,
6632 any narrowing the buffer may have is undone by the @code{widen}
6633 command that immediately follows the @code{save-restriction} command.
6634 Any original narrowing is restored just before the completion of the
6638 The template for a @code{save-restriction} expression is simple:
6648 The body of the @code{save-restriction} is one or more expressions that
6649 will be evaluated in sequence by the Lisp interpreter.
6651 Finally, a point to note: when you use both @code{save-excursion} and
6652 @code{save-restriction}, one right after the other, you should use
6653 @code{save-excursion} outermost. If you write them in reverse order,
6654 you may fail to record narrowing in the buffer to which Emacs switches
6655 after calling @code{save-excursion}. Thus, when written together,
6656 @code{save-excursion} and @code{save-restriction} should be written
6667 In other circumstances, when not written together, the
6668 @code{save-excursion} and @code{save-restriction} special forms must
6669 be written in the order appropriate to the function.
6685 /usr/local/src/emacs/lisp/simple.el
6688 "Print the current buffer line number and narrowed line number of point."
6690 (let ((start (point-min))
6691 (n (line-number-at-pos)))
6693 (message "Line %d" n)
6697 (message "line %d (narrowed line %d)"
6698 (+ n (line-number-at-pos start) -1) n))))))
6700 (defun line-number-at-pos (&optional pos)
6701 "Return (narrowed) buffer line number at position POS.
6702 If POS is nil, use current buffer location.
6703 Counting starts at (point-min), so the value refers
6704 to the contents of the accessible portion of the buffer."
6705 (let ((opoint (or pos (point))) start)
6707 (goto-char (point-min))
6708 (setq start (point))
6711 (1+ (count-lines start (point))))))
6713 (defun count-lines (start end)
6714 "Return number of lines between START and END.
6715 This is usually the number of newlines between them,
6716 but can be one more if START is not equal to END
6717 and the greater of them is not at the start of a line."
6720 (narrow-to-region start end)
6721 (goto-char (point-min))
6722 (if (eq selective-display t)
6725 (while (re-search-forward "[\n\C-m]" nil t 40)
6726 (setq done (+ 40 done)))
6727 (while (re-search-forward "[\n\C-m]" nil t 1)
6728 (setq done (+ 1 done)))
6729 (goto-char (point-max))
6730 (if (and (/= start end)
6734 (- (buffer-size) (forward-line (buffer-size)))))))
6738 @section @code{what-line}
6740 @cindex Widening, example of
6742 The @code{what-line} command tells you the number of the line in which
6743 the cursor is located. The function illustrates the use of the
6744 @code{save-restriction} and @code{save-excursion} commands. Here is the
6745 original text of the function:
6750 "Print the current line number (in the buffer) of point."
6757 (1+ (count-lines 1 (point)))))))
6761 (In recent versions of GNU Emacs, the @code{what-line} function has
6762 been expanded to tell you your line number in a narrowed buffer as
6763 well as your line number in a widened buffer. The recent version is
6764 more complex than the version shown here. If you feel adventurous,
6765 you might want to look at it after figuring out how this version
6766 works. You will probably need to use @kbd{C-h f}
6767 (@code{describe-function}). The newer version uses a conditional to
6768 determine whether the buffer has been narrowed.
6770 (Also, it uses @code{line-number-at-pos}, which among other simple
6771 expressions, such as @code{(goto-char (point-min))}, moves point to
6772 the beginning of the current line with @code{(forward-line 0)} rather
6773 than @code{beginning-of-line}.)
6775 The @code{what-line} function as shown here has a documentation line
6776 and is interactive, as you would expect. The next two lines use the
6777 functions @code{save-restriction} and @code{widen}.
6779 The @code{save-restriction} special form notes whatever narrowing is in
6780 effect, if any, in the current buffer and restores that narrowing after
6781 the code in the body of the @code{save-restriction} has been evaluated.
6783 The @code{save-restriction} special form is followed by @code{widen}.
6784 This function undoes any narrowing the current buffer may have had
6785 when @code{what-line} was called. (The narrowing that was there is
6786 the narrowing that @code{save-restriction} remembers.) This widening
6787 makes it possible for the line counting commands to count from the
6788 beginning of the buffer. Otherwise, they would have been limited to
6789 counting within the accessible region. Any original narrowing is
6790 restored just before the completion of the function by the
6791 @code{save-restriction} special form.
6793 The call to @code{widen} is followed by @code{save-excursion}, which
6794 saves the location of the cursor (i.e., of point) and of the mark, and
6795 restores them after the code in the body of the @code{save-excursion}
6796 uses the @code{beginning-of-line} function to move point.
6798 (Note that the @code{(widen)} expression comes between the
6799 @code{save-restriction} and @code{save-excursion} special forms. When
6800 you write the two @code{save- @dots{}} expressions in sequence, write
6801 @code{save-excursion} outermost.)
6804 The last two lines of the @code{what-line} function are functions to
6805 count the number of lines in the buffer and then print the number in the
6811 (1+ (count-lines 1 (point)))))))
6815 The @code{message} function prints a one-line message at the bottom of
6816 the Emacs screen. The first argument is inside of quotation marks and
6817 is printed as a string of characters. However, it may contain a
6818 @samp{%d} expression to print a following argument. @samp{%d} prints
6819 the argument as a decimal, so the message will say something such as
6823 The number that is printed in place of the @samp{%d} is computed by the
6824 last line of the function:
6827 (1+ (count-lines 1 (point)))
6833 (defun count-lines (start end)
6834 "Return number of lines between START and END.
6835 This is usually the number of newlines between them,
6836 but can be one more if START is not equal to END
6837 and the greater of them is not at the start of a line."
6840 (narrow-to-region start end)
6841 (goto-char (point-min))
6842 (if (eq selective-display t)
6845 (while (re-search-forward "[\n\C-m]" nil t 40)
6846 (setq done (+ 40 done)))
6847 (while (re-search-forward "[\n\C-m]" nil t 1)
6848 (setq done (+ 1 done)))
6849 (goto-char (point-max))
6850 (if (and (/= start end)
6854 (- (buffer-size) (forward-line (buffer-size)))))))
6858 What this does is count the lines from the first position of the
6859 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6860 one to that number. (The @code{1+} function adds one to its
6861 argument.) We add one to it because line 2 has only one line before
6862 it, and @code{count-lines} counts only the lines @emph{before} the
6865 After @code{count-lines} has done its job, and the message has been
6866 printed in the echo area, the @code{save-excursion} restores point and
6867 mark to their original positions; and @code{save-restriction} restores
6868 the original narrowing, if any.
6870 @node narrow Exercise
6871 @section Exercise with Narrowing
6873 Write a function that will display the first 60 characters of the
6874 current buffer, even if you have narrowed the buffer to its latter
6875 half so that the first line is inaccessible. Restore point, mark, and
6876 narrowing. For this exercise, you need to use a whole potpourri of
6877 functions, including @code{save-restriction}, @code{widen},
6878 @code{goto-char}, @code{point-min}, @code{message}, and
6879 @code{buffer-substring}.
6881 @cindex Properties, mention of @code{buffer-substring-no-properties}
6882 (@code{buffer-substring} is a previously unmentioned function you will
6883 have to investigate yourself; or perhaps you will have to use
6884 @code{buffer-substring-no-properties} or
6885 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6886 properties are a feature otherwise not discussed here. @xref{Text
6887 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6890 Additionally, do you really need @code{goto-char} or @code{point-min}?
6891 Or can you write the function without them?
6893 @node car cdr & cons
6894 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6895 @findex car, @r{introduced}
6896 @findex cdr, @r{introduced}
6898 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6899 functions. The @code{cons} function is used to construct lists, and
6900 the @code{car} and @code{cdr} functions are used to take them apart.
6902 In the walk through of the @code{copy-region-as-kill} function, we
6903 will see @code{cons} as well as two variants on @code{cdr},
6904 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6907 * Strange Names:: An historical aside: why the strange names?
6908 * car & cdr:: Functions for extracting part of a list.
6909 * cons:: Constructing a list.
6910 * nthcdr:: Calling @code{cdr} repeatedly.
6912 * setcar:: Changing the first element of a list.
6913 * setcdr:: Changing the rest of a list.
6919 @unnumberedsec Strange Names
6922 The name of the @code{cons} function is not unreasonable: it is an
6923 abbreviation of the word `construct'. The origins of the names for
6924 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6925 is an acronym from the phrase `Contents of the Address part of the
6926 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6927 the phrase `Contents of the Decrement part of the Register'. These
6928 phrases refer to specific pieces of hardware on the very early
6929 computer on which the original Lisp was developed. Besides being
6930 obsolete, the phrases have been completely irrelevant for more than 25
6931 years to anyone thinking about Lisp. Nonetheless, although a few
6932 brave scholars have begun to use more reasonable names for these
6933 functions, the old terms are still in use. In particular, since the
6934 terms are used in the Emacs Lisp source code, we will use them in this
6938 @section @code{car} and @code{cdr}
6940 The @sc{car} of a list is, quite simply, the first item in the list.
6941 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
6945 If you are reading this in Info in GNU Emacs, you can see this by
6946 evaluating the following:
6949 (car '(rose violet daisy buttercup))
6953 After evaluating the expression, @code{rose} will appear in the echo
6956 Clearly, a more reasonable name for the @code{car} function would be
6957 @code{first} and this is often suggested.
6959 @code{car} does not remove the first item from the list; it only reports
6960 what it is. After @code{car} has been applied to a list, the list is
6961 still the same as it was. In the jargon, @code{car} is
6962 `non-destructive'. This feature turns out to be important.
6964 The @sc{cdr} of a list is the rest of the list, that is, the
6965 @code{cdr} function returns the part of the list that follows the
6966 first item. Thus, while the @sc{car} of the list @code{'(rose violet
6967 daisy buttercup)} is @code{rose}, the rest of the list, the value
6968 returned by the @code{cdr} function, is @code{(violet daisy
6972 You can see this by evaluating the following in the usual way:
6975 (cdr '(rose violet daisy buttercup))
6979 When you evaluate this, @code{(violet daisy buttercup)} will appear in
6982 Like @code{car}, @code{cdr} does not remove any elements from the
6983 list---it just returns a report of what the second and subsequent
6986 Incidentally, in the example, the list of flowers is quoted. If it were
6987 not, the Lisp interpreter would try to evaluate the list by calling
6988 @code{rose} as a function. In this example, we do not want to do that.
6990 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
6992 (There is a lesson here: when you name new functions, consider very
6993 carefully what you are doing, since you may be stuck with the names
6994 for far longer than you expect. The reason this document perpetuates
6995 these names is that the Emacs Lisp source code uses them, and if I did
6996 not use them, you would have a hard time reading the code; but do,
6997 please, try to avoid using these terms yourself. The people who come
6998 after you will be grateful to you.)
7000 When @code{car} and @code{cdr} are applied to a list made up of symbols,
7001 such as the list @code{(pine fir oak maple)}, the element of the list
7002 returned by the function @code{car} is the symbol @code{pine} without
7003 any parentheses around it. @code{pine} is the first element in the
7004 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7005 oak maple)}, as you can see by evaluating the following expressions in
7010 (car '(pine fir oak maple))
7012 (cdr '(pine fir oak maple))
7016 On the other hand, in a list of lists, the first element is itself a
7017 list. @code{car} returns this first element as a list. For example,
7018 the following list contains three sub-lists, a list of carnivores, a
7019 list of herbivores and a list of sea mammals:
7023 (car '((lion tiger cheetah)
7024 (gazelle antelope zebra)
7025 (whale dolphin seal)))
7030 In this example, the first element or @sc{car} of the list is the list of
7031 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7032 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7036 (cdr '((lion tiger cheetah)
7037 (gazelle antelope zebra)
7038 (whale dolphin seal)))
7042 It is worth saying again that @code{car} and @code{cdr} are
7043 non-destructive---that is, they do not modify or change lists to which
7044 they are applied. This is very important for how they are used.
7046 Also, in the first chapter, in the discussion about atoms, I said that
7047 in Lisp, ``certain kinds of atom, such as an array, can be separated
7048 into parts; but the mechanism for doing this is different from the
7049 mechanism for splitting a list. As far as Lisp is concerned, the
7050 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7051 @code{car} and @code{cdr} functions are used for splitting lists and
7052 are considered fundamental to Lisp. Since they cannot split or gain
7053 access to the parts of an array, an array is considered an atom.
7054 Conversely, the other fundamental function, @code{cons}, can put
7055 together or construct a list, but not an array. (Arrays are handled
7056 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7057 Emacs Lisp Reference Manual}.)
7060 @section @code{cons}
7061 @findex cons, @r{introduced}
7063 The @code{cons} function constructs lists; it is the inverse of
7064 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7065 a four element list from the three element list, @code{(fir oak maple)}:
7068 (cons 'pine '(fir oak maple))
7073 After evaluating this list, you will see
7076 (pine fir oak maple)
7080 appear in the echo area. @code{cons} causes the creation of a new
7081 list in which the element is followed by the elements of the original
7084 We often say that `@code{cons} puts a new element at the beginning of
7085 a list; it attaches or pushes elements onto the list', but this
7086 phrasing can be misleading, since @code{cons} does not change an
7087 existing list, but creates a new one.
7089 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7093 * length:: How to find the length of a list.
7098 @unnumberedsubsec Build a list
7101 @code{cons} must have a list to attach to.@footnote{Actually, you can
7102 @code{cons} an element to an atom to produce a dotted pair. Dotted
7103 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7104 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7105 cannot start from absolutely nothing. If you are building a list, you
7106 need to provide at least an empty list at the beginning. Here is a
7107 series of @code{cons} expressions that build up a list of flowers. If
7108 you are reading this in Info in GNU Emacs, you can evaluate each of
7109 the expressions in the usual way; the value is printed in this text
7110 after @samp{@result{}}, which you may read as `evaluates to'.
7114 (cons 'buttercup ())
7115 @result{} (buttercup)
7119 (cons 'daisy '(buttercup))
7120 @result{} (daisy buttercup)
7124 (cons 'violet '(daisy buttercup))
7125 @result{} (violet daisy buttercup)
7129 (cons 'rose '(violet daisy buttercup))
7130 @result{} (rose violet daisy buttercup)
7135 In the first example, the empty list is shown as @code{()} and a list
7136 made up of @code{buttercup} followed by the empty list is constructed.
7137 As you can see, the empty list is not shown in the list that was
7138 constructed. All that you see is @code{(buttercup)}. The empty list is
7139 not counted as an element of a list because there is nothing in an empty
7140 list. Generally speaking, an empty list is invisible.
7142 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7143 two element list by putting @code{daisy} in front of @code{buttercup};
7144 and the third example constructs a three element list by putting
7145 @code{violet} in front of @code{daisy} and @code{buttercup}.
7148 @subsection Find the Length of a List: @code{length}
7151 You can find out how many elements there are in a list by using the Lisp
7152 function @code{length}, as in the following examples:
7156 (length '(buttercup))
7161 (length '(daisy buttercup))
7166 (length (cons 'violet '(daisy buttercup)))
7172 In the third example, the @code{cons} function is used to construct a
7173 three element list which is then passed to the @code{length} function as
7177 We can also use @code{length} to count the number of elements in an
7188 As you would expect, the number of elements in an empty list is zero.
7190 An interesting experiment is to find out what happens if you try to find
7191 the length of no list at all; that is, if you try to call @code{length}
7192 without giving it an argument, not even an empty list:
7200 What you see, if you evaluate this, is the error message
7203 Lisp error: (wrong-number-of-arguments length 0)
7207 This means that the function receives the wrong number of
7208 arguments, zero, when it expects some other number of arguments. In
7209 this case, one argument is expected, the argument being a list whose
7210 length the function is measuring. (Note that @emph{one} list is
7211 @emph{one} argument, even if the list has many elements inside it.)
7213 The part of the error message that says @samp{length} is the name of
7217 @code{length} is still a subroutine, but you need C-h f to discover that.
7219 In an earlier version:
7220 This is written with a special notation, @samp{#<subr},
7221 that indicates that the function @code{length} is one of the primitive
7222 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7223 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7224 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7229 @section @code{nthcdr}
7232 The @code{nthcdr} function is associated with the @code{cdr} function.
7233 What it does is take the @sc{cdr} of a list repeatedly.
7235 If you take the @sc{cdr} of the list @code{(pine fir
7236 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7237 repeat this on what was returned, you will be returned the list
7238 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7239 list will just give you the original @sc{cdr} since the function does
7240 not change the list. You need to evaluate the @sc{cdr} of the
7241 @sc{cdr} and so on.) If you continue this, eventually you will be
7242 returned an empty list, which in this case, instead of being shown as
7243 @code{()} is shown as @code{nil}.
7246 For review, here is a series of repeated @sc{cdr}s, the text following
7247 the @samp{@result{}} shows what is returned.
7251 (cdr '(pine fir oak maple))
7252 @result{}(fir oak maple)
7256 (cdr '(fir oak maple))
7257 @result{} (oak maple)
7282 You can also do several @sc{cdr}s without printing the values in
7287 (cdr (cdr '(pine fir oak maple)))
7288 @result{} (oak maple)
7293 In this example, the Lisp interpreter evaluates the innermost list first.
7294 The innermost list is quoted, so it just passes the list as it is to the
7295 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7296 second and subsequent elements of the list to the outermost @code{cdr},
7297 which produces a list composed of the third and subsequent elements of
7298 the original list. In this example, the @code{cdr} function is repeated
7299 and returns a list that consists of the original list without its
7302 The @code{nthcdr} function does the same as repeating the call to
7303 @code{cdr}. In the following example, the argument 2 is passed to the
7304 function @code{nthcdr}, along with the list, and the value returned is
7305 the list without its first two items, which is exactly the same
7306 as repeating @code{cdr} twice on the list:
7310 (nthcdr 2 '(pine fir oak maple))
7311 @result{} (oak maple)
7316 Using the original four element list, we can see what happens when
7317 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7322 ;; @r{Leave the list as it was.}
7323 (nthcdr 0 '(pine fir oak maple))
7324 @result{} (pine fir oak maple)
7328 ;; @r{Return a copy without the first element.}
7329 (nthcdr 1 '(pine fir oak maple))
7330 @result{} (fir oak maple)
7334 ;; @r{Return a copy of the list without three elements.}
7335 (nthcdr 3 '(pine fir oak maple))
7340 ;; @r{Return a copy lacking all four elements.}
7341 (nthcdr 4 '(pine fir oak maple))
7346 ;; @r{Return a copy lacking all elements.}
7347 (nthcdr 5 '(pine fir oak maple))
7356 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7357 The @code{nth} function takes the @sc{car} of the result returned by
7358 @code{nthcdr}. It returns the Nth element of the list.
7361 Thus, if it were not defined in C for speed, the definition of
7362 @code{nth} would be:
7367 "Returns the Nth element of LIST.
7368 N counts from zero. If LIST is not that long, nil is returned."
7369 (car (nthcdr n list)))
7374 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7375 but its definition was redone in C in the 1980s.)
7377 The @code{nth} function returns a single element of a list.
7378 This can be very convenient.
7380 Note that the elements are numbered from zero, not one. That is to
7381 say, the first element of a list, its @sc{car} is the zeroth element.
7382 This is called `zero-based' counting and often bothers people who
7383 are accustomed to the first element in a list being number one, which
7391 (nth 0 '("one" "two" "three"))
7394 (nth 1 '("one" "two" "three"))
7399 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7400 @code{cdr}, does not change the original list---the function is
7401 non-destructive. This is in sharp contrast to the @code{setcar} and
7402 @code{setcdr} functions.
7405 @section @code{setcar}
7408 As you might guess from their names, the @code{setcar} and @code{setcdr}
7409 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7410 They actually change the original list, unlike @code{car} and @code{cdr}
7411 which leave the original list as it was. One way to find out how this
7412 works is to experiment. We will start with the @code{setcar} function.
7415 First, we can make a list and then set the value of a variable to the
7416 list, using the @code{setq} function. Here is a list of animals:
7419 (setq animals '(antelope giraffe lion tiger))
7423 If you are reading this in Info inside of GNU Emacs, you can evaluate
7424 this expression in the usual fashion, by positioning the cursor after
7425 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7426 as I write this. This is one of the advantages of having the
7427 interpreter built into the computing environment. Incidentally, when
7428 there is nothing on the line after the final parentheses, such as a
7429 comment, point can be on the next line. Thus, if your cursor is in
7430 the first column of the next line, you do not need to move it.
7431 Indeed, Emacs permits any amount of white space after the final
7435 When we evaluate the variable @code{animals}, we see that it is bound to
7436 the list @code{(antelope giraffe lion tiger)}:
7441 @result{} (antelope giraffe lion tiger)
7446 Put another way, the variable @code{animals} points to the list
7447 @code{(antelope giraffe lion tiger)}.
7449 Next, evaluate the function @code{setcar} while passing it two
7450 arguments, the variable @code{animals} and the quoted symbol
7451 @code{hippopotamus}; this is done by writing the three element list
7452 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7456 (setcar animals 'hippopotamus)
7461 After evaluating this expression, evaluate the variable @code{animals}
7462 again. You will see that the list of animals has changed:
7467 @result{} (hippopotamus giraffe lion tiger)
7472 The first element on the list, @code{antelope} is replaced by
7473 @code{hippopotamus}.
7475 So we can see that @code{setcar} did not add a new element to the list
7476 as @code{cons} would have; it replaced @code{antelope} with
7477 @code{hippopotamus}; it @emph{changed} the list.
7480 @section @code{setcdr}
7483 The @code{setcdr} function is similar to the @code{setcar} function,
7484 except that the function replaces the second and subsequent elements of
7485 a list rather than the first element.
7487 (To see how to change the last element of a list, look ahead to
7488 @ref{kill-new function, , The @code{kill-new} function}, which uses
7489 the @code{nthcdr} and @code{setcdr} functions.)
7492 To see how this works, set the value of the variable to a list of
7493 domesticated animals by evaluating the following expression:
7496 (setq domesticated-animals '(horse cow sheep goat))
7501 If you now evaluate the list, you will be returned the list
7502 @code{(horse cow sheep goat)}:
7506 domesticated-animals
7507 @result{} (horse cow sheep goat)
7512 Next, evaluate @code{setcdr} with two arguments, the name of the
7513 variable which has a list as its value, and the list to which the
7514 @sc{cdr} of the first list will be set;
7517 (setcdr domesticated-animals '(cat dog))
7521 If you evaluate this expression, the list @code{(cat dog)} will appear
7522 in the echo area. This is the value returned by the function. The
7523 result we are interested in is the ``side effect'', which we can see by
7524 evaluating the variable @code{domesticated-animals}:
7528 domesticated-animals
7529 @result{} (horse cat dog)
7534 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7535 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7536 @code{(cow sheep goat)} to @code{(cat dog)}.
7541 Construct a list of four birds by evaluating several expressions with
7542 @code{cons}. Find out what happens when you @code{cons} a list onto
7543 itself. Replace the first element of the list of four birds with a
7544 fish. Replace the rest of that list with a list of other fish.
7546 @node Cutting & Storing Text
7547 @chapter Cutting and Storing Text
7548 @cindex Cutting and storing text
7549 @cindex Storing and cutting text
7550 @cindex Killing text
7551 @cindex Clipping text
7552 @cindex Erasing text
7553 @cindex Deleting text
7555 Whenever you cut or clip text out of a buffer with a `kill' command in
7556 GNU Emacs, it is stored in a list and you can bring it back with a
7559 (The use of the word `kill' in Emacs for processes which specifically
7560 @emph{do not} destroy the values of the entities is an unfortunate
7561 historical accident. A much more appropriate word would be `clip' since
7562 that is what the kill commands do; they clip text out of a buffer and
7563 put it into storage from which it can be brought back. I have often
7564 been tempted to replace globally all occurrences of `kill' in the Emacs
7565 sources with `clip' and all occurrences of `killed' with `clipped'.)
7568 * Storing Text:: Text is stored in a list.
7569 * zap-to-char:: Cutting out text up to a character.
7570 * kill-region:: Cutting text out of a region.
7571 * copy-region-as-kill:: A definition for copying text.
7572 * Digression into C:: Minor note on C programming language macros.
7573 * defvar:: How to give a variable an initial value.
7574 * cons & search-fwd Review::
7575 * search Exercises::
7580 @unnumberedsec Storing Text in a List
7583 When text is cut out of a buffer, it is stored on a list. Successive
7584 pieces of text are stored on the list successively, so the list might
7588 ("a piece of text" "previous piece")
7593 The function @code{cons} can be used to create a new list from a piece
7594 of text (an `atom', to use the jargon) and an existing list, like
7599 (cons "another piece"
7600 '("a piece of text" "previous piece"))
7606 If you evaluate this expression, a list of three elements will appear in
7610 ("another piece" "a piece of text" "previous piece")
7613 With the @code{car} and @code{nthcdr} functions, you can retrieve
7614 whichever piece of text you want. For example, in the following code,
7615 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7616 and the @code{car} returns the first element of that remainder---the
7617 second element of the original list:
7621 (car (nthcdr 1 '("another piece"
7624 @result{} "a piece of text"
7628 The actual functions in Emacs are more complex than this, of course.
7629 The code for cutting and retrieving text has to be written so that
7630 Emacs can figure out which element in the list you want---the first,
7631 second, third, or whatever. In addition, when you get to the end of
7632 the list, Emacs should give you the first element of the list, rather
7633 than nothing at all.
7635 The list that holds the pieces of text is called the @dfn{kill ring}.
7636 This chapter leads up to a description of the kill ring and how it is
7637 used by first tracing how the @code{zap-to-char} function works. This
7638 function uses (or `calls') a function that invokes a function that
7639 manipulates the kill ring. Thus, before reaching the mountains, we
7640 climb the foothills.
7642 A subsequent chapter describes how text that is cut from the buffer is
7643 retrieved. @xref{Yanking, , Yanking Text Back}.
7646 @section @code{zap-to-char}
7649 @c FIXME remove obsolete stuff
7650 The @code{zap-to-char} function changed little between GNU Emacs
7651 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7652 calls another function, @code{kill-region}, which enjoyed a major
7655 The @code{kill-region} function in Emacs 19 is complex, but does not
7656 use code that is important at this time. We will skip it.
7658 The @code{kill-region} function in Emacs 22 is easier to read than the
7659 same function in Emacs 19 and introduces a very important concept,
7660 that of error handling. We will walk through the function.
7662 But first, let us look at the interactive @code{zap-to-char} function.
7665 * Complete zap-to-char:: The complete implementation.
7666 * zap-to-char interactive:: A three part interactive expression.
7667 * zap-to-char body:: A short overview.
7668 * search-forward:: How to search for a string.
7669 * progn:: The @code{progn} special form.
7670 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7674 @node Complete zap-to-char
7675 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7678 The @code{zap-to-char} function removes the text in the region between
7679 the location of the cursor (i.e., of point) up to and including the
7680 next occurrence of a specified character. The text that
7681 @code{zap-to-char} removes is put in the kill ring; and it can be
7682 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7683 the command is given an argument, it removes text through that number
7684 of occurrences. Thus, if the cursor were at the beginning of this
7685 sentence and the character were @samp{s}, @samp{Thus} would be
7686 removed. If the argument were two, @samp{Thus, if the curs} would be
7687 removed, up to and including the @samp{s} in @samp{cursor}.
7689 If the specified character is not found, @code{zap-to-char} will say
7690 ``Search failed'', tell you the character you typed, and not remove
7693 In order to determine how much text to remove, @code{zap-to-char} uses
7694 a search function. Searches are used extensively in code that
7695 manipulates text, and we will focus attention on them as well as on the
7699 @c GNU Emacs version 19
7700 (defun zap-to-char (arg char) ; version 19 implementation
7701 "Kill up to and including ARG'th occurrence of CHAR.
7702 Goes backward if ARG is negative; error if CHAR not found."
7703 (interactive "*p\ncZap to char: ")
7704 (kill-region (point)
7707 (char-to-string char) nil nil arg)
7712 Here is the complete text of the version 22 implementation of the function:
7717 (defun zap-to-char (arg char)
7718 "Kill up to and including ARG'th occurrence of CHAR.
7719 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7720 Goes backward if ARG is negative; error if CHAR not found."
7721 (interactive "p\ncZap to char: ")
7722 (if (char-table-p translation-table-for-input)
7723 (setq char (or (aref translation-table-for-input char) char)))
7724 (kill-region (point) (progn
7725 (search-forward (char-to-string char)
7731 The documentation is thorough. You do need to know the jargon meaning
7734 @node zap-to-char interactive
7735 @subsection The @code{interactive} Expression
7738 The interactive expression in the @code{zap-to-char} command looks like
7742 (interactive "p\ncZap to char: ")
7745 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7746 two different things. First, and most simply, is the @samp{p}.
7747 This part is separated from the next part by a newline, @samp{\n}.
7748 The @samp{p} means that the first argument to the function will be
7749 passed the value of a `processed prefix'. The prefix argument is
7750 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7751 the function is called interactively without a prefix, 1 is passed to
7754 The second part of @code{"p\ncZap to char:@: "} is
7755 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7756 indicates that @code{interactive} expects a prompt and that the
7757 argument will be a character. The prompt follows the @samp{c} and is
7758 the string @samp{Zap to char:@: } (with a space after the colon to
7761 What all this does is prepare the arguments to @code{zap-to-char} so they
7762 are of the right type, and give the user a prompt.
7764 In a read-only buffer, the @code{zap-to-char} function copies the text
7765 to the kill ring, but does not remove it. The echo area displays a
7766 message saying that the buffer is read-only. Also, the terminal may
7767 beep or blink at you.
7769 @node zap-to-char body
7770 @subsection The Body of @code{zap-to-char}
7772 The body of the @code{zap-to-char} function contains the code that
7773 kills (that is, removes) the text in the region from the current
7774 position of the cursor up to and including the specified character.
7776 The first part of the code looks like this:
7779 (if (char-table-p translation-table-for-input)
7780 (setq char (or (aref translation-table-for-input char) char)))
7781 (kill-region (point) (progn
7782 (search-forward (char-to-string char) nil nil arg)
7787 @code{char-table-p} is an hitherto unseen function. It determines
7788 whether its argument is a character table. When it is, it sets the
7789 character passed to @code{zap-to-char} to one of them, if that
7790 character exists, or to the character itself. (This becomes important
7791 for certain characters in non-European languages. The @code{aref}
7792 function extracts an element from an array. It is an array-specific
7793 function that is not described in this document. @xref{Arrays, ,
7794 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7797 @code{(point)} is the current position of the cursor.
7799 The next part of the code is an expression using @code{progn}. The body
7800 of the @code{progn} consists of calls to @code{search-forward} and
7803 It is easier to understand how @code{progn} works after learning about
7804 @code{search-forward}, so we will look at @code{search-forward} and
7805 then at @code{progn}.
7807 @node search-forward
7808 @subsection The @code{search-forward} Function
7809 @findex search-forward
7811 The @code{search-forward} function is used to locate the
7812 zapped-for-character in @code{zap-to-char}. If the search is
7813 successful, @code{search-forward} leaves point immediately after the
7814 last character in the target string. (In @code{zap-to-char}, the
7815 target string is just one character long. @code{zap-to-char} uses the
7816 function @code{char-to-string} to ensure that the computer treats that
7817 character as a string.) If the search is backwards,
7818 @code{search-forward} leaves point just before the first character in
7819 the target. Also, @code{search-forward} returns @code{t} for true.
7820 (Moving point is therefore a `side effect'.)
7823 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7826 (search-forward (char-to-string char) nil nil arg)
7829 The @code{search-forward} function takes four arguments:
7833 The first argument is the target, what is searched for. This must be a
7834 string, such as @samp{"z"}.
7836 As it happens, the argument passed to @code{zap-to-char} is a single
7837 character. Because of the way computers are built, the Lisp
7838 interpreter may treat a single character as being different from a
7839 string of characters. Inside the computer, a single character has a
7840 different electronic format than a string of one character. (A single
7841 character can often be recorded in the computer using exactly one
7842 byte; but a string may be longer, and the computer needs to be ready
7843 for this.) Since the @code{search-forward} function searches for a
7844 string, the character that the @code{zap-to-char} function receives as
7845 its argument must be converted inside the computer from one format to
7846 the other; otherwise the @code{search-forward} function will fail.
7847 The @code{char-to-string} function is used to make this conversion.
7850 The second argument bounds the search; it is specified as a position in
7851 the buffer. In this case, the search can go to the end of the buffer,
7852 so no bound is set and the second argument is @code{nil}.
7855 The third argument tells the function what it should do if the search
7856 fails---it can signal an error (and print a message) or it can return
7857 @code{nil}. A @code{nil} as the third argument causes the function to
7858 signal an error when the search fails.
7861 The fourth argument to @code{search-forward} is the repeat count---how
7862 many occurrences of the string to look for. This argument is optional
7863 and if the function is called without a repeat count, this argument is
7864 passed the value 1. If this argument is negative, the search goes
7869 In template form, a @code{search-forward} expression looks like this:
7873 (search-forward "@var{target-string}"
7874 @var{limit-of-search}
7875 @var{what-to-do-if-search-fails}
7880 We will look at @code{progn} next.
7883 @subsection The @code{progn} Special Form
7886 @code{progn} is a special form that causes each of its arguments to be
7887 evaluated in sequence and then returns the value of the last one. The
7888 preceding expressions are evaluated only for the side effects they
7889 perform. The values produced by them are discarded.
7892 The template for a @code{progn} expression is very simple:
7901 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7902 put point in exactly the right position; and return the location of
7903 point so that @code{kill-region} will know how far to kill to.
7905 The first argument to the @code{progn} is @code{search-forward}. When
7906 @code{search-forward} finds the string, the function leaves point
7907 immediately after the last character in the target string. (In this
7908 case the target string is just one character long.) If the search is
7909 backwards, @code{search-forward} leaves point just before the first
7910 character in the target. The movement of point is a side effect.
7912 The second and last argument to @code{progn} is the expression
7913 @code{(point)}. This expression returns the value of point, which in
7914 this case will be the location to which it has been moved by
7915 @code{search-forward}. (In the source, a line that tells the function
7916 to go to the previous character, if it is going forward, was commented
7917 out in 1999; I don't remember whether that feature or mis-feature was
7918 ever a part of the distributed source.) The value of @code{point} is
7919 returned by the @code{progn} expression and is passed to
7920 @code{kill-region} as @code{kill-region}'s second argument.
7922 @node Summing up zap-to-char
7923 @subsection Summing up @code{zap-to-char}
7925 Now that we have seen how @code{search-forward} and @code{progn} work,
7926 we can see how the @code{zap-to-char} function works as a whole.
7928 The first argument to @code{kill-region} is the position of the cursor
7929 when the @code{zap-to-char} command is given---the value of point at
7930 that time. Within the @code{progn}, the search function then moves
7931 point to just after the zapped-to-character and @code{point} returns the
7932 value of this location. The @code{kill-region} function puts together
7933 these two values of point, the first one as the beginning of the region
7934 and the second one as the end of the region, and removes the region.
7936 The @code{progn} special form is necessary because the
7937 @code{kill-region} command takes two arguments; and it would fail if
7938 @code{search-forward} and @code{point} expressions were written in
7939 sequence as two additional arguments. The @code{progn} expression is
7940 a single argument to @code{kill-region} and returns the one value that
7941 @code{kill-region} needs for its second argument.
7944 @section @code{kill-region}
7947 The @code{zap-to-char} function uses the @code{kill-region} function.
7948 This function clips text from a region and copies that text to
7949 the kill ring, from which it may be retrieved.
7954 (defun kill-region (beg end &optional yank-handler)
7955 "Kill (\"cut\") text between point and mark.
7956 This deletes the text from the buffer and saves it in the kill ring.
7957 The command \\[yank] can retrieve it from there.
7958 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
7960 If you want to append the killed region to the last killed text,
7961 use \\[append-next-kill] before \\[kill-region].
7963 If the buffer is read-only, Emacs will beep and refrain from deleting
7964 the text, but put the text in the kill ring anyway. This means that
7965 you can use the killing commands to copy text from a read-only buffer.
7967 This is the primitive for programs to kill text (as opposed to deleting it).
7968 Supply two arguments, character positions indicating the stretch of text
7970 Any command that calls this function is a \"kill command\".
7971 If the previous command was also a kill command,
7972 the text killed this time appends to the text killed last time
7973 to make one entry in the kill ring.
7975 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
7976 specifies the yank-handler text property to be set on the killed
7977 text. See `insert-for-yank'."
7978 ;; Pass point first, then mark, because the order matters
7979 ;; when calling kill-append.
7980 (interactive (list (point) (mark)))
7981 (unless (and beg end)
7982 (error "The mark is not set now, so there is no region"))
7984 (let ((string (filter-buffer-substring beg end t)))
7985 (when string ;STRING is nil if BEG = END
7986 ;; Add that string to the kill ring, one way or another.
7987 (if (eq last-command 'kill-region)
7988 (kill-append string (< end beg) yank-handler)
7989 (kill-new string nil yank-handler)))
7990 (when (or string (eq last-command 'kill-region))
7991 (setq this-command 'kill-region))
7993 ((buffer-read-only text-read-only)
7994 ;; The code above failed because the buffer, or some of the characters
7995 ;; in the region, are read-only.
7996 ;; We should beep, in case the user just isn't aware of this.
7997 ;; However, there's no harm in putting
7998 ;; the region's text in the kill ring, anyway.
7999 (copy-region-as-kill beg end)
8000 ;; Set this-command now, so it will be set even if we get an error.
8001 (setq this-command 'kill-region)
8002 ;; This should barf, if appropriate, and give us the correct error.
8003 (if kill-read-only-ok
8004 (progn (message "Read only text copied to kill ring") nil)
8005 ;; Signal an error if the buffer is read-only.
8006 (barf-if-buffer-read-only)
8007 ;; If the buffer isn't read-only, the text is.
8008 (signal 'text-read-only (list (current-buffer)))))))
8011 The Emacs 22 version of that function uses @code{condition-case} and
8012 @code{copy-region-as-kill}, both of which we will explain.
8013 @code{condition-case} is an important special form.
8015 In essence, the @code{kill-region} function calls
8016 @code{condition-case}, which takes three arguments. In this function,
8017 the first argument does nothing. The second argument contains the
8018 code that does the work when all goes well. The third argument
8019 contains the code that is called in the event of an error.
8022 * Complete kill-region:: The function definition.
8023 * condition-case:: Dealing with a problem.
8028 @node Complete kill-region
8029 @unnumberedsubsec The Complete @code{kill-region} Definition
8033 We will go through the @code{condition-case} code in a moment. First,
8034 let us look at the definition of @code{kill-region}, with comments
8040 (defun kill-region (beg end)
8041 "Kill (\"cut\") text between point and mark.
8042 This deletes the text from the buffer and saves it in the kill ring.
8043 The command \\[yank] can retrieve it from there. @dots{} "
8047 ;; @bullet{} Since order matters, pass point first.
8048 (interactive (list (point) (mark)))
8049 ;; @bullet{} And tell us if we cannot cut the text.
8050 ;; `unless' is an `if' without a then-part.
8051 (unless (and beg end)
8052 (error "The mark is not set now, so there is no region"))
8056 ;; @bullet{} `condition-case' takes three arguments.
8057 ;; If the first argument is nil, as it is here,
8058 ;; information about the error signal is not
8059 ;; stored for use by another function.
8064 ;; @bullet{} The second argument to `condition-case' tells the
8065 ;; Lisp interpreter what to do when all goes well.
8069 ;; It starts with a `let' function that extracts the string
8070 ;; and tests whether it exists. If so (that is what the
8071 ;; `when' checks), it calls an `if' function that determines
8072 ;; whether the previous command was another call to
8073 ;; `kill-region'; if it was, then the new text is appended to
8074 ;; the previous text; if not, then a different function,
8075 ;; `kill-new', is called.
8079 ;; The `kill-append' function concatenates the new string and
8080 ;; the old. The `kill-new' function inserts text into a new
8081 ;; item in the kill ring.
8085 ;; `when' is an `if' without an else-part. The second `when'
8086 ;; again checks whether the current string exists; in
8087 ;; addition, it checks whether the previous command was
8088 ;; another call to `kill-region'. If one or the other
8089 ;; condition is true, then it sets the current command to
8090 ;; be `kill-region'.
8093 (let ((string (filter-buffer-substring beg end t)))
8094 (when string ;STRING is nil if BEG = END
8095 ;; Add that string to the kill ring, one way or another.
8096 (if (eq last-command 'kill-region)
8099 ;; @minus{} `yank-handler' is an optional argument to
8100 ;; `kill-region' that tells the `kill-append' and
8101 ;; `kill-new' functions how deal with properties
8102 ;; added to the text, such as `bold' or `italics'.
8103 (kill-append string (< end beg) yank-handler)
8104 (kill-new string nil yank-handler)))
8105 (when (or string (eq last-command 'kill-region))
8106 (setq this-command 'kill-region))
8111 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8112 ;; what to do with an error.
8115 ;; The third argument has a conditions part and a body part.
8116 ;; If the conditions are met (in this case,
8117 ;; if text or buffer are read-only)
8118 ;; then the body is executed.
8121 ;; The first part of the third argument is the following:
8122 ((buffer-read-only text-read-only) ;; the if-part
8123 ;; @dots{} the then-part
8124 (copy-region-as-kill beg end)
8127 ;; Next, also as part of the then-part, set this-command, so
8128 ;; it will be set in an error
8129 (setq this-command 'kill-region)
8130 ;; Finally, in the then-part, send a message if you may copy
8131 ;; the text to the kill ring without signaling an error, but
8132 ;; don't if you may not.
8135 (if kill-read-only-ok
8136 (progn (message "Read only text copied to kill ring") nil)
8137 (barf-if-buffer-read-only)
8138 ;; If the buffer isn't read-only, the text is.
8139 (signal 'text-read-only (list (current-buffer)))))
8147 (defun kill-region (beg end)
8148 "Kill between point and mark.
8149 The text is deleted but saved in the kill ring."
8154 ;; 1. `condition-case' takes three arguments.
8155 ;; If the first argument is nil, as it is here,
8156 ;; information about the error signal is not
8157 ;; stored for use by another function.
8162 ;; 2. The second argument to `condition-case'
8163 ;; tells the Lisp interpreter what to do when all goes well.
8167 ;; The `delete-and-extract-region' function usually does the
8168 ;; work. If the beginning and ending of the region are both
8169 ;; the same, then the variable `string' will be empty, or nil
8170 (let ((string (delete-and-extract-region beg end)))
8174 ;; `when' is an `if' clause that cannot take an `else-part'.
8175 ;; Emacs normally sets the value of `last-command' to the
8176 ;; previous command.
8179 ;; `kill-append' concatenates the new string and the old.
8180 ;; `kill-new' inserts text into a new item in the kill ring.
8182 (if (eq last-command 'kill-region)
8183 ;; if true, prepend string
8184 (kill-append string (< end beg))
8186 (setq this-command 'kill-region))
8190 ;; 3. The third argument to `condition-case' tells the interpreter
8191 ;; what to do with an error.
8194 ;; The third argument has a conditions part and a body part.
8195 ;; If the conditions are met (in this case,
8196 ;; if text or buffer are read-only)
8197 ;; then the body is executed.
8200 ((buffer-read-only text-read-only) ;; this is the if-part
8202 (copy-region-as-kill beg end)
8205 (if kill-read-only-ok ;; usually this variable is nil
8206 (message "Read only text copied to kill ring")
8207 ;; or else, signal an error if the buffer is read-only;
8208 (barf-if-buffer-read-only)
8209 ;; and, in any case, signal that the text is read-only.
8210 (signal 'text-read-only (list (current-buffer)))))))
8215 @node condition-case
8216 @subsection @code{condition-case}
8217 @findex condition-case
8219 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8220 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8221 expression, it provides you with help; in the jargon, this is called
8222 ``signaling an error''. Usually, the computer stops the program and
8223 shows you a message.
8225 However, some programs undertake complicated actions. They should not
8226 simply stop on an error. In the @code{kill-region} function, the most
8227 likely error is that you will try to kill text that is read-only and
8228 cannot be removed. So the @code{kill-region} function contains code
8229 to handle this circumstance. This code, which makes up the body of
8230 the @code{kill-region} function, is inside of a @code{condition-case}
8234 The template for @code{condition-case} looks like this:
8241 @var{error-handler}@dots{})
8245 The second argument, @var{bodyform}, is straightforward. The
8246 @code{condition-case} special form causes the Lisp interpreter to
8247 evaluate the code in @var{bodyform}. If no error occurs, the special
8248 form returns the code's value and produces the side-effects, if any.
8250 In short, the @var{bodyform} part of a @code{condition-case}
8251 expression determines what should happen when everything works
8254 However, if an error occurs, among its other actions, the function
8255 generating the error signal will define one or more error condition
8258 An error handler is the third argument to @code{condition case}.
8259 An error handler has two parts, a @var{condition-name} and a
8260 @var{body}. If the @var{condition-name} part of an error handler
8261 matches a condition name generated by an error, then the @var{body}
8262 part of the error handler is run.
8264 As you will expect, the @var{condition-name} part of an error handler
8265 may be either a single condition name or a list of condition names.
8267 Also, a complete @code{condition-case} expression may contain more
8268 than one error handler. When an error occurs, the first applicable
8271 Lastly, the first argument to the @code{condition-case} expression,
8272 the @var{var} argument, is sometimes bound to a variable that
8273 contains information about the error. However, if that argument is
8274 nil, as is the case in @code{kill-region}, that information is
8278 In brief, in the @code{kill-region} function, the code
8279 @code{condition-case} works like this:
8283 @var{If no errors}, @var{run only this code}
8284 @var{but}, @var{if errors}, @var{run this other code}.
8291 copy-region-as-kill is short, 12 lines, and uses
8292 filter-buffer-substring, which is longer, 39 lines
8293 and has delete-and-extract-region in it.
8294 delete-and-extract-region is written in C.
8296 see Initializing a Variable with @code{defvar}
8298 Initializing a Variable with @code{defvar} includes line 8350
8302 @subsection Lisp macro
8306 The part of the @code{condition-case} expression that is evaluated in
8307 the expectation that all goes well has a @code{when}. The code uses
8308 @code{when} to determine whether the @code{string} variable points to
8311 A @code{when} expression is simply a programmers' convenience. It is
8312 an @code{if} without the possibility of an else clause. In your mind,
8313 you can replace @code{when} with @code{if} and understand what goes
8314 on. That is what the Lisp interpreter does.
8316 Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro}
8317 enables you to define new control constructs and other language
8318 features. It tells the interpreter how to compute another Lisp
8319 expression which will in turn compute the value. In this case, the
8320 `other expression' is an @code{if} expression.
8322 The @code{kill-region} function definition also has an @code{unless}
8323 macro; it is the converse of @code{when}. The @code{unless} macro is
8324 an @code{if} without a then clause
8326 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8327 Emacs Lisp Reference Manual}. The C programming language also
8328 provides macros. These are different, but also useful.
8331 We will briefly look at C macros in
8332 @ref{Digression into C}.
8336 Regarding the @code{when} macro, in the @code{condition-case}
8337 expression, when the string has content, then another conditional
8338 expression is executed. This is an @code{if} with both a then-part
8343 (if (eq last-command 'kill-region)
8344 (kill-append string (< end beg) yank-handler)
8345 (kill-new string nil yank-handler))
8349 The then-part is evaluated if the previous command was another call to
8350 @code{kill-region}; if not, the else-part is evaluated.
8352 @code{yank-handler} is an optional argument to @code{kill-region} that
8353 tells the @code{kill-append} and @code{kill-new} functions how deal
8354 with properties added to the text, such as `bold' or `italics'.
8356 @code{last-command} is a variable that comes with Emacs that we have
8357 not seen before. Normally, whenever a function is executed, Emacs
8358 sets the value of @code{last-command} to the previous command.
8361 In this segment of the definition, the @code{if} expression checks
8362 whether the previous command was @code{kill-region}. If it was,
8365 (kill-append string (< end beg) yank-handler)
8369 concatenates a copy of the newly clipped text to the just previously
8370 clipped text in the kill ring.
8372 @node copy-region-as-kill
8373 @section @code{copy-region-as-kill}
8374 @findex copy-region-as-kill
8377 The @code{copy-region-as-kill} function copies a region of text from a
8378 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8379 in the @code{kill-ring}.
8381 If you call @code{copy-region-as-kill} immediately after a
8382 @code{kill-region} command, Emacs appends the newly copied text to the
8383 previously copied text. This means that if you yank back the text, you
8384 get it all, from both this and the previous operation. On the other
8385 hand, if some other command precedes the @code{copy-region-as-kill},
8386 the function copies the text into a separate entry in the kill ring.
8389 * Complete copy-region-as-kill:: The complete function definition.
8390 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8394 @node Complete copy-region-as-kill
8395 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8399 Here is the complete text of the version 22 @code{copy-region-as-kill}
8404 (defun copy-region-as-kill (beg end)
8405 "Save the region as if killed, but don't kill it.
8406 In Transient Mark mode, deactivate the mark.
8407 If `interprogram-cut-function' is non-nil, also save the text for a window
8408 system cut and paste."
8412 (if (eq last-command 'kill-region)
8413 (kill-append (filter-buffer-substring beg end) (< end beg))
8414 (kill-new (filter-buffer-substring beg end)))
8417 (if transient-mark-mode
8418 (setq deactivate-mark t))
8424 As usual, this function can be divided into its component parts:
8428 (defun copy-region-as-kill (@var{argument-list})
8429 "@var{documentation}@dots{}"
8435 The arguments are @code{beg} and @code{end} and the function is
8436 interactive with @code{"r"}, so the two arguments must refer to the
8437 beginning and end of the region. If you have been reading though this
8438 document from the beginning, understanding these parts of a function is
8439 almost becoming routine.
8441 The documentation is somewhat confusing unless you remember that the
8442 word `kill' has a meaning different from usual. The `Transient Mark'
8443 and @code{interprogram-cut-function} comments explain certain
8446 After you once set a mark, a buffer always contains a region. If you
8447 wish, you can use Transient Mark mode to highlight the region
8448 temporarily. (No one wants to highlight the region all the time, so
8449 Transient Mark mode highlights it only at appropriate times. Many
8450 people turn off Transient Mark mode, so the region is never
8453 Also, a windowing system allows you to copy, cut, and paste among
8454 different programs. In the X windowing system, for example, the
8455 @code{interprogram-cut-function} function is @code{x-select-text},
8456 which works with the windowing system's equivalent of the Emacs kill
8459 The body of the @code{copy-region-as-kill} function starts with an
8460 @code{if} clause. What this clause does is distinguish between two
8461 different situations: whether or not this command is executed
8462 immediately after a previous @code{kill-region} command. In the first
8463 case, the new region is appended to the previously copied text.
8464 Otherwise, it is inserted into the beginning of the kill ring as a
8465 separate piece of text from the previous piece.
8467 The last two lines of the function prevent the region from lighting up
8468 if Transient Mark mode is turned on.
8470 The body of @code{copy-region-as-kill} merits discussion in detail.
8472 @node copy-region-as-kill body
8473 @subsection The Body of @code{copy-region-as-kill}
8475 The @code{copy-region-as-kill} function works in much the same way as
8476 the @code{kill-region} function. Both are written so that two or more
8477 kills in a row combine their text into a single entry. If you yank
8478 back the text from the kill ring, you get it all in one piece.
8479 Moreover, kills that kill forward from the current position of the
8480 cursor are added to the end of the previously copied text and commands
8481 that copy text backwards add it to the beginning of the previously
8482 copied text. This way, the words in the text stay in the proper
8485 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8486 use of the @code{last-command} variable that keeps track of the
8487 previous Emacs command.
8490 * last-command & this-command::
8491 * kill-append function::
8492 * kill-new function::
8496 @node last-command & this-command
8497 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8500 Normally, whenever a function is executed, Emacs sets the value of
8501 @code{this-command} to the function being executed (which in this case
8502 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8503 the value of @code{last-command} to the previous value of
8504 @code{this-command}.
8506 In the first part of the body of the @code{copy-region-as-kill}
8507 function, an @code{if} expression determines whether the value of
8508 @code{last-command} is @code{kill-region}. If so, the then-part of
8509 the @code{if} expression is evaluated; it uses the @code{kill-append}
8510 function to concatenate the text copied at this call to the function
8511 with the text already in the first element (the @sc{car}) of the kill
8512 ring. On the other hand, if the value of @code{last-command} is not
8513 @code{kill-region}, then the @code{copy-region-as-kill} function
8514 attaches a new element to the kill ring using the @code{kill-new}
8518 The @code{if} expression reads as follows; it uses @code{eq}:
8522 (if (eq last-command 'kill-region)
8524 (kill-append (filter-buffer-substring beg end) (< end beg))
8526 (kill-new (filter-buffer-substring beg end)))
8530 @findex filter-buffer-substring
8531 (The @code{filter-buffer-substring} function returns a filtered
8532 substring of the buffer, if any. Optionally---the arguments are not
8533 here, so neither is done---the function may delete the initial text or
8534 return the text without its properties; this function is a replacement
8535 for the older @code{buffer-substring} function, which came before text
8536 properties were implemented.)
8538 @findex eq @r{(example of use)}
8540 The @code{eq} function tests whether its first argument is the same Lisp
8541 object as its second argument. The @code{eq} function is similar to the
8542 @code{equal} function in that it is used to test for equality, but
8543 differs in that it determines whether two representations are actually
8544 the same object inside the computer, but with different names.
8545 @code{equal} determines whether the structure and contents of two
8546 expressions are the same.
8548 If the previous command was @code{kill-region}, then the Emacs Lisp
8549 interpreter calls the @code{kill-append} function
8551 @node kill-append function
8552 @unnumberedsubsubsec The @code{kill-append} function
8556 The @code{kill-append} function looks like this:
8561 (defun kill-append (string before-p &optional yank-handler)
8562 "Append STRING to the end of the latest kill in the kill ring.
8563 If BEFORE-P is non-nil, prepend STRING to the kill.
8565 (let* ((cur (car kill-ring)))
8566 (kill-new (if before-p (concat string cur) (concat cur string))
8567 (or (= (length cur) 0)
8569 (get-text-property 0 'yank-handler cur)))
8576 (defun kill-append (string before-p)
8577 "Append STRING to the end of the latest kill in the kill ring.
8578 If BEFORE-P is non-nil, prepend STRING to the kill.
8579 If `interprogram-cut-function' is set, pass the resulting kill to
8581 (kill-new (if before-p
8582 (concat string (car kill-ring))
8583 (concat (car kill-ring) string))
8588 The @code{kill-append} function is fairly straightforward. It uses
8589 the @code{kill-new} function, which we will discuss in more detail in
8592 (Also, the function provides an optional argument called
8593 @code{yank-handler}; when invoked, this argument tells the function
8594 how to deal with properties added to the text, such as `bold' or
8597 @c !!! bug in GNU Emacs 22 version of kill-append ?
8598 It has a @code{let*} function to set the value of the first element of
8599 the kill ring to @code{cur}. (I do not know why the function does not
8600 use @code{let} instead; only one value is set in the expression.
8601 Perhaps this is a bug that produces no problems?)
8603 Consider the conditional that is one of the two arguments to
8604 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8605 the @sc{car} of the kill ring. Whether it prepends or appends the
8606 text depends on the results of an @code{if} expression:
8610 (if before-p ; @r{if-part}
8611 (concat string cur) ; @r{then-part}
8612 (concat cur string)) ; @r{else-part}
8617 If the region being killed is before the region that was killed in the
8618 last command, then it should be prepended before the material that was
8619 saved in the previous kill; and conversely, if the killed text follows
8620 what was just killed, it should be appended after the previous text.
8621 The @code{if} expression depends on the predicate @code{before-p} to
8622 decide whether the newly saved text should be put before or after the
8623 previously saved text.
8625 The symbol @code{before-p} is the name of one of the arguments to
8626 @code{kill-append}. When the @code{kill-append} function is
8627 evaluated, it is bound to the value returned by evaluating the actual
8628 argument. In this case, this is the expression @code{(< end beg)}.
8629 This expression does not directly determine whether the killed text in
8630 this command is located before or after the kill text of the last
8631 command; what it does is determine whether the value of the variable
8632 @code{end} is less than the value of the variable @code{beg}. If it
8633 is, it means that the user is most likely heading towards the
8634 beginning of the buffer. Also, the result of evaluating the predicate
8635 expression, @code{(< end beg)}, will be true and the text will be
8636 prepended before the previous text. On the other hand, if the value of
8637 the variable @code{end} is greater than the value of the variable
8638 @code{beg}, the text will be appended after the previous text.
8641 When the newly saved text will be prepended, then the string with the new
8642 text will be concatenated before the old text:
8650 But if the text will be appended, it will be concatenated
8654 (concat cur string))
8657 To understand how this works, we first need to review the
8658 @code{concat} function. The @code{concat} function links together or
8659 unites two strings of text. The result is a string. For example:
8663 (concat "abc" "def")
8669 (car '("first element" "second element")))
8670 @result{} "new first element"
8673 '("first element" "second element")) " modified")
8674 @result{} "first element modified"
8678 We can now make sense of @code{kill-append}: it modifies the contents
8679 of the kill ring. The kill ring is a list, each element of which is
8680 saved text. The @code{kill-append} function uses the @code{kill-new}
8681 function which in turn uses the @code{setcar} function.
8683 @node kill-new function
8684 @unnumberedsubsubsec The @code{kill-new} function
8687 @c in GNU Emacs 22, additional documentation to kill-new:
8689 Optional third arguments YANK-HANDLER controls how the STRING is later
8690 inserted into a buffer; see `insert-for-yank' for details.
8691 When a yank handler is specified, STRING must be non-empty (the yank
8692 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8694 When the yank handler has a non-nil PARAM element, the original STRING
8695 argument is not used by `insert-for-yank'. However, since Lisp code
8696 may access and use elements from the kill ring directly, the STRING
8697 argument should still be a \"useful\" string for such uses."
8700 The @code{kill-new} function looks like this:
8704 (defun kill-new (string &optional replace yank-handler)
8705 "Make STRING the latest kill in the kill ring.
8706 Set `kill-ring-yank-pointer' to point to it.
8708 If `interprogram-cut-function' is non-nil, apply it to STRING.
8709 Optional second argument REPLACE non-nil means that STRING will replace
8710 the front of the kill ring, rather than being added to the list.
8714 (if (> (length string) 0)
8716 (put-text-property 0 (length string)
8717 'yank-handler yank-handler string))
8719 (signal 'args-out-of-range
8720 (list string "yank-handler specified for empty string"))))
8723 (if (fboundp 'menu-bar-update-yank-menu)
8724 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8727 (if (and replace kill-ring)
8728 (setcar kill-ring string)
8729 (push string kill-ring)
8730 (if (> (length kill-ring) kill-ring-max)
8731 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8734 (setq kill-ring-yank-pointer kill-ring)
8735 (if interprogram-cut-function
8736 (funcall interprogram-cut-function string (not replace))))
8741 (defun kill-new (string &optional replace)
8742 "Make STRING the latest kill in the kill ring.
8743 Set the kill-ring-yank pointer to point to it.
8744 If `interprogram-cut-function' is non-nil, apply it to STRING.
8745 Optional second argument REPLACE non-nil means that STRING will replace
8746 the front of the kill ring, rather than being added to the list."
8747 (and (fboundp 'menu-bar-update-yank-menu)
8748 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8749 (if (and replace kill-ring)
8750 (setcar kill-ring string)
8751 (setq kill-ring (cons string kill-ring))
8752 (if (> (length kill-ring) kill-ring-max)
8753 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8754 (setq kill-ring-yank-pointer kill-ring)
8755 (if interprogram-cut-function
8756 (funcall interprogram-cut-function string (not replace))))
8759 (Notice that the function is not interactive.)
8761 As usual, we can look at this function in parts.
8763 The function definition has an optional @code{yank-handler} argument,
8764 which when invoked tells the function how to deal with properties
8765 added to the text, such as `bold' or `italics'. We will skip that.
8768 The first line of the documentation makes sense:
8771 Make STRING the latest kill in the kill ring.
8775 Let's skip over the rest of the documentation for the moment.
8778 Also, let's skip over the initial @code{if} expression and those lines
8779 of code involving @code{menu-bar-update-yank-menu}. We will explain
8783 The critical lines are these:
8787 (if (and replace kill-ring)
8789 (setcar kill-ring string)
8793 (push string kill-ring)
8796 (setq kill-ring (cons string kill-ring))
8797 (if (> (length kill-ring) kill-ring-max)
8798 ;; @r{avoid overly long kill ring}
8799 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8802 (setq kill-ring-yank-pointer kill-ring)
8803 (if interprogram-cut-function
8804 (funcall interprogram-cut-function string (not replace))))
8808 The conditional test is @w{@code{(and replace kill-ring)}}.
8809 This will be true when two conditions are met: the kill ring has
8810 something in it, and the @code{replace} variable is true.
8813 When the @code{kill-append} function sets @code{replace} to be true
8814 and when the kill ring has at least one item in it, the @code{setcar}
8815 expression is executed:
8818 (setcar kill-ring string)
8821 The @code{setcar} function actually changes the first element of the
8822 @code{kill-ring} list to the value of @code{string}. It replaces the
8826 On the other hand, if the kill ring is empty, or replace is false, the
8827 else-part of the condition is executed:
8830 (push string kill-ring)
8835 @code{push} puts its first argument onto the second. It is similar to
8839 (setq kill-ring (cons string kill-ring))
8847 (add-to-list kill-ring string)
8851 When it is false, the expression first constructs a new version of the
8852 kill ring by prepending @code{string} to the existing kill ring as a
8853 new element (that is what the @code{push} does). Then it executes a
8854 second @code{if} clause. This second @code{if} clause keeps the kill
8855 ring from growing too long.
8857 Let's look at these two expressions in order.
8859 The @code{push} line of the else-part sets the new value of the kill
8860 ring to what results from adding the string being killed to the old
8863 We can see how this works with an example.
8869 (setq example-list '("here is a clause" "another clause"))
8874 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8875 @code{example-list} and see what it returns:
8880 @result{} ("here is a clause" "another clause")
8886 Now, we can add a new element on to this list by evaluating the
8887 following expression:
8888 @findex push, @r{example}
8891 (push "a third clause" example-list)
8896 When we evaluate @code{example-list}, we find its value is:
8901 @result{} ("a third clause" "here is a clause" "another clause")
8906 Thus, the third clause is added to the list by @code{push}.
8909 Now for the second part of the @code{if} clause. This expression
8910 keeps the kill ring from growing too long. It looks like this:
8914 (if (> (length kill-ring) kill-ring-max)
8915 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8919 The code checks whether the length of the kill ring is greater than
8920 the maximum permitted length. This is the value of
8921 @code{kill-ring-max} (which is 60, by default). If the length of the
8922 kill ring is too long, then this code sets the last element of the
8923 kill ring to @code{nil}. It does this by using two functions,
8924 @code{nthcdr} and @code{setcdr}.
8926 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
8927 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
8928 @sc{car} of a list. In this case, however, @code{setcdr} will not be
8929 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
8930 function is used to cause it to set the @sc{cdr} of the next to last
8931 element of the kill ring---this means that since the @sc{cdr} of the
8932 next to last element is the last element of the kill ring, it will set
8933 the last element of the kill ring.
8935 @findex nthcdr, @r{example}
8936 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
8937 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
8938 @dots{} It does this @var{N} times and returns the results.
8939 (@xref{nthcdr, , @code{nthcdr}}.)
8941 @findex setcdr, @r{example}
8942 Thus, if we had a four element list that was supposed to be three
8943 elements long, we could set the @sc{cdr} of the next to last element
8944 to @code{nil}, and thereby shorten the list. (If you set the last
8945 element to some other value than @code{nil}, which you could do, then
8946 you would not have shortened the list. @xref{setcdr, ,
8949 You can see shortening by evaluating the following three expressions
8950 in turn. First set the value of @code{trees} to @code{(maple oak pine
8951 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
8952 and then find the value of @code{trees}:
8956 (setq trees '(maple oak pine birch))
8957 @result{} (maple oak pine birch)
8961 (setcdr (nthcdr 2 trees) nil)
8965 @result{} (maple oak pine)
8970 (The value returned by the @code{setcdr} expression is @code{nil} since
8971 that is what the @sc{cdr} is set to.)
8973 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
8974 @sc{cdr} a number of times that is one less than the maximum permitted
8975 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
8976 element (which will be the rest of the elements in the kill ring) to
8977 @code{nil}. This prevents the kill ring from growing too long.
8980 The next to last expression in the @code{kill-new} function is
8983 (setq kill-ring-yank-pointer kill-ring)
8986 The @code{kill-ring-yank-pointer} is a global variable that is set to be
8987 the @code{kill-ring}.
8989 Even though the @code{kill-ring-yank-pointer} is called a
8990 @samp{pointer}, it is a variable just like the kill ring. However, the
8991 name has been chosen to help humans understand how the variable is used.
8994 Now, to return to an early expression in the body of the function:
8998 (if (fboundp 'menu-bar-update-yank-menu)
8999 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9004 It starts with an @code{if} expression
9006 In this case, the expression tests first to see whether
9007 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9008 calls it. The @code{fboundp} function returns true if the symbol it
9009 is testing has a function definition that `is not void'. If the
9010 symbol's function definition were void, we would receive an error
9011 message, as we did when we created errors intentionally (@pxref{Making
9012 Errors, , Generate an Error Message}).
9015 The then-part contains an expression whose first element is the
9016 function @code{and}.
9019 The @code{and} special form evaluates each of its arguments until one
9020 of the arguments returns a value of @code{nil}, in which case the
9021 @code{and} expression returns @code{nil}; however, if none of the
9022 arguments returns a value of @code{nil}, the value resulting from
9023 evaluating the last argument is returned. (Since such a value is not
9024 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9025 @code{and} expression returns a true value only if all its arguments
9026 are true. (@xref{Second Buffer Related Review}.)
9028 The expression determines whether the second argument to
9029 @code{menu-bar-update-yank-menu} is true or not.
9031 ;; If we're supposed to be extending an existing string, and that
9032 ;; string really is at the front of the menu, then update it in place.
9035 @code{menu-bar-update-yank-menu} is one of the functions that make it
9036 possible to use the `Select and Paste' menu in the Edit item of a menu
9037 bar; using a mouse, you can look at the various pieces of text you
9038 have saved and select one piece to paste.
9040 The last expression in the @code{kill-new} function adds the newly
9041 copied string to whatever facility exists for copying and pasting
9042 among different programs running in a windowing system. In the X
9043 Windowing system, for example, the @code{x-select-text} function takes
9044 the string and stores it in memory operated by X@. You can paste the
9045 string in another program, such as an Xterm.
9048 The expression looks like this:
9052 (if interprogram-cut-function
9053 (funcall interprogram-cut-function string (not replace))))
9057 If an @code{interprogram-cut-function} exists, then Emacs executes
9058 @code{funcall}, which in turn calls its first argument as a function
9059 and passes the remaining arguments to it. (Incidentally, as far as I
9060 can see, this @code{if} expression could be replaced by an @code{and}
9061 expression similar to the one in the first part of the function.)
9063 We are not going to discuss windowing systems and other programs
9064 further, but merely note that this is a mechanism that enables GNU
9065 Emacs to work easily and well with other programs.
9067 This code for placing text in the kill ring, either concatenated with
9068 an existing element or as a new element, leads us to the code for
9069 bringing back text that has been cut out of the buffer---the yank
9070 commands. However, before discussing the yank commands, it is better
9071 to learn how lists are implemented in a computer. This will make
9072 clear such mysteries as the use of the term `pointer'. But before
9073 that, we will digress into C.
9076 @c is this true in Emacs 22? Does not seems to be
9078 (If the @w{@code{(< end beg))}}
9079 expression is true, @code{kill-append} prepends the string to the just
9080 previously clipped text. For a detailed discussion, see
9081 @ref{kill-append function, , The @code{kill-append} function}.)
9083 If you then yank back the text, i.e., `paste' it, you get both
9084 pieces of text at once. That way, if you delete two words in a row,
9085 and then yank them back, you get both words, in their proper order,
9086 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9089 On the other hand, if the previous command is not @code{kill-region},
9090 then the @code{kill-new} function is called, which adds the text to
9091 the kill ring as the latest item, and sets the
9092 @code{kill-ring-yank-pointer} variable to point to it.
9096 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9097 @c use the delete-and-extract-region function
9099 2006 Oct 26, the Digression into C is now OK but should come after
9100 copy-region-as-kill and filter-buffer-substring
9104 copy-region-as-kill is short, 12 lines, and uses
9105 filter-buffer-substring, which is longer, 39 lines
9106 and has delete-and-extract-region in it.
9107 delete-and-extract-region is written in C.
9109 see Initializing a Variable with @code{defvar}
9112 @node Digression into C
9113 @section Digression into C
9114 @findex delete-and-extract-region
9115 @cindex C, a digression into
9116 @cindex Digression into C
9118 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9119 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9120 function, which in turn uses the @code{delete-and-extract-region}
9121 function. It removes the contents of a region and you cannot get them
9124 Unlike the other code discussed here, the
9125 @code{delete-and-extract-region} function is not written in Emacs
9126 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9127 system. Since it is very simple, I will digress briefly from Lisp and
9130 @c GNU Emacs 24 in src/editfns.c
9131 @c the DEFUN for delete-and-extract-region
9134 Like many of the other Emacs primitives,
9135 @code{delete-and-extract-region} is written as an instance of a C
9136 macro, a macro being a template for code. The complete macro looks
9141 DEFUN ("delete-and-extract-region", Fdelete_and_extract_region,
9142 Sdelete_and_extract_region, 2, 2, 0,
9143 doc: /* Delete the text between START and END and return it. */)
9144 (Lisp_Object start, Lisp_Object end)
9146 validate_region (&start, &end);
9147 if (XINT (start) == XINT (end))
9148 return empty_unibyte_string;
9149 return del_range_1 (XINT (start), XINT (end), 1, 1);
9154 Without going into the details of the macro writing process, let me
9155 point out that this macro starts with the word @code{DEFUN}. The word
9156 @code{DEFUN} was chosen since the code serves the same purpose as
9157 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9158 @file{emacs/src/lisp.h}.)
9160 The word @code{DEFUN} is followed by seven parts inside of
9165 The first part is the name given to the function in Lisp,
9166 @code{delete-and-extract-region}.
9169 The second part is the name of the function in C,
9170 @code{Fdelete_and_extract_region}. By convention, it starts with
9171 @samp{F}. Since C does not use hyphens in names, underscores are used
9175 The third part is the name for the C constant structure that records
9176 information on this function for internal use. It is the name of the
9177 function in C but begins with an @samp{S} instead of an @samp{F}.
9180 The fourth and fifth parts specify the minimum and maximum number of
9181 arguments the function can have. This function demands exactly 2
9185 The sixth part is nearly like the argument that follows the
9186 @code{interactive} declaration in a function written in Lisp: a letter
9187 followed, perhaps, by a prompt. The only difference from the Lisp is
9188 when the macro is called with no arguments. Then you write a @code{0}
9189 (which is a `null string'), as in this macro.
9191 If you were to specify arguments, you would place them between
9192 quotation marks. The C macro for @code{goto-char} includes
9193 @code{"NGoto char: "} in this position to indicate that the function
9194 expects a raw prefix, in this case, a numerical location in a buffer,
9195 and provides a prompt.
9198 The seventh part is a documentation string, just like the one for a
9199 function written in Emacs Lisp. This is written as a C comment. (When
9200 you build Emacs, the program @command{lib-src/make-docfile} extracts
9201 these comments and uses them to make the ``real'' documentation.)
9205 In a C macro, the formal parameters come next, with a statement of
9206 what kind of object they are, followed by what might be called the `body'
9207 of the macro. For @code{delete-and-extract-region} the `body'
9208 consists of the following four lines:
9212 validate_region (&start, &end);
9213 if (XINT (start) == XINT (end))
9214 return empty_unibyte_string;
9215 return del_range_1 (XINT (start), XINT (end), 1, 1);
9219 The @code{validate_region} function checks whether the values
9220 passed as the beginning and end of the region are the proper type and
9221 are within range. If the beginning and end positions are the same,
9222 then return an empty string.
9224 The @code{del_range_1} function actually deletes the text. It is a
9225 complex function we will not look into. It updates the buffer and
9226 does other things. However, it is worth looking at the two arguments
9227 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9228 @w{@code{XINT (end)}}.
9230 As far as the C language is concerned, @code{start} and @code{end} are
9231 two integers that mark the beginning and end of the region to be
9232 deleted@footnote{More precisely, and requiring more expert knowledge
9233 to understand, the two integers are of type `Lisp_Object', which can
9234 also be a C union instead of an integer type.}.
9236 In early versions of Emacs, these two numbers were thirty-two bits
9237 long, but the code is slowly being generalized to handle other
9238 lengths. Three of the available bits are used to specify the type of
9239 information; the remaining bits are used as `content'.
9241 @samp{XINT} is a C macro that extracts the relevant number from the
9242 longer collection of bits; the three other bits are discarded.
9245 The command in @code{delete-and-extract-region} looks like this:
9248 del_range_1 (XINT (start), XINT (end), 1, 1);
9252 It deletes the region between the beginning position, @code{start},
9253 and the ending position, @code{end}.
9255 From the point of view of the person writing Lisp, Emacs is all very
9256 simple; but hidden underneath is a great deal of complexity to make it
9260 @section Initializing a Variable with @code{defvar}
9262 @cindex Initializing a variable
9263 @cindex Variable initialization
9268 copy-region-as-kill is short, 12 lines, and uses
9269 filter-buffer-substring, which is longer, 39 lines
9270 and has delete-and-extract-region in it.
9271 delete-and-extract-region is written in C.
9273 see Initializing a Variable with @code{defvar}
9277 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9278 functions within it, @code{kill-append} and @code{kill-new}, copy a
9279 region in a buffer and save it in a variable called the
9280 @code{kill-ring}. This section describes how the @code{kill-ring}
9281 variable is created and initialized using the @code{defvar} special
9284 (Again we note that the term @code{kill-ring} is a misnomer. The text
9285 that is clipped out of the buffer can be brought back; it is not a ring
9286 of corpses, but a ring of resurrectable text.)
9288 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9289 given an initial value by using the @code{defvar} special form. The
9290 name comes from ``define variable''.
9292 The @code{defvar} special form is similar to @code{setq} in that it sets
9293 the value of a variable. It is unlike @code{setq} in two ways: first,
9294 it only sets the value of the variable if the variable does not already
9295 have a value. If the variable already has a value, @code{defvar} does
9296 not override the existing value. Second, @code{defvar} has a
9297 documentation string.
9299 (There is a related macro, @code{defcustom}, designed for variables
9300 that people customize. It has more features than @code{defvar}.
9301 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9304 * See variable current value::
9305 * defvar and asterisk::
9309 @node See variable current value
9310 @unnumberedsubsec Seeing the Current Value of a Variable
9313 You can see the current value of a variable, any variable, by using
9314 the @code{describe-variable} function, which is usually invoked by
9315 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9316 (followed by @key{RET}) when prompted, you will see what is in your
9317 current kill ring---this may be quite a lot! Conversely, if you have
9318 been doing nothing this Emacs session except read this document, you
9319 may have nothing in it. Also, you will see the documentation for
9325 List of killed text sequences.
9326 Since the kill ring is supposed to interact nicely with cut-and-paste
9327 facilities offered by window systems, use of this variable should
9330 interact nicely with `interprogram-cut-function' and
9331 `interprogram-paste-function'. The functions `kill-new',
9332 `kill-append', and `current-kill' are supposed to implement this
9333 interaction; you may want to use them instead of manipulating the kill
9339 The kill ring is defined by a @code{defvar} in the following way:
9343 (defvar kill-ring nil
9344 "List of killed text sequences.
9350 In this variable definition, the variable is given an initial value of
9351 @code{nil}, which makes sense, since if you have saved nothing, you want
9352 nothing back if you give a @code{yank} command. The documentation
9353 string is written just like the documentation string of a @code{defun}.
9354 As with the documentation string of the @code{defun}, the first line of
9355 the documentation should be a complete sentence, since some commands,
9356 like @code{apropos}, print only the first line of documentation.
9357 Succeeding lines should not be indented; otherwise they look odd when
9358 you use @kbd{C-h v} (@code{describe-variable}).
9360 @node defvar and asterisk
9361 @subsection @code{defvar} and an asterisk
9362 @findex defvar @r{for a user customizable variable}
9363 @findex defvar @r{with an asterisk}
9365 In the past, Emacs used the @code{defvar} special form both for
9366 internal variables that you would not expect a user to change and for
9367 variables that you do expect a user to change. Although you can still
9368 use @code{defvar} for user customizable variables, please use
9369 @code{defcustom} instead, since that special form provides a path into
9370 the Customization commands. (@xref{defcustom, , Specifying Variables
9371 using @code{defcustom}}.)
9373 When you specified a variable using the @code{defvar} special form,
9374 you could distinguish a variable that a user might want to change from
9375 others by typing an asterisk, @samp{*}, in the first column of its
9376 documentation string. For example:
9380 (defvar shell-command-default-error-buffer nil
9381 "*Buffer name for `shell-command' @dots{} error output.
9386 @findex set-variable
9388 You could (and still can) use the @code{set-variable} command to
9389 change the value of @code{shell-command-default-error-buffer}
9390 temporarily. However, options set using @code{set-variable} are set
9391 only for the duration of your editing session. The new values are not
9392 saved between sessions. Each time Emacs starts, it reads the original
9393 value, unless you change the value within your @file{.emacs} file,
9394 either by setting it manually or by using @code{customize}.
9395 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9397 For me, the major use of the @code{set-variable} command is to suggest
9398 variables that I might want to set in my @file{.emacs} file. There
9399 are now more than 700 such variables, far too many to remember
9400 readily. Fortunately, you can press @key{TAB} after calling the
9401 @code{M-x set-variable} command to see the list of variables.
9402 (@xref{Examining, , Examining and Setting Variables, emacs,
9403 The GNU Emacs Manual}.)
9406 @node cons & search-fwd Review
9409 Here is a brief summary of some recently introduced functions.
9414 @code{car} returns the first element of a list; @code{cdr} returns the
9415 second and subsequent elements of a list.
9422 (car '(1 2 3 4 5 6 7))
9424 (cdr '(1 2 3 4 5 6 7))
9425 @result{} (2 3 4 5 6 7)
9430 @code{cons} constructs a list by prepending its first argument to its
9444 @code{funcall} evaluates its first argument as a function. It passes
9445 its remaining arguments to its first argument.
9448 Return the result of taking @sc{cdr} `n' times on a list.
9456 The `rest of the rest', as it were.
9463 (nthcdr 3 '(1 2 3 4 5 6 7))
9470 @code{setcar} changes the first element of a list; @code{setcdr}
9471 changes the second and subsequent elements of a list.
9478 (setq triple '(1 2 3))
9485 (setcdr triple '("foo" "bar"))
9488 @result{} (37 "foo" "bar")
9493 Evaluate each argument in sequence and then return the value of the
9506 @item save-restriction
9507 Record whatever narrowing is in effect in the current buffer, if any,
9508 and restore that narrowing after evaluating the arguments.
9510 @item search-forward
9511 Search for a string, and if the string is found, move point. With a
9512 regular expression, use the similar @code{re-search-forward}.
9513 (@xref{Regexp Search, , Regular Expression Searches}, for an
9514 explanation of regular expression patterns and searches.)
9518 @code{search-forward} and @code{re-search-forward} take four
9523 The string or regular expression to search for.
9526 Optionally, the limit of the search.
9529 Optionally, what to do if the search fails, return @code{nil} or an
9533 Optionally, how many times to repeat the search; if negative, the
9534 search goes backwards.
9538 @itemx delete-and-extract-region
9539 @itemx copy-region-as-kill
9541 @code{kill-region} cuts the text between point and mark from the
9542 buffer and stores that text in the kill ring, so you can get it back
9545 @code{copy-region-as-kill} copies the text between point and mark into
9546 the kill ring, from which you can get it by yanking. The function
9547 does not cut or remove the text from the buffer.
9550 @code{delete-and-extract-region} removes the text between point and
9551 mark from the buffer and throws it away. You cannot get it back.
9552 (This is not an interactive command.)
9555 @node search Exercises
9556 @section Searching Exercises
9560 Write an interactive function that searches for a string. If the
9561 search finds the string, leave point after it and display a message
9562 that says ``Found!''. (Do not use @code{search-forward} for the name
9563 of this function; if you do, you will overwrite the existing version of
9564 @code{search-forward} that comes with Emacs. Use a name such as
9565 @code{test-search} instead.)
9568 Write a function that prints the third element of the kill ring in the
9569 echo area, if any; if the kill ring does not contain a third element,
9570 print an appropriate message.
9573 @node List Implementation
9574 @chapter How Lists are Implemented
9575 @cindex Lists in a computer
9577 In Lisp, atoms are recorded in a straightforward fashion; if the
9578 implementation is not straightforward in practice, it is, nonetheless,
9579 straightforward in theory. The atom @samp{rose}, for example, is
9580 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9581 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9582 is equally simple, but it takes a moment to get used to the idea. A
9583 list is kept using a series of pairs of pointers. In the series, the
9584 first pointer in each pair points to an atom or to another list, and the
9585 second pointer in each pair points to the next pair, or to the symbol
9586 @code{nil}, which marks the end of the list.
9588 A pointer itself is quite simply the electronic address of what is
9589 pointed to. Hence, a list is kept as a series of electronic addresses.
9592 * Lists diagrammed::
9593 * Symbols as Chest:: Exploring a powerful metaphor.
9598 @node Lists diagrammed
9599 @unnumberedsec Lists diagrammed
9602 For example, the list @code{(rose violet buttercup)} has three elements,
9603 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9604 electronic address of @samp{rose} is recorded in a segment of computer
9605 memory along with the address that gives the electronic address of where
9606 the atom @samp{violet} is located; and that address (the one that tells
9607 where @samp{violet} is located) is kept along with an address that tells
9608 where the address for the atom @samp{buttercup} is located.
9611 This sounds more complicated than it is and is easier seen in a diagram:
9613 @c clear print-postscript-figures
9614 @c !!! cons-cell-diagram #1
9618 ___ ___ ___ ___ ___ ___
9619 |___|___|--> |___|___|--> |___|___|--> nil
9622 --> rose --> violet --> buttercup
9626 @ifset print-postscript-figures
9629 @center @image{cons-1}
9630 %%%% old method of including an image
9631 % \input /usr/local/lib/tex/inputs/psfig.tex
9632 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9637 @ifclear print-postscript-figures
9641 ___ ___ ___ ___ ___ ___
9642 |___|___|--> |___|___|--> |___|___|--> nil
9645 --> rose --> violet --> buttercup
9652 In the diagram, each box represents a word of computer memory that
9653 holds a Lisp object, usually in the form of a memory address. The boxes,
9654 i.e., the addresses, are in pairs. Each arrow points to what the address
9655 is the address of, either an atom or another pair of addresses. The
9656 first box is the electronic address of @samp{rose} and the arrow points
9657 to @samp{rose}; the second box is the address of the next pair of boxes,
9658 the first part of which is the address of @samp{violet} and the second
9659 part of which is the address of the next pair. The very last box
9660 points to the symbol @code{nil}, which marks the end of the list.
9663 When a variable is set to a list with a function such as @code{setq},
9664 it stores the address of the first box in the variable. Thus,
9665 evaluation of the expression
9668 (setq bouquet '(rose violet buttercup))
9673 creates a situation like this:
9675 @c cons-cell-diagram #2
9681 | ___ ___ ___ ___ ___ ___
9682 --> |___|___|--> |___|___|--> |___|___|--> nil
9685 --> rose --> violet --> buttercup
9689 @ifset print-postscript-figures
9692 @center @image{cons-2}
9693 %%%% old method of including an image
9694 % \input /usr/local/lib/tex/inputs/psfig.tex
9695 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9700 @ifclear print-postscript-figures
9706 | ___ ___ ___ ___ ___ ___
9707 --> |___|___|--> |___|___|--> |___|___|--> nil
9710 --> rose --> violet --> buttercup
9717 In this example, the symbol @code{bouquet} holds the address of the first
9721 This same list can be illustrated in a different sort of box notation
9724 @c cons-cell-diagram #2a
9730 | -------------- --------------- ----------------
9731 | | car | cdr | | car | cdr | | car | cdr |
9732 -->| rose | o------->| violet | o------->| butter- | nil |
9733 | | | | | | | cup | |
9734 -------------- --------------- ----------------
9738 @ifset print-postscript-figures
9741 @center @image{cons-2a}
9742 %%%% old method of including an image
9743 % \input /usr/local/lib/tex/inputs/psfig.tex
9744 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9749 @ifclear print-postscript-figures
9755 | -------------- --------------- ----------------
9756 | | car | cdr | | car | cdr | | car | cdr |
9757 -->| rose | o------->| violet | o------->| butter- | nil |
9758 | | | | | | | cup | |
9759 -------------- --------------- ----------------
9765 (Symbols consist of more than pairs of addresses, but the structure of
9766 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9767 consists of a group of address-boxes, one of which is the address of
9768 the printed word @samp{bouquet}, a second of which is the address of a
9769 function definition attached to the symbol, if any, a third of which
9770 is the address of the first pair of address-boxes for the list
9771 @code{(rose violet buttercup)}, and so on. Here we are showing that
9772 the symbol's third address-box points to the first pair of
9773 address-boxes for the list.)
9775 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9776 changed; the symbol simply has an address further down the list. (In
9777 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9778 evaluation of the following expression
9781 (setq flowers (cdr bouquet))
9788 @c cons-cell-diagram #3
9795 | ___ ___ | ___ ___ ___ ___
9796 --> | | | --> | | | | | |
9797 |___|___|----> |___|___|--> |___|___|--> nil
9800 --> rose --> violet --> buttercup
9805 @ifset print-postscript-figures
9808 @center @image{cons-3}
9809 %%%% old method of including an image
9810 % \input /usr/local/lib/tex/inputs/psfig.tex
9811 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9816 @ifclear print-postscript-figures
9823 | ___ ___ | ___ ___ ___ ___
9824 --> | | | --> | | | | | |
9825 |___|___|----> |___|___|--> |___|___|--> nil
9828 --> rose --> violet --> buttercup
9836 The value of @code{flowers} is @code{(violet buttercup)}, which is
9837 to say, the symbol @code{flowers} holds the address of the pair of
9838 address-boxes, the first of which holds the address of @code{violet},
9839 and the second of which holds the address of @code{buttercup}.
9841 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9842 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9843 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9844 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9845 information about cons cells and dotted pairs.
9848 The function @code{cons} adds a new pair of addresses to the front of
9849 a series of addresses like that shown above. For example, evaluating
9853 (setq bouquet (cons 'lily bouquet))
9860 @c cons-cell-diagram #4
9867 | ___ ___ ___ ___ | ___ ___ ___ ___
9868 --> | | | | | | --> | | | | | |
9869 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9872 --> lily --> rose --> violet --> buttercup
9877 @ifset print-postscript-figures
9880 @center @image{cons-4}
9881 %%%% old method of including an image
9882 % \input /usr/local/lib/tex/inputs/psfig.tex
9883 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9888 @ifclear print-postscript-figures
9895 | ___ ___ ___ ___ | ___ ___ ___ ___
9896 --> | | | | | | --> | | | | | |
9897 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9900 --> lily --> rose --> violet --> buttercup
9909 However, this does not change the value of the symbol
9910 @code{flowers}, as you can see by evaluating the following,
9913 (eq (cdr (cdr bouquet)) flowers)
9917 which returns @code{t} for true.
9919 Until it is reset, @code{flowers} still has the value
9920 @code{(violet buttercup)}; that is, it has the address of the cons
9921 cell whose first address is of @code{violet}. Also, this does not
9922 alter any of the pre-existing cons cells; they are all still there.
9924 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
9925 of the next cons cell in the series; to get the @sc{car} of a list,
9926 you get the address of the first element of the list; to @code{cons} a
9927 new element on a list, you add a new cons cell to the front of the list.
9928 That is all there is to it! The underlying structure of Lisp is
9931 And what does the last address in a series of cons cells refer to? It
9932 is the address of the empty list, of @code{nil}.
9934 In summary, when a Lisp variable is set to a value, it is provided with
9935 the address of the list to which the variable refers.
9937 @node Symbols as Chest
9938 @section Symbols as a Chest of Drawers
9939 @cindex Symbols as a Chest of Drawers
9940 @cindex Chest of Drawers, metaphor for a symbol
9941 @cindex Drawers, Chest of, metaphor for a symbol
9943 In an earlier section, I suggested that you might imagine a symbol as
9944 being a chest of drawers. The function definition is put in one
9945 drawer, the value in another, and so on. What is put in the drawer
9946 holding the value can be changed without affecting the contents of the
9947 drawer holding the function definition, and vice-verse.
9949 Actually, what is put in each drawer is the address of the value or
9950 function definition. It is as if you found an old chest in the attic,
9951 and in one of its drawers you found a map giving you directions to
9952 where the buried treasure lies.
9954 (In addition to its name, symbol definition, and variable value, a
9955 symbol has a `drawer' for a @dfn{property list} which can be used to
9956 record other information. Property lists are not discussed here; see
9957 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
9961 Here is a fanciful representation:
9963 @c chest-of-drawers diagram
9968 Chest of Drawers Contents of Drawers
9972 ---------------------
9973 | directions to | [map to]
9974 | symbol name | bouquet
9976 +---------------------+
9978 | symbol definition | [none]
9980 +---------------------+
9981 | directions to | [map to]
9982 | variable value | (rose violet buttercup)
9984 +---------------------+
9986 | property list | [not described here]
9988 +---------------------+
9994 @ifset print-postscript-figures
9997 @center @image{drawers}
9998 %%%% old method of including an image
9999 % \input /usr/local/lib/tex/inputs/psfig.tex
10000 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10005 @ifclear print-postscript-figures
10010 Chest of Drawers Contents of Drawers
10014 ---------------------
10015 | directions to | [map to]
10016 | symbol name | bouquet
10018 +---------------------+
10020 | symbol definition | [none]
10022 +---------------------+
10023 | directions to | [map to]
10024 | variable value | (rose violet buttercup)
10026 +---------------------+
10028 | property list | [not described here]
10030 +---------------------+
10038 @node List Exercise
10041 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10042 more flowers on to this list and set this new list to
10043 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10044 What does the @code{more-flowers} list now contain?
10047 @chapter Yanking Text Back
10049 @cindex Text retrieval
10050 @cindex Retrieving text
10051 @cindex Pasting text
10053 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10054 you can bring it back with a `yank' command. The text that is cut out of
10055 the buffer is put in the kill ring and the yank commands insert the
10056 appropriate contents of the kill ring back into a buffer (not necessarily
10057 the original buffer).
10059 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10060 the kill ring into the current buffer. If the @kbd{C-y} command is
10061 followed immediately by @kbd{M-y}, the first element is replaced by
10062 the second element. Successive @kbd{M-y} commands replace the second
10063 element with the third, fourth, or fifth element, and so on. When the
10064 last element in the kill ring is reached, it is replaced by the first
10065 element and the cycle is repeated. (Thus the kill ring is called a
10066 `ring' rather than just a `list'. However, the actual data structure
10067 that holds the text is a list.
10068 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10069 list is handled as a ring.)
10072 * Kill Ring Overview::
10073 * kill-ring-yank-pointer:: The kill ring is a list.
10074 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
10077 @node Kill Ring Overview
10078 @section Kill Ring Overview
10079 @cindex Kill ring overview
10081 The kill ring is a list of textual strings. This is what it looks like:
10084 ("some text" "a different piece of text" "yet more text")
10087 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10088 string of characters saying @samp{some text} would be inserted in this
10089 buffer where my cursor is located.
10091 The @code{yank} command is also used for duplicating text by copying it.
10092 The copied text is not cut from the buffer, but a copy of it is put on the
10093 kill ring and is inserted by yanking it back.
10095 Three functions are used for bringing text back from the kill ring:
10096 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10097 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10098 which is used by the two other functions.
10100 These functions refer to the kill ring through a variable called the
10101 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10102 @code{yank} and @code{yank-pop} functions is:
10105 (insert (car kill-ring-yank-pointer))
10109 (Well, no more. In GNU Emacs 22, the function has been replaced by
10110 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10111 repetitively for each @code{yank-handler} segment. In turn,
10112 @code{insert-for-yank-1} strips text properties from the inserted text
10113 according to @code{yank-excluded-properties}. Otherwise, it is just
10114 like @code{insert}. We will stick with plain @code{insert} since it
10115 is easier to understand.)
10117 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10118 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10120 @node kill-ring-yank-pointer
10121 @section The @code{kill-ring-yank-pointer} Variable
10123 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10124 a variable. It points to something by being bound to the value of what
10125 it points to, like any other Lisp variable.
10128 Thus, if the value of the kill ring is:
10131 ("some text" "a different piece of text" "yet more text")
10136 and the @code{kill-ring-yank-pointer} points to the second clause, the
10137 value of @code{kill-ring-yank-pointer} is:
10140 ("a different piece of text" "yet more text")
10143 As explained in the previous chapter (@pxref{List Implementation}), the
10144 computer does not keep two different copies of the text being pointed to
10145 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10146 words ``a different piece of text'' and ``yet more text'' are not
10147 duplicated. Instead, the two Lisp variables point to the same pieces of
10148 text. Here is a diagram:
10150 @c cons-cell-diagram #5
10154 kill-ring kill-ring-yank-pointer
10156 | ___ ___ | ___ ___ ___ ___
10157 ---> | | | --> | | | | | |
10158 |___|___|----> |___|___|--> |___|___|--> nil
10161 | | --> "yet more text"
10163 | --> "a different piece of text"
10170 @ifset print-postscript-figures
10173 @center @image{cons-5}
10174 %%%% old method of including an image
10175 % \input /usr/local/lib/tex/inputs/psfig.tex
10176 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10181 @ifclear print-postscript-figures
10185 kill-ring kill-ring-yank-pointer
10187 | ___ ___ | ___ ___ ___ ___
10188 ---> | | | --> | | | | | |
10189 |___|___|----> |___|___|--> |___|___|--> nil
10192 | | --> "yet more text"
10194 | --> "a different piece of text
10203 Both the variable @code{kill-ring} and the variable
10204 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10205 usually described as if it were actually what it is composed of. The
10206 @code{kill-ring} is spoken of as if it were the list rather than that it
10207 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10208 spoken of as pointing to a list.
10210 These two ways of talking about the same thing sound confusing at first but
10211 make sense on reflection. The kill ring is generally thought of as the
10212 complete structure of data that holds the information of what has recently
10213 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10214 on the other hand, serves to indicate---that is, to `point to'---that part
10215 of the kill ring of which the first element (the @sc{car}) will be
10219 In GNU Emacs 22, the @code{kill-new} function calls
10221 @code{(setq kill-ring-yank-pointer kill-ring)}
10223 (defun rotate-yank-pointer (arg)
10224 "Rotate the yanking point in the kill ring.
10225 With argument, rotate that many kills forward (or backward, if negative)."
10227 (current-kill arg))
10229 (defun current-kill (n &optional do-not-move)
10230 "Rotate the yanking point by N places, and then return that kill.
10231 If N is zero, `interprogram-paste-function' is set, and calling it
10232 returns a string, then that string is added to the front of the
10233 kill ring and returned as the latest kill.
10234 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10235 yanking point; just return the Nth kill forward."
10236 (let ((interprogram-paste (and (= n 0)
10237 interprogram-paste-function
10238 (funcall interprogram-paste-function))))
10239 (if interprogram-paste
10241 ;; Disable the interprogram cut function when we add the new
10242 ;; text to the kill ring, so Emacs doesn't try to own the
10243 ;; selection, with identical text.
10244 (let ((interprogram-cut-function nil))
10245 (kill-new interprogram-paste))
10246 interprogram-paste)
10247 (or kill-ring (error "Kill ring is empty"))
10248 (let ((ARGth-kill-element
10249 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10250 (length kill-ring))
10253 (setq kill-ring-yank-pointer ARGth-kill-element))
10254 (car ARGth-kill-element)))))
10259 @node yank nthcdr Exercises
10260 @section Exercises with @code{yank} and @code{nthcdr}
10264 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10265 your kill ring. Add several items to your kill ring; look at its
10266 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10267 around the kill ring. How many items were in your kill ring? Find
10268 the value of @code{kill-ring-max}. Was your kill ring full, or could
10269 you have kept more blocks of text within it?
10272 Using @code{nthcdr} and @code{car}, construct a series of expressions
10273 to return the first, second, third, and fourth elements of a list.
10276 @node Loops & Recursion
10277 @chapter Loops and Recursion
10278 @cindex Loops and recursion
10279 @cindex Recursion and loops
10280 @cindex Repetition (loops)
10282 Emacs Lisp has two primary ways to cause an expression, or a series of
10283 expressions, to be evaluated repeatedly: one uses a @code{while}
10284 loop, and the other uses @dfn{recursion}.
10286 Repetition can be very valuable. For example, to move forward four
10287 sentences, you need only write a program that will move forward one
10288 sentence and then repeat the process four times. Since a computer does
10289 not get bored or tired, such repetitive action does not have the
10290 deleterious effects that excessive or the wrong kinds of repetition can
10293 People mostly write Emacs Lisp functions using @code{while} loops and
10294 their kin; but you can use recursion, which provides a very powerful
10295 way to think about and then to solve problems@footnote{You can write
10296 recursive functions to be frugal or wasteful of mental or computer
10297 resources; as it happens, methods that people find easy---that are
10298 frugal of `mental resources'---sometimes use considerable computer
10299 resources. Emacs was designed to run on machines that we now consider
10300 limited and its default settings are conservative. You may want to
10301 increase the values of @code{max-specpdl-size} and
10302 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10303 15 and 30 times their default value.}.
10306 * while:: Causing a stretch of code to repeat.
10308 * Recursion:: Causing a function to call itself.
10309 * Looping exercise::
10313 @section @code{while}
10317 The @code{while} special form tests whether the value returned by
10318 evaluating its first argument is true or false. This is similar to what
10319 the Lisp interpreter does with an @code{if}; what the interpreter does
10320 next, however, is different.
10322 In a @code{while} expression, if the value returned by evaluating the
10323 first argument is false, the Lisp interpreter skips the rest of the
10324 expression (the @dfn{body} of the expression) and does not evaluate it.
10325 However, if the value is true, the Lisp interpreter evaluates the body
10326 of the expression and then again tests whether the first argument to
10327 @code{while} is true or false. If the value returned by evaluating the
10328 first argument is again true, the Lisp interpreter again evaluates the
10329 body of the expression.
10332 The template for a @code{while} expression looks like this:
10336 (while @var{true-or-false-test}
10342 * Looping with while:: Repeat so long as test returns true.
10343 * Loop Example:: A @code{while} loop that uses a list.
10344 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10345 * Incrementing Loop:: A loop with an incrementing counter.
10346 * Incrementing Loop Details::
10347 * Decrementing Loop:: A loop with a decrementing counter.
10351 @node Looping with while
10352 @unnumberedsubsec Looping with @code{while}
10355 So long as the true-or-false-test of the @code{while} expression
10356 returns a true value when it is evaluated, the body is repeatedly
10357 evaluated. This process is called a loop since the Lisp interpreter
10358 repeats the same thing again and again, like an airplane doing a loop.
10359 When the result of evaluating the true-or-false-test is false, the
10360 Lisp interpreter does not evaluate the rest of the @code{while}
10361 expression and `exits the loop'.
10363 Clearly, if the value returned by evaluating the first argument to
10364 @code{while} is always true, the body following will be evaluated
10365 again and again @dots{} and again @dots{} forever. Conversely, if the
10366 value returned is never true, the expressions in the body will never
10367 be evaluated. The craft of writing a @code{while} loop consists of
10368 choosing a mechanism such that the true-or-false-test returns true
10369 just the number of times that you want the subsequent expressions to
10370 be evaluated, and then have the test return false.
10372 The value returned by evaluating a @code{while} is the value of the
10373 true-or-false-test. An interesting consequence of this is that a
10374 @code{while} loop that evaluates without error will return @code{nil}
10375 or false regardless of whether it has looped 1 or 100 times or none at
10376 all. A @code{while} expression that evaluates successfully never
10377 returns a true value! What this means is that @code{while} is always
10378 evaluated for its side effects, which is to say, the consequences of
10379 evaluating the expressions within the body of the @code{while} loop.
10380 This makes sense. It is not the mere act of looping that is desired,
10381 but the consequences of what happens when the expressions in the loop
10382 are repeatedly evaluated.
10385 @subsection A @code{while} Loop and a List
10387 A common way to control a @code{while} loop is to test whether a list
10388 has any elements. If it does, the loop is repeated; but if it does not,
10389 the repetition is ended. Since this is an important technique, we will
10390 create a short example to illustrate it.
10392 A simple way to test whether a list has elements is to evaluate the
10393 list: if it has no elements, it is an empty list and will return the
10394 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10395 the other hand, a list with elements will return those elements when it
10396 is evaluated. Since Emacs Lisp considers as true any value that is not
10397 @code{nil}, a list that returns elements will test true in a
10401 For example, you can set the variable @code{empty-list} to @code{nil} by
10402 evaluating the following @code{setq} expression:
10405 (setq empty-list ())
10409 After evaluating the @code{setq} expression, you can evaluate the
10410 variable @code{empty-list} in the usual way, by placing the cursor after
10411 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10418 On the other hand, if you set a variable to be a list with elements, the
10419 list will appear when you evaluate the variable, as you can see by
10420 evaluating the following two expressions:
10424 (setq animals '(gazelle giraffe lion tiger))
10430 Thus, to create a @code{while} loop that tests whether there are any
10431 items in the list @code{animals}, the first part of the loop will be
10442 When the @code{while} tests its first argument, the variable
10443 @code{animals} is evaluated. It returns a list. So long as the list
10444 has elements, the @code{while} considers the results of the test to be
10445 true; but when the list is empty, it considers the results of the test
10448 To prevent the @code{while} loop from running forever, some mechanism
10449 needs to be provided to empty the list eventually. An oft-used
10450 technique is to have one of the subsequent forms in the @code{while}
10451 expression set the value of the list to be the @sc{cdr} of the list.
10452 Each time the @code{cdr} function is evaluated, the list will be made
10453 shorter, until eventually only the empty list will be left. At this
10454 point, the test of the @code{while} loop will return false, and the
10455 arguments to the @code{while} will no longer be evaluated.
10457 For example, the list of animals bound to the variable @code{animals}
10458 can be set to be the @sc{cdr} of the original list with the
10459 following expression:
10462 (setq animals (cdr animals))
10466 If you have evaluated the previous expressions and then evaluate this
10467 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10468 area. If you evaluate the expression again, @code{(lion tiger)} will
10469 appear in the echo area. If you evaluate it again and yet again,
10470 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10472 A template for a @code{while} loop that uses the @code{cdr} function
10473 repeatedly to cause the true-or-false-test eventually to test false
10478 (while @var{test-whether-list-is-empty}
10480 @var{set-list-to-cdr-of-list})
10484 This test and use of @code{cdr} can be put together in a function that
10485 goes through a list and prints each element of the list on a line of its
10488 @node print-elements-of-list
10489 @subsection An Example: @code{print-elements-of-list}
10490 @findex print-elements-of-list
10492 The @code{print-elements-of-list} function illustrates a @code{while}
10495 @cindex @file{*scratch*} buffer
10496 The function requires several lines for its output. If you are
10497 reading this in a recent instance of GNU Emacs,
10498 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10499 you can evaluate the following expression inside of Info, as usual.
10501 If you are using an earlier version of Emacs, you need to copy the
10502 necessary expressions to your @file{*scratch*} buffer and evaluate
10503 them there. This is because the echo area had only one line in the
10506 You can copy the expressions by marking the beginning of the region
10507 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10508 the end of the region and then copying the region using @kbd{M-w}
10509 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10510 then provides visual feedback). In the @file{*scratch*}
10511 buffer, you can yank the expressions back by typing @kbd{C-y}
10514 After you have copied the expressions to the @file{*scratch*} buffer,
10515 evaluate each expression in turn. Be sure to evaluate the last
10516 expression, @code{(print-elements-of-list animals)}, by typing
10517 @kbd{C-u C-x C-e}, that is, by giving an argument to
10518 @code{eval-last-sexp}. This will cause the result of the evaluation
10519 to be printed in the @file{*scratch*} buffer instead of being printed
10520 in the echo area. (Otherwise you will see something like this in your
10521 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10522 each @samp{^J} stands for a `newline'.)
10525 In a recent instance of GNU Emacs, you can evaluate these expressions
10526 directly in the Info buffer, and the echo area will grow to show the
10531 (setq animals '(gazelle giraffe lion tiger))
10533 (defun print-elements-of-list (list)
10534 "Print each element of LIST on a line of its own."
10537 (setq list (cdr list))))
10539 (print-elements-of-list animals)
10545 When you evaluate the three expressions in sequence, you will see
10561 Each element of the list is printed on a line of its own (that is what
10562 the function @code{print} does) and then the value returned by the
10563 function is printed. Since the last expression in the function is the
10564 @code{while} loop, and since @code{while} loops always return
10565 @code{nil}, a @code{nil} is printed after the last element of the list.
10567 @node Incrementing Loop
10568 @subsection A Loop with an Incrementing Counter
10570 A loop is not useful unless it stops when it ought. Besides
10571 controlling a loop with a list, a common way of stopping a loop is to
10572 write the first argument as a test that returns false when the correct
10573 number of repetitions are complete. This means that the loop must
10574 have a counter---an expression that counts how many times the loop
10578 @node Incrementing Loop Details
10579 @unnumberedsubsec Details of an Incrementing Loop
10582 The test for a loop with an incrementing counter can be an expression
10583 such as @code{(< count desired-number)} which returns @code{t} for
10584 true if the value of @code{count} is less than the
10585 @code{desired-number} of repetitions and @code{nil} for false if the
10586 value of @code{count} is equal to or is greater than the
10587 @code{desired-number}. The expression that increments the count can
10588 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10589 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10590 argument. (The expression @w{@code{(1+ count)}} has the same result
10591 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10594 The template for a @code{while} loop controlled by an incrementing
10595 counter looks like this:
10599 @var{set-count-to-initial-value}
10600 (while (< count desired-number) ; @r{true-or-false-test}
10602 (setq count (1+ count))) ; @r{incrementer}
10607 Note that you need to set the initial value of @code{count}; usually it
10611 * Incrementing Example:: Counting pebbles in a triangle.
10612 * Inc Example parts:: The parts of the function definition.
10613 * Inc Example altogether:: Putting the function definition together.
10616 @node Incrementing Example
10617 @unnumberedsubsubsec Example with incrementing counter
10619 Suppose you are playing on the beach and decide to make a triangle of
10620 pebbles, putting one pebble in the first row, two in the second row,
10621 three in the third row and so on, like this:
10639 @bullet{} @bullet{}
10640 @bullet{} @bullet{} @bullet{}
10641 @bullet{} @bullet{} @bullet{} @bullet{}
10648 (About 2500 years ago, Pythagoras and others developed the beginnings of
10649 number theory by considering questions such as this.)
10651 Suppose you want to know how many pebbles you will need to make a
10652 triangle with 7 rows?
10654 Clearly, what you need to do is add up the numbers from 1 to 7. There
10655 are two ways to do this; start with the smallest number, one, and add up
10656 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10657 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10658 mechanisms illustrate common ways of writing @code{while} loops, we will
10659 create two examples, one counting up and the other counting down. In
10660 this first example, we will start with 1 and add 2, 3, 4 and so on.
10662 If you are just adding up a short list of numbers, the easiest way to do
10663 it is to add up all the numbers at once. However, if you do not know
10664 ahead of time how many numbers your list will have, or if you want to be
10665 prepared for a very long list, then you need to design your addition so
10666 that what you do is repeat a simple process many times instead of doing
10667 a more complex process once.
10669 For example, instead of adding up all the pebbles all at once, what you
10670 can do is add the number of pebbles in the first row, 1, to the number
10671 in the second row, 2, and then add the total of those two rows to the
10672 third row, 3. Then you can add the number in the fourth row, 4, to the
10673 total of the first three rows; and so on.
10675 The critical characteristic of the process is that each repetitive
10676 action is simple. In this case, at each step we add only two numbers,
10677 the number of pebbles in the row and the total already found. This
10678 process of adding two numbers is repeated again and again until the last
10679 row has been added to the total of all the preceding rows. In a more
10680 complex loop the repetitive action might not be so simple, but it will
10681 be simpler than doing everything all at once.
10683 @node Inc Example parts
10684 @unnumberedsubsubsec The parts of the function definition
10686 The preceding analysis gives us the bones of our function definition:
10687 first, we will need a variable that we can call @code{total} that will
10688 be the total number of pebbles. This will be the value returned by
10691 Second, we know that the function will require an argument: this
10692 argument will be the total number of rows in the triangle. It can be
10693 called @code{number-of-rows}.
10695 Finally, we need a variable to use as a counter. We could call this
10696 variable @code{counter}, but a better name is @code{row-number}. That
10697 is because what the counter does in this function is count rows, and a
10698 program should be written to be as understandable as possible.
10700 When the Lisp interpreter first starts evaluating the expressions in the
10701 function, the value of @code{total} should be set to zero, since we have
10702 not added anything to it. Then the function should add the number of
10703 pebbles in the first row to the total, and then add the number of
10704 pebbles in the second to the total, and then add the number of
10705 pebbles in the third row to the total, and so on, until there are no
10706 more rows left to add.
10708 Both @code{total} and @code{row-number} are used only inside the
10709 function, so they can be declared as local variables with @code{let}
10710 and given initial values. Clearly, the initial value for @code{total}
10711 should be 0. The initial value of @code{row-number} should be 1,
10712 since we start with the first row. This means that the @code{let}
10713 statement will look like this:
10723 After the internal variables are declared and bound to their initial
10724 values, we can begin the @code{while} loop. The expression that serves
10725 as the test should return a value of @code{t} for true so long as the
10726 @code{row-number} is less than or equal to the @code{number-of-rows}.
10727 (If the expression tests true only so long as the row number is less
10728 than the number of rows in the triangle, the last row will never be
10729 added to the total; hence the row number has to be either less than or
10730 equal to the number of rows.)
10733 @findex <= @r{(less than or equal)}
10734 Lisp provides the @code{<=} function that returns true if the value of
10735 its first argument is less than or equal to the value of its second
10736 argument and false otherwise. So the expression that the @code{while}
10737 will evaluate as its test should look like this:
10740 (<= row-number number-of-rows)
10743 The total number of pebbles can be found by repeatedly adding the number
10744 of pebbles in a row to the total already found. Since the number of
10745 pebbles in the row is equal to the row number, the total can be found by
10746 adding the row number to the total. (Clearly, in a more complex
10747 situation, the number of pebbles in the row might be related to the row
10748 number in a more complicated way; if this were the case, the row number
10749 would be replaced by the appropriate expression.)
10752 (setq total (+ total row-number))
10756 What this does is set the new value of @code{total} to be equal to the
10757 sum of adding the number of pebbles in the row to the previous total.
10759 After setting the value of @code{total}, the conditions need to be
10760 established for the next repetition of the loop, if there is one. This
10761 is done by incrementing the value of the @code{row-number} variable,
10762 which serves as a counter. After the @code{row-number} variable has
10763 been incremented, the true-or-false-test at the beginning of the
10764 @code{while} loop tests whether its value is still less than or equal to
10765 the value of the @code{number-of-rows} and if it is, adds the new value
10766 of the @code{row-number} variable to the @code{total} of the previous
10767 repetition of the loop.
10770 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10771 @code{row-number} variable can be incremented with this expression:
10774 (setq row-number (1+ row-number))
10777 @node Inc Example altogether
10778 @unnumberedsubsubsec Putting the function definition together
10780 We have created the parts for the function definition; now we need to
10784 First, the contents of the @code{while} expression:
10788 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10789 (setq total (+ total row-number))
10790 (setq row-number (1+ row-number))) ; @r{incrementer}
10794 Along with the @code{let} expression varlist, this very nearly
10795 completes the body of the function definition. However, it requires
10796 one final element, the need for which is somewhat subtle.
10798 The final touch is to place the variable @code{total} on a line by
10799 itself after the @code{while} expression. Otherwise, the value returned
10800 by the whole function is the value of the last expression that is
10801 evaluated in the body of the @code{let}, and this is the value
10802 returned by the @code{while}, which is always @code{nil}.
10804 This may not be evident at first sight. It almost looks as if the
10805 incrementing expression is the last expression of the whole function.
10806 But that expression is part of the body of the @code{while}; it is the
10807 last element of the list that starts with the symbol @code{while}.
10808 Moreover, the whole of the @code{while} loop is a list within the body
10812 In outline, the function will look like this:
10816 (defun @var{name-of-function} (@var{argument-list})
10817 "@var{documentation}@dots{}"
10818 (let (@var{varlist})
10819 (while (@var{true-or-false-test})
10820 @var{body-of-while}@dots{} )
10821 @dots{} )) ; @r{Need final expression here.}
10825 The result of evaluating the @code{let} is what is going to be returned
10826 by the @code{defun} since the @code{let} is not embedded within any
10827 containing list, except for the @code{defun} as a whole. However, if
10828 the @code{while} is the last element of the @code{let} expression, the
10829 function will always return @code{nil}. This is not what we want!
10830 Instead, what we want is the value of the variable @code{total}. This
10831 is returned by simply placing the symbol as the last element of the list
10832 starting with @code{let}. It gets evaluated after the preceding
10833 elements of the list are evaluated, which means it gets evaluated after
10834 it has been assigned the correct value for the total.
10836 It may be easier to see this by printing the list starting with
10837 @code{let} all on one line. This format makes it evident that the
10838 @var{varlist} and @code{while} expressions are the second and third
10839 elements of the list starting with @code{let}, and the @code{total} is
10844 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10849 Putting everything together, the @code{triangle} function definition
10854 (defun triangle (number-of-rows) ; @r{Version with}
10855 ; @r{ incrementing counter.}
10856 "Add up the number of pebbles in a triangle.
10857 The first row has one pebble, the second row two pebbles,
10858 the third row three pebbles, and so on.
10859 The argument is NUMBER-OF-ROWS."
10864 (while (<= row-number number-of-rows)
10865 (setq total (+ total row-number))
10866 (setq row-number (1+ row-number)))
10872 After you have installed @code{triangle} by evaluating the function, you
10873 can try it out. Here are two examples:
10884 The sum of the first four numbers is 10 and the sum of the first seven
10887 @node Decrementing Loop
10888 @subsection Loop with a Decrementing Counter
10890 Another common way to write a @code{while} loop is to write the test
10891 so that it determines whether a counter is greater than zero. So long
10892 as the counter is greater than zero, the loop is repeated. But when
10893 the counter is equal to or less than zero, the loop is stopped. For
10894 this to work, the counter has to start out greater than zero and then
10895 be made smaller and smaller by a form that is evaluated
10898 The test will be an expression such as @code{(> counter 0)} which
10899 returns @code{t} for true if the value of @code{counter} is greater
10900 than zero, and @code{nil} for false if the value of @code{counter} is
10901 equal to or less than zero. The expression that makes the number
10902 smaller and smaller can be a simple @code{setq} such as @code{(setq
10903 counter (1- counter))}, where @code{1-} is a built-in function in
10904 Emacs Lisp that subtracts 1 from its argument.
10907 The template for a decrementing @code{while} loop looks like this:
10911 (while (> counter 0) ; @r{true-or-false-test}
10913 (setq counter (1- counter))) ; @r{decrementer}
10918 * Decrementing Example:: More pebbles on the beach.
10919 * Dec Example parts:: The parts of the function definition.
10920 * Dec Example altogether:: Putting the function definition together.
10923 @node Decrementing Example
10924 @unnumberedsubsubsec Example with decrementing counter
10926 To illustrate a loop with a decrementing counter, we will rewrite the
10927 @code{triangle} function so the counter decreases to zero.
10929 This is the reverse of the earlier version of the function. In this
10930 case, to find out how many pebbles are needed to make a triangle with
10931 3 rows, add the number of pebbles in the third row, 3, to the number
10932 in the preceding row, 2, and then add the total of those two rows to
10933 the row that precedes them, which is 1.
10935 Likewise, to find the number of pebbles in a triangle with 7 rows, add
10936 the number of pebbles in the seventh row, 7, to the number in the
10937 preceding row, which is 6, and then add the total of those two rows to
10938 the row that precedes them, which is 5, and so on. As in the previous
10939 example, each addition only involves adding two numbers, the total of
10940 the rows already added up and the number of pebbles in the row that is
10941 being added to the total. This process of adding two numbers is
10942 repeated again and again until there are no more pebbles to add.
10944 We know how many pebbles to start with: the number of pebbles in the
10945 last row is equal to the number of rows. If the triangle has seven
10946 rows, the number of pebbles in the last row is 7. Likewise, we know how
10947 many pebbles are in the preceding row: it is one less than the number in
10950 @node Dec Example parts
10951 @unnumberedsubsubsec The parts of the function definition
10953 We start with three variables: the total number of rows in the
10954 triangle; the number of pebbles in a row; and the total number of
10955 pebbles, which is what we want to calculate. These variables can be
10956 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
10957 @code{total}, respectively.
10959 Both @code{total} and @code{number-of-pebbles-in-row} are used only
10960 inside the function and are declared with @code{let}. The initial
10961 value of @code{total} should, of course, be zero. However, the
10962 initial value of @code{number-of-pebbles-in-row} should be equal to
10963 the number of rows in the triangle, since the addition will start with
10967 This means that the beginning of the @code{let} expression will look
10973 (number-of-pebbles-in-row number-of-rows))
10978 The total number of pebbles can be found by repeatedly adding the number
10979 of pebbles in a row to the total already found, that is, by repeatedly
10980 evaluating the following expression:
10983 (setq total (+ total number-of-pebbles-in-row))
10987 After the @code{number-of-pebbles-in-row} is added to the @code{total},
10988 the @code{number-of-pebbles-in-row} should be decremented by one, since
10989 the next time the loop repeats, the preceding row will be
10990 added to the total.
10992 The number of pebbles in a preceding row is one less than the number of
10993 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
10994 used to compute the number of pebbles in the preceding row. This can be
10995 done with the following expression:
10999 (setq number-of-pebbles-in-row
11000 (1- number-of-pebbles-in-row))
11004 Finally, we know that the @code{while} loop should stop making repeated
11005 additions when there are no pebbles in a row. So the test for
11006 the @code{while} loop is simply:
11009 (while (> number-of-pebbles-in-row 0)
11012 @node Dec Example altogether
11013 @unnumberedsubsubsec Putting the function definition together
11015 We can put these expressions together to create a function definition
11016 that works. However, on examination, we find that one of the local
11017 variables is unneeded!
11020 The function definition looks like this:
11024 ;;; @r{First subtractive version.}
11025 (defun triangle (number-of-rows)
11026 "Add up the number of pebbles in a triangle."
11028 (number-of-pebbles-in-row number-of-rows))
11029 (while (> number-of-pebbles-in-row 0)
11030 (setq total (+ total number-of-pebbles-in-row))
11031 (setq number-of-pebbles-in-row
11032 (1- number-of-pebbles-in-row)))
11037 As written, this function works.
11039 However, we do not need @code{number-of-pebbles-in-row}.
11041 @cindex Argument as local variable
11042 When the @code{triangle} function is evaluated, the symbol
11043 @code{number-of-rows} will be bound to a number, giving it an initial
11044 value. That number can be changed in the body of the function as if
11045 it were a local variable, without any fear that such a change will
11046 effect the value of the variable outside of the function. This is a
11047 very useful characteristic of Lisp; it means that the variable
11048 @code{number-of-rows} can be used anywhere in the function where
11049 @code{number-of-pebbles-in-row} is used.
11052 Here is a second version of the function written a bit more cleanly:
11056 (defun triangle (number) ; @r{Second version.}
11057 "Return sum of numbers 1 through NUMBER inclusive."
11059 (while (> number 0)
11060 (setq total (+ total number))
11061 (setq number (1- number)))
11066 In brief, a properly written @code{while} loop will consist of three parts:
11070 A test that will return false after the loop has repeated itself the
11071 correct number of times.
11074 An expression the evaluation of which will return the value desired
11075 after being repeatedly evaluated.
11078 An expression to change the value passed to the true-or-false-test so
11079 that the test returns false after the loop has repeated itself the right
11083 @node dolist dotimes
11084 @section Save your time: @code{dolist} and @code{dotimes}
11086 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11087 provide for looping. Sometimes these are quicker to write than the
11088 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11089 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11091 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11092 list': @code{dolist} automatically shortens the list each time it
11093 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11094 each shorter version of the list to the first of its arguments.
11096 @code{dotimes} loops a specific number of times: you specify the number.
11104 @unnumberedsubsec The @code{dolist} Macro
11107 Suppose, for example, you want to reverse a list, so that
11108 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11111 In practice, you would use the @code{reverse} function, like this:
11115 (setq animals '(gazelle giraffe lion tiger))
11123 Here is how you could reverse the list using a @code{while} loop:
11127 (setq animals '(gazelle giraffe lion tiger))
11129 (defun reverse-list-with-while (list)
11130 "Using while, reverse the order of LIST."
11131 (let (value) ; make sure list starts empty
11133 (setq value (cons (car list) value))
11134 (setq list (cdr list)))
11137 (reverse-list-with-while animals)
11143 And here is how you could use the @code{dolist} macro:
11147 (setq animals '(gazelle giraffe lion tiger))
11149 (defun reverse-list-with-dolist (list)
11150 "Using dolist, reverse the order of LIST."
11151 (let (value) ; make sure list starts empty
11152 (dolist (element list value)
11153 (setq value (cons element value)))))
11155 (reverse-list-with-dolist animals)
11161 In Info, you can place your cursor after the closing parenthesis of
11162 each expression and type @kbd{C-x C-e}; in each case, you should see
11165 (tiger lion giraffe gazelle)
11171 For this example, the existing @code{reverse} function is obviously best.
11172 The @code{while} loop is just like our first example (@pxref{Loop
11173 Example, , A @code{while} Loop and a List}). The @code{while} first
11174 checks whether the list has elements; if so, it constructs a new list
11175 by adding the first element of the list to the existing list (which in
11176 the first iteration of the loop is @code{nil}). Since the second
11177 element is prepended in front of the first element, and the third
11178 element is prepended in front of the second element, the list is reversed.
11180 In the expression using a @code{while} loop,
11181 the @w{@code{(setq list (cdr list))}}
11182 expression shortens the list, so the @code{while} loop eventually
11183 stops. In addition, it provides the @code{cons} expression with a new
11184 first element by creating a new and shorter list at each repetition of
11187 The @code{dolist} expression does very much the same as the
11188 @code{while} expression, except that the @code{dolist} macro does some
11189 of the work you have to do when writing a @code{while} expression.
11191 Like a @code{while} loop, a @code{dolist} loops. What is different is
11192 that it automatically shortens the list each time it loops---it
11193 `@sc{cdr}s down the list' on its own---and it automatically binds
11194 the @sc{car} of each shorter version of the list to the first of its
11197 In the example, the @sc{car} of each shorter version of the list is
11198 referred to using the symbol @samp{element}, the list itself is called
11199 @samp{list}, and the value returned is called @samp{value}. The
11200 remainder of the @code{dolist} expression is the body.
11202 The @code{dolist} expression binds the @sc{car} of each shorter
11203 version of the list to @code{element} and then evaluates the body of
11204 the expression; and repeats the loop. The result is returned in
11208 @unnumberedsubsec The @code{dotimes} Macro
11211 The @code{dotimes} macro is similar to @code{dolist}, except that it
11212 loops a specific number of times.
11214 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11215 and so forth each time around the loop, and the value of the third
11216 argument is returned. You need to provide the value of the second
11217 argument, which is how many times the macro loops.
11220 For example, the following binds the numbers from 0 up to, but not
11221 including, the number 3 to the first argument, @var{number}, and then
11222 constructs a list of the three numbers. (The first number is 0, the
11223 second number is 1, and the third number is 2; this makes a total of
11224 three numbers in all, starting with zero as the first number.)
11228 (let (value) ; otherwise a value is a void variable
11229 (dotimes (number 3 value)
11230 (setq value (cons number value))))
11237 @code{dotimes} returns @code{value}, so the way to use
11238 @code{dotimes} is to operate on some expression @var{number} number of
11239 times and then return the result, either as a list or an atom.
11242 Here is an example of a @code{defun} that uses @code{dotimes} to add
11243 up the number of pebbles in a triangle.
11247 (defun triangle-using-dotimes (number-of-rows)
11248 "Using dotimes, add up the number of pebbles in a triangle."
11249 (let ((total 0)) ; otherwise a total is a void variable
11250 (dotimes (number number-of-rows total)
11251 (setq total (+ total (1+ number))))))
11253 (triangle-using-dotimes 4)
11261 A recursive function contains code that tells the Lisp interpreter to
11262 call a program that runs exactly like itself, but with slightly
11263 different arguments. The code runs exactly the same because it has
11264 the same name. However, even though the program has the same name, it
11265 is not the same entity. It is different. In the jargon, it is a
11266 different `instance'.
11268 Eventually, if the program is written correctly, the `slightly
11269 different arguments' will become sufficiently different from the first
11270 arguments that the final instance will stop.
11273 * Building Robots:: Same model, different serial number ...
11274 * Recursive Definition Parts:: Walk until you stop ...
11275 * Recursion with list:: Using a list as the test whether to recurse.
11276 * Recursive triangle function::
11277 * Recursion with cond::
11278 * Recursive Patterns:: Often used templates.
11279 * No Deferment:: Don't store up work ...
11280 * No deferment solution::
11283 @node Building Robots
11284 @subsection Building Robots: Extending the Metaphor
11285 @cindex Building robots
11286 @cindex Robots, building
11288 It is sometimes helpful to think of a running program as a robot that
11289 does a job. In doing its job, a recursive function calls on a second
11290 robot to help it. The second robot is identical to the first in every
11291 way, except that the second robot helps the first and has been
11292 passed different arguments than the first.
11294 In a recursive function, the second robot may call a third; and the
11295 third may call a fourth, and so on. Each of these is a different
11296 entity; but all are clones.
11298 Since each robot has slightly different instructions---the arguments
11299 will differ from one robot to the next---the last robot should know
11302 Let's expand on the metaphor in which a computer program is a robot.
11304 A function definition provides the blueprints for a robot. When you
11305 install a function definition, that is, when you evaluate a
11306 @code{defun} macro, you install the necessary equipment to build
11307 robots. It is as if you were in a factory, setting up an assembly
11308 line. Robots with the same name are built according to the same
11309 blueprints. So they have, as it were, the same `model number', but a
11310 different `serial number'.
11312 We often say that a recursive function `calls itself'. What we mean
11313 is that the instructions in a recursive function cause the Lisp
11314 interpreter to run a different function that has the same name and
11315 does the same job as the first, but with different arguments.
11317 It is important that the arguments differ from one instance to the
11318 next; otherwise, the process will never stop.
11320 @node Recursive Definition Parts
11321 @subsection The Parts of a Recursive Definition
11322 @cindex Parts of a Recursive Definition
11323 @cindex Recursive Definition Parts
11325 A recursive function typically contains a conditional expression which
11330 A true-or-false-test that determines whether the function is called
11331 again, here called the @dfn{do-again-test}.
11334 The name of the function. When this name is called, a new instance of
11335 the function---a new robot, as it were---is created and told what to do.
11338 An expression that returns a different value each time the function is
11339 called, here called the @dfn{next-step-expression}. Consequently, the
11340 argument (or arguments) passed to the new instance of the function
11341 will be different from that passed to the previous instance. This
11342 causes the conditional expression, the @dfn{do-again-test}, to test
11343 false after the correct number of repetitions.
11346 Recursive functions can be much simpler than any other kind of
11347 function. Indeed, when people first start to use them, they often look
11348 so mysteriously simple as to be incomprehensible. Like riding a
11349 bicycle, reading a recursive function definition takes a certain knack
11350 which is hard at first but then seems simple.
11353 There are several different common recursive patterns. A very simple
11354 pattern looks like this:
11358 (defun @var{name-of-recursive-function} (@var{argument-list})
11359 "@var{documentation}@dots{}"
11360 (if @var{do-again-test}
11362 (@var{name-of-recursive-function}
11363 @var{next-step-expression})))
11367 Each time a recursive function is evaluated, a new instance of it is
11368 created and told what to do. The arguments tell the instance what to do.
11370 An argument is bound to the value of the next-step-expression. Each
11371 instance runs with a different value of the next-step-expression.
11373 The value in the next-step-expression is used in the do-again-test.
11375 The value returned by the next-step-expression is passed to the new
11376 instance of the function, which evaluates it (or some
11377 transmogrification of it) to determine whether to continue or stop.
11378 The next-step-expression is designed so that the do-again-test returns
11379 false when the function should no longer be repeated.
11381 The do-again-test is sometimes called the @dfn{stop condition},
11382 since it stops the repetitions when it tests false.
11384 @node Recursion with list
11385 @subsection Recursion with a List
11387 The example of a @code{while} loop that printed the elements of a list
11388 of numbers can be written recursively. Here is the code, including
11389 an expression to set the value of the variable @code{animals} to a list.
11391 If you are reading this in Info in Emacs, you can evaluate this
11392 expression directly in Info. Otherwise, you must copy the example
11393 to the @file{*scratch*} buffer and evaluate each expression there.
11394 Use @kbd{C-u C-x C-e} to evaluate the
11395 @code{(print-elements-recursively animals)} expression so that the
11396 results are printed in the buffer; otherwise the Lisp interpreter will
11397 try to squeeze the results into the one line of the echo area.
11399 Also, place your cursor immediately after the last closing parenthesis
11400 of the @code{print-elements-recursively} function, before the comment.
11401 Otherwise, the Lisp interpreter will try to evaluate the comment.
11403 @findex print-elements-recursively
11406 (setq animals '(gazelle giraffe lion tiger))
11408 (defun print-elements-recursively (list)
11409 "Print each element of LIST on a line of its own.
11411 (when list ; @r{do-again-test}
11412 (print (car list)) ; @r{body}
11413 (print-elements-recursively ; @r{recursive call}
11414 (cdr list)))) ; @r{next-step-expression}
11416 (print-elements-recursively animals)
11420 The @code{print-elements-recursively} function first tests whether
11421 there is any content in the list; if there is, the function prints the
11422 first element of the list, the @sc{car} of the list. Then the
11423 function `invokes itself', but gives itself as its argument, not the
11424 whole list, but the second and subsequent elements of the list, the
11425 @sc{cdr} of the list.
11427 Put another way, if the list is not empty, the function invokes
11428 another instance of code that is similar to the initial code, but is a
11429 different thread of execution, with different arguments than the first
11432 Put in yet another way, if the list is not empty, the first robot
11433 assembles a second robot and tells it what to do; the second robot is
11434 a different individual from the first, but is the same model.
11436 When the second evaluation occurs, the @code{when} expression is
11437 evaluated and if true, prints the first element of the list it
11438 receives as its argument (which is the second element of the original
11439 list). Then the function `calls itself' with the @sc{cdr} of the list
11440 it is invoked with, which (the second time around) is the @sc{cdr} of
11441 the @sc{cdr} of the original list.
11443 Note that although we say that the function `calls itself', what we
11444 mean is that the Lisp interpreter assembles and instructs a new
11445 instance of the program. The new instance is a clone of the first,
11446 but is a separate individual.
11448 Each time the function `invokes itself', it invokes itself on a
11449 shorter version of the original list. It creates a new instance that
11450 works on a shorter list.
11452 Eventually, the function invokes itself on an empty list. It creates
11453 a new instance whose argument is @code{nil}. The conditional expression
11454 tests the value of @code{list}. Since the value of @code{list} is
11455 @code{nil}, the @code{when} expression tests false so the then-part is
11456 not evaluated. The function as a whole then returns @code{nil}.
11459 When you evaluate the expression @code{(print-elements-recursively
11460 animals)} in the @file{*scratch*} buffer, you see this result:
11476 @node Recursive triangle function
11477 @subsection Recursion in Place of a Counter
11478 @findex triangle-recursively
11481 The @code{triangle} function described in a previous section can also
11482 be written recursively. It looks like this:
11486 (defun triangle-recursively (number)
11487 "Return the sum of the numbers 1 through NUMBER inclusive.
11489 (if (= number 1) ; @r{do-again-test}
11491 (+ number ; @r{else-part}
11492 (triangle-recursively ; @r{recursive call}
11493 (1- number))))) ; @r{next-step-expression}
11495 (triangle-recursively 7)
11500 You can install this function by evaluating it and then try it by
11501 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11502 cursor immediately after the last parenthesis of the function
11503 definition, before the comment.) The function evaluates to 28.
11505 To understand how this function works, let's consider what happens in the
11506 various cases when the function is passed 1, 2, 3, or 4 as the value of
11510 * Recursive Example arg of 1 or 2::
11511 * Recursive Example arg of 3 or 4::
11515 @node Recursive Example arg of 1 or 2
11516 @unnumberedsubsubsec An argument of 1 or 2
11519 First, what happens if the value of the argument is 1?
11521 The function has an @code{if} expression after the documentation
11522 string. It tests whether the value of @code{number} is equal to 1; if
11523 so, Emacs evaluates the then-part of the @code{if} expression, which
11524 returns the number 1 as the value of the function. (A triangle with
11525 one row has one pebble in it.)
11527 Suppose, however, that the value of the argument is 2. In this case,
11528 Emacs evaluates the else-part of the @code{if} expression.
11531 The else-part consists of an addition, the recursive call to
11532 @code{triangle-recursively} and a decrementing action; and it looks like
11536 (+ number (triangle-recursively (1- number)))
11539 When Emacs evaluates this expression, the innermost expression is
11540 evaluated first; then the other parts in sequence. Here are the steps
11544 @item Step 1 @w{ } Evaluate the innermost expression.
11546 The innermost expression is @code{(1- number)} so Emacs decrements the
11547 value of @code{number} from 2 to 1.
11549 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11551 The Lisp interpreter creates an individual instance of
11552 @code{triangle-recursively}. It does not matter that this function is
11553 contained within itself. Emacs passes the result Step 1 as the
11554 argument used by this instance of the @code{triangle-recursively}
11557 In this case, Emacs evaluates @code{triangle-recursively} with an
11558 argument of 1. This means that this evaluation of
11559 @code{triangle-recursively} returns 1.
11561 @item Step 3 @w{ } Evaluate the value of @code{number}.
11563 The variable @code{number} is the second element of the list that
11564 starts with @code{+}; its value is 2.
11566 @item Step 4 @w{ } Evaluate the @code{+} expression.
11568 The @code{+} expression receives two arguments, the first
11569 from the evaluation of @code{number} (Step 3) and the second from the
11570 evaluation of @code{triangle-recursively} (Step 2).
11572 The result of the addition is the sum of 2 plus 1, and the number 3 is
11573 returned, which is correct. A triangle with two rows has three
11577 @node Recursive Example arg of 3 or 4
11578 @unnumberedsubsubsec An argument of 3 or 4
11580 Suppose that @code{triangle-recursively} is called with an argument of
11584 @item Step 1 @w{ } Evaluate the do-again-test.
11586 The @code{if} expression is evaluated first. This is the do-again
11587 test and returns false, so the else-part of the @code{if} expression
11588 is evaluated. (Note that in this example, the do-again-test causes
11589 the function to call itself when it tests false, not when it tests
11592 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11594 The innermost expression of the else-part is evaluated, which decrements
11595 3 to 2. This is the next-step-expression.
11597 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11599 The number 2 is passed to the @code{triangle-recursively} function.
11601 We already know what happens when Emacs evaluates @code{triangle-recursively} with
11602 an argument of 2. After going through the sequence of actions described
11603 earlier, it returns a value of 3. So that is what will happen here.
11605 @item Step 4 @w{ } Evaluate the addition.
11607 3 will be passed as an argument to the addition and will be added to the
11608 number with which the function was called, which is 3.
11612 The value returned by the function as a whole will be 6.
11614 Now that we know what will happen when @code{triangle-recursively} is
11615 called with an argument of 3, it is evident what will happen if it is
11616 called with an argument of 4:
11620 In the recursive call, the evaluation of
11623 (triangle-recursively (1- 4))
11628 will return the value of evaluating
11631 (triangle-recursively 3)
11635 which is 6 and this value will be added to 4 by the addition in the
11640 The value returned by the function as a whole will be 10.
11642 Each time @code{triangle-recursively} is evaluated, it evaluates a
11643 version of itself---a different instance of itself---with a smaller
11644 argument, until the argument is small enough so that it does not
11647 Note that this particular design for a recursive function
11648 requires that operations be deferred.
11650 Before @code{(triangle-recursively 7)} can calculate its answer, it
11651 must call @code{(triangle-recursively 6)}; and before
11652 @code{(triangle-recursively 6)} can calculate its answer, it must call
11653 @code{(triangle-recursively 5)}; and so on. That is to say, the
11654 calculation that @code{(triangle-recursively 7)} makes must be
11655 deferred until @code{(triangle-recursively 6)} makes its calculation;
11656 and @code{(triangle-recursively 6)} must defer until
11657 @code{(triangle-recursively 5)} completes; and so on.
11659 If each of these instances of @code{triangle-recursively} are thought
11660 of as different robots, the first robot must wait for the second to
11661 complete its job, which must wait until the third completes, and so
11664 There is a way around this kind of waiting, which we will discuss in
11665 @ref{No Deferment, , Recursion without Deferments}.
11667 @node Recursion with cond
11668 @subsection Recursion Example Using @code{cond}
11671 The version of @code{triangle-recursively} described earlier is written
11672 with the @code{if} special form. It can also be written using another
11673 special form called @code{cond}. The name of the special form
11674 @code{cond} is an abbreviation of the word @samp{conditional}.
11676 Although the @code{cond} special form is not used as often in the
11677 Emacs Lisp sources as @code{if}, it is used often enough to justify
11681 The template for a @code{cond} expression looks like this:
11691 where the @var{body} is a series of lists.
11694 Written out more fully, the template looks like this:
11699 (@var{first-true-or-false-test} @var{first-consequent})
11700 (@var{second-true-or-false-test} @var{second-consequent})
11701 (@var{third-true-or-false-test} @var{third-consequent})
11706 When the Lisp interpreter evaluates the @code{cond} expression, it
11707 evaluates the first element (the @sc{car} or true-or-false-test) of
11708 the first expression in a series of expressions within the body of the
11711 If the true-or-false-test returns @code{nil} the rest of that
11712 expression, the consequent, is skipped and the true-or-false-test of the
11713 next expression is evaluated. When an expression is found whose
11714 true-or-false-test returns a value that is not @code{nil}, the
11715 consequent of that expression is evaluated. The consequent can be one
11716 or more expressions. If the consequent consists of more than one
11717 expression, the expressions are evaluated in sequence and the value of
11718 the last one is returned. If the expression does not have a consequent,
11719 the value of the true-or-false-test is returned.
11721 If none of the true-or-false-tests test true, the @code{cond} expression
11722 returns @code{nil}.
11725 Written using @code{cond}, the @code{triangle} function looks like this:
11729 (defun triangle-using-cond (number)
11730 (cond ((<= number 0) 0)
11733 (+ number (triangle-using-cond (1- number))))))
11738 In this example, the @code{cond} returns 0 if the number is less than or
11739 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11740 number (triangle-using-cond (1- number)))} if the number is greater than
11743 @node Recursive Patterns
11744 @subsection Recursive Patterns
11745 @cindex Recursive Patterns
11747 Here are three common recursive patterns. Each involves a list.
11748 Recursion does not need to involve lists, but Lisp is designed for lists
11749 and this provides a sense of its primal capabilities.
11758 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11759 @cindex Every, type of recursive pattern
11760 @cindex Recursive pattern: every
11762 In the @code{every} recursive pattern, an action is performed on every
11766 The basic pattern is:
11770 If a list be empty, return @code{nil}.
11772 Else, act on the beginning of the list (the @sc{car} of the list)
11775 through a recursive call by the function on the rest (the
11776 @sc{cdr}) of the list,
11778 and, optionally, combine the acted-on element, using @code{cons},
11779 with the results of acting on the rest.
11788 (defun square-each (numbers-list)
11789 "Square each of a NUMBERS LIST, recursively."
11790 (if (not numbers-list) ; do-again-test
11793 (* (car numbers-list) (car numbers-list))
11794 (square-each (cdr numbers-list))))) ; next-step-expression
11798 (square-each '(1 2 3))
11805 If @code{numbers-list} is empty, do nothing. But if it has content,
11806 construct a list combining the square of the first number in the list
11807 with the result of the recursive call.
11809 (The example follows the pattern exactly: @code{nil} is returned if
11810 the numbers' list is empty. In practice, you would write the
11811 conditional so it carries out the action when the numbers' list is not
11814 The @code{print-elements-recursively} function (@pxref{Recursion with
11815 list, , Recursion with a List}) is another example of an @code{every}
11816 pattern, except in this case, rather than bring the results together
11817 using @code{cons}, we print each element of output.
11820 The @code{print-elements-recursively} function looks like this:
11824 (setq animals '(gazelle giraffe lion tiger))
11828 (defun print-elements-recursively (list)
11829 "Print each element of LIST on a line of its own.
11831 (when list ; @r{do-again-test}
11832 (print (car list)) ; @r{body}
11833 (print-elements-recursively ; @r{recursive call}
11834 (cdr list)))) ; @r{next-step-expression}
11836 (print-elements-recursively animals)
11841 The pattern for @code{print-elements-recursively} is:
11845 When the list is empty, do nothing.
11847 But when the list has at least one element,
11850 act on the beginning of the list (the @sc{car} of the list),
11852 and make a recursive call on the rest (the @sc{cdr}) of the list.
11857 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11858 @cindex Accumulate, type of recursive pattern
11859 @cindex Recursive pattern: accumulate
11861 Another recursive pattern is called the @code{accumulate} pattern. In
11862 the @code{accumulate} recursive pattern, an action is performed on
11863 every element of a list and the result of that action is accumulated
11864 with the results of performing the action on the other elements.
11866 This is very like the `every' pattern using @code{cons}, except that
11867 @code{cons} is not used, but some other combiner.
11874 If a list be empty, return zero or some other constant.
11876 Else, act on the beginning of the list (the @sc{car} of the list),
11879 and combine that acted-on element, using @code{+} or
11880 some other combining function, with
11882 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11887 Here is an example:
11891 (defun add-elements (numbers-list)
11892 "Add the elements of NUMBERS-LIST together."
11893 (if (not numbers-list)
11895 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
11899 (add-elements '(1 2 3 4))
11904 @xref{Files List, , Making a List of Files}, for an example of the
11905 accumulate pattern.
11908 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
11909 @cindex Keep, type of recursive pattern
11910 @cindex Recursive pattern: keep
11912 A third recursive pattern is called the @code{keep} pattern.
11913 In the @code{keep} recursive pattern, each element of a list is tested;
11914 the element is acted on and the results are kept only if the element
11917 Again, this is very like the `every' pattern, except the element is
11918 skipped unless it meets a criterion.
11921 The pattern has three parts:
11925 If a list be empty, return @code{nil}.
11927 Else, if the beginning of the list (the @sc{car} of the list) passes
11931 act on that element and combine it, using @code{cons} with
11933 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11936 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
11940 skip on that element,
11942 and, recursively call the function on the rest (the @sc{cdr}) of the list.
11947 Here is an example that uses @code{cond}:
11951 (defun keep-three-letter-words (word-list)
11952 "Keep three letter words in WORD-LIST."
11954 ;; First do-again-test: stop-condition
11955 ((not word-list) nil)
11957 ;; Second do-again-test: when to act
11958 ((eq 3 (length (symbol-name (car word-list))))
11959 ;; combine acted-on element with recursive call on shorter list
11960 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
11962 ;; Third do-again-test: when to skip element;
11963 ;; recursively call shorter list with next-step expression
11964 (t (keep-three-letter-words (cdr word-list)))))
11968 (keep-three-letter-words '(one two three four five six))
11969 @result{} (one two six)
11973 It goes without saying that you need not use @code{nil} as the test for
11974 when to stop; and you can, of course, combine these patterns.
11977 @subsection Recursion without Deferments
11978 @cindex Deferment in recursion
11979 @cindex Recursion without Deferments
11981 Let's consider again what happens with the @code{triangle-recursively}
11982 function. We will find that the intermediate calculations are
11983 deferred until all can be done.
11986 Here is the function definition:
11990 (defun triangle-recursively (number)
11991 "Return the sum of the numbers 1 through NUMBER inclusive.
11993 (if (= number 1) ; @r{do-again-test}
11995 (+ number ; @r{else-part}
11996 (triangle-recursively ; @r{recursive call}
11997 (1- number))))) ; @r{next-step-expression}
12001 What happens when we call this function with a argument of 7?
12003 The first instance of the @code{triangle-recursively} function adds
12004 the number 7 to the value returned by a second instance of
12005 @code{triangle-recursively}, an instance that has been passed an
12006 argument of 6. That is to say, the first calculation is:
12009 (+ 7 (triangle-recursively 6))
12013 The first instance of @code{triangle-recursively}---you may want to
12014 think of it as a little robot---cannot complete its job. It must hand
12015 off the calculation for @code{(triangle-recursively 6)} to a second
12016 instance of the program, to a second robot. This second individual is
12017 completely different from the first one; it is, in the jargon, a
12018 `different instantiation'. Or, put another way, it is a different
12019 robot. It is the same model as the first; it calculates triangle
12020 numbers recursively; but it has a different serial number.
12022 And what does @code{(triangle-recursively 6)} return? It returns the
12023 number 6 added to the value returned by evaluating
12024 @code{triangle-recursively} with an argument of 5. Using the robot
12025 metaphor, it asks yet another robot to help it.
12031 (+ 7 6 (triangle-recursively 5))
12035 And what happens next?
12038 (+ 7 6 5 (triangle-recursively 4))
12041 Each time @code{triangle-recursively} is called, except for the last
12042 time, it creates another instance of the program---another robot---and
12043 asks it to make a calculation.
12046 Eventually, the full addition is set up and performed:
12052 This design for the function defers the calculation of the first step
12053 until the second can be done, and defers that until the third can be
12054 done, and so on. Each deferment means the computer must remember what
12055 is being waited on. This is not a problem when there are only a few
12056 steps, as in this example. But it can be a problem when there are
12059 @node No deferment solution
12060 @subsection No Deferment Solution
12061 @cindex No deferment solution
12062 @cindex Defermentless solution
12063 @cindex Solution without deferment
12065 The solution to the problem of deferred operations is to write in a
12066 manner that does not defer operations@footnote{The phrase @dfn{tail
12067 recursive} is used to describe such a process, one that uses
12068 `constant space'.}. This requires
12069 writing to a different pattern, often one that involves writing two
12070 function definitions, an `initialization' function and a `helper'
12073 The `initialization' function sets up the job; the `helper' function
12077 Here are the two function definitions for adding up numbers. They are
12078 so simple, I find them hard to understand.
12082 (defun triangle-initialization (number)
12083 "Return the sum of the numbers 1 through NUMBER inclusive.
12084 This is the `initialization' component of a two function
12085 duo that uses recursion."
12086 (triangle-recursive-helper 0 0 number))
12092 (defun triangle-recursive-helper (sum counter number)
12093 "Return SUM, using COUNTER, through NUMBER inclusive.
12094 This is the `helper' component of a two function duo
12095 that uses recursion."
12096 (if (> counter number)
12098 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12099 (1+ counter) ; @r{counter}
12100 number))) ; @r{number}
12105 Install both function definitions by evaluating them, then call
12106 @code{triangle-initialization} with 2 rows:
12110 (triangle-initialization 2)
12115 The `initialization' function calls the first instance of the `helper'
12116 function with three arguments: zero, zero, and a number which is the
12117 number of rows in the triangle.
12119 The first two arguments passed to the `helper' function are
12120 initialization values. These values are changed when
12121 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12122 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12123 process that is iterative in a procedure that is recursive. The
12124 process is called iterative because the computer need only record the
12125 three values, @code{sum}, @code{counter}, and @code{number}; the
12126 procedure is recursive because the function `calls itself'. On the
12127 other hand, both the process and the procedure used by
12128 @code{triangle-recursively} are called recursive. The word
12129 `recursive' has different meanings in the two contexts.}
12131 Let's see what happens when we have a triangle that has one row. (This
12132 triangle will have one pebble in it!)
12135 @code{triangle-initialization} will call its helper with
12136 the arguments @w{@code{0 0 1}}. That function will run the conditional
12137 test whether @code{(> counter number)}:
12145 and find that the result is false, so it will invoke
12146 the else-part of the @code{if} clause:
12150 (triangle-recursive-helper
12151 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12152 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12153 number) ; @r{number stays the same}
12159 which will first compute:
12163 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12164 (1+ 0) ; @r{counter}
12168 (triangle-recursive-helper 0 1 1)
12172 Again, @code{(> counter number)} will be false, so again, the Lisp
12173 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12174 new instance with new arguments.
12177 This new instance will be;
12181 (triangle-recursive-helper
12182 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12183 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12184 number) ; @r{number stays the same}
12188 (triangle-recursive-helper 1 2 1)
12192 In this case, the @code{(> counter number)} test will be true! So the
12193 instance will return the value of the sum, which will be 1, as
12196 Now, let's pass @code{triangle-initialization} an argument
12197 of 2, to find out how many pebbles there are in a triangle with two rows.
12199 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12202 In stages, the instances called will be:
12206 @r{sum counter number}
12207 (triangle-recursive-helper 0 1 2)
12209 (triangle-recursive-helper 1 2 2)
12211 (triangle-recursive-helper 3 3 2)
12215 When the last instance is called, the @code{(> counter number)} test
12216 will be true, so the instance will return the value of @code{sum},
12219 This kind of pattern helps when you are writing functions that can use
12220 many resources in a computer.
12223 @node Looping exercise
12224 @section Looping Exercise
12228 Write a function similar to @code{triangle} in which each row has a
12229 value which is the square of the row number. Use a @code{while} loop.
12232 Write a function similar to @code{triangle} that multiplies instead of
12236 Rewrite these two functions recursively. Rewrite these functions
12239 @c comma in printed title causes problem in Info cross reference
12241 Write a function for Texinfo mode that creates an index entry at the
12242 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12243 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12244 written in Texinfo.)
12246 Many of the functions you will need are described in two of the
12247 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12248 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12249 @code{forward-paragraph} to put the index entry at the beginning of
12250 the paragraph, you will have to use @w{@kbd{C-h f}}
12251 (@code{describe-function}) to find out how to make the command go
12254 For more information, see
12256 @ref{Indicating, , Indicating Definitions, texinfo}.
12259 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12260 a Texinfo manual in the current directory. Or, if you are on the
12262 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12265 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12266 Documentation Format}.
12270 @node Regexp Search
12271 @chapter Regular Expression Searches
12272 @cindex Searches, illustrating
12273 @cindex Regular expression searches
12274 @cindex Patterns, searching for
12275 @cindex Motion by sentence and paragraph
12276 @cindex Sentences, movement by
12277 @cindex Paragraphs, movement by
12279 Regular expression searches are used extensively in GNU Emacs. The
12280 two functions, @code{forward-sentence} and @code{forward-paragraph},
12281 illustrate these searches well. They use regular expressions to find
12282 where to move point. The phrase `regular expression' is often written
12285 Regular expression searches are described in @ref{Regexp Search, ,
12286 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12287 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12288 Manual}. In writing this chapter, I am presuming that you have at
12289 least a mild acquaintance with them. The major point to remember is
12290 that regular expressions permit you to search for patterns as well as
12291 for literal strings of characters. For example, the code in
12292 @code{forward-sentence} searches for the pattern of possible
12293 characters that could mark the end of a sentence, and moves point to
12296 Before looking at the code for the @code{forward-sentence} function, it
12297 is worth considering what the pattern that marks the end of a sentence
12298 must be. The pattern is discussed in the next section; following that
12299 is a description of the regular expression search function,
12300 @code{re-search-forward}. The @code{forward-sentence} function
12301 is described in the section following. Finally, the
12302 @code{forward-paragraph} function is described in the last section of
12303 this chapter. @code{forward-paragraph} is a complex function that
12304 introduces several new features.
12307 * sentence-end:: The regular expression for @code{sentence-end}.
12308 * re-search-forward:: Very similar to @code{search-forward}.
12309 * forward-sentence:: A straightforward example of regexp search.
12310 * forward-paragraph:: A somewhat complex example.
12311 * etags:: How to create your own @file{TAGS} table.
12313 * re-search Exercises::
12317 @section The Regular Expression for @code{sentence-end}
12318 @findex sentence-end
12320 The symbol @code{sentence-end} is bound to the pattern that marks the
12321 end of a sentence. What should this regular expression be?
12323 Clearly, a sentence may be ended by a period, a question mark, or an
12324 exclamation mark. Indeed, in English, only clauses that end with one
12325 of those three characters should be considered the end of a sentence.
12326 This means that the pattern should include the character set:
12332 However, we do not want @code{forward-sentence} merely to jump to a
12333 period, a question mark, or an exclamation mark, because such a character
12334 might be used in the middle of a sentence. A period, for example, is
12335 used after abbreviations. So other information is needed.
12337 According to convention, you type two spaces after every sentence, but
12338 only one space after a period, a question mark, or an exclamation mark in
12339 the body of a sentence. So a period, a question mark, or an exclamation
12340 mark followed by two spaces is a good indicator of an end of sentence.
12341 However, in a file, the two spaces may instead be a tab or the end of a
12342 line. This means that the regular expression should include these three
12343 items as alternatives.
12346 This group of alternatives will look like this:
12357 Here, @samp{$} indicates the end of the line, and I have pointed out
12358 where the tab and two spaces are inserted in the expression. Both are
12359 inserted by putting the actual characters into the expression.
12361 Two backslashes, @samp{\\}, are required before the parentheses and
12362 vertical bars: the first backslash quotes the following backslash in
12363 Emacs; and the second indicates that the following character, the
12364 parenthesis or the vertical bar, is special.
12367 Also, a sentence may be followed by one or more carriage returns, like
12378 Like tabs and spaces, a carriage return is inserted into a regular
12379 expression by inserting it literally. The asterisk indicates that the
12380 @key{RET} is repeated zero or more times.
12382 But a sentence end does not consist only of a period, a question mark or
12383 an exclamation mark followed by appropriate space: a closing quotation
12384 mark or a closing brace of some kind may precede the space. Indeed more
12385 than one such mark or brace may precede the space. These require a
12386 expression that looks like this:
12392 In this expression, the first @samp{]} is the first character in the
12393 expression; the second character is @samp{"}, which is preceded by a
12394 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12395 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12397 All this suggests what the regular expression pattern for matching the
12398 end of a sentence should be; and, indeed, if we evaluate
12399 @code{sentence-end} we find that it returns the following value:
12404 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12410 (Well, not in GNU Emacs 22; that is because of an effort to make the
12411 process simpler and to handle more glyphs and languages. When the
12412 value of @code{sentence-end} is @code{nil}, then use the value defined
12413 by the function @code{sentence-end}. (Here is a use of the difference
12414 between a value and a function in Emacs Lisp.) The function returns a
12415 value constructed from the variables @code{sentence-end-base},
12416 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12417 and @code{sentence-end-without-space}. The critical variable is
12418 @code{sentence-end-base}; its global value is similar to the one
12419 described above but it also contains two additional quotation marks.
12420 These have differing degrees of curliness. The
12421 @code{sentence-end-without-period} variable, when true, tells Emacs
12422 that a sentence may end without a period, such as text in Thai.)
12426 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12427 literally in the pattern.)
12429 This regular expression can be deciphered as follows:
12433 The first part of the pattern is the three characters, a period, a question
12434 mark and an exclamation mark, within square brackets. The pattern must
12435 begin with one or other of these characters.
12438 The second part of the pattern is the group of closing braces and
12439 quotation marks, which can appear zero or more times. These may follow
12440 the period, question mark or exclamation mark. In a regular expression,
12441 the backslash, @samp{\}, followed by the double quotation mark,
12442 @samp{"}, indicates the class of string-quote characters. Usually, the
12443 double quotation mark is the only character in this class. The
12444 asterisk, @samp{*}, indicates that the items in the previous group (the
12445 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12448 @item \\($\\| \\| \\)
12449 The third part of the pattern is one or other of: either the end of a
12450 line, or two blank spaces, or a tab. The double back-slashes are used
12451 to prevent Emacs from reading the parentheses and vertical bars as part
12452 of the search pattern; the parentheses are used to mark the group and
12453 the vertical bars are used to indicated that the patterns to either side
12454 of them are alternatives. The dollar sign is used to indicate the end
12455 of a line and both the two spaces and the tab are each inserted as is to
12456 indicate what they are.
12459 Finally, the last part of the pattern indicates that the end of the line
12460 or the whitespace following the period, question mark or exclamation
12461 mark may, but need not, be followed by one or more carriage returns. In
12462 the pattern, the carriage return is inserted as an actual carriage
12463 return between square brackets but here it is shown as @key{RET}.
12467 @node re-search-forward
12468 @section The @code{re-search-forward} Function
12469 @findex re-search-forward
12471 The @code{re-search-forward} function is very like the
12472 @code{search-forward} function. (@xref{search-forward, , The
12473 @code{search-forward} Function}.)
12475 @code{re-search-forward} searches for a regular expression. If the
12476 search is successful, it leaves point immediately after the last
12477 character in the target. If the search is backwards, it leaves point
12478 just before the first character in the target. You may tell
12479 @code{re-search-forward} to return @code{t} for true. (Moving point
12480 is therefore a `side effect'.)
12482 Like @code{search-forward}, the @code{re-search-forward} function takes
12487 The first argument is the regular expression that the function searches
12488 for. The regular expression will be a string between quotation marks.
12491 The optional second argument limits how far the function will search; it is a
12492 bound, which is specified as a position in the buffer.
12495 The optional third argument specifies how the function responds to
12496 failure: @code{nil} as the third argument causes the function to
12497 signal an error (and print a message) when the search fails; any other
12498 value causes it to return @code{nil} if the search fails and @code{t}
12499 if the search succeeds.
12502 The optional fourth argument is the repeat count. A negative repeat
12503 count causes @code{re-search-forward} to search backwards.
12507 The template for @code{re-search-forward} looks like this:
12511 (re-search-forward "@var{regular-expression}"
12512 @var{limit-of-search}
12513 @var{what-to-do-if-search-fails}
12514 @var{repeat-count})
12518 The second, third, and fourth arguments are optional. However, if you
12519 want to pass a value to either or both of the last two arguments, you
12520 must also pass a value to all the preceding arguments. Otherwise, the
12521 Lisp interpreter will mistake which argument you are passing the value
12525 In the @code{forward-sentence} function, the regular expression will be
12526 the value of the variable @code{sentence-end}. In simple form, that is:
12530 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12536 The limit of the search will be the end of the paragraph (since a
12537 sentence cannot go beyond a paragraph). If the search fails, the
12538 function will return @code{nil}; and the repeat count will be provided
12539 by the argument to the @code{forward-sentence} function.
12541 @node forward-sentence
12542 @section @code{forward-sentence}
12543 @findex forward-sentence
12545 The command to move the cursor forward a sentence is a straightforward
12546 illustration of how to use regular expression searches in Emacs Lisp.
12547 Indeed, the function looks longer and more complicated than it is; this
12548 is because the function is designed to go backwards as well as forwards;
12549 and, optionally, over more than one sentence. The function is usually
12550 bound to the key command @kbd{M-e}.
12553 * Complete forward-sentence::
12554 * fwd-sentence while loops:: Two @code{while} loops.
12555 * fwd-sentence re-search:: A regular expression search.
12559 @node Complete forward-sentence
12560 @unnumberedsubsec Complete @code{forward-sentence} function definition
12564 Here is the code for @code{forward-sentence}:
12569 (defun forward-sentence (&optional arg)
12570 "Move forward to next `sentence-end'. With argument, repeat.
12571 With negative argument, move backward repeatedly to `sentence-beginning'.
12573 The variable `sentence-end' is a regular expression that matches ends of
12574 sentences. Also, every paragraph boundary terminates sentences as well."
12578 (or arg (setq arg 1))
12579 (let ((opoint (point))
12580 (sentence-end (sentence-end)))
12582 (let ((pos (point))
12583 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12584 (if (and (re-search-backward sentence-end par-beg t)
12585 (or (< (match-end 0) pos)
12586 (re-search-backward sentence-end par-beg t)))
12587 (goto-char (match-end 0))
12588 (goto-char par-beg)))
12589 (setq arg (1+ arg)))
12593 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12594 (if (re-search-forward sentence-end par-end t)
12595 (skip-chars-backward " \t\n")
12596 (goto-char par-end)))
12597 (setq arg (1- arg)))
12598 (constrain-to-field nil opoint t)))
12606 (defun forward-sentence (&optional arg)
12607 "Move forward to next sentence-end. With argument, repeat.
12608 With negative argument, move backward repeatedly to sentence-beginning.
12609 Sentence ends are identified by the value of sentence-end
12610 treated as a regular expression. Also, every paragraph boundary
12611 terminates sentences as well."
12615 (or arg (setq arg 1))
12618 (save-excursion (start-of-paragraph-text) (point))))
12619 (if (re-search-backward
12620 (concat sentence-end "[^ \t\n]") par-beg t)
12621 (goto-char (1- (match-end 0)))
12622 (goto-char par-beg)))
12623 (setq arg (1+ arg)))
12626 (save-excursion (end-of-paragraph-text) (point))))
12627 (if (re-search-forward sentence-end par-end t)
12628 (skip-chars-backward " \t\n")
12629 (goto-char par-end)))
12630 (setq arg (1- arg))))
12635 The function looks long at first sight and it is best to look at its
12636 skeleton first, and then its muscle. The way to see the skeleton is to
12637 look at the expressions that start in the left-most columns:
12641 (defun forward-sentence (&optional arg)
12642 "@var{documentation}@dots{}"
12644 (or arg (setq arg 1))
12645 (let ((opoint (point)) (sentence-end (sentence-end)))
12647 (let ((pos (point))
12648 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12649 @var{rest-of-body-of-while-loop-when-going-backwards}
12651 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12652 @var{rest-of-body-of-while-loop-when-going-forwards}
12653 @var{handle-forms-and-equivalent}
12657 This looks much simpler! The function definition consists of
12658 documentation, an @code{interactive} expression, an @code{or}
12659 expression, a @code{let} expression, and @code{while} loops.
12661 Let's look at each of these parts in turn.
12663 We note that the documentation is thorough and understandable.
12665 The function has an @code{interactive "p"} declaration. This means
12666 that the processed prefix argument, if any, is passed to the
12667 function as its argument. (This will be a number.) If the function
12668 is not passed an argument (it is optional) then the argument
12669 @code{arg} will be bound to 1.
12671 When @code{forward-sentence} is called non-interactively without an
12672 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12673 handles this. What it does is either leave the value of @code{arg} as
12674 it is, but only if @code{arg} is bound to a value; or it sets the
12675 value of @code{arg} to 1, in the case when @code{arg} is bound to
12678 Next is a @code{let}. That specifies the values of two local
12679 variables, @code{point} and @code{sentence-end}. The local value of
12680 point, from before the search, is used in the
12681 @code{constrain-to-field} function which handles forms and
12682 equivalents. The @code{sentence-end} variable is set by the
12683 @code{sentence-end} function.
12685 @node fwd-sentence while loops
12686 @unnumberedsubsec The @code{while} loops
12688 Two @code{while} loops follow. The first @code{while} has a
12689 true-or-false-test that tests true if the prefix argument for
12690 @code{forward-sentence} is a negative number. This is for going
12691 backwards. The body of this loop is similar to the body of the second
12692 @code{while} clause, but it is not exactly the same. We will skip
12693 this @code{while} loop and concentrate on the second @code{while}
12697 The second @code{while} loop is for moving point forward. Its skeleton
12702 (while (> arg 0) ; @r{true-or-false-test}
12704 (if (@var{true-or-false-test})
12707 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12711 The @code{while} loop is of the decrementing kind.
12712 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12713 has a true-or-false-test that tests true so long as the counter (in
12714 this case, the variable @code{arg}) is greater than zero; and it has a
12715 decrementer that subtracts 1 from the value of the counter every time
12718 If no prefix argument is given to @code{forward-sentence}, which is
12719 the most common way the command is used, this @code{while} loop will
12720 run once, since the value of @code{arg} will be 1.
12722 The body of the @code{while} loop consists of a @code{let} expression,
12723 which creates and binds a local variable, and has, as its body, an
12724 @code{if} expression.
12727 The body of the @code{while} loop looks like this:
12732 (save-excursion (end-of-paragraph-text) (point))))
12733 (if (re-search-forward sentence-end par-end t)
12734 (skip-chars-backward " \t\n")
12735 (goto-char par-end)))
12739 The @code{let} expression creates and binds the local variable
12740 @code{par-end}. As we shall see, this local variable is designed to
12741 provide a bound or limit to the regular expression search. If the
12742 search fails to find a proper sentence ending in the paragraph, it will
12743 stop on reaching the end of the paragraph.
12745 But first, let us examine how @code{par-end} is bound to the value of
12746 the end of the paragraph. What happens is that the @code{let} sets the
12747 value of @code{par-end} to the value returned when the Lisp interpreter
12748 evaluates the expression
12752 (save-excursion (end-of-paragraph-text) (point))
12757 In this expression, @code{(end-of-paragraph-text)} moves point to the
12758 end of the paragraph, @code{(point)} returns the value of point, and then
12759 @code{save-excursion} restores point to its original position. Thus,
12760 the @code{let} binds @code{par-end} to the value returned by the
12761 @code{save-excursion} expression, which is the position of the end of
12762 the paragraph. (The @code{end-of-paragraph-text} function uses
12763 @code{forward-paragraph}, which we will discuss shortly.)
12766 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12767 expression that looks like this:
12771 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12772 (skip-chars-backward " \t\n") ; @r{then-part}
12773 (goto-char par-end))) ; @r{else-part}
12777 The @code{if} tests whether its first argument is true and if so,
12778 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12779 evaluates the else-part. The true-or-false-test of the @code{if}
12780 expression is the regular expression search.
12782 It may seem odd to have what looks like the `real work' of
12783 the @code{forward-sentence} function buried here, but this is a common
12784 way this kind of operation is carried out in Lisp.
12786 @node fwd-sentence re-search
12787 @unnumberedsubsec The regular expression search
12789 The @code{re-search-forward} function searches for the end of the
12790 sentence, that is, for the pattern defined by the @code{sentence-end}
12791 regular expression. If the pattern is found---if the end of the sentence is
12792 found---then the @code{re-search-forward} function does two things:
12796 The @code{re-search-forward} function carries out a side effect, which
12797 is to move point to the end of the occurrence found.
12800 The @code{re-search-forward} function returns a value of true. This is
12801 the value received by the @code{if}, and means that the search was
12806 The side effect, the movement of point, is completed before the
12807 @code{if} function is handed the value returned by the successful
12808 conclusion of the search.
12810 When the @code{if} function receives the value of true from a successful
12811 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12812 which is the expression @code{(skip-chars-backward " \t\n")}. This
12813 expression moves backwards over any blank spaces, tabs or carriage
12814 returns until a printed character is found and then leaves point after
12815 the character. Since point has already been moved to the end of the
12816 pattern that marks the end of the sentence, this action leaves point
12817 right after the closing printed character of the sentence, which is
12820 On the other hand, if the @code{re-search-forward} function fails to
12821 find a pattern marking the end of the sentence, the function returns
12822 false. The false then causes the @code{if} to evaluate its third
12823 argument, which is @code{(goto-char par-end)}: it moves point to the
12824 end of the paragraph.
12826 (And if the text is in a form or equivalent, and point may not move
12827 fully, then the @code{constrain-to-field} function comes into play.)
12829 Regular expression searches are exceptionally useful and the pattern
12830 illustrated by @code{re-search-forward}, in which the search is the
12831 test of an @code{if} expression, is handy. You will see or write code
12832 incorporating this pattern often.
12834 @node forward-paragraph
12835 @section @code{forward-paragraph}: a Goldmine of Functions
12836 @findex forward-paragraph
12840 (defun forward-paragraph (&optional arg)
12841 "Move forward to end of paragraph.
12842 With argument ARG, do it ARG times;
12843 a negative argument ARG = -N means move backward N paragraphs.
12845 A line which `paragraph-start' matches either separates paragraphs
12846 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12847 A paragraph end is the beginning of a line which is not part of the paragraph
12848 to which the end of the previous line belongs, or the end of the buffer.
12849 Returns the count of paragraphs left to move."
12851 (or arg (setq arg 1))
12852 (let* ((opoint (point))
12853 (fill-prefix-regexp
12854 (and fill-prefix (not (equal fill-prefix ""))
12855 (not paragraph-ignore-fill-prefix)
12856 (regexp-quote fill-prefix)))
12857 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12858 ;; These regexps shouldn't be anchored, because we look for them
12859 ;; starting at the left-margin. This allows paragraph commands to
12860 ;; work normally with indented text.
12861 ;; This hack will not find problem cases like "whatever\\|^something".
12862 (parstart (if (and (not (equal "" paragraph-start))
12863 (equal ?^ (aref paragraph-start 0)))
12864 (substring paragraph-start 1)
12866 (parsep (if (and (not (equal "" paragraph-separate))
12867 (equal ?^ (aref paragraph-separate 0)))
12868 (substring paragraph-separate 1)
12869 paragraph-separate))
12871 (if fill-prefix-regexp
12872 (concat parsep "\\|"
12873 fill-prefix-regexp "[ \t]*$")
12875 ;; This is used for searching.
12876 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
12878 (while (and (< arg 0) (not (bobp)))
12879 (if (and (not (looking-at parsep))
12880 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
12881 (looking-at parsep))
12882 (setq arg (1+ arg))
12883 (setq start (point))
12884 ;; Move back over paragraph-separating lines.
12885 (forward-char -1) (beginning-of-line)
12886 (while (and (not (bobp))
12887 (progn (move-to-left-margin)
12888 (looking-at parsep)))
12892 (setq arg (1+ arg))
12893 ;; Go to end of the previous (non-separating) line.
12895 ;; Search back for line that starts or separates paragraphs.
12896 (if (if fill-prefix-regexp
12897 ;; There is a fill prefix; it overrides parstart.
12898 (let (multiple-lines)
12899 (while (and (progn (beginning-of-line) (not (bobp)))
12900 (progn (move-to-left-margin)
12901 (not (looking-at parsep)))
12902 (looking-at fill-prefix-regexp))
12903 (unless (= (point) start)
12904 (setq multiple-lines t))
12906 (move-to-left-margin)
12907 ;; This deleted code caused a long hanging-indent line
12908 ;; not to be filled together with the following lines.
12909 ;; ;; Don't move back over a line before the paragraph
12910 ;; ;; which doesn't start with fill-prefix
12911 ;; ;; unless that is the only line we've moved over.
12912 ;; (and (not (looking-at fill-prefix-regexp))
12914 ;; (forward-line 1))
12916 (while (and (re-search-backward sp-parstart nil 1)
12917 (setq found-start t)
12918 ;; Found a candidate, but need to check if it is a
12920 (progn (setq start (point))
12921 (move-to-left-margin)
12922 (not (looking-at parsep)))
12923 (not (and (looking-at parstart)
12924 (or (not use-hard-newlines)
12927 (1- start) 'hard)))))
12928 (setq found-start nil)
12933 ;; Move forward over paragraph separators.
12934 ;; We know this cannot reach the place we started
12935 ;; because we know we moved back over a non-separator.
12936 (while (and (not (eobp))
12937 (progn (move-to-left-margin)
12938 (looking-at parsep)))
12940 ;; If line before paragraph is just margin, back up to there.
12942 (if (> (current-column) (current-left-margin))
12944 (skip-chars-backward " \t")
12946 (forward-line 1))))
12947 ;; No starter or separator line => use buffer beg.
12948 (goto-char (point-min))))))
12950 (while (and (> arg 0) (not (eobp)))
12951 ;; Move forward over separator lines...
12952 (while (and (not (eobp))
12953 (progn (move-to-left-margin) (not (eobp)))
12954 (looking-at parsep))
12956 (unless (eobp) (setq arg (1- arg)))
12957 ;; ... and one more line.
12959 (if fill-prefix-regexp
12960 ;; There is a fill prefix; it overrides parstart.
12961 (while (and (not (eobp))
12962 (progn (move-to-left-margin) (not (eobp)))
12963 (not (looking-at parsep))
12964 (looking-at fill-prefix-regexp))
12966 (while (and (re-search-forward sp-parstart nil 1)
12967 (progn (setq start (match-beginning 0))
12970 (progn (move-to-left-margin)
12971 (not (looking-at parsep)))
12972 (or (not (looking-at parstart))
12973 (and use-hard-newlines
12974 (not (get-text-property (1- start) 'hard)))))
12976 (if (< (point) (point-max))
12977 (goto-char start))))
12978 (constrain-to-field nil opoint t)
12979 ;; Return the number of steps that could not be done.
12983 The @code{forward-paragraph} function moves point forward to the end
12984 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
12985 number of functions that are important in themselves, including
12986 @code{let*}, @code{match-beginning}, and @code{looking-at}.
12988 The function definition for @code{forward-paragraph} is considerably
12989 longer than the function definition for @code{forward-sentence}
12990 because it works with a paragraph, each line of which may begin with a
12993 A fill prefix consists of a string of characters that are repeated at
12994 the beginning of each line. For example, in Lisp code, it is a
12995 convention to start each line of a paragraph-long comment with
12996 @samp{;;; }. In Text mode, four blank spaces make up another common
12997 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
12998 emacs, The GNU Emacs Manual}, for more information about fill
13001 The existence of a fill prefix means that in addition to being able to
13002 find the end of a paragraph whose lines begin on the left-most
13003 column, the @code{forward-paragraph} function must be able to find the
13004 end of a paragraph when all or many of the lines in the buffer begin
13005 with the fill prefix.
13007 Moreover, it is sometimes practical to ignore a fill prefix that
13008 exists, especially when blank lines separate paragraphs.
13009 This is an added complication.
13012 * forward-paragraph in brief:: Key parts of the function definition.
13013 * fwd-para let:: The @code{let*} expression.
13014 * fwd-para while:: The forward motion @code{while} loop.
13018 @node forward-paragraph in brief
13019 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13022 Rather than print all of the @code{forward-paragraph} function, we
13023 will only print parts of it. Read without preparation, the function
13027 In outline, the function looks like this:
13031 (defun forward-paragraph (&optional arg)
13032 "@var{documentation}@dots{}"
13034 (or arg (setq arg 1))
13037 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13039 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13044 The first parts of the function are routine: the function's argument
13045 list consists of one optional argument. Documentation follows.
13047 The lower case @samp{p} in the @code{interactive} declaration means
13048 that the processed prefix argument, if any, is passed to the function.
13049 This will be a number, and is the repeat count of how many paragraphs
13050 point will move. The @code{or} expression in the next line handles
13051 the common case when no argument is passed to the function, which occurs
13052 if the function is called from other code rather than interactively.
13053 This case was described earlier. (@xref{forward-sentence, The
13054 @code{forward-sentence} function}.) Now we reach the end of the
13055 familiar part of this function.
13058 @unnumberedsubsec The @code{let*} expression
13060 The next line of the @code{forward-paragraph} function begins a
13061 @code{let*} expression. This is a different than @code{let}. The
13062 symbol is @code{let*} not @code{let}.
13064 The @code{let*} special form is like @code{let} except that Emacs sets
13065 each variable in sequence, one after another, and variables in the
13066 latter part of the varlist can make use of the values to which Emacs
13067 set variables in the earlier part of the varlist.
13070 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13073 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13075 In the @code{let*} expression in this function, Emacs binds a total of
13076 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13077 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13078 @code{found-start}.
13080 The variable @code{parsep} appears twice, first, to remove instances
13081 of @samp{^}, and second, to handle fill prefixes.
13083 The variable @code{opoint} is just the value of @code{point}. As you
13084 can guess, it is used in a @code{constrain-to-field} expression, just
13085 as in @code{forward-sentence}.
13087 The variable @code{fill-prefix-regexp} is set to the value returned by
13088 evaluating the following list:
13093 (not (equal fill-prefix ""))
13094 (not paragraph-ignore-fill-prefix)
13095 (regexp-quote fill-prefix))
13100 This is an expression whose first element is the @code{and} special form.
13102 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13103 function}), the @code{and} special form evaluates each of its
13104 arguments until one of the arguments returns a value of @code{nil}, in
13105 which case the @code{and} expression returns @code{nil}; however, if
13106 none of the arguments returns a value of @code{nil}, the value
13107 resulting from evaluating the last argument is returned. (Since such
13108 a value is not @code{nil}, it is considered true in Lisp.) In other
13109 words, an @code{and} expression returns a true value only if all its
13110 arguments are true.
13113 In this case, the variable @code{fill-prefix-regexp} is bound to a
13114 non-@code{nil} value only if the following four expressions produce a
13115 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13116 @code{fill-prefix-regexp} is bound to @code{nil}.
13120 When this variable is evaluated, the value of the fill prefix, if any,
13121 is returned. If there is no fill prefix, this variable returns
13124 @item (not (equal fill-prefix "")
13125 This expression checks whether an existing fill prefix is an empty
13126 string, that is, a string with no characters in it. An empty string is
13127 not a useful fill prefix.
13129 @item (not paragraph-ignore-fill-prefix)
13130 This expression returns @code{nil} if the variable
13131 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13132 true value such as @code{t}.
13134 @item (regexp-quote fill-prefix)
13135 This is the last argument to the @code{and} special form. If all the
13136 arguments to the @code{and} are true, the value resulting from
13137 evaluating this expression will be returned by the @code{and} expression
13138 and bound to the variable @code{fill-prefix-regexp},
13141 @findex regexp-quote
13143 The result of evaluating this @code{and} expression successfully is that
13144 @code{fill-prefix-regexp} will be bound to the value of
13145 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13146 What @code{regexp-quote} does is read a string and return a regular
13147 expression that will exactly match the string and match nothing else.
13148 This means that @code{fill-prefix-regexp} will be set to a value that
13149 will exactly match the fill prefix if the fill prefix exists.
13150 Otherwise, the variable will be set to @code{nil}.
13152 The next two local variables in the @code{let*} expression are
13153 designed to remove instances of @samp{^} from @code{parstart} and
13154 @code{parsep}, the local variables which indicate the paragraph start
13155 and the paragraph separator. The next expression sets @code{parsep}
13156 again. That is to handle fill prefixes.
13158 This is the setting that requires the definition call @code{let*}
13159 rather than @code{let}. The true-or-false-test for the @code{if}
13160 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13161 @code{nil} or some other value.
13163 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13164 the else-part of the @code{if} expression and binds @code{parsep} to
13165 its local value. (@code{parsep} is a regular expression that matches
13166 what separates paragraphs.)
13168 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13169 the then-part of the @code{if} expression and binds @code{parsep} to a
13170 regular expression that includes the @code{fill-prefix-regexp} as part
13173 Specifically, @code{parsep} is set to the original value of the
13174 paragraph separate regular expression concatenated with an alternative
13175 expression that consists of the @code{fill-prefix-regexp} followed by
13176 optional whitespace to the end of the line. The whitespace is defined
13177 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13178 regexp as an alternative to @code{parsep}.
13180 According to a comment in the code, the next local variable,
13181 @code{sp-parstart}, is used for searching, and then the final two,
13182 @code{start} and @code{found-start}, are set to @code{nil}.
13184 Now we get into the body of the @code{let*}. The first part of the body
13185 of the @code{let*} deals with the case when the function is given a
13186 negative argument and is therefore moving backwards. We will skip this
13189 @node fwd-para while
13190 @unnumberedsubsec The forward motion @code{while} loop
13192 The second part of the body of the @code{let*} deals with forward
13193 motion. It is a @code{while} loop that repeats itself so long as the
13194 value of @code{arg} is greater than zero. In the most common use of
13195 the function, the value of the argument is 1, so the body of the
13196 @code{while} loop is evaluated exactly once, and the cursor moves
13197 forward one paragraph.
13200 (while (and (> arg 0) (not (eobp)))
13202 ;; Move forward over separator lines...
13203 (while (and (not (eobp))
13204 (progn (move-to-left-margin) (not (eobp)))
13205 (looking-at parsep))
13207 (unless (eobp) (setq arg (1- arg)))
13208 ;; ... and one more line.
13211 (if fill-prefix-regexp
13212 ;; There is a fill prefix; it overrides parstart.
13213 (while (and (not (eobp))
13214 (progn (move-to-left-margin) (not (eobp)))
13215 (not (looking-at parsep))
13216 (looking-at fill-prefix-regexp))
13219 (while (and (re-search-forward sp-parstart nil 1)
13220 (progn (setq start (match-beginning 0))
13223 (progn (move-to-left-margin)
13224 (not (looking-at parsep)))
13225 (or (not (looking-at parstart))
13226 (and use-hard-newlines
13227 (not (get-text-property (1- start) 'hard)))))
13230 (if (< (point) (point-max))
13231 (goto-char start))))
13234 This part handles three situations: when point is between paragraphs,
13235 when there is a fill prefix and when there is no fill prefix.
13238 The @code{while} loop looks like this:
13242 ;; @r{going forwards and not at the end of the buffer}
13243 (while (and (> arg 0) (not (eobp)))
13245 ;; @r{between paragraphs}
13246 ;; Move forward over separator lines...
13247 (while (and (not (eobp))
13248 (progn (move-to-left-margin) (not (eobp)))
13249 (looking-at parsep))
13251 ;; @r{This decrements the loop}
13252 (unless (eobp) (setq arg (1- arg)))
13253 ;; ... and one more line.
13258 (if fill-prefix-regexp
13259 ;; There is a fill prefix; it overrides parstart;
13260 ;; we go forward line by line
13261 (while (and (not (eobp))
13262 (progn (move-to-left-margin) (not (eobp)))
13263 (not (looking-at parsep))
13264 (looking-at fill-prefix-regexp))
13269 ;; There is no fill prefix;
13270 ;; we go forward character by character
13271 (while (and (re-search-forward sp-parstart nil 1)
13272 (progn (setq start (match-beginning 0))
13275 (progn (move-to-left-margin)
13276 (not (looking-at parsep)))
13277 (or (not (looking-at parstart))
13278 (and use-hard-newlines
13279 (not (get-text-property (1- start) 'hard)))))
13284 ;; and if there is no fill prefix and if we are not at the end,
13285 ;; go to whatever was found in the regular expression search
13287 (if (< (point) (point-max))
13288 (goto-char start))))
13293 We can see that this is a decrementing counter @code{while} loop,
13294 using the expression @code{(setq arg (1- arg))} as the decrementer.
13295 That expression is not far from the @code{while}, but is hidden in
13296 another Lisp macro, an @code{unless} macro. Unless we are at the end
13297 of the buffer---that is what the @code{eobp} function determines; it
13298 is an abbreviation of @samp{End Of Buffer P}---we decrease the value
13299 of @code{arg} by one.
13301 (If we are at the end of the buffer, we cannot go forward any more and
13302 the next loop of the @code{while} expression will test false since the
13303 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13304 function means exactly as you expect; it is another name for
13305 @code{null}, a function that returns true when its argument is false.)
13307 Interestingly, the loop count is not decremented until we leave the
13308 space between paragraphs, unless we come to the end of buffer or stop
13309 seeing the local value of the paragraph separator.
13311 That second @code{while} also has a @code{(move-to-left-margin)}
13312 expression. The function is self-explanatory. It is inside a
13313 @code{progn} expression and not the last element of its body, so it is
13314 only invoked for its side effect, which is to move point to the left
13315 margin of the current line.
13318 The @code{looking-at} function is also self-explanatory; it returns
13319 true if the text after point matches the regular expression given as
13322 The rest of the body of the loop looks difficult at first, but makes
13323 sense as you come to understand it.
13326 First consider what happens if there is a fill prefix:
13330 (if fill-prefix-regexp
13331 ;; There is a fill prefix; it overrides parstart;
13332 ;; we go forward line by line
13333 (while (and (not (eobp))
13334 (progn (move-to-left-margin) (not (eobp)))
13335 (not (looking-at parsep))
13336 (looking-at fill-prefix-regexp))
13342 This expression moves point forward line by line so long
13343 as four conditions are true:
13347 Point is not at the end of the buffer.
13350 We can move to the left margin of the text and are
13351 not at the end of the buffer.
13354 The text following point does not separate paragraphs.
13357 The pattern following point is the fill prefix regular expression.
13360 The last condition may be puzzling, until you remember that point was
13361 moved to the beginning of the line early in the @code{forward-paragraph}
13362 function. This means that if the text has a fill prefix, the
13363 @code{looking-at} function will see it.
13366 Consider what happens when there is no fill prefix.
13370 (while (and (re-search-forward sp-parstart nil 1)
13371 (progn (setq start (match-beginning 0))
13374 (progn (move-to-left-margin)
13375 (not (looking-at parsep)))
13376 (or (not (looking-at parstart))
13377 (and use-hard-newlines
13378 (not (get-text-property (1- start) 'hard)))))
13384 This @code{while} loop has us searching forward for
13385 @code{sp-parstart}, which is the combination of possible whitespace
13386 with a the local value of the start of a paragraph or of a paragraph
13387 separator. (The latter two are within an expression starting
13388 @code{\(?:} so that they are not referenced by the
13389 @code{match-beginning} function.)
13392 The two expressions,
13396 (setq start (match-beginning 0))
13402 mean go to the start of the text matched by the regular expression
13405 The @code{(match-beginning 0)} expression is new. It returns a number
13406 specifying the location of the start of the text that was matched by
13409 The @code{match-beginning} function is used here because of a
13410 characteristic of a forward search: a successful forward search,
13411 regardless of whether it is a plain search or a regular expression
13412 search, moves point to the end of the text that is found. In this
13413 case, a successful search moves point to the end of the pattern for
13414 @code{sp-parstart}.
13416 However, we want to put point at the end of the current paragraph, not
13417 somewhere else. Indeed, since the search possibly includes the
13418 paragraph separator, point may end up at the beginning of the next one
13419 unless we use an expression that includes @code{match-beginning}.
13421 @findex match-beginning
13422 When given an argument of 0, @code{match-beginning} returns the
13423 position that is the start of the text matched by the most recent
13424 search. In this case, the most recent search looks for
13425 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13426 the beginning position of that pattern, rather than the end position
13429 (Incidentally, when passed a positive number as an argument, the
13430 @code{match-beginning} function returns the location of point at that
13431 parenthesized expression in the last search unless that parenthesized
13432 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13433 appears here since the argument is 0.)
13436 The last expression when there is no fill prefix is
13440 (if (< (point) (point-max))
13441 (goto-char start))))
13446 This says that if there is no fill prefix and if we are not at the
13447 end, point should move to the beginning of whatever was found by the
13448 regular expression search for @code{sp-parstart}.
13450 The full definition for the @code{forward-paragraph} function not only
13451 includes code for going forwards, but also code for going backwards.
13453 If you are reading this inside of GNU Emacs and you want to see the
13454 whole function, you can type @kbd{C-h f} (@code{describe-function})
13455 and the name of the function. This gives you the function
13456 documentation and the name of the library containing the function's
13457 source. Place point over the name of the library and press the RET
13458 key; you will be taken directly to the source. (Be sure to install
13459 your sources! Without them, you are like a person who tries to drive
13460 a car with his eyes shut!)
13463 @section Create Your Own @file{TAGS} File
13465 @cindex @file{TAGS} file, create own
13467 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13468 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13469 name of the function when prompted for it. This is a good habit to
13470 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13471 to the source for a function, variable, or node. The function depends
13472 on tags tables to tell it where to go.
13474 If the @code{find-tag} function first asks you for the name of a
13475 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13476 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13477 @file{TAGS} file depends on how your copy of Emacs was installed. I
13478 just told you the location that provides both my C and my Emacs Lisp
13481 You can also create your own @file{TAGS} file for directories that
13484 You often need to build and install tags tables yourself. They are
13485 not built automatically. A tags table is called a @file{TAGS} file;
13486 the name is in upper case letters.
13488 You can create a @file{TAGS} file by calling the @code{etags} program
13489 that comes as a part of the Emacs distribution. Usually, @code{etags}
13490 is compiled and installed when Emacs is built. (@code{etags} is not
13491 an Emacs Lisp function or a part of Emacs; it is a C program.)
13494 To create a @file{TAGS} file, first switch to the directory in which
13495 you want to create the file. In Emacs you can do this with the
13496 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13497 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13498 compile command, with @w{@code{etags *.el}} as the command to execute
13501 M-x compile RET etags *.el RET
13505 to create a @file{TAGS} file for Emacs Lisp.
13507 For example, if you have a large number of files in your
13508 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13509 of which I load 12---you can create a @file{TAGS} file for the Emacs
13510 Lisp files in that directory.
13513 The @code{etags} program takes all the usual shell `wildcards'. For
13514 example, if you have two directories for which you want a single
13515 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13516 @file{../elisp/} is the second directory:
13519 M-x compile RET etags *.el ../elisp/*.el RET
13526 M-x compile RET etags --help RET
13530 to see a list of the options accepted by @code{etags} as well as a
13531 list of supported languages.
13533 The @code{etags} program handles more than 20 languages, including
13534 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13535 LaTeX, Pascal, Perl, PostScript, Python, TeX, Texinfo, makefiles, and
13536 most assemblers. The program has no switches for specifying the
13537 language; it recognizes the language in an input file according to its
13538 file name and contents.
13540 @file{etags} is very helpful when you are writing code yourself and
13541 want to refer back to functions you have already written. Just run
13542 @code{etags} again at intervals as you write new functions, so they
13543 become part of the @file{TAGS} file.
13545 If you think an appropriate @file{TAGS} file already exists for what
13546 you want, but do not know where it is, you can use the @code{locate}
13547 program to attempt to find it.
13549 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13550 for you the full path names of all your @file{TAGS} files. On my
13551 system, this command lists 34 @file{TAGS} files. On the other hand, a
13552 `plain vanilla' system I recently installed did not contain any
13555 If the tags table you want has been created, you can use the @code{M-x
13556 visit-tags-table} command to specify it. Otherwise, you will need to
13557 create the tag table yourself and then use @code{M-x
13560 @subsubheading Building Tags in the Emacs sources
13561 @cindex Building Tags in the Emacs sources
13562 @cindex Tags in the Emacs sources
13565 The GNU Emacs sources come with a @file{Makefile} that contains a
13566 sophisticated @code{etags} command that creates, collects, and merges
13567 tags tables from all over the Emacs sources and puts the information
13568 into one @file{TAGS} file in the @file{src/} directory. (The
13569 @file{src/} directory is below the top level of your Emacs directory.)
13572 To build this @file{TAGS} file, go to the top level of your Emacs
13573 source directory and run the compile command @code{make tags}:
13576 M-x compile RET make tags RET
13580 (The @code{make tags} command works well with the GNU Emacs sources,
13581 as well as with some other source packages.)
13583 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13586 @node Regexp Review
13589 Here is a brief summary of some recently introduced functions.
13593 Repeatedly evaluate the body of the expression so long as the first
13594 element of the body tests true. Then return @code{nil}. (The
13595 expression is evaluated only for its side effects.)
13604 (insert (format "foo is %d.\n" foo))
13605 (setq foo (1- foo))))
13607 @result{} foo is 2.
13614 (The @code{insert} function inserts its arguments at point; the
13615 @code{format} function returns a string formatted from its arguments
13616 the way @code{message} formats its arguments; @code{\n} produces a new
13619 @item re-search-forward
13620 Search for a pattern, and if the pattern is found, move point to rest
13624 Takes four arguments, like @code{search-forward}:
13628 A regular expression that specifies the pattern to search for.
13629 (Remember to put quotation marks around this argument!)
13632 Optionally, the limit of the search.
13635 Optionally, what to do if the search fails, return @code{nil} or an
13639 Optionally, how many times to repeat the search; if negative, the
13640 search goes backwards.
13644 Bind some variables locally to particular values,
13645 and then evaluate the remaining arguments, returning the value of the
13646 last one. While binding the local variables, use the local values of
13647 variables bound earlier, if any.
13656 (message "`bar' is %d." bar))
13657 @result{} `bar' is 21.
13661 @item match-beginning
13662 Return the position of the start of the text found by the last regular
13666 Return @code{t} for true if the text after point matches the argument,
13667 which should be a regular expression.
13670 Return @code{t} for true if point is at the end of the accessible part
13671 of a buffer. The end of the accessible part is the end of the buffer
13672 if the buffer is not narrowed; it is the end of the narrowed part if
13673 the buffer is narrowed.
13677 @node re-search Exercises
13678 @section Exercises with @code{re-search-forward}
13682 Write a function to search for a regular expression that matches two
13683 or more blank lines in sequence.
13686 Write a function to search for duplicated words, such as `the the'.
13687 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13688 Manual}, for information on how to write a regexp (a regular
13689 expression) to match a string that is composed of two identical
13690 halves. You can devise several regexps; some are better than others.
13691 The function I use is described in an appendix, along with several
13692 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13695 @node Counting Words
13696 @chapter Counting: Repetition and Regexps
13697 @cindex Repetition for word counting
13698 @cindex Regular expressions for word counting
13700 Repetition and regular expression searches are powerful tools that you
13701 often use when you write code in Emacs Lisp. This chapter illustrates
13702 the use of regular expression searches through the construction of
13703 word count commands using @code{while} loops and recursion.
13706 * Why Count Words::
13707 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
13708 * recursive-count-words:: Start with case of no words in region.
13709 * Counting Exercise::
13713 @node Why Count Words
13714 @unnumberedsec Counting words
13717 The standard Emacs distribution contains functions for counting the
13718 number of lines and words within a region.
13720 Certain types of writing ask you to count words. Thus, if you write
13721 an essay, you may be limited to 800 words; if you write a novel, you
13722 may discipline yourself to write 1000 words a day. It seems odd, but
13723 for a long time, Emacs lacked a word count command. Perhaps people used
13724 Emacs mostly for code or types of documentation that did not require
13725 word counts; or perhaps they restricted themselves to the operating
13726 system word count command, @code{wc}. Alternatively, people may have
13727 followed the publishers' convention and computed a word count by
13728 dividing the number of characters in a document by five.
13730 There are many ways to implement a command to count words. Here are
13731 some examples, which you may wish to compare with the standard Emacs
13732 command, @code{count-words-region}.
13734 @node @value{COUNT-WORDS}
13735 @section The @code{@value{COUNT-WORDS}} Function
13736 @findex @value{COUNT-WORDS}
13738 A word count command could count words in a line, paragraph, region,
13739 or buffer. What should the command cover? You could design the
13740 command to count the number of words in a complete buffer. However,
13741 the Emacs tradition encourages flexibility---you may want to count
13742 words in just a section, rather than all of a buffer. So it makes
13743 more sense to design the command to count the number of words in a
13744 region. Once you have a command to count words in a region, you can,
13745 if you wish, count words in a whole buffer by marking it with
13746 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13748 Clearly, counting words is a repetitive act: starting from the
13749 beginning of the region, you count the first word, then the second
13750 word, then the third word, and so on, until you reach the end of the
13751 region. This means that word counting is ideally suited to recursion
13752 or to a @code{while} loop.
13755 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
13756 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
13760 @node Design @value{COUNT-WORDS}
13761 @unnumberedsubsec Designing @code{@value{COUNT-WORDS}}
13764 First, we will implement the word count command with a @code{while}
13765 loop, then with recursion. The command will, of course, be
13769 The template for an interactive function definition is, as always:
13773 (defun @var{name-of-function} (@var{argument-list})
13774 "@var{documentation}@dots{}"
13775 (@var{interactive-expression}@dots{})
13780 What we need to do is fill in the slots.
13782 The name of the function should be self-explanatory and similar to the
13783 existing @code{count-lines-region} name. This makes the name easier
13784 to remember. @code{count-words-region} is the obvious choice. Since
13785 that name is now used for the standard Emacs command to count words, we
13786 will name our implementation @code{@value{COUNT-WORDS}}.
13788 The function counts words within a region. This means that the
13789 argument list must contain symbols that are bound to the two
13790 positions, the beginning and end of the region. These two positions
13791 can be called @samp{beginning} and @samp{end} respectively. The first
13792 line of the documentation should be a single sentence, since that is
13793 all that is printed as documentation by a command such as
13794 @code{apropos}. The interactive expression will be of the form
13795 @samp{(interactive "r")}, since that will cause Emacs to pass the
13796 beginning and end of the region to the function's argument list. All
13799 The body of the function needs to be written to do three tasks:
13800 first, to set up conditions under which the @code{while} loop can
13801 count words, second, to run the @code{while} loop, and third, to send
13802 a message to the user.
13804 When a user calls @code{@value{COUNT-WORDS}}, point may be at the
13805 beginning or the end of the region. However, the counting process
13806 must start at the beginning of the region. This means we will want
13807 to put point there if it is not already there. Executing
13808 @code{(goto-char beginning)} ensures this. Of course, we will want to
13809 return point to its expected position when the function finishes its
13810 work. For this reason, the body must be enclosed in a
13811 @code{save-excursion} expression.
13813 The central part of the body of the function consists of a
13814 @code{while} loop in which one expression jumps point forward word by
13815 word, and another expression counts those jumps. The true-or-false-test
13816 of the @code{while} loop should test true so long as point should jump
13817 forward, and false when point is at the end of the region.
13819 We could use @code{(forward-word 1)} as the expression for moving point
13820 forward word by word, but it is easier to see what Emacs identifies as a
13821 `word' if we use a regular expression search.
13823 A regular expression search that finds the pattern for which it is
13824 searching leaves point after the last character matched. This means
13825 that a succession of successful word searches will move point forward
13828 As a practical matter, we want the regular expression search to jump
13829 over whitespace and punctuation between words as well as over the
13830 words themselves. A regexp that refuses to jump over interword
13831 whitespace would never jump more than one word! This means that
13832 the regexp should include the whitespace and punctuation that follows
13833 a word, if any, as well as the word itself. (A word may end a buffer
13834 and not have any following whitespace or punctuation, so that part of
13835 the regexp must be optional.)
13837 Thus, what we want for the regexp is a pattern defining one or more
13838 word constituent characters followed, optionally, by one or more
13839 characters that are not word constituents. The regular expression for
13847 The buffer's syntax table determines which characters are and are not
13848 word constituents. For more information about syntax,
13849 @pxref{Syntax Tables, , Syntax Tables, elisp, The GNU Emacs Lisp
13853 The search expression looks like this:
13856 (re-search-forward "\\w+\\W*")
13860 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13861 single backslash has special meaning to the Emacs Lisp interpreter.
13862 It indicates that the following character is interpreted differently
13863 than usual. For example, the two characters, @samp{\n}, stand for
13864 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13865 backslashes in a row stand for an ordinary, `unspecial' backslash, so
13866 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13867 letter. So it discovers the letter is special.)
13869 We need a counter to count how many words there are; this variable
13870 must first be set to 0 and then incremented each time Emacs goes
13871 around the @code{while} loop. The incrementing expression is simply:
13874 (setq count (1+ count))
13877 Finally, we want to tell the user how many words there are in the
13878 region. The @code{message} function is intended for presenting this
13879 kind of information to the user. The message has to be phrased so
13880 that it reads properly regardless of how many words there are in the
13881 region: we don't want to say that ``there are 1 words in the region''.
13882 The conflict between singular and plural is ungrammatical. We can
13883 solve this problem by using a conditional expression that evaluates
13884 different messages depending on the number of words in the region.
13885 There are three possibilities: no words in the region, one word in the
13886 region, and more than one word. This means that the @code{cond}
13887 special form is appropriate.
13890 All this leads to the following function definition:
13894 ;;; @r{First version; has bugs!}
13895 (defun @value{COUNT-WORDS} (beginning end)
13896 "Print number of words in the region.
13897 Words are defined as at least one word-constituent
13898 character followed by at least one character that
13899 is not a word-constituent. The buffer's syntax
13900 table determines which characters these are."
13902 (message "Counting words in region ... ")
13906 ;;; @r{1. Set up appropriate conditions.}
13908 (goto-char beginning)
13913 ;;; @r{2. Run the} while @r{loop.}
13914 (while (< (point) end)
13915 (re-search-forward "\\w+\\W*")
13916 (setq count (1+ count)))
13920 ;;; @r{3. Send a message to the user.}
13921 (cond ((zerop count)
13923 "The region does NOT have any words."))
13926 "The region has 1 word."))
13929 "The region has %d words." count))))))
13934 As written, the function works, but not in all circumstances.
13936 @node Whitespace Bug
13937 @subsection The Whitespace Bug in @code{@value{COUNT-WORDS}}
13939 The @code{@value{COUNT-WORDS}} command described in the preceding
13940 section has two bugs, or rather, one bug with two manifestations.
13941 First, if you mark a region containing only whitespace in the middle
13942 of some text, the @code{@value{COUNT-WORDS}} command tells you that the
13943 region contains one word! Second, if you mark a region containing
13944 only whitespace at the end of the buffer or the accessible portion of
13945 a narrowed buffer, the command displays an error message that looks
13949 Search failed: "\\w+\\W*"
13952 If you are reading this in Info in GNU Emacs, you can test for these
13955 First, evaluate the function in the usual manner to install it.
13957 Here is a copy of the definition. Place your cursor after the closing
13958 parenthesis and type @kbd{C-x C-e} to install it.
13962 ;; @r{First version; has bugs!}
13963 (defun @value{COUNT-WORDS} (beginning end)
13964 "Print number of words in the region.
13965 Words are defined as at least one word-constituent character followed
13966 by at least one character that is not a word-constituent. The buffer's
13967 syntax table determines which characters these are."
13971 (message "Counting words in region ... ")
13975 ;;; @r{1. Set up appropriate conditions.}
13977 (goto-char beginning)
13982 ;;; @r{2. Run the} while @r{loop.}
13983 (while (< (point) end)
13984 (re-search-forward "\\w+\\W*")
13985 (setq count (1+ count)))
13989 ;;; @r{3. Send a message to the user.}
13990 (cond ((zerop count)
13991 (message "The region does NOT have any words."))
13992 ((= 1 count) (message "The region has 1 word."))
13993 (t (message "The region has %d words." count))))))
13999 If you wish, you can also install this keybinding by evaluating it:
14002 (global-set-key "\C-c=" '@value{COUNT-WORDS})
14005 To conduct the first test, set mark and point to the beginning and end
14006 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14007 @value{COUNT-WORDS}} if you have not bound @kbd{C-c =}):
14014 Emacs will tell you, correctly, that the region has three words.
14016 Repeat the test, but place mark at the beginning of the line and place
14017 point just @emph{before} the word @samp{one}. Again type the command
14018 @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}). Emacs should tell you
14019 that the region has no words, since it is composed only of the
14020 whitespace at the beginning of the line. But instead Emacs tells you
14021 that the region has one word!
14023 For the third test, copy the sample line to the end of the
14024 @file{*scratch*} buffer and then type several spaces at the end of the
14025 line. Place mark right after the word @samp{three} and point at the
14026 end of line. (The end of the line will be the end of the buffer.)
14027 Type @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}) as you did before.
14028 Again, Emacs should tell you that the region has no words, since it is
14029 composed only of the whitespace at the end of the line. Instead,
14030 Emacs displays an error message saying @samp{Search failed}.
14032 The two bugs stem from the same problem.
14034 Consider the first manifestation of the bug, in which the command
14035 tells you that the whitespace at the beginning of the line contains
14036 one word. What happens is this: The @code{M-x @value{COUNT-WORDS}}
14037 command moves point to the beginning of the region. The @code{while}
14038 tests whether the value of point is smaller than the value of
14039 @code{end}, which it is. Consequently, the regular expression search
14040 looks for and finds the first word. It leaves point after the word.
14041 @code{count} is set to one. The @code{while} loop repeats; but this
14042 time the value of point is larger than the value of @code{end}, the
14043 loop is exited; and the function displays a message saying the number
14044 of words in the region is one. In brief, the regular expression
14045 search looks for and finds the word even though it is outside
14048 In the second manifestation of the bug, the region is whitespace at
14049 the end of the buffer. Emacs says @samp{Search failed}. What happens
14050 is that the true-or-false-test in the @code{while} loop tests true, so
14051 the search expression is executed. But since there are no more words
14052 in the buffer, the search fails.
14054 In both manifestations of the bug, the search extends or attempts to
14055 extend outside of the region.
14057 The solution is to limit the search to the region---this is a fairly
14058 simple action, but as you may have come to expect, it is not quite as
14059 simple as you might think.
14061 As we have seen, the @code{re-search-forward} function takes a search
14062 pattern as its first argument. But in addition to this first,
14063 mandatory argument, it accepts three optional arguments. The optional
14064 second argument bounds the search. The optional third argument, if
14065 @code{t}, causes the function to return @code{nil} rather than signal
14066 an error if the search fails. The optional fourth argument is a
14067 repeat count. (In Emacs, you can see a function's documentation by
14068 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14070 In the @code{@value{COUNT-WORDS}} definition, the value of the end of
14071 the region is held by the variable @code{end} which is passed as an
14072 argument to the function. Thus, we can add @code{end} as an argument
14073 to the regular expression search expression:
14076 (re-search-forward "\\w+\\W*" end)
14079 However, if you make only this change to the @code{@value{COUNT-WORDS}}
14080 definition and then test the new version of the definition on a
14081 stretch of whitespace, you will receive an error message saying
14082 @samp{Search failed}.
14084 What happens is this: the search is limited to the region, and fails
14085 as you expect because there are no word-constituent characters in the
14086 region. Since it fails, we receive an error message. But we do not
14087 want to receive an error message in this case; we want to receive the
14088 message that "The region does NOT have any words."
14090 The solution to this problem is to provide @code{re-search-forward}
14091 with a third argument of @code{t}, which causes the function to return
14092 @code{nil} rather than signal an error if the search fails.
14094 However, if you make this change and try it, you will see the message
14095 ``Counting words in region ... '' and @dots{} you will keep on seeing
14096 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14098 Here is what happens: the search is limited to the region, as before,
14099 and it fails because there are no word-constituent characters in the
14100 region, as expected. Consequently, the @code{re-search-forward}
14101 expression returns @code{nil}. It does nothing else. In particular,
14102 it does not move point, which it does as a side effect if it finds the
14103 search target. After the @code{re-search-forward} expression returns
14104 @code{nil}, the next expression in the @code{while} loop is evaluated.
14105 This expression increments the count. Then the loop repeats. The
14106 true-or-false-test tests true because the value of point is still less
14107 than the value of end, since the @code{re-search-forward} expression
14108 did not move point. @dots{} and the cycle repeats @dots{}
14110 The @code{@value{COUNT-WORDS}} definition requires yet another
14111 modification, to cause the true-or-false-test of the @code{while} loop
14112 to test false if the search fails. Put another way, there are two
14113 conditions that must be satisfied in the true-or-false-test before the
14114 word count variable is incremented: point must still be within the
14115 region and the search expression must have found a word to count.
14117 Since both the first condition and the second condition must be true
14118 together, the two expressions, the region test and the search
14119 expression, can be joined with an @code{and} special form and embedded in
14120 the @code{while} loop as the true-or-false-test, like this:
14123 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14126 @c colon in printed section title causes problem in Info cross reference
14127 @c also trouble with an overfull hbox
14130 (For information about @code{and}, see
14131 @ref{kill-new function, , The @code{kill-new} function}.)
14135 (@xref{kill-new function, , The @code{kill-new} function}, for
14136 information about @code{and}.)
14139 The @code{re-search-forward} expression returns @code{t} if the search
14140 succeeds and as a side effect moves point. Consequently, as words are
14141 found, point is moved through the region. When the search expression
14142 fails to find another word, or when point reaches the end of the
14143 region, the true-or-false-test tests false, the @code{while} loop
14144 exits, and the @code{@value{COUNT-WORDS}} function displays one or
14145 other of its messages.
14147 After incorporating these final changes, the @code{@value{COUNT-WORDS}}
14148 works without bugs (or at least, without bugs that I have found!).
14149 Here is what it looks like:
14153 ;;; @r{Final version:} @code{while}
14154 (defun @value{COUNT-WORDS} (beginning end)
14155 "Print number of words in the region."
14157 (message "Counting words in region ... ")
14161 ;;; @r{1. Set up appropriate conditions.}
14164 (goto-char beginning)
14168 ;;; @r{2. Run the} while @r{loop.}
14169 (while (and (< (point) end)
14170 (re-search-forward "\\w+\\W*" end t))
14171 (setq count (1+ count)))
14175 ;;; @r{3. Send a message to the user.}
14176 (cond ((zerop count)
14178 "The region does NOT have any words."))
14181 "The region has 1 word."))
14184 "The region has %d words." count))))))
14188 @node recursive-count-words
14189 @section Count Words Recursively
14190 @cindex Count words recursively
14191 @cindex Recursively counting words
14192 @cindex Words, counted recursively
14194 You can write the function for counting words recursively as well as
14195 with a @code{while} loop. Let's see how this is done.
14197 First, we need to recognize that the @code{@value{COUNT-WORDS}}
14198 function has three jobs: it sets up the appropriate conditions for
14199 counting to occur; it counts the words in the region; and it sends a
14200 message to the user telling how many words there are.
14202 If we write a single recursive function to do everything, we will
14203 receive a message for every recursive call. If the region contains 13
14204 words, we will receive thirteen messages, one right after the other.
14205 We don't want this! Instead, we must write two functions to do the
14206 job, one of which (the recursive function) will be used inside of the
14207 other. One function will set up the conditions and display the
14208 message; the other will return the word count.
14210 Let us start with the function that causes the message to be displayed.
14211 We can continue to call this @code{@value{COUNT-WORDS}}.
14213 This is the function that the user will call. It will be interactive.
14214 Indeed, it will be similar to our previous versions of this
14215 function, except that it will call @code{recursive-count-words} to
14216 determine how many words are in the region.
14219 We can readily construct a template for this function, based on our
14224 ;; @r{Recursive version; uses regular expression search}
14225 (defun @value{COUNT-WORDS} (beginning end)
14226 "@var{documentation}@dots{}"
14227 (@var{interactive-expression}@dots{})
14231 ;;; @r{1. Set up appropriate conditions.}
14232 (@var{explanatory message})
14233 (@var{set-up functions}@dots{}
14237 ;;; @r{2. Count the words.}
14238 @var{recursive call}
14242 ;;; @r{3. Send a message to the user.}
14243 @var{message providing word count}))
14247 The definition looks straightforward, except that somehow the count
14248 returned by the recursive call must be passed to the message
14249 displaying the word count. A little thought suggests that this can be
14250 done by making use of a @code{let} expression: we can bind a variable
14251 in the varlist of a @code{let} expression to the number of words in
14252 the region, as returned by the recursive call; and then the
14253 @code{cond} expression, using binding, can display the value to the
14256 Often, one thinks of the binding within a @code{let} expression as
14257 somehow secondary to the `primary' work of a function. But in this
14258 case, what you might consider the `primary' job of the function,
14259 counting words, is done within the @code{let} expression.
14262 Using @code{let}, the function definition looks like this:
14266 (defun @value{COUNT-WORDS} (beginning end)
14267 "Print number of words in the region."
14272 ;;; @r{1. Set up appropriate conditions.}
14273 (message "Counting words in region ... ")
14275 (goto-char beginning)
14279 ;;; @r{2. Count the words.}
14280 (let ((count (recursive-count-words end)))
14284 ;;; @r{3. Send a message to the user.}
14285 (cond ((zerop count)
14287 "The region does NOT have any words."))
14290 "The region has 1 word."))
14293 "The region has %d words." count))))))
14297 Next, we need to write the recursive counting function.
14299 A recursive function has at least three parts: the `do-again-test', the
14300 `next-step-expression', and the recursive call.
14302 The do-again-test determines whether the function will or will not be
14303 called again. Since we are counting words in a region and can use a
14304 function that moves point forward for every word, the do-again-test
14305 can check whether point is still within the region. The do-again-test
14306 should find the value of point and determine whether point is before,
14307 at, or after the value of the end of the region. We can use the
14308 @code{point} function to locate point. Clearly, we must pass the
14309 value of the end of the region to the recursive counting function as an
14312 In addition, the do-again-test should also test whether the search finds a
14313 word. If it does not, the function should not call itself again.
14315 The next-step-expression changes a value so that when the recursive
14316 function is supposed to stop calling itself, it stops. More
14317 precisely, the next-step-expression changes a value so that at the
14318 right time, the do-again-test stops the recursive function from
14319 calling itself again. In this case, the next-step-expression can be
14320 the expression that moves point forward, word by word.
14322 The third part of a recursive function is the recursive call.
14324 Somewhere, also, we also need a part that does the `work' of the
14325 function, a part that does the counting. A vital part!
14328 But already, we have an outline of the recursive counting function:
14332 (defun recursive-count-words (region-end)
14333 "@var{documentation}@dots{}"
14334 @var{do-again-test}
14335 @var{next-step-expression}
14336 @var{recursive call})
14340 Now we need to fill in the slots. Let's start with the simplest cases
14341 first: if point is at or beyond the end of the region, there cannot
14342 be any words in the region, so the function should return zero.
14343 Likewise, if the search fails, there are no words to count, so the
14344 function should return zero.
14346 On the other hand, if point is within the region and the search
14347 succeeds, the function should call itself again.
14350 Thus, the do-again-test should look like this:
14354 (and (< (point) region-end)
14355 (re-search-forward "\\w+\\W*" region-end t))
14359 Note that the search expression is part of the do-again-test---the
14360 function returns @code{t} if its search succeeds and @code{nil} if it
14361 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14362 @code{@value{COUNT-WORDS}}}, for an explanation of how
14363 @code{re-search-forward} works.)
14365 The do-again-test is the true-or-false test of an @code{if} clause.
14366 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14367 clause should call the function again; but if it fails, the else-part
14368 should return zero since either point is outside the region or the
14369 search failed because there were no words to find.
14371 But before considering the recursive call, we need to consider the
14372 next-step-expression. What is it? Interestingly, it is the search
14373 part of the do-again-test.
14375 In addition to returning @code{t} or @code{nil} for the
14376 do-again-test, @code{re-search-forward} moves point forward as a side
14377 effect of a successful search. This is the action that changes the
14378 value of point so that the recursive function stops calling itself
14379 when point completes its movement through the region. Consequently,
14380 the @code{re-search-forward} expression is the next-step-expression.
14383 In outline, then, the body of the @code{recursive-count-words}
14384 function looks like this:
14388 (if @var{do-again-test-and-next-step-combined}
14390 @var{recursive-call-returning-count}
14396 How to incorporate the mechanism that counts?
14398 If you are not used to writing recursive functions, a question like
14399 this can be troublesome. But it can and should be approached
14402 We know that the counting mechanism should be associated in some way
14403 with the recursive call. Indeed, since the next-step-expression moves
14404 point forward by one word, and since a recursive call is made for
14405 each word, the counting mechanism must be an expression that adds one
14406 to the value returned by a call to @code{recursive-count-words}.
14409 Consider several cases:
14413 If there are two words in the region, the function should return
14414 a value resulting from adding one to the value returned when it counts
14415 the first word, plus the number returned when it counts the remaining
14416 words in the region, which in this case is one.
14419 If there is one word in the region, the function should return
14420 a value resulting from adding one to the value returned when it counts
14421 that word, plus the number returned when it counts the remaining
14422 words in the region, which in this case is zero.
14425 If there are no words in the region, the function should return zero.
14428 From the sketch we can see that the else-part of the @code{if} returns
14429 zero for the case of no words. This means that the then-part of the
14430 @code{if} must return a value resulting from adding one to the value
14431 returned from a count of the remaining words.
14434 The expression will look like this, where @code{1+} is a function that
14435 adds one to its argument.
14438 (1+ (recursive-count-words region-end))
14442 The whole @code{recursive-count-words} function will then look like
14447 (defun recursive-count-words (region-end)
14448 "@var{documentation}@dots{}"
14450 ;;; @r{1. do-again-test}
14451 (if (and (< (point) region-end)
14452 (re-search-forward "\\w+\\W*" region-end t))
14456 ;;; @r{2. then-part: the recursive call}
14457 (1+ (recursive-count-words region-end))
14459 ;;; @r{3. else-part}
14465 Let's examine how this works:
14467 If there are no words in the region, the else part of the @code{if}
14468 expression is evaluated and consequently the function returns zero.
14470 If there is one word in the region, the value of point is less than
14471 the value of @code{region-end} and the search succeeds. In this case,
14472 the true-or-false-test of the @code{if} expression tests true, and the
14473 then-part of the @code{if} expression is evaluated. The counting
14474 expression is evaluated. This expression returns a value (which will
14475 be the value returned by the whole function) that is the sum of one
14476 added to the value returned by a recursive call.
14478 Meanwhile, the next-step-expression has caused point to jump over the
14479 first (and in this case only) word in the region. This means that
14480 when @code{(recursive-count-words region-end)} is evaluated a second
14481 time, as a result of the recursive call, the value of point will be
14482 equal to or greater than the value of region end. So this time,
14483 @code{recursive-count-words} will return zero. The zero will be added
14484 to one, and the original evaluation of @code{recursive-count-words}
14485 will return one plus zero, which is one, which is the correct amount.
14487 Clearly, if there are two words in the region, the first call to
14488 @code{recursive-count-words} returns one added to the value returned
14489 by calling @code{recursive-count-words} on a region containing the
14490 remaining word---that is, it adds one to one, producing two, which is
14491 the correct amount.
14493 Similarly, if there are three words in the region, the first call to
14494 @code{recursive-count-words} returns one added to the value returned
14495 by calling @code{recursive-count-words} on a region containing the
14496 remaining two words---and so on and so on.
14500 With full documentation the two functions look like this:
14504 The recursive function:
14506 @findex recursive-count-words
14509 (defun recursive-count-words (region-end)
14510 "Number of words between point and REGION-END."
14514 ;;; @r{1. do-again-test}
14515 (if (and (< (point) region-end)
14516 (re-search-forward "\\w+\\W*" region-end t))
14520 ;;; @r{2. then-part: the recursive call}
14521 (1+ (recursive-count-words region-end))
14523 ;;; @r{3. else-part}
14534 ;;; @r{Recursive version}
14535 (defun @value{COUNT-WORDS} (beginning end)
14536 "Print number of words in the region.
14540 Words are defined as at least one word-constituent
14541 character followed by at least one character that is
14542 not a word-constituent. The buffer's syntax table
14543 determines which characters these are."
14547 (message "Counting words in region ... ")
14549 (goto-char beginning)
14550 (let ((count (recursive-count-words end)))
14553 (cond ((zerop count)
14555 "The region does NOT have any words."))
14559 (message "The region has 1 word."))
14562 "The region has %d words." count))))))
14566 @node Counting Exercise
14567 @section Exercise: Counting Punctuation
14569 Using a @code{while} loop, write a function to count the number of
14570 punctuation marks in a region---period, comma, semicolon, colon,
14571 exclamation mark, and question mark. Do the same using recursion.
14573 @node Words in a defun
14574 @chapter Counting Words in a @code{defun}
14575 @cindex Counting words in a @code{defun}
14576 @cindex Word counting in a @code{defun}
14578 Our next project is to count the number of words in a function
14579 definition. Clearly, this can be done using some variant of
14580 @code{@value{COUNT-WORDS}}. @xref{Counting Words, , Counting Words:
14581 Repetition and Regexps}. If we are just going to count the words in
14582 one definition, it is easy enough to mark the definition with the
14583 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14584 @code{@value{COUNT-WORDS}}.
14586 However, I am more ambitious: I want to count the words and symbols in
14587 every definition in the Emacs sources and then print a graph that
14588 shows how many functions there are of each length: how many contain 40
14589 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14590 and so on. I have often been curious how long a typical function is,
14591 and this will tell.
14594 * Divide and Conquer::
14595 * Words and Symbols:: What to count?
14596 * Syntax:: What constitutes a word or symbol?
14597 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
14598 * Several defuns:: Counting several defuns in a file.
14599 * Find a File:: Do you want to look at a file?
14600 * lengths-list-file:: A list of the lengths of many definitions.
14601 * Several files:: Counting in definitions in different files.
14602 * Several files recursively:: Recursively counting in different files.
14603 * Prepare the data:: Prepare the data for display in a graph.
14607 @node Divide and Conquer
14608 @unnumberedsec Divide and Conquer
14611 Described in one phrase, the histogram project is daunting; but
14612 divided into numerous small steps, each of which we can take one at a
14613 time, the project becomes less fearsome. Let us consider what the
14618 First, write a function to count the words in one definition. This
14619 includes the problem of handling symbols as well as words.
14622 Second, write a function to list the numbers of words in each function
14623 in a file. This function can use the @code{count-words-in-defun}
14627 Third, write a function to list the numbers of words in each function
14628 in each of several files. This entails automatically finding the
14629 various files, switching to them, and counting the words in the
14630 definitions within them.
14633 Fourth, write a function to convert the list of numbers that we
14634 created in step three to a form that will be suitable for printing as
14638 Fifth, write a function to print the results as a graph.
14641 This is quite a project! But if we take each step slowly, it will not
14644 @node Words and Symbols
14645 @section What to Count?
14646 @cindex Words and symbols in defun
14648 When we first start thinking about how to count the words in a
14649 function definition, the first question is (or ought to be) what are
14650 we going to count? When we speak of `words' with respect to a Lisp
14651 function definition, we are actually speaking, in large part, of
14652 `symbols'. For example, the following @code{multiply-by-seven}
14653 function contains the five symbols @code{defun},
14654 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14655 addition, in the documentation string, it contains the four words
14656 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14657 symbol @samp{number} is repeated, so the definition contains a total
14658 of ten words and symbols.
14662 (defun multiply-by-seven (number)
14663 "Multiply NUMBER by seven."
14669 However, if we mark the @code{multiply-by-seven} definition with
14670 @kbd{C-M-h} (@code{mark-defun}), and then call
14671 @code{@value{COUNT-WORDS}} on it, we will find that
14672 @code{@value{COUNT-WORDS}} claims the definition has eleven words, not
14673 ten! Something is wrong!
14675 The problem is twofold: @code{@value{COUNT-WORDS}} does not count the
14676 @samp{*} as a word, and it counts the single symbol,
14677 @code{multiply-by-seven}, as containing three words. The hyphens are
14678 treated as if they were interword spaces rather than intraword
14679 connectors: @samp{multiply-by-seven} is counted as if it were written
14680 @samp{multiply by seven}.
14682 The cause of this confusion is the regular expression search within
14683 the @code{@value{COUNT-WORDS}} definition that moves point forward word
14684 by word. In the canonical version of @code{@value{COUNT-WORDS}}, the
14692 This regular expression is a pattern defining one or more word
14693 constituent characters possibly followed by one or more characters
14694 that are not word constituents. What is meant by `word constituent
14695 characters' brings us to the issue of syntax, which is worth a section
14699 @section What Constitutes a Word or Symbol?
14700 @cindex Syntax categories and tables
14702 Emacs treats different characters as belonging to different
14703 @dfn{syntax categories}. For example, the regular expression,
14704 @samp{\\w+}, is a pattern specifying one or more @emph{word
14705 constituent} characters. Word constituent characters are members of
14706 one syntax category. Other syntax categories include the class of
14707 punctuation characters, such as the period and the comma, and the
14708 class of whitespace characters, such as the blank space and the tab
14709 character. (For more information, @pxref{Syntax Tables, , Syntax
14710 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14712 Syntax tables specify which characters belong to which categories.
14713 Usually, a hyphen is not specified as a `word constituent character'.
14714 Instead, it is specified as being in the `class of characters that are
14715 part of symbol names but not words.' This means that the
14716 @code{@value{COUNT-WORDS}} function treats it in the same way it treats
14717 an interword white space, which is why @code{@value{COUNT-WORDS}}
14718 counts @samp{multiply-by-seven} as three words.
14720 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14721 one symbol: modify the syntax table or modify the regular expression.
14723 We could redefine a hyphen as a word constituent character by
14724 modifying the syntax table that Emacs keeps for each mode. This
14725 action would serve our purpose, except that a hyphen is merely the
14726 most common character within symbols that is not typically a word
14727 constituent character; there are others, too.
14729 Alternatively, we can redefine the regexp used in the
14730 @code{@value{COUNT-WORDS}} definition so as to include symbols. This
14731 procedure has the merit of clarity, but the task is a little tricky.
14734 The first part is simple enough: the pattern must match ``at least one
14735 character that is a word or symbol constituent''. Thus:
14738 "\\(\\w\\|\\s_\\)+"
14742 The @samp{\\(} is the first part of the grouping construct that
14743 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14744 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14745 character and the @samp{\\s_} matches any character that is part of a
14746 symbol name but not a word-constituent character. The @samp{+}
14747 following the group indicates that the word or symbol constituent
14748 characters must be matched at least once.
14750 However, the second part of the regexp is more difficult to design.
14751 What we want is to follow the first part with ``optionally one or more
14752 characters that are not constituents of a word or symbol''. At first,
14753 I thought I could define this with the following:
14756 "\\(\\W\\|\\S_\\)*"
14760 The upper case @samp{W} and @samp{S} match characters that are
14761 @emph{not} word or symbol constituents. Unfortunately, this
14762 expression matches any character that is either not a word constituent
14763 or not a symbol constituent. This matches any character!
14765 I then noticed that every word or symbol in my test region was
14766 followed by white space (blank space, tab, or newline). So I tried
14767 placing a pattern to match one or more blank spaces after the pattern
14768 for one or more word or symbol constituents. This failed, too. Words
14769 and symbols are often separated by whitespace, but in actual code
14770 parentheses may follow symbols and punctuation may follow words. So
14771 finally, I designed a pattern in which the word or symbol constituents
14772 are followed optionally by characters that are not white space and
14773 then followed optionally by white space.
14776 Here is the full regular expression:
14779 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14782 @node count-words-in-defun
14783 @section The @code{count-words-in-defun} Function
14784 @cindex Counting words in a @code{defun}
14786 We have seen that there are several ways to write a
14787 @code{count-words-region} function. To write a
14788 @code{count-words-in-defun}, we need merely adapt one of these
14791 The version that uses a @code{while} loop is easy to understand, so I
14792 am going to adapt that. Because @code{count-words-in-defun} will be
14793 part of a more complex program, it need not be interactive and it need
14794 not display a message but just return the count. These considerations
14795 simplify the definition a little.
14797 On the other hand, @code{count-words-in-defun} will be used within a
14798 buffer that contains function definitions. Consequently, it is
14799 reasonable to ask that the function determine whether it is called
14800 when point is within a function definition, and if it is, to return
14801 the count for that definition. This adds complexity to the
14802 definition, but saves us from needing to pass arguments to the
14806 These considerations lead us to prepare the following template:
14810 (defun count-words-in-defun ()
14811 "@var{documentation}@dots{}"
14812 (@var{set up}@dots{}
14813 (@var{while loop}@dots{})
14814 @var{return count})
14819 As usual, our job is to fill in the slots.
14823 We are presuming that this function will be called within a buffer
14824 containing function definitions. Point will either be within a
14825 function definition or not. For @code{count-words-in-defun} to work,
14826 point must move to the beginning of the definition, a counter must
14827 start at zero, and the counting loop must stop when point reaches the
14828 end of the definition.
14830 The @code{beginning-of-defun} function searches backwards for an
14831 opening delimiter such as a @samp{(} at the beginning of a line, and
14832 moves point to that position, or else to the limit of the search. In
14833 practice, this means that @code{beginning-of-defun} moves point to the
14834 beginning of an enclosing or preceding function definition, or else to
14835 the beginning of the buffer. We can use @code{beginning-of-defun} to
14836 place point where we wish to start.
14838 The @code{while} loop requires a counter to keep track of the words or
14839 symbols being counted. A @code{let} expression can be used to create
14840 a local variable for this purpose, and bind it to an initial value of zero.
14842 The @code{end-of-defun} function works like @code{beginning-of-defun}
14843 except that it moves point to the end of the definition.
14844 @code{end-of-defun} can be used as part of an expression that
14845 determines the position of the end of the definition.
14847 The set up for @code{count-words-in-defun} takes shape rapidly: first
14848 we move point to the beginning of the definition, then we create a
14849 local variable to hold the count, and finally, we record the position
14850 of the end of the definition so the @code{while} loop will know when to stop
14854 The code looks like this:
14858 (beginning-of-defun)
14860 (end (save-excursion (end-of-defun) (point))))
14865 The code is simple. The only slight complication is likely to concern
14866 @code{end}: it is bound to the position of the end of the definition
14867 by a @code{save-excursion} expression that returns the value of point
14868 after @code{end-of-defun} temporarily moves it to the end of the
14871 The second part of the @code{count-words-in-defun}, after the set up,
14872 is the @code{while} loop.
14874 The loop must contain an expression that jumps point forward word by
14875 word and symbol by symbol, and another expression that counts the
14876 jumps. The true-or-false-test for the @code{while} loop should test
14877 true so long as point should jump forward, and false when point is at
14878 the end of the definition. We have already redefined the regular
14879 expression for this, so the loop is straightforward:
14883 (while (and (< (point) end)
14885 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t))
14886 (setq count (1+ count)))
14890 The third part of the function definition returns the count of words
14891 and symbols. This part is the last expression within the body of the
14892 @code{let} expression, and can be, very simply, the local variable
14893 @code{count}, which when evaluated returns the count.
14896 Put together, the @code{count-words-in-defun} definition looks like this:
14898 @findex count-words-in-defun
14901 (defun count-words-in-defun ()
14902 "Return the number of words and symbols in a defun."
14903 (beginning-of-defun)
14905 (end (save-excursion (end-of-defun) (point))))
14909 (and (< (point) end)
14911 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14913 (setq count (1+ count)))
14918 How to test this? The function is not interactive, but it is easy to
14919 put a wrapper around the function to make it interactive; we can use
14920 almost the same code as for the recursive version of
14921 @code{@value{COUNT-WORDS}}:
14925 ;;; @r{Interactive version.}
14926 (defun count-words-defun ()
14927 "Number of words and symbols in a function definition."
14930 "Counting words and symbols in function definition ... ")
14933 (let ((count (count-words-in-defun)))
14937 "The definition does NOT have any words or symbols."))
14942 "The definition has 1 word or symbol."))
14945 "The definition has %d words or symbols." count)))))
14951 Let's re-use @kbd{C-c =} as a convenient keybinding:
14954 (global-set-key "\C-c=" 'count-words-defun)
14957 Now we can try out @code{count-words-defun}: install both
14958 @code{count-words-in-defun} and @code{count-words-defun}, and set the
14959 keybinding, and then place the cursor within the following definition:
14963 (defun multiply-by-seven (number)
14964 "Multiply NUMBER by seven."
14971 Success! The definition has 10 words and symbols.
14973 The next problem is to count the numbers of words and symbols in
14974 several definitions within a single file.
14976 @node Several defuns
14977 @section Count Several @code{defuns} Within a File
14979 A file such as @file{simple.el} may have a hundred or more function
14980 definitions within it. Our long term goal is to collect statistics on
14981 many files, but as a first step, our immediate goal is to collect
14982 statistics on one file.
14984 The information will be a series of numbers, each number being the
14985 length of a function definition. We can store the numbers in a list.
14987 We know that we will want to incorporate the information regarding one
14988 file with information about many other files; this means that the
14989 function for counting definition lengths within one file need only
14990 return the list of lengths. It need not and should not display any
14993 The word count commands contain one expression to jump point forward
14994 word by word and another expression to count the jumps. The function
14995 to return the lengths of definitions can be designed to work the same
14996 way, with one expression to jump point forward definition by
14997 definition and another expression to construct the lengths' list.
14999 This statement of the problem makes it elementary to write the
15000 function definition. Clearly, we will start the count at the
15001 beginning of the file, so the first command will be @code{(goto-char
15002 (point-min))}. Next, we start the @code{while} loop; and the
15003 true-or-false test of the loop can be a regular expression search for
15004 the next function definition---so long as the search succeeds, point
15005 is moved forward and then the body of the loop is evaluated. The body
15006 needs an expression that constructs the lengths' list. @code{cons},
15007 the list construction command, can be used to create the list. That
15008 is almost all there is to it.
15011 Here is what this fragment of code looks like:
15015 (goto-char (point-min))
15016 (while (re-search-forward "^(defun" nil t)
15018 (cons (count-words-in-defun) lengths-list)))
15022 What we have left out is the mechanism for finding the file that
15023 contains the function definitions.
15025 In previous examples, we either used this, the Info file, or we
15026 switched back and forth to some other buffer, such as the
15027 @file{*scratch*} buffer.
15029 Finding a file is a new process that we have not yet discussed.
15032 @section Find a File
15033 @cindex Find a File
15035 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15036 command. This command is almost, but not quite right for the lengths
15040 Let's look at the source for @code{find-file}:
15044 (defun find-file (filename)
15045 "Edit file FILENAME.
15046 Switch to a buffer visiting file FILENAME,
15047 creating one if none already exists."
15048 (interactive "FFind file: ")
15049 (switch-to-buffer (find-file-noselect filename)))
15054 (The most recent version of the @code{find-file} function definition
15055 permits you to specify optional wildcards to visit multiple files; that
15056 makes the definition more complex and we will not discuss it here,
15057 since it is not relevant. You can see its source using either
15058 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15062 (defun find-file (filename &optional wildcards)
15063 "Edit file FILENAME.
15064 Switch to a buffer visiting file FILENAME,
15065 creating one if none already exists.
15066 Interactively, the default if you just type RET is the current directory,
15067 but the visited file name is available through the minibuffer history:
15068 type M-n to pull it into the minibuffer.
15070 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15071 expand wildcards (if any) and visit multiple files. You can
15072 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15074 To visit a file without any kind of conversion and without
15075 automatically choosing a major mode, use \\[find-file-literally]."
15076 (interactive (find-file-read-args "Find file: " nil))
15077 (let ((value (find-file-noselect filename nil nil wildcards)))
15079 (mapcar 'switch-to-buffer (nreverse value))
15080 (switch-to-buffer value))))
15083 The definition I am showing possesses short but complete documentation
15084 and an interactive specification that prompts you for a file name when
15085 you use the command interactively. The body of the definition
15086 contains two functions, @code{find-file-noselect} and
15087 @code{switch-to-buffer}.
15089 According to its documentation as shown by @kbd{C-h f} (the
15090 @code{describe-function} command), the @code{find-file-noselect}
15091 function reads the named file into a buffer and returns the buffer.
15092 (Its most recent version includes an optional wildcards argument,
15093 too, as well as another to read a file literally and an other you
15094 suppress warning messages. These optional arguments are irrelevant.)
15096 However, the @code{find-file-noselect} function does not select the
15097 buffer in which it puts the file. Emacs does not switch its attention
15098 (or yours if you are using @code{find-file-noselect}) to the selected
15099 buffer. That is what @code{switch-to-buffer} does: it switches the
15100 buffer to which Emacs attention is directed; and it switches the
15101 buffer displayed in the window to the new buffer. We have discussed
15102 buffer switching elsewhere. (@xref{Switching Buffers}.)
15104 In this histogram project, we do not need to display each file on the
15105 screen as the program determines the length of each definition within
15106 it. Instead of employing @code{switch-to-buffer}, we can work with
15107 @code{set-buffer}, which redirects the attention of the computer
15108 program to a different buffer but does not redisplay it on the screen.
15109 So instead of calling on @code{find-file} to do the job, we must write
15110 our own expression.
15112 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15114 @node lengths-list-file
15115 @section @code{lengths-list-file} in Detail
15117 The core of the @code{lengths-list-file} function is a @code{while}
15118 loop containing a function to move point forward `defun by defun' and
15119 a function to count the number of words and symbols in each defun.
15120 This core must be surrounded by functions that do various other tasks,
15121 including finding the file, and ensuring that point starts out at the
15122 beginning of the file. The function definition looks like this:
15123 @findex lengths-list-file
15127 (defun lengths-list-file (filename)
15128 "Return list of definitions' lengths within FILE.
15129 The returned list is a list of numbers.
15130 Each number is the number of words or
15131 symbols in one function definition."
15134 (message "Working on `%s' ... " filename)
15136 (let ((buffer (find-file-noselect filename))
15138 (set-buffer buffer)
15139 (setq buffer-read-only t)
15141 (goto-char (point-min))
15142 (while (re-search-forward "^(defun" nil t)
15144 (cons (count-words-in-defun) lengths-list)))
15145 (kill-buffer buffer)
15151 The function is passed one argument, the name of the file on which it
15152 will work. It has four lines of documentation, but no interactive
15153 specification. Since people worry that a computer is broken if they
15154 don't see anything going on, the first line of the body is a
15157 The next line contains a @code{save-excursion} that returns Emacs's
15158 attention to the current buffer when the function completes. This is
15159 useful in case you embed this function in another function that
15160 presumes point is restored to the original buffer.
15162 In the varlist of the @code{let} expression, Emacs finds the file and
15163 binds the local variable @code{buffer} to the buffer containing the
15164 file. At the same time, Emacs creates @code{lengths-list} as a local
15167 Next, Emacs switches its attention to the buffer.
15169 In the following line, Emacs makes the buffer read-only. Ideally,
15170 this line is not necessary. None of the functions for counting words
15171 and symbols in a function definition should change the buffer.
15172 Besides, the buffer is not going to be saved, even if it were changed.
15173 This line is entirely the consequence of great, perhaps excessive,
15174 caution. The reason for the caution is that this function and those
15175 it calls work on the sources for Emacs and it is inconvenient if they
15176 are inadvertently modified. It goes without saying that I did not
15177 realize a need for this line until an experiment went awry and started
15178 to modify my Emacs source files @dots{}
15180 Next comes a call to widen the buffer if it is narrowed. This
15181 function is usually not needed---Emacs creates a fresh buffer if none
15182 already exists; but if a buffer visiting the file already exists Emacs
15183 returns that one. In this case, the buffer may be narrowed and must
15184 be widened. If we wanted to be fully `user-friendly', we would
15185 arrange to save the restriction and the location of point, but we
15188 The @code{(goto-char (point-min))} expression moves point to the
15189 beginning of the buffer.
15191 Then comes a @code{while} loop in which the `work' of the function is
15192 carried out. In the loop, Emacs determines the length of each
15193 definition and constructs a lengths' list containing the information.
15195 Emacs kills the buffer after working through it. This is to save
15196 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15197 source files of interest; GNU Emacs 22 contains over a thousand source
15198 files. Another function will apply @code{lengths-list-file} to each
15201 Finally, the last expression within the @code{let} expression is the
15202 @code{lengths-list} variable; its value is returned as the value of
15203 the whole function.
15205 You can try this function by installing it in the usual fashion. Then
15206 place your cursor after the following expression and type @kbd{C-x
15207 C-e} (@code{eval-last-sexp}).
15209 @c !!! 22.1.1 lisp sources location here
15212 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15216 (You may need to change the pathname of the file; the one here is for
15217 GNU Emacs version 22.1.1. To change the expression, copy it to
15218 the @file{*scratch*} buffer and edit it.
15222 (Also, to see the full length of the list, rather than a truncated
15223 version, you may have to evaluate the following:
15226 (custom-set-variables '(eval-expression-print-length nil))
15230 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15231 Then evaluate the @code{lengths-list-file} expression.)
15234 The lengths' list for @file{debug.el} takes less than a second to
15235 produce and looks like this in GNU Emacs 22:
15238 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15242 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15243 took seven seconds to produce and looked like this:
15246 (75 41 80 62 20 45 44 68 45 12 34 235)
15249 (The newer version of @file{debug.el} contains more defuns than the
15250 earlier one; and my new machine is much faster than the old one.)
15252 Note that the length of the last definition in the file is first in
15255 @node Several files
15256 @section Count Words in @code{defuns} in Different Files
15258 In the previous section, we created a function that returns a list of
15259 the lengths of each definition in a file. Now, we want to define a
15260 function to return a master list of the lengths of the definitions in
15263 Working on each of a list of files is a repetitious act, so we can use
15264 either a @code{while} loop or recursion.
15267 * lengths-list-many-files:: Return a list of the lengths of defuns.
15268 * append:: Attach one list to another.
15272 @node lengths-list-many-files
15273 @unnumberedsubsec Determine the lengths of @code{defuns}
15276 The design using a @code{while} loop is routine. The argument passed
15277 the function is a list of files. As we saw earlier (@pxref{Loop
15278 Example}), you can write a @code{while} loop so that the body of the
15279 loop is evaluated if such a list contains elements, but to exit the
15280 loop if the list is empty. For this design to work, the body of the
15281 loop must contain an expression that shortens the list each time the
15282 body is evaluated, so that eventually the list is empty. The usual
15283 technique is to set the value of the list to the value of the @sc{cdr}
15284 of the list each time the body is evaluated.
15287 The template looks like this:
15291 (while @var{test-whether-list-is-empty}
15293 @var{set-list-to-cdr-of-list})
15297 Also, we remember that a @code{while} loop returns @code{nil} (the
15298 result of evaluating the true-or-false-test), not the result of any
15299 evaluation within its body. (The evaluations within the body of the
15300 loop are done for their side effects.) However, the expression that
15301 sets the lengths' list is part of the body---and that is the value
15302 that we want returned by the function as a whole. To do this, we
15303 enclose the @code{while} loop within a @code{let} expression, and
15304 arrange that the last element of the @code{let} expression contains
15305 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15306 Example with an Incrementing Counter}.)
15308 @findex lengths-list-many-files
15310 These considerations lead us directly to the function itself:
15314 ;;; @r{Use @code{while} loop.}
15315 (defun lengths-list-many-files (list-of-files)
15316 "Return list of lengths of defuns in LIST-OF-FILES."
15319 (let (lengths-list)
15321 ;;; @r{true-or-false-test}
15322 (while list-of-files
15327 ;;; @r{Generate a lengths' list.}
15329 (expand-file-name (car list-of-files)))))
15333 ;;; @r{Make files' list shorter.}
15334 (setq list-of-files (cdr list-of-files)))
15336 ;;; @r{Return final value of lengths' list.}
15341 @code{expand-file-name} is a built-in function that converts a file
15342 name to the absolute, long, path name form. The function employs the
15343 name of the directory in which the function is called.
15345 @c !!! 22.1.1 lisp sources location here
15347 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15348 Emacs is visiting the
15349 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15359 @c !!! 22.1.1 lisp sources location here
15361 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15364 The only other new element of this function definition is the as yet
15365 unstudied function @code{append}, which merits a short section for
15369 @subsection The @code{append} Function
15372 The @code{append} function attaches one list to another. Thus,
15375 (append '(1 2 3 4) '(5 6 7 8))
15386 This is exactly how we want to attach two lengths' lists produced by
15387 @code{lengths-list-file} to each other. The results contrast with
15391 (cons '(1 2 3 4) '(5 6 7 8))
15396 which constructs a new list in which the first argument to @code{cons}
15397 becomes the first element of the new list:
15400 ((1 2 3 4) 5 6 7 8)
15403 @node Several files recursively
15404 @section Recursively Count Words in Different Files
15406 Besides a @code{while} loop, you can work on each of a list of files
15407 with recursion. A recursive version of @code{lengths-list-many-files}
15408 is short and simple.
15410 The recursive function has the usual parts: the `do-again-test', the
15411 `next-step-expression', and the recursive call. The `do-again-test'
15412 determines whether the function should call itself again, which it
15413 will do if the @code{list-of-files} contains any remaining elements;
15414 the `next-step-expression' resets the @code{list-of-files} to the
15415 @sc{cdr} of itself, so eventually the list will be empty; and the
15416 recursive call calls itself on the shorter list. The complete
15417 function is shorter than this description!
15418 @findex recursive-lengths-list-many-files
15422 (defun recursive-lengths-list-many-files (list-of-files)
15423 "Return list of lengths of each defun in LIST-OF-FILES."
15424 (if list-of-files ; @r{do-again-test}
15427 (expand-file-name (car list-of-files)))
15428 (recursive-lengths-list-many-files
15429 (cdr list-of-files)))))
15434 In a sentence, the function returns the lengths' list for the first of
15435 the @code{list-of-files} appended to the result of calling itself on
15436 the rest of the @code{list-of-files}.
15438 Here is a test of @code{recursive-lengths-list-many-files}, along with
15439 the results of running @code{lengths-list-file} on each of the files
15442 Install @code{recursive-lengths-list-many-files} and
15443 @code{lengths-list-file}, if necessary, and then evaluate the
15444 following expressions. You may need to change the files' pathnames;
15445 those here work when this Info file and the Emacs sources are located
15446 in their customary places. To change the expressions, copy them to
15447 the @file{*scratch*} buffer, edit them, and then evaluate them.
15449 The results are shown after the @samp{@result{}}. (These results are
15450 for files from Emacs version 22.1.1; files from other versions of
15451 Emacs may produce different results.)
15453 @c !!! 22.1.1 lisp sources location here
15456 (cd "/usr/local/share/emacs/22.1.1/")
15458 (lengths-list-file "./lisp/macros.el")
15459 @result{} (283 263 480 90)
15463 (lengths-list-file "./lisp/mail/mailalias.el")
15464 @result{} (38 32 29 95 178 180 321 218 324)
15468 (lengths-list-file "./lisp/makesum.el")
15473 (recursive-lengths-list-many-files
15474 '("./lisp/macros.el"
15475 "./lisp/mail/mailalias.el"
15476 "./lisp/makesum.el"))
15477 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15481 The @code{recursive-lengths-list-many-files} function produces the
15484 The next step is to prepare the data in the list for display in a graph.
15486 @node Prepare the data
15487 @section Prepare the Data for Display in a Graph
15489 The @code{recursive-lengths-list-many-files} function returns a list
15490 of numbers. Each number records the length of a function definition.
15491 What we need to do now is transform this data into a list of numbers
15492 suitable for generating a graph. The new list will tell how many
15493 functions definitions contain less than 10 words and
15494 symbols, how many contain between 10 and 19 words and symbols, how
15495 many contain between 20 and 29 words and symbols, and so on.
15497 In brief, we need to go through the lengths' list produced by the
15498 @code{recursive-lengths-list-many-files} function and count the number
15499 of defuns within each range of lengths, and produce a list of those
15503 * Data for Display in Detail::
15504 * Sorting:: Sorting lists.
15505 * Files List:: Making a list of files.
15506 * Counting function definitions::
15510 @node Data for Display in Detail
15511 @unnumberedsubsec The Data for Display in Detail
15514 Based on what we have done before, we can readily foresee that it
15515 should not be too hard to write a function that `@sc{cdr}s' down the
15516 lengths' list, looks at each element, determines which length range it
15517 is in, and increments a counter for that range.
15519 However, before beginning to write such a function, we should consider
15520 the advantages of sorting the lengths' list first, so the numbers are
15521 ordered from smallest to largest. First, sorting will make it easier
15522 to count the numbers in each range, since two adjacent numbers will
15523 either be in the same length range or in adjacent ranges. Second, by
15524 inspecting a sorted list, we can discover the highest and lowest
15525 number, and thereby determine the largest and smallest length range
15529 @subsection Sorting Lists
15532 Emacs contains a function to sort lists, called (as you might guess)
15533 @code{sort}. The @code{sort} function takes two arguments, the list
15534 to be sorted, and a predicate that determines whether the first of
15535 two list elements is ``less'' than the second.
15537 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15538 Type Object as an Argument}), a predicate is a function that
15539 determines whether some property is true or false. The @code{sort}
15540 function will reorder a list according to whatever property the
15541 predicate uses; this means that @code{sort} can be used to sort
15542 non-numeric lists by non-numeric criteria---it can, for example,
15543 alphabetize a list.
15546 The @code{<} function is used when sorting a numeric list. For example,
15549 (sort '(4 8 21 17 33 7 21 7) '<)
15557 (4 7 7 8 17 21 21 33)
15561 (Note that in this example, both the arguments are quoted so that the
15562 symbols are not evaluated before being passed to @code{sort} as
15565 Sorting the list returned by the
15566 @code{recursive-lengths-list-many-files} function is straightforward;
15567 it uses the @code{<} function:
15571 In GNU Emacs 22, eval
15573 (cd "/usr/local/share/emacs/22.0.50/")
15575 (recursive-lengths-list-many-files
15576 '("./lisp/macros.el"
15577 "./lisp/mail/mailalias.el"
15578 "./lisp/makesum.el"))
15586 (recursive-lengths-list-many-files
15587 '("./lisp/macros.el"
15588 "./lisp/mailalias.el"
15589 "./lisp/makesum.el"))
15599 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15603 (Note that in this example, the first argument to @code{sort} is not
15604 quoted, since the expression must be evaluated so as to produce the
15605 list that is passed to @code{sort}.)
15608 @subsection Making a List of Files
15610 The @code{recursive-lengths-list-many-files} function requires a list
15611 of files as its argument. For our test examples, we constructed such
15612 a list by hand; but the Emacs Lisp source directory is too large for
15613 us to do for that. Instead, we will write a function to do the job
15614 for us. In this function, we will use both a @code{while} loop and a
15617 @findex directory-files
15618 We did not have to write a function like this for older versions of
15619 GNU Emacs, since they placed all the @samp{.el} files in one
15620 directory. Instead, we were able to use the @code{directory-files}
15621 function, which lists the names of files that match a specified
15622 pattern within a single directory.
15624 However, recent versions of Emacs place Emacs Lisp files in
15625 sub-directories of the top level @file{lisp} directory. This
15626 re-arrangement eases navigation. For example, all the mail related
15627 files are in a @file{lisp} sub-directory called @file{mail}. But at
15628 the same time, this arrangement forces us to create a file listing
15629 function that descends into the sub-directories.
15631 @findex files-in-below-directory
15632 We can create this function, called @code{files-in-below-directory},
15633 using familiar functions such as @code{car}, @code{nthcdr}, and
15634 @code{substring} in conjunction with an existing function called
15635 @code{directory-files-and-attributes}. This latter function not only
15636 lists all the filenames in a directory, including the names
15637 of sub-directories, but also their attributes.
15639 To restate our goal: to create a function that will enable us
15640 to feed filenames to @code{recursive-lengths-list-many-files}
15641 as a list that looks like this (but with more elements):
15645 ("./lisp/macros.el"
15646 "./lisp/mail/rmail.el"
15647 "./lisp/makesum.el")
15651 The @code{directory-files-and-attributes} function returns a list of
15652 lists. Each of the lists within the main list consists of 13
15653 elements. The first element is a string that contains the name of the
15654 file---which, in GNU/Linux, may be a `directory file', that is to
15655 say, a file with the special attributes of a directory. The second
15656 element of the list is @code{t} for a directory, a string
15657 for symbolic link (the string is the name linked to), or @code{nil}.
15659 For example, the first @samp{.el} file in the @file{lisp/} directory
15660 is @file{abbrev.el}. Its name is
15661 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15662 directory or a symbolic link.
15665 This is how @code{directory-files-and-attributes} lists that file and
15677 (20615 27034 579989 697000)
15679 (20615 26327 734791 805000)
15691 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15692 directory. The beginning of its listing looks like this:
15703 (To learn about the different attributes, look at the documentation of
15704 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15705 function does not list the filename, so its first element is
15706 @code{directory-files-and-attributes}'s second element.)
15708 We will want our new function, @code{files-in-below-directory}, to
15709 list the @samp{.el} files in the directory it is told to check, and in
15710 any directories below that directory.
15712 This gives us a hint on how to construct
15713 @code{files-in-below-directory}: within a directory, the function
15714 should add @samp{.el} filenames to a list; and if, within a directory,
15715 the function comes upon a sub-directory, it should go into that
15716 sub-directory and repeat its actions.
15718 However, we should note that every directory contains a name that
15719 refers to itself, called @file{.}, (``dot'') and a name that refers to
15720 its parent directory, called @file{..} (``double dot''). (In
15721 @file{/}, the root directory, @file{..} refers to itself, since
15722 @file{/} has no parent.) Clearly, we do not want our
15723 @code{files-in-below-directory} function to enter those directories,
15724 since they always lead us, directly or indirectly, to the current
15727 Consequently, our @code{files-in-below-directory} function must do
15732 Check to see whether it is looking at a filename that ends in
15733 @samp{.el}; and if so, add its name to a list.
15736 Check to see whether it is looking at a filename that is the name of a
15737 directory; and if so,
15741 Check to see whether it is looking at @file{.} or @file{..}; and if
15745 Or else, go into that directory and repeat the process.
15749 Let's write a function definition to do these tasks. We will use a
15750 @code{while} loop to move from one filename to another within a
15751 directory, checking what needs to be done; and we will use a recursive
15752 call to repeat the actions on each sub-directory. The recursive
15753 pattern is `accumulate'
15754 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15755 using @code{append} as the combiner.
15758 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15759 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15761 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15762 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15765 @c /usr/local/share/emacs/22.1.1/lisp/
15768 Here is the function:
15772 (defun files-in-below-directory (directory)
15773 "List the .el files in DIRECTORY and in its sub-directories."
15774 ;; Although the function will be used non-interactively,
15775 ;; it will be easier to test if we make it interactive.
15776 ;; The directory will have a name such as
15777 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15778 (interactive "DDirectory name: ")
15781 (let (el-files-list
15782 (current-directory-list
15783 (directory-files-and-attributes directory t)))
15784 ;; while we are in the current directory
15785 (while current-directory-list
15789 ;; check to see whether filename ends in `.el'
15790 ;; and if so, append its name to a list.
15791 ((equal ".el" (substring (car (car current-directory-list)) -3))
15792 (setq el-files-list
15793 (cons (car (car current-directory-list)) el-files-list)))
15796 ;; check whether filename is that of a directory
15797 ((eq t (car (cdr (car current-directory-list))))
15798 ;; decide whether to skip or recurse
15801 (substring (car (car current-directory-list)) -1))
15802 ;; then do nothing since filename is that of
15803 ;; current directory or parent, "." or ".."
15807 ;; else descend into the directory and repeat the process
15808 (setq el-files-list
15810 (files-in-below-directory
15811 (car (car current-directory-list)))
15813 ;; move to the next filename in the list; this also
15814 ;; shortens the list so the while loop eventually comes to an end
15815 (setq current-directory-list (cdr current-directory-list)))
15816 ;; return the filenames
15821 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15822 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15824 The @code{files-in-below-directory} @code{directory-files} function
15825 takes one argument, the name of a directory.
15828 Thus, on my system,
15830 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15832 @c !!! 22.1.1 lisp sources location here
15836 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15841 tells me that in and below my Lisp sources directory are 1031
15844 @code{files-in-below-directory} returns a list in reverse alphabetical
15845 order. An expression to sort the list in alphabetical order looks
15851 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15858 "Test how long it takes to find lengths of all sorted elisp defuns."
15859 (insert "\n" (current-time-string) "\n")
15862 (recursive-lengths-list-many-files
15863 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15865 (insert (format "%s" (current-time-string))))
15868 @node Counting function definitions
15869 @subsection Counting function definitions
15871 Our immediate goal is to generate a list that tells us how many
15872 function definitions contain fewer than 10 words and symbols, how many
15873 contain between 10 and 19 words and symbols, how many contain between
15874 20 and 29 words and symbols, and so on.
15876 With a sorted list of numbers, this is easy: count how many elements
15877 of the list are smaller than 10, then, after moving past the numbers
15878 just counted, count how many are smaller than 20, then, after moving
15879 past the numbers just counted, count how many are smaller than 30, and
15880 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
15881 larger than the top of that range. We can call the list of such
15882 numbers the @code{top-of-ranges} list.
15885 If we wished, we could generate this list automatically, but it is
15886 simpler to write a list manually. Here it is:
15887 @vindex top-of-ranges
15891 (defvar top-of-ranges
15894 110 120 130 140 150
15895 160 170 180 190 200
15896 210 220 230 240 250
15897 260 270 280 290 300)
15898 "List specifying ranges for `defuns-per-range'.")
15902 To change the ranges, we edit this list.
15904 Next, we need to write the function that creates the list of the
15905 number of definitions within each range. Clearly, this function must
15906 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
15909 The @code{defuns-per-range} function must do two things again and
15910 again: it must count the number of definitions within a range
15911 specified by the current top-of-range value; and it must shift to the
15912 next higher value in the @code{top-of-ranges} list after counting the
15913 number of definitions in the current range. Since each of these
15914 actions is repetitive, we can use @code{while} loops for the job.
15915 One loop counts the number of definitions in the range defined by the
15916 current top-of-range value, and the other loop selects each of the
15917 top-of-range values in turn.
15919 Several entries of the @code{sorted-lengths} list are counted for each
15920 range; this means that the loop for the @code{sorted-lengths} list
15921 will be inside the loop for the @code{top-of-ranges} list, like a
15922 small gear inside a big gear.
15924 The inner loop counts the number of definitions within the range. It
15925 is a simple counting loop of the type we have seen before.
15926 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
15927 The true-or-false test of the loop tests whether the value from the
15928 @code{sorted-lengths} list is smaller than the current value of the
15929 top of the range. If it is, the function increments the counter and
15930 tests the next value from the @code{sorted-lengths} list.
15933 The inner loop looks like this:
15937 (while @var{length-element-smaller-than-top-of-range}
15938 (setq number-within-range (1+ number-within-range))
15939 (setq sorted-lengths (cdr sorted-lengths)))
15943 The outer loop must start with the lowest value of the
15944 @code{top-of-ranges} list, and then be set to each of the succeeding
15945 higher values in turn. This can be done with a loop like this:
15949 (while top-of-ranges
15950 @var{body-of-loop}@dots{}
15951 (setq top-of-ranges (cdr top-of-ranges)))
15956 Put together, the two loops look like this:
15960 (while top-of-ranges
15962 ;; @r{Count the number of elements within the current range.}
15963 (while @var{length-element-smaller-than-top-of-range}
15964 (setq number-within-range (1+ number-within-range))
15965 (setq sorted-lengths (cdr sorted-lengths)))
15967 ;; @r{Move to next range.}
15968 (setq top-of-ranges (cdr top-of-ranges)))
15972 In addition, in each circuit of the outer loop, Emacs should record
15973 the number of definitions within that range (the value of
15974 @code{number-within-range}) in a list. We can use @code{cons} for
15975 this purpose. (@xref{cons, , @code{cons}}.)
15977 The @code{cons} function works fine, except that the list it
15978 constructs will contain the number of definitions for the highest
15979 range at its beginning and the number of definitions for the lowest
15980 range at its end. This is because @code{cons} attaches new elements
15981 of the list to the beginning of the list, and since the two loops are
15982 working their way through the lengths' list from the lower end first,
15983 the @code{defuns-per-range-list} will end up largest number first.
15984 But we will want to print our graph with smallest values first and the
15985 larger later. The solution is to reverse the order of the
15986 @code{defuns-per-range-list}. We can do this using the
15987 @code{nreverse} function, which reverses the order of a list.
15994 (nreverse '(1 2 3 4))
16005 Note that the @code{nreverse} function is ``destructive''---that is,
16006 it changes the list to which it is applied; this contrasts with the
16007 @code{car} and @code{cdr} functions, which are non-destructive. In
16008 this case, we do not want the original @code{defuns-per-range-list},
16009 so it does not matter that it is destroyed. (The @code{reverse}
16010 function provides a reversed copy of a list, leaving the original list
16015 Put all together, the @code{defuns-per-range} looks like this:
16019 (defun defuns-per-range (sorted-lengths top-of-ranges)
16020 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16021 (let ((top-of-range (car top-of-ranges))
16022 (number-within-range 0)
16023 defuns-per-range-list)
16028 (while top-of-ranges
16034 ;; @r{Need number for numeric test.}
16035 (car sorted-lengths)
16036 (< (car sorted-lengths) top-of-range))
16040 ;; @r{Count number of definitions within current range.}
16041 (setq number-within-range (1+ number-within-range))
16042 (setq sorted-lengths (cdr sorted-lengths)))
16044 ;; @r{Exit inner loop but remain within outer loop.}
16048 (setq defuns-per-range-list
16049 (cons number-within-range defuns-per-range-list))
16050 (setq number-within-range 0) ; @r{Reset count to zero.}
16054 ;; @r{Move to next range.}
16055 (setq top-of-ranges (cdr top-of-ranges))
16056 ;; @r{Specify next top of range value.}
16057 (setq top-of-range (car top-of-ranges)))
16061 ;; @r{Exit outer loop and count the number of defuns larger than}
16062 ;; @r{ the largest top-of-range value.}
16063 (setq defuns-per-range-list
16065 (length sorted-lengths)
16066 defuns-per-range-list))
16070 ;; @r{Return a list of the number of definitions within each range,}
16071 ;; @r{ smallest to largest.}
16072 (nreverse defuns-per-range-list)))
16078 The function is straightforward except for one subtle feature. The
16079 true-or-false test of the inner loop looks like this:
16083 (and (car sorted-lengths)
16084 (< (car sorted-lengths) top-of-range))
16090 instead of like this:
16093 (< (car sorted-lengths) top-of-range)
16096 The purpose of the test is to determine whether the first item in the
16097 @code{sorted-lengths} list is less than the value of the top of the
16100 The simple version of the test works fine unless the
16101 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16102 @code{(car sorted-lengths)} expression function returns
16103 @code{nil}. The @code{<} function cannot compare a number to
16104 @code{nil}, which is an empty list, so Emacs signals an error and
16105 stops the function from attempting to continue to execute.
16107 The @code{sorted-lengths} list always becomes @code{nil} when the
16108 counter reaches the end of the list. This means that any attempt to
16109 use the @code{defuns-per-range} function with the simple version of
16110 the test will fail.
16112 We solve the problem by using the @code{(car sorted-lengths)}
16113 expression in conjunction with the @code{and} expression. The
16114 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16115 value so long as the list has at least one number within it, but
16116 returns @code{nil} if the list is empty. The @code{and} expression
16117 first evaluates the @code{(car sorted-lengths)} expression, and
16118 if it is @code{nil}, returns false @emph{without} evaluating the
16119 @code{<} expression. But if the @code{(car sorted-lengths)}
16120 expression returns a non-@code{nil} value, the @code{and} expression
16121 evaluates the @code{<} expression, and returns that value as the value
16122 of the @code{and} expression.
16124 @c colon in printed section title causes problem in Info cross reference
16125 This way, we avoid an error.
16128 (For information about @code{and}, see
16129 @ref{kill-new function, , The @code{kill-new} function}.)
16133 (@xref{kill-new function, , The @code{kill-new} function}, for
16134 information about @code{and}.)
16137 Here is a short test of the @code{defuns-per-range} function. First,
16138 evaluate the expression that binds (a shortened)
16139 @code{top-of-ranges} list to the list of values, then evaluate the
16140 expression for binding the @code{sorted-lengths} list, and then
16141 evaluate the @code{defuns-per-range} function.
16145 ;; @r{(Shorter list than we will use later.)}
16146 (setq top-of-ranges
16147 '(110 120 130 140 150
16148 160 170 180 190 200))
16150 (setq sorted-lengths
16151 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16153 (defuns-per-range sorted-lengths top-of-ranges)
16159 The list returned looks like this:
16162 (2 2 2 0 0 1 0 2 0 0 4)
16166 Indeed, there are two elements of the @code{sorted-lengths} list
16167 smaller than 110, two elements between 110 and 119, two elements
16168 between 120 and 129, and so on. There are four elements with a value
16171 @c The next step is to turn this numbers' list into a graph.
16172 @node Readying a Graph
16173 @chapter Readying a Graph
16174 @cindex Readying a graph
16175 @cindex Graph prototype
16176 @cindex Prototype graph
16177 @cindex Body of graph
16179 Our goal is to construct a graph showing the numbers of function
16180 definitions of various lengths in the Emacs lisp sources.
16182 As a practical matter, if you were creating a graph, you would
16183 probably use a program such as @code{gnuplot} to do the job.
16184 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16185 however, we create one from scratch, and in the process we will
16186 re-acquaint ourselves with some of what we learned before and learn
16189 In this chapter, we will first write a simple graph printing function.
16190 This first definition will be a @dfn{prototype}, a rapidly written
16191 function that enables us to reconnoiter this unknown graph-making
16192 territory. We will discover dragons, or find that they are myth.
16193 After scouting the terrain, we will feel more confident and enhance
16194 the function to label the axes automatically.
16197 * Columns of a graph::
16198 * graph-body-print:: How to print the body of a graph.
16199 * recursive-graph-body-print::
16201 * Line Graph Exercise::
16205 @node Columns of a graph
16206 @unnumberedsec Printing the Columns of a Graph
16209 Since Emacs is designed to be flexible and work with all kinds of
16210 terminals, including character-only terminals, the graph will need to
16211 be made from one of the `typewriter' symbols. An asterisk will do; as
16212 we enhance the graph-printing function, we can make the choice of
16213 symbol a user option.
16215 We can call this function @code{graph-body-print}; it will take a
16216 @code{numbers-list} as its only argument. At this stage, we will not
16217 label the graph, but only print its body.
16219 The @code{graph-body-print} function inserts a vertical column of
16220 asterisks for each element in the @code{numbers-list}. The height of
16221 each line is determined by the value of that element of the
16222 @code{numbers-list}.
16224 Inserting columns is a repetitive act; that means that this function can
16225 be written either with a @code{while} loop or recursively.
16227 Our first challenge is to discover how to print a column of asterisks.
16228 Usually, in Emacs, we print characters onto a screen horizontally,
16229 line by line, by typing. We have two routes we can follow: write our
16230 own column-insertion function or discover whether one exists in Emacs.
16232 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16233 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16234 command, except that the latter finds only those functions that are
16235 commands. The @kbd{M-x apropos} command lists all symbols that match
16236 a regular expression, including functions that are not interactive.
16239 What we want to look for is some command that prints or inserts
16240 columns. Very likely, the name of the function will contain either
16241 the word `print' or the word `insert' or the word `column'.
16242 Therefore, we can simply type @kbd{M-x apropos RET
16243 print\|insert\|column RET} and look at the result. On my system, this
16244 command once too takes quite some time, and then produced a list of 79
16245 functions and variables. Now it does not take much time at all and
16246 produces a list of 211 functions and variables. Scanning down the
16247 list, the only function that looks as if it might do the job is
16248 @code{insert-rectangle}.
16251 Indeed, this is the function we want; its documentation says:
16256 Insert text of RECTANGLE with upper left corner at point.
16257 RECTANGLE's first line is inserted at point,
16258 its second line is inserted at a point vertically under point, etc.
16259 RECTANGLE should be a list of strings.
16260 After this command, the mark is at the upper left corner
16261 and point is at the lower right corner.
16265 We can run a quick test, to make sure it does what we expect of it.
16267 Here is the result of placing the cursor after the
16268 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16269 (@code{eval-last-sexp}). The function inserts the strings
16270 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16271 point. Also the function returns @code{nil}.
16275 (insert-rectangle '("first" "second" "third"))first
16282 Of course, we won't be inserting the text of the
16283 @code{insert-rectangle} expression itself into the buffer in which we
16284 are making the graph, but will call the function from our program. We
16285 shall, however, have to make sure that point is in the buffer at the
16286 place where the @code{insert-rectangle} function will insert its
16289 If you are reading this in Info, you can see how this works by
16290 switching to another buffer, such as the @file{*scratch*} buffer,
16291 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16292 @code{insert-rectangle} expression into the minibuffer at the prompt,
16293 and then typing @key{RET}. This causes Emacs to evaluate the
16294 expression in the minibuffer, but to use as the value of point the
16295 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16296 keybinding for @code{eval-expression}. Also, @code{nil} does not
16297 appear in the @file{*scratch*} buffer since the expression is
16298 evaluated in the minibuffer.)
16300 We find when we do this that point ends up at the end of the last
16301 inserted line---that is to say, this function moves point as a
16302 side-effect. If we were to repeat the command, with point at this
16303 position, the next insertion would be below and to the right of the
16304 previous insertion. We don't want this! If we are going to make a
16305 bar graph, the columns need to be beside each other.
16307 So we discover that each cycle of the column-inserting @code{while}
16308 loop must reposition point to the place we want it, and that place
16309 will be at the top, not the bottom, of the column. Moreover, we
16310 remember that when we print a graph, we do not expect all the columns
16311 to be the same height. This means that the top of each column may be
16312 at a different height from the previous one. We cannot simply
16313 reposition point to the same line each time, but moved over to the
16314 right---or perhaps we can@dots{}
16316 We are planning to make the columns of the bar graph out of asterisks.
16317 The number of asterisks in the column is the number specified by the
16318 current element of the @code{numbers-list}. We need to construct a
16319 list of asterisks of the right length for each call to
16320 @code{insert-rectangle}. If this list consists solely of the requisite
16321 number of asterisks, then we will have position point the right number
16322 of lines above the base for the graph to print correctly. This could
16325 Alternatively, if we can figure out some way to pass
16326 @code{insert-rectangle} a list of the same length each time, then we
16327 can place point on the same line each time, but move it over one
16328 column to the right for each new column. If we do this, however, some
16329 of the entries in the list passed to @code{insert-rectangle} must be
16330 blanks rather than asterisks. For example, if the maximum height of
16331 the graph is 5, but the height of the column is 3, then
16332 @code{insert-rectangle} requires an argument that looks like this:
16335 (" " " " "*" "*" "*")
16338 This last proposal is not so difficult, so long as we can determine
16339 the column height. There are two ways for us to specify the column
16340 height: we can arbitrarily state what it will be, which would work
16341 fine for graphs of that height; or we can search through the list of
16342 numbers and use the maximum height of the list as the maximum height
16343 of the graph. If the latter operation were difficult, then the former
16344 procedure would be easiest, but there is a function built into Emacs
16345 that determines the maximum of its arguments. We can use that
16346 function. The function is called @code{max} and it returns the
16347 largest of all its arguments, which must be numbers. Thus, for
16355 returns 7. (A corresponding function called @code{min} returns the
16356 smallest of all its arguments.)
16360 However, we cannot simply call @code{max} on the @code{numbers-list};
16361 the @code{max} function expects numbers as its argument, not a list of
16362 numbers. Thus, the following expression,
16365 (max '(3 4 6 5 7 3))
16370 produces the following error message;
16373 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16377 We need a function that passes a list of arguments to a function.
16378 This function is @code{apply}. This function `applies' its first
16379 argument (a function) to its remaining arguments, the last of which
16386 (apply 'max 3 4 7 3 '(4 8 5))
16392 (Incidentally, I don't know how you would learn of this function
16393 without a book such as this. It is possible to discover other
16394 functions, like @code{search-forward} or @code{insert-rectangle}, by
16395 guessing at a part of their names and then using @code{apropos}. Even
16396 though its base in metaphor is clear---`apply' its first argument to
16397 the rest---I doubt a novice would come up with that particular word
16398 when using @code{apropos} or other aid. Of course, I could be wrong;
16399 after all, the function was first named by someone who had to invent
16402 The second and subsequent arguments to @code{apply} are optional, so
16403 we can use @code{apply} to call a function and pass the elements of a
16404 list to it, like this, which also returns 8:
16407 (apply 'max '(4 8 5))
16410 This latter way is how we will use @code{apply}. The
16411 @code{recursive-lengths-list-many-files} function returns a numbers'
16412 list to which we can apply @code{max} (we could also apply @code{max} to
16413 the sorted numbers' list; it does not matter whether the list is
16417 Hence, the operation for finding the maximum height of the graph is this:
16420 (setq max-graph-height (apply 'max numbers-list))
16423 Now we can return to the question of how to create a list of strings
16424 for a column of the graph. Told the maximum height of the graph
16425 and the number of asterisks that should appear in the column, the
16426 function should return a list of strings for the
16427 @code{insert-rectangle} command to insert.
16429 Each column is made up of asterisks or blanks. Since the function is
16430 passed the value of the height of the column and the number of
16431 asterisks in the column, the number of blanks can be found by
16432 subtracting the number of asterisks from the height of the column.
16433 Given the number of blanks and the number of asterisks, two
16434 @code{while} loops can be used to construct the list:
16438 ;;; @r{First version.}
16439 (defun column-of-graph (max-graph-height actual-height)
16440 "Return list of strings that is one column of a graph."
16441 (let ((insert-list nil)
16442 (number-of-top-blanks
16443 (- max-graph-height actual-height)))
16447 ;; @r{Fill in asterisks.}
16448 (while (> actual-height 0)
16449 (setq insert-list (cons "*" insert-list))
16450 (setq actual-height (1- actual-height)))
16454 ;; @r{Fill in blanks.}
16455 (while (> number-of-top-blanks 0)
16456 (setq insert-list (cons " " insert-list))
16457 (setq number-of-top-blanks
16458 (1- number-of-top-blanks)))
16462 ;; @r{Return whole list.}
16467 If you install this function and then evaluate the following
16468 expression you will see that it returns the list as desired:
16471 (column-of-graph 5 3)
16479 (" " " " "*" "*" "*")
16482 As written, @code{column-of-graph} contains a major flaw: the symbols
16483 used for the blank and for the marked entries in the column are
16484 `hard-coded' as a space and asterisk. This is fine for a prototype,
16485 but you, or another user, may wish to use other symbols. For example,
16486 in testing the graph function, you many want to use a period in place
16487 of the space, to make sure the point is being repositioned properly
16488 each time the @code{insert-rectangle} function is called; or you might
16489 want to substitute a @samp{+} sign or other symbol for the asterisk.
16490 You might even want to make a graph-column that is more than one
16491 display column wide. The program should be more flexible. The way to
16492 do that is to replace the blank and the asterisk with two variables
16493 that we can call @code{graph-blank} and @code{graph-symbol} and define
16494 those variables separately.
16496 Also, the documentation is not well written. These considerations
16497 lead us to the second version of the function:
16501 (defvar graph-symbol "*"
16502 "String used as symbol in graph, usually an asterisk.")
16506 (defvar graph-blank " "
16507 "String used as blank in graph, usually a blank space.
16508 graph-blank must be the same number of columns wide
16514 (For an explanation of @code{defvar}, see
16515 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16519 ;;; @r{Second version.}
16520 (defun column-of-graph (max-graph-height actual-height)
16521 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16525 The graph-symbols are contiguous entries at the end
16527 The list will be inserted as one column of a graph.
16528 The strings are either graph-blank or graph-symbol."
16532 (let ((insert-list nil)
16533 (number-of-top-blanks
16534 (- max-graph-height actual-height)))
16538 ;; @r{Fill in @code{graph-symbols}.}
16539 (while (> actual-height 0)
16540 (setq insert-list (cons graph-symbol insert-list))
16541 (setq actual-height (1- actual-height)))
16545 ;; @r{Fill in @code{graph-blanks}.}
16546 (while (> number-of-top-blanks 0)
16547 (setq insert-list (cons graph-blank insert-list))
16548 (setq number-of-top-blanks
16549 (1- number-of-top-blanks)))
16551 ;; @r{Return whole list.}
16556 If we wished, we could rewrite @code{column-of-graph} a third time to
16557 provide optionally for a line graph as well as for a bar graph. This
16558 would not be hard to do. One way to think of a line graph is that it
16559 is no more than a bar graph in which the part of each bar that is
16560 below the top is blank. To construct a column for a line graph, the
16561 function first constructs a list of blanks that is one shorter than
16562 the value, then it uses @code{cons} to attach a graph symbol to the
16563 list; then it uses @code{cons} again to attach the `top blanks' to
16566 It is easy to see how to write such a function, but since we don't
16567 need it, we will not do it. But the job could be done, and if it were
16568 done, it would be done with @code{column-of-graph}. Even more
16569 important, it is worth noting that few changes would have to be made
16570 anywhere else. The enhancement, if we ever wish to make it, is
16573 Now, finally, we come to our first actual graph printing function.
16574 This prints the body of a graph, not the labels for the vertical and
16575 horizontal axes, so we can call this @code{graph-body-print}.
16577 @node graph-body-print
16578 @section The @code{graph-body-print} Function
16579 @findex graph-body-print
16581 After our preparation in the preceding section, the
16582 @code{graph-body-print} function is straightforward. The function
16583 will print column after column of asterisks and blanks, using the
16584 elements of a numbers' list to specify the number of asterisks in each
16585 column. This is a repetitive act, which means we can use a
16586 decrementing @code{while} loop or recursive function for the job. In
16587 this section, we will write the definition using a @code{while} loop.
16589 The @code{column-of-graph} function requires the height of the graph
16590 as an argument, so we should determine and record that as a local variable.
16592 This leads us to the following template for the @code{while} loop
16593 version of this function:
16597 (defun graph-body-print (numbers-list)
16598 "@var{documentation}@dots{}"
16599 (let ((height @dots{}
16604 (while numbers-list
16605 @var{insert-columns-and-reposition-point}
16606 (setq numbers-list (cdr numbers-list)))))
16611 We need to fill in the slots of the template.
16613 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16614 determine the height of the graph.
16616 The @code{while} loop will cycle through the @code{numbers-list} one
16617 element at a time. As it is shortened by the @code{(setq numbers-list
16618 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16619 list is the value of the argument for @code{column-of-graph}.
16621 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16622 function inserts the list returned by @code{column-of-graph}. Since
16623 the @code{insert-rectangle} function moves point to the lower right of
16624 the inserted rectangle, we need to save the location of point at the
16625 time the rectangle is inserted, move back to that position after the
16626 rectangle is inserted, and then move horizontally to the next place
16627 from which @code{insert-rectangle} is called.
16629 If the inserted columns are one character wide, as they will be if
16630 single blanks and asterisks are used, the repositioning command is
16631 simply @code{(forward-char 1)}; however, the width of a column may be
16632 greater than one. This means that the repositioning command should be
16633 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16634 itself is the length of a @code{graph-blank} and can be found using
16635 the expression @code{(length graph-blank)}. The best place to bind
16636 the @code{symbol-width} variable to the value of the width of graph
16637 column is in the varlist of the @code{let} expression.
16640 These considerations lead to the following function definition:
16644 (defun graph-body-print (numbers-list)
16645 "Print a bar graph of the NUMBERS-LIST.
16646 The numbers-list consists of the Y-axis values."
16648 (let ((height (apply 'max numbers-list))
16649 (symbol-width (length graph-blank))
16654 (while numbers-list
16655 (setq from-position (point))
16657 (column-of-graph height (car numbers-list)))
16658 (goto-char from-position)
16659 (forward-char symbol-width)
16662 ;; @r{Draw graph column by column.}
16664 (setq numbers-list (cdr numbers-list)))
16667 ;; @r{Place point for X axis labels.}
16668 (forward-line height)
16675 The one unexpected expression in this function is the
16676 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16677 expression makes the graph printing operation more interesting to
16678 watch than it would be otherwise. The expression causes Emacs to
16679 `sit' or do nothing for a zero length of time and then redraw the
16680 screen. Placed here, it causes Emacs to redraw the screen column by
16681 column. Without it, Emacs would not redraw the screen until the
16684 We can test @code{graph-body-print} with a short list of numbers.
16688 Install @code{graph-symbol}, @code{graph-blank},
16689 @code{column-of-graph}, which are in
16691 @ref{Readying a Graph, , Readying a Graph},
16694 @ref{Columns of a graph},
16696 and @code{graph-body-print}.
16700 Copy the following expression:
16703 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16707 Switch to the @file{*scratch*} buffer and place the cursor where you
16708 want the graph to start.
16711 Type @kbd{M-:} (@code{eval-expression}).
16714 Yank the @code{graph-body-print} expression into the minibuffer
16715 with @kbd{C-y} (@code{yank)}.
16718 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16722 Emacs will print a graph like this:
16736 @node recursive-graph-body-print
16737 @section The @code{recursive-graph-body-print} Function
16738 @findex recursive-graph-body-print
16740 The @code{graph-body-print} function may also be written recursively.
16741 The recursive solution is divided into two parts: an outside `wrapper'
16742 that uses a @code{let} expression to determine the values of several
16743 variables that need only be found once, such as the maximum height of
16744 the graph, and an inside function that is called recursively to print
16748 The `wrapper' is uncomplicated:
16752 (defun recursive-graph-body-print (numbers-list)
16753 "Print a bar graph of the NUMBERS-LIST.
16754 The numbers-list consists of the Y-axis values."
16755 (let ((height (apply 'max numbers-list))
16756 (symbol-width (length graph-blank))
16758 (recursive-graph-body-print-internal
16765 The recursive function is a little more difficult. It has four parts:
16766 the `do-again-test', the printing code, the recursive call, and the
16767 `next-step-expression'. The `do-again-test' is a @code{when}
16768 expression that determines whether the @code{numbers-list} contains
16769 any remaining elements; if it does, the function prints one column of
16770 the graph using the printing code and calls itself again. The
16771 function calls itself again according to the value produced by the
16772 `next-step-expression' which causes the call to act on a shorter
16773 version of the @code{numbers-list}.
16777 (defun recursive-graph-body-print-internal
16778 (numbers-list height symbol-width)
16779 "Print a bar graph.
16780 Used within recursive-graph-body-print function."
16785 (setq from-position (point))
16787 (column-of-graph height (car numbers-list)))
16790 (goto-char from-position)
16791 (forward-char symbol-width)
16792 (sit-for 0) ; @r{Draw graph column by column.}
16793 (recursive-graph-body-print-internal
16794 (cdr numbers-list) height symbol-width)))
16799 After installation, this expression can be tested; here is a sample:
16802 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16806 Here is what @code{recursive-graph-body-print} produces:
16820 Either of these two functions, @code{graph-body-print} or
16821 @code{recursive-graph-body-print}, create the body of a graph.
16824 @section Need for Printed Axes
16826 A graph needs printed axes, so you can orient yourself. For a do-once
16827 project, it may be reasonable to draw the axes by hand using Emacs's
16828 Picture mode; but a graph drawing function may be used more than once.
16830 For this reason, I have written enhancements to the basic
16831 @code{print-graph-body} function that automatically print labels for
16832 the horizontal and vertical axes. Since the label printing functions
16833 do not contain much new material, I have placed their description in
16834 an appendix. @xref{Full Graph, , A Graph with Labeled Axes}.
16836 @node Line Graph Exercise
16839 Write a line graph version of the graph printing functions.
16841 @node Emacs Initialization
16842 @chapter Your @file{.emacs} File
16843 @cindex @file{.emacs} file
16844 @cindex Customizing your @file{.emacs} file
16845 @cindex Initialization file
16847 ``You don't have to like Emacs to like it''---this seemingly
16848 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16849 the box' Emacs is a generic tool. Most people who use it, customize
16850 it to suit themselves.
16852 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16853 expressions in Emacs Lisp you can change or extend Emacs.
16856 * Default Configuration::
16857 * Site-wide Init:: You can write site-wide init files.
16858 * defcustom:: Emacs will write code for you.
16859 * Beginning a .emacs File:: How to write a @code{.emacs file}.
16860 * Text and Auto-fill:: Automatically wrap lines.
16861 * Mail Aliases:: Use abbreviations for email addresses.
16862 * Indent Tabs Mode:: Don't use tabs with @TeX{}
16863 * Keybindings:: Create some personal keybindings.
16864 * Keymaps:: More about key binding.
16865 * Loading Files:: Load (i.e., evaluate) files automatically.
16866 * Autoload:: Make functions available.
16867 * Simple Extension:: Define a function; bind it to a key.
16868 * X11 Colors:: Colors in X.
16870 * Mode Line:: How to customize your mode line.
16874 @node Default Configuration
16875 @unnumberedsec Emacs's Default Configuration
16878 There are those who appreciate Emacs's default configuration. After
16879 all, Emacs starts you in C mode when you edit a C file, starts you in
16880 Fortran mode when you edit a Fortran file, and starts you in
16881 Fundamental mode when you edit an unadorned file. This all makes
16882 sense, if you do not know who is going to use Emacs. Who knows what a
16883 person hopes to do with an unadorned file? Fundamental mode is the
16884 right default for such a file, just as C mode is the right default for
16885 editing C code. (Enough programming languages have syntaxes
16886 that enable them to share or nearly share features, so C mode is
16887 now provided by CC mode, the `C Collection'.)
16889 But when you do know who is going to use Emacs---you,
16890 yourself---then it makes sense to customize Emacs.
16892 For example, I seldom want Fundamental mode when I edit an
16893 otherwise undistinguished file; I want Text mode. This is why I
16894 customize Emacs: so it suits me.
16896 You can customize and extend Emacs by writing or adapting a
16897 @file{~/.emacs} file. This is your personal initialization file; its
16898 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
16899 may also add @file{.el} to @file{~/.emacs} and call it a
16900 @file{~/.emacs.el} file. In the past, you were forbidden to type the
16901 extra keystrokes that the name @file{~/.emacs.el} requires, but now
16902 you may. The new format is consistent with the Emacs Lisp file
16903 naming conventions; the old format saves typing.}
16905 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
16906 code yourself; or you can use Emacs's @code{customize} feature to write
16907 the code for you. You can combine your own expressions and
16908 auto-written Customize expressions in your @file{.emacs} file.
16910 (I myself prefer to write my own expressions, except for those,
16911 particularly fonts, that I find easier to manipulate using the
16912 @code{customize} command. I combine the two methods.)
16914 Most of this chapter is about writing expressions yourself. It
16915 describes a simple @file{.emacs} file; for more information, see
16916 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
16917 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
16920 @node Site-wide Init
16921 @section Site-wide Initialization Files
16923 @cindex @file{default.el} init file
16924 @cindex @file{site-init.el} init file
16925 @cindex @file{site-load.el} init file
16926 In addition to your personal initialization file, Emacs automatically
16927 loads various site-wide initialization files, if they exist. These
16928 have the same form as your @file{.emacs} file, but are loaded by
16931 Two site-wide initialization files, @file{site-load.el} and
16932 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
16933 `dumped' version of Emacs is created, as is most common. (Dumped
16934 copies of Emacs load more quickly. However, once a file is loaded and
16935 dumped, a change to it does not lead to a change in Emacs unless you
16936 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
16937 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
16938 @file{INSTALL} file.)
16940 Three other site-wide initialization files are loaded automatically
16941 each time you start Emacs, if they exist. These are
16942 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
16943 file, and @file{default.el}, and the terminal type file, which are both
16944 loaded @emph{after} your @file{.emacs} file.
16946 Settings and definitions in your @file{.emacs} file will overwrite
16947 conflicting settings and definitions in a @file{site-start.el} file,
16948 if it exists; but the settings and definitions in a @file{default.el}
16949 or terminal type file will overwrite those in your @file{.emacs} file.
16950 (You can prevent interference from a terminal type file by setting
16951 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
16952 Simple Extension}.)
16954 @c Rewritten to avoid overfull hbox.
16955 The @file{INSTALL} file that comes in the distribution contains
16956 descriptions of the @file{site-init.el} and @file{site-load.el} files.
16958 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
16959 control loading. These files are in the @file{lisp} directory of the
16960 Emacs distribution and are worth perusing.
16962 The @file{loaddefs.el} file contains a good many suggestions as to
16963 what to put into your own @file{.emacs} file, or into a site-wide
16964 initialization file.
16967 @section Specifying Variables using @code{defcustom}
16970 You can specify variables using @code{defcustom} so that you and
16971 others can then use Emacs's @code{customize} feature to set their
16972 values. (You cannot use @code{customize} to write function
16973 definitions; but you can write @code{defuns} in your @file{.emacs}
16974 file. Indeed, you can write any Lisp expression in your @file{.emacs}
16977 The @code{customize} feature depends on the @code{defcustom} macro.
16978 Although you can use @code{defvar} or @code{setq} for variables that
16979 users set, the @code{defcustom} macro is designed for the job.
16981 You can use your knowledge of @code{defvar} for writing the
16982 first three arguments for @code{defcustom}. The first argument to
16983 @code{defcustom} is the name of the variable. The second argument is
16984 the variable's initial value, if any; and this value is set only if
16985 the value has not already been set. The third argument is the
16988 The fourth and subsequent arguments to @code{defcustom} specify types
16989 and options; these are not featured in @code{defvar}. (These
16990 arguments are optional.)
16992 Each of these arguments consists of a keyword followed by a value.
16993 Each keyword starts with the colon character @samp{:}.
16996 For example, the customizable user option variable
16997 @code{text-mode-hook} looks like this:
17001 (defcustom text-mode-hook nil
17002 "Normal hook run when entering Text mode and many related modes."
17004 :options '(turn-on-auto-fill flyspell-mode)
17010 The name of the variable is @code{text-mode-hook}; it has no default
17011 value; and its documentation string tells you what it does.
17013 The @code{:type} keyword tells Emacs the kind of data to which
17014 @code{text-mode-hook} should be set and how to display the value in a
17015 Customization buffer.
17017 The @code{:options} keyword specifies a suggested list of values for
17018 the variable. Usually, @code{:options} applies to a hook.
17019 The list is only a suggestion; it is not exclusive; a person who sets
17020 the variable may set it to other values; the list shown following the
17021 @code{:options} keyword is intended to offer convenient choices to a
17024 Finally, the @code{:group} keyword tells the Emacs Customization
17025 command in which group the variable is located. This tells where to
17028 The @code{defcustom} function recognizes more than a dozen keywords.
17029 For more information, see @ref{Customization, , Writing Customization
17030 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17032 Consider @code{text-mode-hook} as an example.
17034 There are two ways to customize this variable. You can use the
17035 customization command or write the appropriate expressions yourself.
17038 Using the customization command, you can type:
17045 and find that the group for editing files of data is called `data'.
17046 Enter that group. Text Mode Hook is the first member. You can click
17047 on its various options, such as @code{turn-on-auto-fill}, to set the
17048 values. After you click on the button to
17051 Save for Future Sessions
17055 Emacs will write an expression into your @file{.emacs} file.
17056 It will look like this:
17060 (custom-set-variables
17061 ;; custom-set-variables was added by Custom.
17062 ;; If you edit it by hand, you could mess it up, so be careful.
17063 ;; Your init file should contain only one such instance.
17064 ;; If there is more than one, they won't work right.
17065 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17070 (The @code{text-mode-hook-identify} function tells
17071 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17072 It comes on automatically.)
17074 The @code{custom-set-variables} function works somewhat differently
17075 than a @code{setq}. While I have never learned the differences, I
17076 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17077 file by hand: I make the changes in what appears to me to be a
17078 reasonable manner and have not had any problems. Others prefer to use
17079 the Customization command and let Emacs do the work for them.
17081 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17082 This function sets the various font faces. Over time, I have set a
17083 considerable number of faces. Some of the time, I re-set them using
17084 @code{customize}; other times, I simply edit the
17085 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17087 The second way to customize your @code{text-mode-hook} is to set it
17088 yourself in your @file{.emacs} file using code that has nothing to do
17089 with the @code{custom-set-@dots{}} functions.
17092 When you do this, and later use @code{customize}, you will see a
17096 CHANGED outside Customize; operating on it here may be unreliable.
17100 This message is only a warning. If you click on the button to
17103 Save for Future Sessions
17107 Emacs will write a @code{custom-set-@dots{}} expression near the end
17108 of your @file{.emacs} file that will be evaluated after your
17109 hand-written expression. It will, therefore, overrule your
17110 hand-written expression. No harm will be done. When you do this,
17111 however, be careful to remember which expression is active; if you
17112 forget, you may confuse yourself.
17114 So long as you remember where the values are set, you will have no
17115 trouble. In any event, the values are always set in your
17116 initialization file, which is usually called @file{.emacs}.
17118 I myself use @code{customize} for hardly anything. Mostly, I write
17119 expressions myself.
17123 Incidentally, to be more complete concerning defines: @code{defsubst}
17124 defines an inline function. The syntax is just like that of
17125 @code{defun}. @code{defconst} defines a symbol as a constant. The
17126 intent is that neither programs nor users should ever change a value
17127 set by @code{defconst}. (You can change it; the value set is a
17128 variable; but please do not.)
17130 @node Beginning a .emacs File
17131 @section Beginning a @file{.emacs} File
17132 @cindex @file{.emacs} file, beginning of
17134 When you start Emacs, it loads your @file{.emacs} file unless you tell
17135 it not to by specifying @samp{-q} on the command line. (The
17136 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17138 A @file{.emacs} file contains Lisp expressions. Often, these are no
17139 more than expressions to set values; sometimes they are function
17142 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17143 Manual}, for a short description of initialization files.
17145 This chapter goes over some of the same ground, but is a walk among
17146 extracts from a complete, long-used @file{.emacs} file---my own.
17148 The first part of the file consists of comments: reminders to myself.
17149 By now, of course, I remember these things, but when I started, I did
17155 ;;;; Bob's .emacs file
17156 ; Robert J. Chassell
17157 ; 26 September 1985
17162 Look at that date! I started this file a long time ago. I have been
17163 adding to it ever since.
17167 ; Each section in this file is introduced by a
17168 ; line beginning with four semicolons; and each
17169 ; entry is introduced by a line beginning with
17170 ; three semicolons.
17175 This describes the usual conventions for comments in Emacs Lisp.
17176 Everything on a line that follows a semicolon is a comment. Two,
17177 three, and four semicolons are used as subsection and section markers.
17178 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17179 more about comments.)
17184 ; Control-h is the help key;
17185 ; after typing control-h, type a letter to
17186 ; indicate the subject about which you want help.
17187 ; For an explanation of the help facility,
17188 ; type control-h two times in a row.
17193 Just remember: type @kbd{C-h} two times for help.
17197 ; To find out about any mode, type control-h m
17198 ; while in that mode. For example, to find out
17199 ; about mail mode, enter mail mode and then type
17205 `Mode help', as I call this, is very helpful. Usually, it tells you
17206 all you need to know.
17208 Of course, you don't need to include comments like these in your
17209 @file{.emacs} file. I included them in mine because I kept forgetting
17210 about Mode help or the conventions for comments---but I was able to
17211 remember to look here to remind myself.
17213 @node Text and Auto-fill
17214 @section Text and Auto Fill Mode
17216 Now we come to the part that `turns on' Text mode and
17221 ;;; Text mode and Auto Fill mode
17222 ;; The next two lines put Emacs into Text mode
17223 ;; and Auto Fill mode, and are for writers who
17224 ;; want to start writing prose rather than code.
17225 (setq-default major-mode 'text-mode)
17226 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17230 Here is the first part of this @file{.emacs} file that does something
17231 besides remind a forgetful human!
17233 The first of the two lines in parentheses tells Emacs to turn on Text
17234 mode when you find a file, @emph{unless} that file should go into some
17235 other mode, such as C mode.
17237 @cindex Per-buffer, local variables list
17238 @cindex Local variables list, per-buffer,
17239 @cindex Automatic mode selection
17240 @cindex Mode selection, automatic
17241 When Emacs reads a file, it looks at the extension to the file name,
17242 if any. (The extension is the part that comes after a @samp{.}.) If
17243 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17244 on C mode. Also, Emacs looks at first nonblank line of the file; if
17245 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17246 possesses a list of extensions and specifications that it uses
17247 automatically. In addition, Emacs looks near the last page for a
17248 per-buffer, ``local variables list'', if any.
17251 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17254 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17258 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17259 Files'' in @cite{The GNU Emacs Manual}.
17262 Now, back to the @file{.emacs} file.
17265 Here is the line again; how does it work?
17267 @cindex Text Mode turned on
17269 (setq major-mode 'text-mode)
17273 This line is a short, but complete Emacs Lisp expression.
17275 We are already familiar with @code{setq}. It sets the following variable,
17276 @code{major-mode}, to the subsequent value, which is @code{text-mode}.
17277 The single quote mark before @code{text-mode} tells Emacs to deal directly
17278 with the @code{text-mode} symbol, not with whatever it might stand for.
17279 @xref{set & setq, , Setting the Value of a Variable},
17280 for a reminder of how @code{setq} works.
17281 The main point is that there is no difference between the procedure you
17282 use to set a value in your @file{.emacs} file and the procedure you use
17283 anywhere else in Emacs.
17286 Here is the next line:
17288 @cindex Auto Fill mode turned on
17291 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17295 In this line, the @code{add-hook} command adds
17296 @code{turn-on-auto-fill} to the variable.
17298 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17299 it!, turns on Auto Fill mode.
17301 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17302 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17303 turns on Auto Fill mode.
17305 In brief, the first line causes Emacs to enter Text mode when you edit a
17306 file, unless the file name extension, a first non-blank line, or local
17307 variables to tell Emacs otherwise.
17309 Text mode among other actions, sets the syntax table to work
17310 conveniently for writers. In Text mode, Emacs considers an apostrophe
17311 as part of a word like a letter; but Emacs does not consider a period
17312 or a space as part of a word. Thus, @kbd{M-f} moves you over
17313 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17314 the @samp{t} of @samp{it's}.
17316 The second line causes Emacs to turn on Auto Fill mode when it turns
17317 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17318 that is too wide and brings the excessively wide part of the line down
17319 to the next line. Emacs breaks lines between words, not within them.
17321 When Auto Fill mode is turned off, lines continue to the right as you
17322 type them. Depending on how you set the value of
17323 @code{truncate-lines}, the words you type either disappear off the
17324 right side of the screen, or else are shown, in a rather ugly and
17325 unreadable manner, as a continuation line on the screen.
17328 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17329 fill commands to insert two spaces after a colon:
17332 (setq colon-double-space t)
17336 @section Mail Aliases
17338 Here is a @code{setq} that `turns on' mail aliases, along with more
17344 ; To enter mail mode, type `C-x m'
17345 ; To enter RMAIL (for reading mail),
17347 (setq mail-aliases t)
17351 @cindex Mail aliases
17353 This @code{setq} command sets the value of the variable
17354 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17355 says, in effect, ``Yes, use mail aliases.''
17357 Mail aliases are convenient short names for long email addresses or
17358 for lists of email addresses. The file where you keep your `aliases'
17359 is @file{~/.mailrc}. You write an alias like this:
17362 alias geo george@@foobar.wiz.edu
17366 When you write a message to George, address it to @samp{geo}; the
17367 mailer will automatically expand @samp{geo} to the full address.
17369 @node Indent Tabs Mode
17370 @section Indent Tabs Mode
17371 @cindex Tabs, preventing
17372 @findex indent-tabs-mode
17374 By default, Emacs inserts tabs in place of multiple spaces when it
17375 formats a region. (For example, you might indent many lines of text
17376 all at once with the @code{indent-region} command.) Tabs look fine on
17377 a terminal or with ordinary printing, but they produce badly indented
17378 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17381 The following turns off Indent Tabs mode:
17385 ;;; Prevent Extraneous Tabs
17386 (setq-default indent-tabs-mode nil)
17390 Note that this line uses @code{setq-default} rather than the
17391 @code{setq} command that we have seen before. The @code{setq-default}
17392 command sets values only in buffers that do not have their own local
17393 values for the variable.
17396 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17398 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17402 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17403 Files'' in @cite{The GNU Emacs Manual}.
17408 @section Some Keybindings
17410 Now for some personal keybindings:
17414 ;;; Compare windows
17415 (global-set-key "\C-cw" 'compare-windows)
17419 @findex compare-windows
17420 @code{compare-windows} is a nifty command that compares the text in
17421 your current window with text in the next window. It makes the
17422 comparison by starting at point in each window, moving over text in
17423 each window as far as they match. I use this command all the time.
17425 This also shows how to set a key globally, for all modes.
17427 @cindex Setting a key globally
17428 @cindex Global set key
17429 @cindex Key setting globally
17430 @findex global-set-key
17431 The command is @code{global-set-key}. It is followed by the
17432 keybinding. In a @file{.emacs} file, the keybinding is written as
17433 shown: @code{\C-c} stands for `control-c', which means `press the
17434 control key and the @key{c} key at the same time'. The @code{w} means
17435 `press the @key{w} key'. The keybinding is surrounded by double
17436 quotation marks. In documentation, you would write this as
17437 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17438 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17439 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17440 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17443 The command invoked by the keys is @code{compare-windows}. Note that
17444 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17445 would first try to evaluate the symbol to determine its value.
17447 These three things, the double quotation marks, the backslash before
17448 the @samp{C}, and the single quote mark are necessary parts of
17449 keybinding that I tend to forget. Fortunately, I have come to
17450 remember that I should look at my existing @file{.emacs} file, and
17451 adapt what is there.
17453 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17454 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17455 set of keys, @kbd{C-c} followed by a single character, is strictly
17456 reserved for individuals' own use. (I call these `own' keys, since
17457 these are for my own use.) You should always be able to create such a
17458 keybinding for your own use without stomping on someone else's
17459 keybinding. If you ever write an extension to Emacs, please avoid
17460 taking any of these keys for public use. Create a key like @kbd{C-c
17461 C-w} instead. Otherwise, we will run out of `own' keys.
17464 Here is another keybinding, with a comment:
17468 ;;; Keybinding for `occur'
17469 ; I use occur a lot, so let's bind it to a key:
17470 (global-set-key "\C-co" 'occur)
17475 The @code{occur} command shows all the lines in the current buffer
17476 that contain a match for a regular expression. Matching lines are
17477 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17478 to jump to occurrences.
17480 @findex global-unset-key
17481 @cindex Unbinding key
17482 @cindex Key unbinding
17484 Here is how to unbind a key, so it does not
17490 (global-unset-key "\C-xf")
17494 There is a reason for this unbinding: I found I inadvertently typed
17495 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17496 file, as I intended, I accidentally set the width for filled text,
17497 almost always to a width I did not want. Since I hardly ever reset my
17498 default width, I simply unbound the key.
17500 @findex list-buffers, @r{rebound}
17501 @findex buffer-menu, @r{bound to key}
17503 The following rebinds an existing key:
17507 ;;; Rebind `C-x C-b' for `buffer-menu'
17508 (global-set-key "\C-x\C-b" 'buffer-menu)
17512 By default, @kbd{C-x C-b} runs the
17513 @code{list-buffers} command. This command lists
17514 your buffers in @emph{another} window. Since I
17515 almost always want to do something in that
17516 window, I prefer the @code{buffer-menu}
17517 command, which not only lists the buffers,
17518 but moves point into that window.
17523 @cindex Rebinding keys
17525 Emacs uses @dfn{keymaps} to record which keys call which commands.
17526 When you use @code{global-set-key} to set the keybinding for a single
17527 command in all parts of Emacs, you are specifying the keybinding in
17528 @code{current-global-map}.
17530 Specific modes, such as C mode or Text mode, have their own keymaps;
17531 the mode-specific keymaps override the global map that is shared by
17534 The @code{global-set-key} function binds, or rebinds, the global
17535 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17536 function @code{buffer-menu}:
17539 (global-set-key "\C-x\C-b" 'buffer-menu)
17542 Mode-specific keymaps are bound using the @code{define-key} function,
17543 which takes a specific keymap as an argument, as well as the key and
17544 the command. For example, my @file{.emacs} file contains the
17545 following expression to bind the @code{texinfo-insert-@@group} command
17546 to @kbd{C-c C-c g}:
17550 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17555 The @code{texinfo-insert-@@group} function itself is a little extension
17556 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17557 use this command all the time and prefer to type the three strokes
17558 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17559 (@samp{@@group} and its matching @samp{@@end group} are commands that
17560 keep all enclosed text together on one page; many multi-line examples
17561 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17564 Here is the @code{texinfo-insert-@@group} function definition:
17568 (defun texinfo-insert-@@group ()
17569 "Insert the string @@group in a Texinfo buffer."
17571 (beginning-of-line)
17572 (insert "@@group\n"))
17576 (Of course, I could have used Abbrev mode to save typing, rather than
17577 write a function to insert a word; but I prefer key strokes consistent
17578 with other Texinfo mode key bindings.)
17580 You will see numerous @code{define-key} expressions in
17581 @file{loaddefs.el} as well as in the various mode libraries, such as
17582 @file{cc-mode.el} and @file{lisp-mode.el}.
17584 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17585 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17586 Reference Manual}, for more information about keymaps.
17588 @node Loading Files
17589 @section Loading Files
17590 @cindex Loading files
17593 Many people in the GNU Emacs community have written extensions to
17594 Emacs. As time goes by, these extensions are often included in new
17595 releases. For example, the Calendar and Diary packages are now part
17596 of the standard GNU Emacs, as is Calc.
17598 You can use a @code{load} command to evaluate a complete file and
17599 thereby install all the functions and variables in the file into Emacs.
17602 @c (auto-compression-mode t)
17605 (load "~/emacs/slowsplit")
17608 This evaluates, i.e., loads, the @file{slowsplit.el} file or if it
17609 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17610 @file{emacs} sub-directory of your home directory. The file contains
17611 the function @code{split-window-quietly}, which John Robinson wrote in
17614 The @code{split-window-quietly} function splits a window with the
17615 minimum of redisplay. I installed it in 1989 because it worked well
17616 with the slow 1200 baud terminals I was then using. Nowadays, I only
17617 occasionally come across such a slow connection, but I continue to use
17618 the function because I like the way it leaves the bottom half of a
17619 buffer in the lower of the new windows and the top half in the upper
17623 To replace the key binding for the default
17624 @code{split-window-vertically}, you must also unset that key and bind
17625 the keys to @code{split-window-quietly}, like this:
17629 (global-unset-key "\C-x2")
17630 (global-set-key "\C-x2" 'split-window-quietly)
17635 If you load many extensions, as I do, then instead of specifying the
17636 exact location of the extension file, as shown above, you can specify
17637 that directory as part of Emacs's @code{load-path}. Then, when Emacs
17638 loads a file, it will search that directory as well as its default
17639 list of directories. (The default list is specified in @file{paths.h}
17640 when Emacs is built.)
17643 The following command adds your @file{~/emacs} directory to the
17644 existing load path:
17648 ;;; Emacs Load Path
17649 (setq load-path (cons "~/emacs" load-path))
17653 Incidentally, @code{load-library} is an interactive interface to the
17654 @code{load} function. The complete function looks like this:
17656 @findex load-library
17659 (defun load-library (library)
17660 "Load the library named LIBRARY.
17661 This is an interface to the function `load'."
17663 (list (completing-read "Load library: "
17664 (apply-partially 'locate-file-completion-table
17666 (get-load-suffixes)))))
17671 The name of the function, @code{load-library}, comes from the use of
17672 `library' as a conventional synonym for `file'. The source for the
17673 @code{load-library} command is in the @file{files.el} library.
17675 Another interactive command that does a slightly different job is
17676 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17677 Emacs, emacs, The GNU Emacs Manual}, for information on the
17678 distinction between @code{load-library} and this command.
17681 @section Autoloading
17684 Instead of installing a function by loading the file that contains it,
17685 or by evaluating the function definition, you can make the function
17686 available but not actually install it until it is first called. This
17687 is called @dfn{autoloading}.
17689 When you execute an autoloaded function, Emacs automatically evaluates
17690 the file that contains the definition, and then calls the function.
17692 Emacs starts quicker with autoloaded functions, since their libraries
17693 are not loaded right away; but you need to wait a moment when you
17694 first use such a function, while its containing file is evaluated.
17696 Rarely used functions are frequently autoloaded. The
17697 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17698 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17699 come to use a `rare' function frequently. When you do, you should
17700 load that function's file with a @code{load} expression in your
17701 @file{.emacs} file.
17703 In my @file{.emacs} file, I load 14 libraries that contain functions
17704 that would otherwise be autoloaded. (Actually, it would have been
17705 better to include these files in my `dumped' Emacs, but I forgot.
17706 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17707 Reference Manual}, and the @file{INSTALL} file for more about
17710 You may also want to include autoloaded expressions in your @file{.emacs}
17711 file. @code{autoload} is a built-in function that takes up to five
17712 arguments, the final three of which are optional. The first argument
17713 is the name of the function to be autoloaded; the second is the name
17714 of the file to be loaded. The third argument is documentation for the
17715 function, and the fourth tells whether the function can be called
17716 interactively. The fifth argument tells what type of
17717 object---@code{autoload} can handle a keymap or macro as well as a
17718 function (the default is a function).
17721 Here is a typical example:
17725 (autoload 'html-helper-mode
17726 "html-helper-mode" "Edit HTML documents" t)
17731 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17732 which is a standard part of the distribution.)
17735 This expression autoloads the @code{html-helper-mode} function. It
17736 takes it from the @file{html-helper-mode.el} file (or from the byte
17737 compiled version @file{html-helper-mode.elc}, if that exists.) The
17738 file must be located in a directory specified by @code{load-path}.
17739 The documentation says that this is a mode to help you edit documents
17740 written in the HyperText Markup Language. You can call this mode
17741 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17742 duplicate the function's regular documentation in the autoload
17743 expression because the regular function is not yet loaded, so its
17744 documentation is not available.)
17746 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17747 Manual}, for more information.
17749 @node Simple Extension
17750 @section A Simple Extension: @code{line-to-top-of-window}
17751 @findex line-to-top-of-window
17752 @cindex Simple extension in @file{.emacs} file
17754 Here is a simple extension to Emacs that moves the line point is on to
17755 the top of the window. I use this all the time, to make text easier
17758 You can put the following code into a separate file and then load it
17759 from your @file{.emacs} file, or you can include it within your
17760 @file{.emacs} file.
17763 Here is the definition:
17767 ;;; Line to top of window;
17768 ;;; replace three keystroke sequence C-u 0 C-l
17769 (defun line-to-top-of-window ()
17770 "Move the line point is on to top of window."
17777 Now for the keybinding.
17779 Nowadays, function keys as well as mouse button events and
17780 non-@sc{ascii} characters are written within square brackets, without
17781 quotation marks. (In Emacs version 18 and before, you had to write
17782 different function key bindings for each different make of terminal.)
17784 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17788 (global-set-key [f6] 'line-to-top-of-window)
17791 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17792 Your Init File, emacs, The GNU Emacs Manual}.
17794 @cindex Conditional 'twixt two versions of Emacs
17795 @cindex Version of Emacs, choosing
17796 @cindex Emacs version, choosing
17797 If you run two versions of GNU Emacs, such as versions 22 and 23, and
17798 use one @file{.emacs} file, you can select which code to evaluate with
17799 the following conditional:
17804 ((= 22 emacs-major-version)
17805 ;; evaluate version 22 code
17807 ((= 23 emacs-major-version)
17808 ;; evaluate version 23 code
17813 For example, recent versions blink
17814 their cursors by default. I hate such blinking, as well as other
17815 features, so I placed the following in my @file{.emacs}
17816 file@footnote{When I start instances of Emacs that do not load my
17817 @file{.emacs} file or any site file, I also turn off blinking:
17820 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17822 @exdent Or nowadays, using an even more sophisticated set of options,
17830 (when (>= emacs-major-version 21)
17831 (blink-cursor-mode 0)
17832 ;; Insert newline when you press `C-n' (next-line)
17833 ;; at the end of the buffer
17834 (setq next-line-add-newlines t)
17837 ;; Turn on image viewing
17838 (auto-image-file-mode t)
17841 ;; Turn on menu bar (this bar has text)
17842 ;; (Use numeric argument to turn on)
17846 ;; Turn off tool bar (this bar has icons)
17847 ;; (Use numeric argument to turn on)
17848 (tool-bar-mode nil)
17851 ;; Turn off tooltip mode for tool bar
17852 ;; (This mode causes icon explanations to pop up)
17853 ;; (Use numeric argument to turn on)
17855 ;; If tooltips turned on, make tips appear promptly
17856 (setq tooltip-delay 0.1) ; default is 0.7 second
17862 @section X11 Colors
17864 You can specify colors when you use Emacs with the MIT X Windowing
17867 I dislike the default colors and specify my own.
17870 Here are the expressions in my @file{.emacs}
17871 file that set values:
17875 ;; Set cursor color
17876 (set-cursor-color "white")
17879 (set-mouse-color "white")
17881 ;; Set foreground and background
17882 (set-foreground-color "white")
17883 (set-background-color "darkblue")
17887 ;;; Set highlighting colors for isearch and drag
17888 (set-face-foreground 'highlight "white")
17889 (set-face-background 'highlight "blue")
17893 (set-face-foreground 'region "cyan")
17894 (set-face-background 'region "blue")
17898 (set-face-foreground 'secondary-selection "skyblue")
17899 (set-face-background 'secondary-selection "darkblue")
17903 ;; Set calendar highlighting colors
17904 (setq calendar-load-hook
17906 (set-face-foreground 'diary-face "skyblue")
17907 (set-face-background 'holiday-face "slate blue")
17908 (set-face-foreground 'holiday-face "white")))
17912 The various shades of blue soothe my eye and prevent me from seeing
17913 the screen flicker.
17915 Alternatively, I could have set my specifications in various X
17916 initialization files. For example, I could set the foreground,
17917 background, cursor, and pointer (i.e., mouse) colors in my
17918 @file{~/.Xresources} file like this:
17922 Emacs*foreground: white
17923 Emacs*background: darkblue
17924 Emacs*cursorColor: white
17925 Emacs*pointerColor: white
17929 In any event, since it is not part of Emacs, I set the root color of
17930 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
17931 run more modern window managers, such as Enlightenment, Gnome, or KDE;
17932 in those cases, I often specify an image rather than a plain color.}:
17935 xsetroot -solid Navy -fg white &
17939 @node Miscellaneous
17940 @section Miscellaneous Settings for a @file{.emacs} File
17943 Here are a few miscellaneous settings:
17948 Set the shape and color of the mouse cursor:
17952 ; Cursor shapes are defined in
17953 ; `/usr/include/X11/cursorfont.h';
17954 ; for example, the `target' cursor is number 128;
17955 ; the `top_left_arrow' cursor is number 132.
17959 (let ((mpointer (x-get-resource "*mpointer"
17960 "*emacs*mpointer")))
17961 ;; If you have not set your mouse pointer
17962 ;; then set it, otherwise leave as is:
17963 (if (eq mpointer nil)
17964 (setq mpointer "132")) ; top_left_arrow
17967 (setq x-pointer-shape (string-to-int mpointer))
17968 (set-mouse-color "white"))
17973 Or you can set the values of a variety of features in an alist, like
17979 default-frame-alist
17980 '((cursor-color . "white")
17981 (mouse-color . "white")
17982 (foreground-color . "white")
17983 (background-color . "DodgerBlue4")
17984 ;; (cursor-type . bar)
17985 (cursor-type . box)
17988 (tool-bar-lines . 0)
17989 (menu-bar-lines . 1)
17993 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
17999 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
18000 into @kbd{@key{CTRL}-h}.@*
18001 (Some older keyboards needed this, although I have not seen the
18006 ;; Translate `C-h' to <DEL>.
18007 ; (keyboard-translate ?\C-h ?\C-?)
18009 ;; Translate <DEL> to `C-h'.
18010 (keyboard-translate ?\C-? ?\C-h)
18014 @item Turn off a blinking cursor!
18018 (if (fboundp 'blink-cursor-mode)
18019 (blink-cursor-mode -1))
18024 or start GNU Emacs with the command @code{emacs -nbc}.
18027 @item When using `grep'@*
18028 @samp{-i}@w{ } Ignore case distinctions@*
18029 @samp{-n}@w{ } Prefix each line of output with line number@*
18030 @samp{-H}@w{ } Print the filename for each match.@*
18031 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18034 (setq grep-command "grep -i -nH -e ")
18038 @c Evidently, no longer needed in GNU Emacs 22
18040 item Automatically uncompress compressed files when visiting them
18043 (load "uncompress")
18048 @item Find an existing buffer, even if it has a different name@*
18049 This avoids problems with symbolic links.
18052 (setq find-file-existing-other-name t)
18055 @item Set your language environment and default input method
18059 (set-language-environment "latin-1")
18060 ;; Remember you can enable or disable multilingual text input
18061 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18062 (setq default-input-method "latin-1-prefix")
18066 If you want to write with Chinese `GB' characters, set this instead:
18070 (set-language-environment "Chinese-GB")
18071 (setq default-input-method "chinese-tonepy")
18076 @subsubheading Fixing Unpleasant Key Bindings
18077 @cindex Key bindings, fixing
18078 @cindex Bindings, key, fixing unpleasant
18080 Some systems bind keys unpleasantly. Sometimes, for example, the
18081 @key{CTRL} key appears in an awkward spot rather than at the far left
18084 Usually, when people fix these sorts of keybindings, they do not
18085 change their @file{~/.emacs} file. Instead, they bind the proper keys
18086 on their consoles with the @code{loadkeys} or @code{install-keymap}
18087 commands in their boot script and then include @code{xmodmap} commands
18088 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18096 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18098 install-keymap emacs2
18104 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18105 Lock} key is at the far left of the home row:
18109 # Bind the key labeled `Caps Lock' to `Control'
18110 # (Such a broken user interface suggests that keyboard manufacturers
18111 # think that computers are typewriters from 1885.)
18113 xmodmap -e "clear Lock"
18114 xmodmap -e "add Control = Caps_Lock"
18120 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18121 key to a @key{META} key:
18125 # Some ill designed keyboards have a key labeled ALT and no Meta
18126 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18132 @section A Modified Mode Line
18133 @vindex mode-line-format
18134 @cindex Mode line format
18136 Finally, a feature I really like: a modified mode line.
18138 When I work over a network, I forget which machine I am using. Also,
18139 I tend to I lose track of where I am, and which line point is on.
18141 So I reset my mode line to look like this:
18144 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18147 I am visiting a file called @file{foo.texi}, on my machine
18148 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18149 Texinfo mode, and am at the top of the buffer.
18152 My @file{.emacs} file has a section that looks like this:
18156 ;; Set a Mode Line that tells me which machine, which directory,
18157 ;; and which line I am on, plus the other customary information.
18158 (setq-default mode-line-format
18162 "mouse-1: select window, mouse-2: delete others ..."))
18163 mode-line-mule-info
18165 mode-line-frame-identification
18169 mode-line-buffer-identification
18172 (system-name) 0 (string-match "\\..+" (system-name))))
18177 "mouse-1: select window, mouse-2: delete others ..."))
18178 (line-number-mode " Line %l ")
18184 "mouse-1: select window, mouse-2: delete others ..."))
18185 (:eval (mode-line-mode-name))
18188 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18197 Here, I redefine the default mode line. Most of the parts are from
18198 the original; but I make a few changes. I set the @emph{default} mode
18199 line format so as to permit various modes, such as Info, to override
18202 Many elements in the list are self-explanatory:
18203 @code{mode-line-modified} is a variable that tells whether the buffer
18204 has been modified, @code{mode-name} tells the name of the mode, and so
18205 on. However, the format looks complicated because of two features we
18206 have not discussed.
18208 @cindex Properties, in mode line example
18209 The first string in the mode line is a dash or hyphen, @samp{-}. In
18210 the old days, it would have been specified simply as @code{"-"}. But
18211 nowadays, Emacs can add properties to a string, such as highlighting
18212 or, as in this case, a help feature. If you place your mouse cursor
18213 over the hyphen, some help information appears (By default, you must
18214 wait seven-tenths of a second before the information appears. You can
18215 change that timing by changing the value of @code{tooltip-delay}.)
18218 The new string format has a special syntax:
18221 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18225 The @code{#(} begins a list. The first element of the list is the
18226 string itself, just one @samp{-}. The second and third
18227 elements specify the range over which the fourth element applies. A
18228 range starts @emph{after} a character, so a zero means the range
18229 starts just before the first character; a 1 means that the range ends
18230 just after the first character. The third element is the property for
18231 the range. It consists of a property list, a
18232 property name, in this case, @samp{help-echo}, followed by a value, in this
18233 case, a string. The second, third, and fourth elements of this new
18234 string format can be repeated.
18236 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18237 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18238 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18240 @code{mode-line-buffer-identification}
18241 displays the current buffer name. It is a list
18242 beginning @code{(#("%12b" 0 4 @dots{}}.
18243 The @code{#(} begins the list.
18245 The @samp{"%12b"} displays the current buffer name, using the
18246 @code{buffer-name} function with which we are familiar; the `12'
18247 specifies the maximum number of characters that will be displayed.
18248 When a name has fewer characters, whitespace is added to fill out to
18249 this number. (Buffer names can and often should be longer than 12
18250 characters; this length works well in a typical 80 column wide
18253 @code{:eval} says to evaluate the following form and use the result as
18254 a string to display. In this case, the expression displays the first
18255 component of the full system name. The end of the first component is
18256 a @samp{.} (`period'), so I use the @code{string-match} function to
18257 tell me the length of the first component. The substring from the
18258 zeroth character to that length is the name of the machine.
18261 This is the expression:
18266 (system-name) 0 (string-match "\\..+" (system-name))))
18270 @samp{%[} and @samp{%]} cause a pair of square brackets
18271 to appear for each recursive editing level. @samp{%n} says `Narrow'
18272 when narrowing is in effect. @samp{%P} tells you the percentage of
18273 the buffer that is above the bottom of the window, or `Top', `Bottom',
18274 or `All'. (A lower case @samp{p} tell you the percentage above the
18275 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18278 Remember, ``You don't have to like Emacs to like it''---your own
18279 Emacs can have different colors, different commands, and different
18280 keys than a default Emacs.
18282 On the other hand, if you want to bring up a plain `out of the box'
18283 Emacs, with no customization, type:
18290 This will start an Emacs that does @emph{not} load your
18291 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18298 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18299 first is built into the internals of Emacs and is always with you;
18300 the second requires that you instrument a function before you can use it.
18302 Both debuggers are described extensively in @ref{Debugging, ,
18303 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18304 In this chapter, I will walk through a short example of each.
18307 * debug:: How to use the built-in debugger.
18308 * debug-on-entry:: Start debugging when you call a function.
18309 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18310 * edebug:: How to use Edebug, a source level debugger.
18311 * Debugging Exercises::
18315 @section @code{debug}
18318 Suppose you have written a function definition that is intended to
18319 return the sum of the numbers 1 through a given number. (This is the
18320 @code{triangle} function discussed earlier. @xref{Decrementing
18321 Example, , Example with Decrementing Counter}, for a discussion.)
18322 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18324 However, your function definition has a bug. You have mistyped
18325 @samp{1=} for @samp{1-}. Here is the broken definition:
18327 @findex triangle-bugged
18330 (defun triangle-bugged (number)
18331 "Return sum of numbers 1 through NUMBER inclusive."
18333 (while (> number 0)
18334 (setq total (+ total number))
18335 (setq number (1= number))) ; @r{Error here.}
18340 If you are reading this in Info, you can evaluate this definition in
18341 the normal fashion. You will see @code{triangle-bugged} appear in the
18345 Now evaluate the @code{triangle-bugged} function with an
18349 (triangle-bugged 4)
18353 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18359 ---------- Buffer: *Backtrace* ----------
18360 Debugger entered--Lisp error: (void-function 1=)
18362 (setq number (1= number))
18363 (while (> number 0) (setq total (+ total number))
18364 (setq number (1= number)))
18365 (let ((total 0)) (while (> number 0) (setq total ...)
18366 (setq number ...)) total)
18370 eval((triangle-bugged 4))
18371 eval-last-sexp-1(nil)
18372 eval-last-sexp(nil)
18373 call-interactively(eval-last-sexp)
18374 ---------- Buffer: *Backtrace* ----------
18379 (I have reformatted this example slightly; the debugger does not fold
18380 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18381 the @file{*Backtrace*} buffer.)
18383 In practice, for a bug as simple as this, the `Lisp error' line will
18384 tell you what you need to know to correct the definition. The
18385 function @code{1=} is `void'.
18389 In GNU Emacs 20 and before, you will see:
18392 Symbol's function definition is void:@: 1=
18396 which has the same meaning as the @file{*Backtrace*} buffer line in
18400 However, suppose you are not quite certain what is going on?
18401 You can read the complete backtrace.
18403 In this case, you need to run a recent GNU Emacs, which automatically
18404 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18405 else, you need to start the debugger manually as described below.
18407 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18408 what Emacs did that led to the error. Emacs made an interactive call
18409 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18410 of the @code{triangle-bugged} expression. Each line above tells you
18411 what the Lisp interpreter evaluated next.
18414 The third line from the top of the buffer is
18417 (setq number (1= number))
18421 Emacs tried to evaluate this expression; in order to do so, it tried
18422 to evaluate the inner expression shown on the second line from the
18431 This is where the error occurred; as the top line says:
18434 Debugger entered--Lisp error: (void-function 1=)
18438 You can correct the mistake, re-evaluate the function definition, and
18439 then run your test again.
18441 @node debug-on-entry
18442 @section @code{debug-on-entry}
18443 @findex debug-on-entry
18445 A recent GNU Emacs starts the debugger automatically when your
18446 function has an error.
18449 GNU Emacs version 20 and before did not; it simply
18450 presented you with an error message. You had to start the debugger
18454 Incidentally, you can start the debugger manually for all versions of
18455 Emacs; the advantage is that the debugger runs even if you do not have
18456 a bug in your code. Sometimes your code will be free of bugs!
18458 You can enter the debugger when you call the function by calling
18459 @code{debug-on-entry}.
18466 M-x debug-on-entry RET triangle-bugged RET
18471 Now, evaluate the following:
18474 (triangle-bugged 5)
18478 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18479 you that it is beginning to evaluate the @code{triangle-bugged}
18484 ---------- Buffer: *Backtrace* ----------
18485 Debugger entered--entering a function:
18486 * triangle-bugged(5)
18487 eval((triangle-bugged 5))
18490 eval-last-sexp-1(nil)
18491 eval-last-sexp(nil)
18492 call-interactively(eval-last-sexp)
18493 ---------- Buffer: *Backtrace* ----------
18497 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18498 the first expression in @code{triangle-bugged}; the buffer will look
18503 ---------- Buffer: *Backtrace* ----------
18504 Debugger entered--beginning evaluation of function call form:
18505 * (let ((total 0)) (while (> number 0) (setq total ...)
18506 (setq number ...)) total)
18507 * triangle-bugged(5)
18508 eval((triangle-bugged 5))
18511 eval-last-sexp-1(nil)
18512 eval-last-sexp(nil)
18513 call-interactively(eval-last-sexp)
18514 ---------- Buffer: *Backtrace* ----------
18519 Now, type @kbd{d} again, eight times, slowly. Each time you type
18520 @kbd{d}, Emacs will evaluate another expression in the function
18524 Eventually, the buffer will look like this:
18528 ---------- Buffer: *Backtrace* ----------
18529 Debugger entered--beginning evaluation of function call form:
18530 * (setq number (1= number))
18531 * (while (> number 0) (setq total (+ total number))
18532 (setq number (1= number)))
18535 * (let ((total 0)) (while (> number 0) (setq total ...)
18536 (setq number ...)) total)
18537 * triangle-bugged(5)
18538 eval((triangle-bugged 5))
18541 eval-last-sexp-1(nil)
18542 eval-last-sexp(nil)
18543 call-interactively(eval-last-sexp)
18544 ---------- Buffer: *Backtrace* ----------
18550 Finally, after you type @kbd{d} two more times, Emacs will reach the
18551 error, and the top two lines of the @file{*Backtrace*} buffer will look
18556 ---------- Buffer: *Backtrace* ----------
18557 Debugger entered--Lisp error: (void-function 1=)
18560 ---------- Buffer: *Backtrace* ----------
18564 By typing @kbd{d}, you were able to step through the function.
18566 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18567 quits the trace, but does not cancel @code{debug-on-entry}.
18569 @findex cancel-debug-on-entry
18570 To cancel the effect of @code{debug-on-entry}, call
18571 @code{cancel-debug-on-entry} and the name of the function, like this:
18574 M-x cancel-debug-on-entry RET triangle-bugged RET
18578 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18580 @node debug-on-quit
18581 @section @code{debug-on-quit} and @code{(debug)}
18583 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18584 there are two other ways to start @code{debug}.
18586 @findex debug-on-quit
18587 You can start @code{debug} whenever you type @kbd{C-g}
18588 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18589 @code{t}. This is useful for debugging infinite loops.
18592 @cindex @code{(debug)} in code
18593 Or, you can insert a line that says @code{(debug)} into your code
18594 where you want the debugger to start, like this:
18598 (defun triangle-bugged (number)
18599 "Return sum of numbers 1 through NUMBER inclusive."
18601 (while (> number 0)
18602 (setq total (+ total number))
18603 (debug) ; @r{Start debugger.}
18604 (setq number (1= number))) ; @r{Error here.}
18609 The @code{debug} function is described in detail in @ref{Debugger, ,
18610 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18613 @section The @code{edebug} Source Level Debugger
18614 @cindex Source level debugger
18617 Edebug is a source level debugger. Edebug normally displays the
18618 source of the code you are debugging, with an arrow at the left that
18619 shows which line you are currently executing.
18621 You can walk through the execution of a function, line by line, or run
18622 quickly until reaching a @dfn{breakpoint} where execution stops.
18624 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18625 Lisp Reference Manual}.
18628 Here is a bugged function definition for @code{triangle-recursively}.
18629 @xref{Recursive triangle function, , Recursion in place of a counter},
18630 for a review of it.
18634 (defun triangle-recursively-bugged (number)
18635 "Return sum of numbers 1 through NUMBER inclusive.
18640 (triangle-recursively-bugged
18641 (1= number))))) ; @r{Error here.}
18646 Normally, you would install this definition by positioning your cursor
18647 after the function's closing parenthesis and typing @kbd{C-x C-e}
18648 (@code{eval-last-sexp}) or else by positioning your cursor within the
18649 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18650 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18654 However, to prepare this function definition for Edebug, you must
18655 first @dfn{instrument} the code using a different command. You can do
18656 this by positioning your cursor within or just after the definition
18660 M-x edebug-defun RET
18664 This will cause Emacs to load Edebug automatically if it is not
18665 already loaded, and properly instrument the function.
18667 After instrumenting the function, place your cursor after the
18668 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18671 (triangle-recursively-bugged 3)
18675 You will be jumped back to the source for
18676 @code{triangle-recursively-bugged} and the cursor positioned at the
18677 beginning of the @code{if} line of the function. Also, you will see
18678 an arrowhead at the left hand side of that line. The arrowhead marks
18679 the line where the function is executing. (In the following examples,
18680 we show the arrowhead with @samp{=>}; in a windowing system, you may
18681 see the arrowhead as a solid triangle in the window `fringe'.)
18684 =>@point{}(if (= number 1)
18689 In the example, the location of point is displayed with a star,
18690 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18693 In the example, the location of point is displayed as @samp{@point{}}
18694 (in a printed book, it is displayed with a five pointed star).
18697 If you now press @key{SPC}, point will move to the next expression to
18698 be executed; the line will look like this:
18701 =>(if @point{}(= number 1)
18705 As you continue to press @key{SPC}, point will move from expression to
18706 expression. At the same time, whenever an expression returns a value,
18707 that value will be displayed in the echo area. For example, after you
18708 move point past @code{number}, you will see the following:
18711 Result: 3 (#o3, #x3, ?\C-c)
18715 This means the value of @code{number} is 3, which is octal three,
18716 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18717 alphabet, in case you need to know this information).
18719 You can continue moving through the code until you reach the line with
18720 the error. Before evaluation, that line looks like this:
18723 => @point{}(1= number))))) ; @r{Error here.}
18728 When you press @key{SPC} once again, you will produce an error message
18732 Symbol's function definition is void:@: 1=
18738 Press @kbd{q} to quit Edebug.
18740 To remove instrumentation from a function definition, simply
18741 re-evaluate it with a command that does not instrument it.
18742 For example, you could place your cursor after the definition's
18743 closing parenthesis and type @kbd{C-x C-e}.
18745 Edebug does a great deal more than walk with you through a function.
18746 You can set it so it races through on its own, stopping only at an
18747 error or at specified stopping points; you can cause it to display the
18748 changing values of various expressions; you can find out how many
18749 times a function is called, and more.
18751 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18752 Lisp Reference Manual}.
18755 @node Debugging Exercises
18756 @section Debugging Exercises
18760 Install the @code{@value{COUNT-WORDS}} function and then cause it to
18761 enter the built-in debugger when you call it. Run the command on a
18762 region containing two words. You will need to press @kbd{d} a
18763 remarkable number of times. On your system, is a `hook' called after
18764 the command finishes? (For information on hooks, see @ref{Command
18765 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18769 Copy @code{@value{COUNT-WORDS}} into the @file{*scratch*} buffer,
18770 instrument the function for Edebug, and walk through its execution.
18771 The function does not need to have a bug, although you can introduce
18772 one if you wish. If the function lacks a bug, the walk-through
18773 completes without problems.
18776 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18777 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.,
18778 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18779 for commands made outside of the Edebug debugging buffer.)
18782 In the Edebug debugging buffer, use the @kbd{p}
18783 (@code{edebug-bounce-point}) command to see where in the region the
18784 @code{@value{COUNT-WORDS}} is working.
18787 Move point to some spot further down the function and then type the
18788 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18791 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18792 walk through the function on its own; use an upper case @kbd{T} for
18793 @code{edebug-Trace-fast-mode}.
18796 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18801 @chapter Conclusion
18803 We have now reached the end of this Introduction. You have now
18804 learned enough about programming in Emacs Lisp to set values, to write
18805 simple @file{.emacs} files for yourself and your friends, and write
18806 simple customizations and extensions to Emacs.
18808 This is a place to stop. Or, if you wish, you can now go onward, and
18811 You have learned some of the basic nuts and bolts of programming. But
18812 only some. There are a great many more brackets and hinges that are
18813 easy to use that we have not touched.
18815 A path you can follow right now lies among the sources to GNU Emacs
18818 @cite{The GNU Emacs Lisp Reference Manual}.
18821 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18822 Emacs Lisp Reference Manual}.
18825 The Emacs Lisp sources are an adventure. When you read the sources and
18826 come across a function or expression that is unfamiliar, you need to
18827 figure out or find out what it does.
18829 Go to the Reference Manual. It is a thorough, complete, and fairly
18830 easy-to-read description of Emacs Lisp. It is written not only for
18831 experts, but for people who know what you know. (The @cite{Reference
18832 Manual} comes with the standard GNU Emacs distribution. Like this
18833 introduction, it comes as a Texinfo source file, so you can read it
18834 on-line and as a typeset, printed book.)
18836 Go to the other on-line help that is part of GNU Emacs: the on-line
18837 documentation for all functions and variables, and @code{find-tag},
18838 the program that takes you to sources.
18840 Here is an example of how I explore the sources. Because of its name,
18841 @file{simple.el} is the file I looked at first, a long time ago. As
18842 it happens some of the functions in @file{simple.el} are complicated,
18843 or at least look complicated at first sight. The @code{open-line}
18844 function, for example, looks complicated.
18846 You may want to walk through this function slowly, as we did with the
18847 @code{forward-sentence} function. (@xref{forward-sentence, The
18848 @code{forward-sentence} function}.) Or you may want to skip that
18849 function and look at another, such as @code{split-line}. You don't
18850 need to read all the functions. According to
18851 @code{count-words-in-defun}, the @code{split-line} function contains
18852 102 words and symbols.
18854 Even though it is short, @code{split-line} contains expressions
18855 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18856 @code{current-column} and @code{insert-and-inherit}.
18858 Consider the @code{skip-chars-forward} function. (It is part of the
18859 function definition for @code{back-to-indentation}, which is shown in
18860 @ref{Review, , Review}.)
18862 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18863 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18864 function. This gives you the function documentation.
18866 You may be able to guess what is done by a well named function such as
18867 @code{indent-to}; or you can look it up, too. Incidentally, the
18868 @code{describe-function} function itself is in @file{help.el}; it is
18869 one of those long, but decipherable functions. You can look up
18870 @code{describe-function} using the @kbd{C-h f} command!
18872 In this instance, since the code is Lisp, the @file{*Help*} buffer
18873 contains the name of the library containing the function's source.
18874 You can put point over the name of the library and press the RET key,
18875 which in this situation is bound to @code{help-follow}, and be taken
18876 directly to the source, in the same way as @kbd{M-.}
18879 The definition for @code{describe-function} illustrates how to
18880 customize the @code{interactive} expression without using the standard
18881 character codes; and it shows how to create a temporary buffer.
18883 (The @code{indent-to} function is written in C rather than Emacs Lisp;
18884 it is a `built-in' function. @code{help-follow} takes you to its
18885 source as does @code{find-tag}, when properly set up.)
18887 You can look at a function's source using @code{find-tag}, which is
18888 bound to @kbd{M-.} Finally, you can find out what the Reference
18889 Manual has to say by visiting the manual in Info, and typing @kbd{i}
18890 (@code{Info-index}) and the name of the function, or by looking up the
18891 function in the index to a printed copy of the manual.
18893 Similarly, you can find out what is meant by
18894 @code{insert-and-inherit}.
18896 Other interesting source files include @file{paragraphs.el},
18897 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
18898 file includes short, easily understood functions as well as longer
18899 ones. The @file{loaddefs.el} file contains the many standard
18900 autoloads and many keymaps. I have never looked at it all; only at
18901 parts. @file{loadup.el} is the file that loads the standard parts of
18902 Emacs; it tells you a great deal about how Emacs is built.
18903 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
18904 Reference Manual}, for more about building.)
18906 As I said, you have learned some nuts and bolts; however, and very
18907 importantly, we have hardly touched major aspects of programming; I
18908 have said nothing about how to sort information, except to use the
18909 predefined @code{sort} function; I have said nothing about how to store
18910 information, except to use variables and lists; I have said nothing
18911 about how to write programs that write programs. These are topics for
18912 another, and different kind of book, a different kind of learning.
18914 What you have done is learn enough for much practical work with GNU
18915 Emacs. What you have done is get started. This is the end of a
18918 @c ================ Appendix ================
18921 @appendix The @code{the-the} Function
18923 @cindex Duplicated words function
18924 @cindex Words, duplicated
18926 Sometimes when you you write text, you duplicate words---as with ``you
18927 you'' near the beginning of this sentence. I find that most
18928 frequently, I duplicate ``the''; hence, I call the function for
18929 detecting duplicated words, @code{the-the}.
18932 As a first step, you could use the following regular expression to
18933 search for duplicates:
18936 \\(\\w+[ \t\n]+\\)\\1
18940 This regexp matches one or more word-constituent characters followed
18941 by one or more spaces, tabs, or newlines. However, it does not detect
18942 duplicated words on different lines, since the ending of the first
18943 word, the end of the line, is different from the ending of the second
18944 word, a space. (For more information about regular expressions, see
18945 @ref{Regexp Search, , Regular Expression Searches}, as well as
18946 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
18947 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
18948 The GNU Emacs Lisp Reference Manual}.)
18950 You might try searching just for duplicated word-constituent
18951 characters but that does not work since the pattern detects doubles
18952 such as the two occurrences of `th' in `with the'.
18954 Another possible regexp searches for word-constituent characters
18955 followed by non-word-constituent characters, reduplicated. Here,
18956 @w{@samp{\\w+}} matches one or more word-constituent characters and
18957 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
18960 \\(\\(\\w+\\)\\W*\\)\\1
18966 Here is the pattern that I use. It is not perfect, but good enough.
18967 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
18968 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
18969 any characters that are @emph{not} an @@-sign, space, newline, or tab.
18972 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
18975 One can write more complicated expressions, but I found that this
18976 expression is good enough, so I use it.
18978 Here is the @code{the-the} function, as I include it in my
18979 @file{.emacs} file, along with a handy global key binding:
18984 "Search forward for for a duplicated word."
18986 (message "Searching for for duplicated words ...")
18990 ;; This regexp is not perfect
18991 ;; but is fairly good over all:
18992 (if (re-search-forward
18993 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
18994 (message "Found duplicated word.")
18995 (message "End of buffer")))
18999 ;; Bind `the-the' to C-c \
19000 (global-set-key "\C-c\\" 'the-the)
19009 one two two three four five
19014 You can substitute the other regular expressions shown above in the
19015 function definition and try each of them on this list.
19018 @appendix Handling the Kill Ring
19019 @cindex Kill ring handling
19020 @cindex Handling the kill ring
19021 @cindex Ring, making a list like a
19023 The kill ring is a list that is transformed into a ring by the
19024 workings of the @code{current-kill} function. The @code{yank} and
19025 @code{yank-pop} commands use the @code{current-kill} function.
19027 This appendix describes the @code{current-kill} function as well as
19028 both the @code{yank} and the @code{yank-pop} commands, but first,
19029 consider the workings of the kill ring.
19032 * What the Kill Ring Does::
19034 * yank:: Paste a copy of a clipped element.
19035 * yank-pop:: Insert element pointed to.
19040 @node What the Kill Ring Does
19041 @unnumberedsec What the Kill Ring Does
19045 The kill ring has a default maximum length of sixty items; this number
19046 is too large for an explanation. Instead, set it to four. Please
19047 evaluate the following:
19051 (setq old-kill-ring-max kill-ring-max)
19052 (setq kill-ring-max 4)
19057 Then, please copy each line of the following indented example into the
19058 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19062 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19063 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19064 merely copy it to the kill ring. However, your machine may beep at
19065 you. Alternatively, for silence, you may copy the region of each line
19066 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19067 each line for this command to succeed, but it does not matter at which
19068 end you put point or mark.)
19072 Please invoke the calls in order, so that five elements attempt to
19073 fill the kill ring:
19078 second piece of text
19080 fourth line of text
19087 Then find the value of @code{kill-ring} by evaluating
19099 ("fifth bit of text" "fourth line of text"
19100 "third line" "second piece of text")
19105 The first element, @samp{first some text}, was dropped.
19108 To return to the old value for the length of the kill ring, evaluate:
19111 (setq kill-ring-max old-kill-ring-max)
19115 @appendixsec The @code{current-kill} Function
19116 @findex current-kill
19118 The @code{current-kill} function changes the element in the kill ring
19119 to which @code{kill-ring-yank-pointer} points. (Also, the
19120 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19121 to the latest element of the kill ring. The @code{kill-new}
19122 function is used directly or indirectly by @code{kill-append},
19123 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19124 and @code{kill-region}.)
19127 * Code for current-kill::
19128 * Understanding current-kill::
19132 @node Code for current-kill
19133 @unnumberedsubsec The code for @code{current-kill}
19138 The @code{current-kill} function is used by @code{yank} and by
19139 @code{yank-pop}. Here is the code for @code{current-kill}:
19143 (defun current-kill (n &optional do-not-move)
19144 "Rotate the yanking point by N places, and then return that kill.
19145 If N is zero, `interprogram-paste-function' is set, and calling it
19146 returns a string, then that string is added to the front of the
19147 kill ring and returned as the latest kill.
19150 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19151 yanking point; just return the Nth kill forward."
19152 (let ((interprogram-paste (and (= n 0)
19153 interprogram-paste-function
19154 (funcall interprogram-paste-function))))
19157 (if interprogram-paste
19159 ;; Disable the interprogram cut function when we add the new
19160 ;; text to the kill ring, so Emacs doesn't try to own the
19161 ;; selection, with identical text.
19162 (let ((interprogram-cut-function nil))
19163 (kill-new interprogram-paste))
19164 interprogram-paste)
19167 (or kill-ring (error "Kill ring is empty"))
19168 (let ((ARGth-kill-element
19169 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19170 (length kill-ring))
19173 (setq kill-ring-yank-pointer ARGth-kill-element))
19174 (car ARGth-kill-element)))))
19178 Remember also that the @code{kill-new} function sets
19179 @code{kill-ring-yank-pointer} to the latest element of the kill
19180 ring, which means that all the functions that call it set the value
19181 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19182 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19185 Here is the line in @code{kill-new}, which is explained in
19186 @ref{kill-new function, , The @code{kill-new} function}.
19189 (setq kill-ring-yank-pointer kill-ring)
19193 @node Understanding current-kill
19194 @unnumberedsubsec @code{current-kill} in Outline
19197 The @code{current-kill} function looks complex, but as usual, it can
19198 be understood by taking it apart piece by piece. First look at it in
19203 (defun current-kill (n &optional do-not-move)
19204 "Rotate the yanking point by N places, and then return that kill."
19210 This function takes two arguments, one of which is optional. It has a
19211 documentation string. It is @emph{not} interactive.
19214 * Body of current-kill::
19215 * Digression concerning error:: How to mislead humans, but not computers.
19216 * Determining the Element::
19220 @node Body of current-kill
19221 @unnumberedsubsubsec The Body of @code{current-kill}
19224 The body of the function definition is a @code{let} expression, which
19225 itself has a body as well as a @var{varlist}.
19227 The @code{let} expression declares a variable that will be only usable
19228 within the bounds of this function. This variable is called
19229 @code{interprogram-paste} and is for copying to another program. It
19230 is not for copying within this instance of GNU Emacs. Most window
19231 systems provide a facility for interprogram pasting. Sadly, that
19232 facility usually provides only for the last element. Most windowing
19233 systems have not adopted a ring of many possibilities, even though
19234 Emacs has provided it for decades.
19236 The @code{if} expression has two parts, one if there exists
19237 @code{interprogram-paste} and one if not.
19240 Let us consider the `if not' or else-part of the @code{current-kill}
19241 function. (The then-part uses the @code{kill-new} function, which
19242 we have already described. @xref{kill-new function, , The
19243 @code{kill-new} function}.)
19247 (or kill-ring (error "Kill ring is empty"))
19248 (let ((ARGth-kill-element
19249 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19250 (length kill-ring))
19253 (setq kill-ring-yank-pointer ARGth-kill-element))
19254 (car ARGth-kill-element))
19259 The code first checks whether the kill ring has content; otherwise it
19263 Note that the @code{or} expression is very similar to testing length
19270 (if (zerop (length kill-ring)) ; @r{if-part}
19271 (error "Kill ring is empty")) ; @r{then-part}
19277 If there is not anything in the kill ring, its length must be zero and
19278 an error message sent to the user: @samp{Kill ring is empty}. The
19279 @code{current-kill} function uses an @code{or} expression which is
19280 simpler. But an @code{if} expression reminds us what goes on.
19282 This @code{if} expression uses the function @code{zerop} which returns
19283 true if the value it is testing is zero. When @code{zerop} tests
19284 true, the then-part of the @code{if} is evaluated. The then-part is a
19285 list starting with the function @code{error}, which is a function that
19286 is similar to the @code{message} function
19287 (@pxref{message, , The @code{message} Function}) in that
19288 it prints a one-line message in the echo area. However, in addition
19289 to printing a message, @code{error} also stops evaluation of the
19290 function within which it is embedded. This means that the rest of the
19291 function will not be evaluated if the length of the kill ring is zero.
19293 Then the @code{current-kill} function selects the element to return.
19294 The selection depends on the number of places that @code{current-kill}
19295 rotates and on where @code{kill-ring-yank-pointer} points.
19297 Next, either the optional @code{do-not-move} argument is true or the
19298 current value of @code{kill-ring-yank-pointer} is set to point to the
19299 list. Finally, another expression returns the first element of the
19300 list even if the @code{do-not-move} argument is true.
19303 @node Digression concerning error
19304 @unnumberedsubsubsec Digression about the word `error'
19307 In my opinion, it is slightly misleading, at least to humans, to use
19308 the term `error' as the name of the @code{error} function. A better
19309 term would be `cancel'. Strictly speaking, of course, you cannot
19310 point to, much less rotate a pointer to a list that has no length, so
19311 from the point of view of the computer, the word `error' is correct.
19312 But a human expects to attempt this sort of thing, if only to find out
19313 whether the kill ring is full or empty. This is an act of
19316 From the human point of view, the act of exploration and discovery is
19317 not necessarily an error, and therefore should not be labeled as one,
19318 even in the bowels of a computer. As it is, the code in Emacs implies
19319 that a human who is acting virtuously, by exploring his or her
19320 environment, is making an error. This is bad. Even though the computer
19321 takes the same steps as it does when there is an `error', a term such as
19322 `cancel' would have a clearer connotation.
19325 @node Determining the Element
19326 @unnumberedsubsubsec Determining the Element
19329 Among other actions, the else-part of the @code{if} expression sets
19330 the value of @code{kill-ring-yank-pointer} to
19331 @code{ARGth-kill-element} when the kill ring has something in it and
19332 the value of @code{do-not-move} is @code{nil}.
19335 The code looks like this:
19339 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19340 (length kill-ring))
19345 This needs some examination. Unless it is not supposed to move the
19346 pointer, the @code{current-kill} function changes where
19347 @code{kill-ring-yank-pointer} points.
19349 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19350 expression does. Also, clearly, @code{ARGth-kill-element} is being
19351 set to be equal to some @sc{cdr} of the kill ring, using the
19352 @code{nthcdr} function that is described in an earlier section.
19353 (@xref{copy-region-as-kill}.) How does it do this?
19355 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19356 works by repeatedly taking the @sc{cdr} of a list---it takes the
19357 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19360 The two following expressions produce the same result:
19364 (setq kill-ring-yank-pointer (cdr kill-ring))
19366 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19370 However, the @code{nthcdr} expression is more complicated. It uses
19371 the @code{mod} function to determine which @sc{cdr} to select.
19373 (You will remember to look at inner functions first; indeed, we will
19374 have to go inside the @code{mod}.)
19376 The @code{mod} function returns the value of its first argument modulo
19377 the second; that is to say, it returns the remainder after dividing
19378 the first argument by the second. The value returned has the same
19379 sign as the second argument.
19387 @result{} 0 ;; @r{because there is no remainder}
19394 In this case, the first argument is often smaller than the second.
19406 We can guess what the @code{-} function does. It is like @code{+} but
19407 subtracts instead of adds; the @code{-} function subtracts its second
19408 argument from its first. Also, we already know what the @code{length}
19409 function does (@pxref{length}). It returns the length of a list.
19411 And @code{n} is the name of the required argument to the
19412 @code{current-kill} function.
19415 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19416 expression returns the whole list, as you can see by evaluating the
19421 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19422 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19423 (nthcdr (mod (- 0 4) 4)
19424 '("fourth line of text"
19426 "second piece of text"
19427 "first some text"))
19432 When the first argument to the @code{current-kill} function is one,
19433 the @code{nthcdr} expression returns the list without its first
19438 (nthcdr (mod (- 1 4) 4)
19439 '("fourth line of text"
19441 "second piece of text"
19442 "first some text"))
19446 @cindex @samp{global variable} defined
19447 @cindex @samp{variable, global}, defined
19448 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19449 are @dfn{global variables}. That means that any expression in Emacs
19450 Lisp can access them. They are not like the local variables set by
19451 @code{let} or like the symbols in an argument list.
19452 Local variables can only be accessed
19453 within the @code{let} that defines them or the function that specifies
19454 them in an argument list (and within expressions called by them).
19457 @c texi2dvi fails when the name of the section is within ifnottex ...
19458 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19459 @ref{defun, , The @code{defun} Macro}.)
19463 @appendixsec @code{yank}
19466 After learning about @code{current-kill}, the code for the
19467 @code{yank} function is almost easy.
19469 The @code{yank} function does not use the
19470 @code{kill-ring-yank-pointer} variable directly. It calls
19471 @code{insert-for-yank} which calls @code{current-kill} which sets the
19472 @code{kill-ring-yank-pointer} variable.
19475 The code looks like this:
19480 (defun yank (&optional arg)
19481 "Reinsert (\"paste\") the last stretch of killed text.
19482 More precisely, reinsert the stretch of killed text most recently
19483 killed OR yanked. Put point at end, and set mark at beginning.
19484 With just \\[universal-argument] as argument, same but put point at
19485 beginning (and mark at end). With argument N, reinsert the Nth most
19486 recently killed stretch of killed text.
19488 When this command inserts killed text into the buffer, it honors
19489 `yank-excluded-properties' and `yank-handler' as described in the
19490 doc string for `insert-for-yank-1', which see.
19492 See also the command \\[yank-pop]."
19496 (setq yank-window-start (window-start))
19497 ;; If we don't get all the way thru, make last-command indicate that
19498 ;; for the following command.
19499 (setq this-command t)
19500 (push-mark (point))
19503 (insert-for-yank (current-kill (cond
19508 ;; This is like exchange-point-and-mark,
19509 ;; but doesn't activate the mark.
19510 ;; It is cleaner to avoid activation, even though the command
19511 ;; loop would deactivate the mark because we inserted text.
19512 (goto-char (prog1 (mark t)
19513 (set-marker (mark-marker) (point) (current-buffer)))))
19516 ;; If we do get all the way thru, make this-command indicate that.
19517 (if (eq this-command t)
19518 (setq this-command 'yank))
19523 The key expression is @code{insert-for-yank}, which inserts the string
19524 returned by @code{current-kill}, but removes some text properties from
19527 However, before getting to that expression, the function sets the value
19528 of @code{yank-window-start} to the position returned by the
19529 @code{(window-start)} expression, the position at which the display
19530 currently starts. The @code{yank} function also sets
19531 @code{this-command} and pushes the mark.
19533 After it yanks the appropriate element, if the optional argument is a
19534 @sc{cons} rather than a number or nothing, it puts point at beginning
19535 of the yanked text and mark at its end.
19537 (The @code{prog1} function is like @code{progn} but returns the value
19538 of its first argument rather than the value of its last argument. Its
19539 first argument is forced to return the buffer's mark as an integer.
19540 You can see the documentation for these functions by placing point
19541 over them in this buffer and then typing @kbd{C-h f}
19542 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19545 The last part of the function tells what to do when it succeeds.
19548 @appendixsec @code{yank-pop}
19551 After understanding @code{yank} and @code{current-kill}, you know how
19552 to approach the @code{yank-pop} function. Leaving out the
19553 documentation to save space, it looks like this:
19558 (defun yank-pop (&optional arg)
19561 (if (not (eq last-command 'yank))
19562 (error "Previous command was not a yank"))
19565 (setq this-command 'yank)
19566 (unless arg (setq arg 1))
19567 (let ((inhibit-read-only t)
19568 (before (< (point) (mark t))))
19572 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19573 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19574 (setq yank-undo-function nil)
19577 (set-marker (mark-marker) (point) (current-buffer))
19578 (insert-for-yank (current-kill arg))
19579 ;; Set the window start back where it was in the yank command,
19581 (set-window-start (selected-window) yank-window-start t)
19585 ;; This is like exchange-point-and-mark,
19586 ;; but doesn't activate the mark.
19587 ;; It is cleaner to avoid activation, even though the command
19588 ;; loop would deactivate the mark because we inserted text.
19589 (goto-char (prog1 (mark t)
19590 (set-marker (mark-marker)
19592 (current-buffer))))))
19597 The function is interactive with a small @samp{p} so the prefix
19598 argument is processed and passed to the function. The command can
19599 only be used after a previous yank; otherwise an error message is
19600 sent. This check uses the variable @code{last-command} which is set
19601 by @code{yank} and is discussed elsewhere.
19602 (@xref{copy-region-as-kill}.)
19604 The @code{let} clause sets the variable @code{before} to true or false
19605 depending whether point is before or after mark and then the region
19606 between point and mark is deleted. This is the region that was just
19607 inserted by the previous yank and it is this text that will be
19610 @code{funcall} calls its first argument as a function, passing
19611 remaining arguments to it. The first argument is whatever the
19612 @code{or} expression returns. The two remaining arguments are the
19613 positions of point and mark set by the preceding @code{yank} command.
19615 There is more, but that is the hardest part.
19618 @appendixsec The @file{ring.el} File
19619 @cindex @file{ring.el} file
19621 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19622 provides many of the features we just discussed. But functions such
19623 as @code{kill-ring-yank-pointer} do not use this library, possibly
19624 because they were written earlier.
19627 @appendix A Graph with Labeled Axes
19629 Printed axes help you understand a graph. They convey scale. In an
19630 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19631 wrote the code to print the body of a graph. Here we write the code
19632 for printing and labeling vertical and horizontal axes, along with the
19636 * Labeled Example::
19637 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19638 * print-Y-axis:: Print a label for the vertical axis.
19639 * print-X-axis:: Print a horizontal label.
19640 * Print Whole Graph:: The function to print a complete graph.
19644 @node Labeled Example
19645 @unnumberedsec Labeled Example Graph
19648 Since insertions fill a buffer to the right and below point, the new
19649 graph printing function should first print the Y or vertical axis,
19650 then the body of the graph, and finally the X or horizontal axis.
19651 This sequence lays out for us the contents of the function:
19661 Print body of graph.
19668 Here is an example of how a finished graph should look:
19681 1 - ****************
19688 In this graph, both the vertical and the horizontal axes are labeled
19689 with numbers. However, in some graphs, the horizontal axis is time
19690 and would be better labeled with months, like this:
19704 Indeed, with a little thought, we can easily come up with a variety of
19705 vertical and horizontal labeling schemes. Our task could become
19706 complicated. But complications breed confusion. Rather than permit
19707 this, it is better choose a simple labeling scheme for our first
19708 effort, and to modify or replace it later.
19711 These considerations suggest the following outline for the
19712 @code{print-graph} function:
19716 (defun print-graph (numbers-list)
19717 "@var{documentation}@dots{}"
19718 (let ((height @dots{}
19722 (print-Y-axis height @dots{} )
19723 (graph-body-print numbers-list)
19724 (print-X-axis @dots{} )))
19728 We can work on each part of the @code{print-graph} function definition
19731 @node print-graph Varlist
19732 @appendixsec The @code{print-graph} Varlist
19733 @cindex @code{print-graph} varlist
19735 In writing the @code{print-graph} function, the first task is to write
19736 the varlist in the @code{let} expression. (We will leave aside for the
19737 moment any thoughts about making the function interactive or about the
19738 contents of its documentation string.)
19740 The varlist should set several values. Clearly, the top of the label
19741 for the vertical axis must be at least the height of the graph, which
19742 means that we must obtain this information here. Note that the
19743 @code{print-graph-body} function also requires this information. There
19744 is no reason to calculate the height of the graph in two different
19745 places, so we should change @code{print-graph-body} from the way we
19746 defined it earlier to take advantage of the calculation.
19748 Similarly, both the function for printing the X axis labels and the
19749 @code{print-graph-body} function need to learn the value of the width of
19750 each symbol. We can perform the calculation here and change the
19751 definition for @code{print-graph-body} from the way we defined it in the
19754 The length of the label for the horizontal axis must be at least as long
19755 as the graph. However, this information is used only in the function
19756 that prints the horizontal axis, so it does not need to be calculated here.
19758 These thoughts lead us directly to the following form for the varlist
19759 in the @code{let} for @code{print-graph}:
19763 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19764 (symbol-width (length graph-blank)))
19769 As we shall see, this expression is not quite right.
19773 @appendixsec The @code{print-Y-axis} Function
19774 @cindex Axis, print vertical
19775 @cindex Y axis printing
19776 @cindex Vertical axis printing
19777 @cindex Print vertical axis
19779 The job of the @code{print-Y-axis} function is to print a label for
19780 the vertical axis that looks like this:
19798 The function should be passed the height of the graph, and then should
19799 construct and insert the appropriate numbers and marks.
19802 * print-Y-axis in Detail::
19803 * Height of label:: What height for the Y axis?
19804 * Compute a Remainder:: How to compute the remainder of a division.
19805 * Y Axis Element:: Construct a line for the Y axis.
19806 * Y-axis-column:: Generate a list of Y axis labels.
19807 * print-Y-axis Penultimate:: A not quite final version.
19811 @node print-Y-axis in Detail
19812 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19815 It is easy enough to see in the figure what the Y axis label should
19816 look like; but to say in words, and then to write a function
19817 definition to do the job is another matter. It is not quite true to
19818 say that we want a number and a tic every five lines: there are only
19819 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19820 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19821 and 9). It is better to say that we want a number and a tic mark on
19822 the base line (number 1) and then that we want a number and a tic on
19823 the fifth line from the bottom and on every line that is a multiple of
19827 @node Height of label
19828 @unnumberedsubsec What height should the label be?
19831 The next issue is what height the label should be? Suppose the maximum
19832 height of tallest column of the graph is seven. Should the highest
19833 label on the Y axis be @samp{5 -}, and should the graph stick up above
19834 the label? Or should the highest label be @samp{7 -}, and mark the peak
19835 of the graph? Or should the highest label be @code{10 -}, which is a
19836 multiple of five, and be higher than the topmost value of the graph?
19838 The latter form is preferred. Most graphs are drawn within rectangles
19839 whose sides are an integral number of steps long---5, 10, 15, and so
19840 on for a step distance of five. But as soon as we decide to use a
19841 step height for the vertical axis, we discover that the simple
19842 expression in the varlist for computing the height is wrong. The
19843 expression is @code{(apply 'max numbers-list)}. This returns the
19844 precise height, not the maximum height plus whatever is necessary to
19845 round up to the nearest multiple of five. A more complex expression
19848 As usual in cases like this, a complex problem becomes simpler if it is
19849 divided into several smaller problems.
19851 First, consider the case when the highest value of the graph is an
19852 integral multiple of five---when it is 5, 10, 15, or some higher
19853 multiple of five. We can use this value as the Y axis height.
19855 A fairly simply way to determine whether a number is a multiple of
19856 five is to divide it by five and see if the division results in a
19857 remainder. If there is no remainder, the number is a multiple of
19858 five. Thus, seven divided by five has a remainder of two, and seven
19859 is not an integral multiple of five. Put in slightly different
19860 language, more reminiscent of the classroom, five goes into seven
19861 once, with a remainder of two. However, five goes into ten twice,
19862 with no remainder: ten is an integral multiple of five.
19864 @node Compute a Remainder
19865 @appendixsubsec Side Trip: Compute a Remainder
19867 @findex % @r{(remainder function)}
19868 @cindex Remainder function, @code{%}
19869 In Lisp, the function for computing a remainder is @code{%}. The
19870 function returns the remainder of its first argument divided by its
19871 second argument. As it happens, @code{%} is a function in Emacs Lisp
19872 that you cannot discover using @code{apropos}: you find nothing if you
19873 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
19874 learn of the existence of @code{%} is to read about it in a book such
19875 as this or in the Emacs Lisp sources.
19877 You can try the @code{%} function by evaluating the following two
19889 The first expression returns 2 and the second expression returns 0.
19891 To test whether the returned value is zero or some other number, we
19892 can use the @code{zerop} function. This function returns @code{t} if
19893 its argument, which must be a number, is zero.
19905 Thus, the following expression will return @code{t} if the height
19906 of the graph is evenly divisible by five:
19909 (zerop (% height 5))
19913 (The value of @code{height}, of course, can be found from @code{(apply
19914 'max numbers-list)}.)
19916 On the other hand, if the value of @code{height} is not a multiple of
19917 five, we want to reset the value to the next higher multiple of five.
19918 This is straightforward arithmetic using functions with which we are
19919 already familiar. First, we divide the value of @code{height} by five
19920 to determine how many times five goes into the number. Thus, five
19921 goes into twelve twice. If we add one to this quotient and multiply by
19922 five, we will obtain the value of the next multiple of five that is
19923 larger than the height. Five goes into twelve twice. Add one to two,
19924 and multiply by five; the result is fifteen, which is the next multiple
19925 of five that is higher than twelve. The Lisp expression for this is:
19928 (* (1+ (/ height 5)) 5)
19932 For example, if you evaluate the following, the result is 15:
19935 (* (1+ (/ 12 5)) 5)
19938 All through this discussion, we have been using `five' as the value
19939 for spacing labels on the Y axis; but we may want to use some other
19940 value. For generality, we should replace `five' with a variable to
19941 which we can assign a value. The best name I can think of for this
19942 variable is @code{Y-axis-label-spacing}.
19945 Using this term, and an @code{if} expression, we produce the
19950 (if (zerop (% height Y-axis-label-spacing))
19953 (* (1+ (/ height Y-axis-label-spacing))
19954 Y-axis-label-spacing))
19959 This expression returns the value of @code{height} itself if the height
19960 is an even multiple of the value of the @code{Y-axis-label-spacing} or
19961 else it computes and returns a value of @code{height} that is equal to
19962 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
19964 We can now include this expression in the @code{let} expression of the
19965 @code{print-graph} function (after first setting the value of
19966 @code{Y-axis-label-spacing}):
19967 @vindex Y-axis-label-spacing
19971 (defvar Y-axis-label-spacing 5
19972 "Number of lines from one Y axis label to next.")
19977 (let* ((height (apply 'max numbers-list))
19978 (height-of-top-line
19979 (if (zerop (% height Y-axis-label-spacing))
19984 (* (1+ (/ height Y-axis-label-spacing))
19985 Y-axis-label-spacing)))
19986 (symbol-width (length graph-blank))))
19992 (Note use of the @code{let*} function: the initial value of height is
19993 computed once by the @code{(apply 'max numbers-list)} expression and
19994 then the resulting value of @code{height} is used to compute its
19995 final value. @xref{fwd-para let, , The @code{let*} expression}, for
19996 more about @code{let*}.)
19998 @node Y Axis Element
19999 @appendixsubsec Construct a Y Axis Element
20001 When we print the vertical axis, we want to insert strings such as
20002 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20003 Moreover, we want the numbers and dashes to line up, so shorter
20004 numbers must be padded with leading spaces. If some of the strings
20005 use two digit numbers, the strings with single digit numbers must
20006 include a leading blank space before the number.
20008 @findex number-to-string
20009 To figure out the length of the number, the @code{length} function is
20010 used. But the @code{length} function works only with a string, not with
20011 a number. So the number has to be converted from being a number to
20012 being a string. This is done with the @code{number-to-string} function.
20017 (length (number-to-string 35))
20020 (length (number-to-string 100))
20026 (@code{number-to-string} is also called @code{int-to-string}; you will
20027 see this alternative name in various sources.)
20029 In addition, in each label, each number is followed by a string such
20030 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20031 This variable is defined with @code{defvar}:
20036 (defvar Y-axis-tic " - "
20037 "String that follows number in a Y axis label.")
20041 The length of the Y label is the sum of the length of the Y axis tic
20042 mark and the length of the number of the top of the graph.
20045 (length (concat (number-to-string height) Y-axis-tic)))
20048 This value will be calculated by the @code{print-graph} function in
20049 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20050 did not think to include this in the varlist when we first proposed it.)
20052 To make a complete vertical axis label, a tic mark is concatenated
20053 with a number; and the two together may be preceded by one or more
20054 spaces depending on how long the number is. The label consists of
20055 three parts: the (optional) leading spaces, the number, and the tic
20056 mark. The function is passed the value of the number for the specific
20057 row, and the value of the width of the top line, which is calculated
20058 (just once) by @code{print-graph}.
20062 (defun Y-axis-element (number full-Y-label-width)
20063 "Construct a NUMBERed label element.
20064 A numbered element looks like this ` 5 - ',
20065 and is padded as needed so all line up with
20066 the element for the largest number."
20069 (let* ((leading-spaces
20070 (- full-Y-label-width
20072 (concat (number-to-string number)
20077 (make-string leading-spaces ? )
20078 (number-to-string number)
20083 The @code{Y-axis-element} function concatenates together the leading
20084 spaces, if any; the number, as a string; and the tic mark.
20086 To figure out how many leading spaces the label will need, the
20087 function subtracts the actual length of the label---the length of the
20088 number plus the length of the tic mark---from the desired label width.
20090 @findex make-string
20091 Blank spaces are inserted using the @code{make-string} function. This
20092 function takes two arguments: the first tells it how long the string
20093 will be and the second is a symbol for the character to insert, in a
20094 special format. The format is a question mark followed by a blank
20095 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20096 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20097 syntax for characters. (Of course, you might want to replace the
20098 blank space by some other character @dots{} You know what to do.)
20100 The @code{number-to-string} function is used in the concatenation
20101 expression, to convert the number to a string that is concatenated
20102 with the leading spaces and the tic mark.
20104 @node Y-axis-column
20105 @appendixsubsec Create a Y Axis Column
20107 The preceding functions provide all the tools needed to construct a
20108 function that generates a list of numbered and blank strings to insert
20109 as the label for the vertical axis:
20111 @findex Y-axis-column
20114 (defun Y-axis-column (height width-of-label)
20115 "Construct list of Y axis labels and blank strings.
20116 For HEIGHT of line above base and WIDTH-OF-LABEL."
20120 (while (> height 1)
20121 (if (zerop (% height Y-axis-label-spacing))
20122 ;; @r{Insert label.}
20125 (Y-axis-element height width-of-label)
20129 ;; @r{Else, insert blanks.}
20132 (make-string width-of-label ? )
20134 (setq height (1- height)))
20135 ;; @r{Insert base line.}
20137 (cons (Y-axis-element 1 width-of-label) Y-axis))
20138 (nreverse Y-axis)))
20142 In this function, we start with the value of @code{height} and
20143 repetitively subtract one from its value. After each subtraction, we
20144 test to see whether the value is an integral multiple of the
20145 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20146 using the @code{Y-axis-element} function; if not, we construct a
20147 blank label using the @code{make-string} function. The base line
20148 consists of the number one followed by a tic mark.
20151 @node print-Y-axis Penultimate
20152 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20154 The list constructed by the @code{Y-axis-column} function is passed to
20155 the @code{print-Y-axis} function, which inserts the list as a column.
20157 @findex print-Y-axis
20160 (defun print-Y-axis (height full-Y-label-width)
20161 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20162 Height must be the maximum height of the graph.
20163 Full width is the width of the highest label element."
20164 ;; Value of height and full-Y-label-width
20165 ;; are passed by `print-graph'.
20168 (let ((start (point)))
20170 (Y-axis-column height full-Y-label-width))
20171 ;; @r{Place point ready for inserting graph.}
20173 ;; @r{Move point forward by value of} full-Y-label-width
20174 (forward-char full-Y-label-width)))
20178 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20179 insert the Y axis labels created by the @code{Y-axis-column} function.
20180 In addition, it places point at the correct position for printing the body of
20183 You can test @code{print-Y-axis}:
20191 Y-axis-label-spacing
20200 Copy the following expression:
20203 (print-Y-axis 12 5)
20207 Switch to the @file{*scratch*} buffer and place the cursor where you
20208 want the axis labels to start.
20211 Type @kbd{M-:} (@code{eval-expression}).
20214 Yank the @code{graph-body-print} expression into the minibuffer
20215 with @kbd{C-y} (@code{yank)}.
20218 Press @key{RET} to evaluate the expression.
20221 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20222 }}}. (The @code{print-graph} function will pass the value of
20223 @code{height-of-top-line}, which in this case will end up as 15,
20224 thereby getting rid of what might appear as a bug.)
20228 @appendixsec The @code{print-X-axis} Function
20229 @cindex Axis, print horizontal
20230 @cindex X axis printing
20231 @cindex Print horizontal axis
20232 @cindex Horizontal axis printing
20234 X axis labels are much like Y axis labels, except that the ticks are on a
20235 line above the numbers. Labels should look like this:
20244 The first tic is under the first column of the graph and is preceded by
20245 several blank spaces. These spaces provide room in rows above for the Y
20246 axis labels. The second, third, fourth, and subsequent ticks are all
20247 spaced equally, according to the value of @code{X-axis-label-spacing}.
20249 The second row of the X axis consists of numbers, preceded by several
20250 blank spaces and also separated according to the value of the variable
20251 @code{X-axis-label-spacing}.
20253 The value of the variable @code{X-axis-label-spacing} should itself be
20254 measured in units of @code{symbol-width}, since you may want to change
20255 the width of the symbols that you are using to print the body of the
20256 graph without changing the ways the graph is labeled.
20259 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
20260 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20264 @node Similarities differences
20265 @unnumberedsubsec Similarities and differences
20268 The @code{print-X-axis} function is constructed in more or less the
20269 same fashion as the @code{print-Y-axis} function except that it has
20270 two lines: the line of tic marks and the numbers. We will write a
20271 separate function to print each line and then combine them within the
20272 @code{print-X-axis} function.
20274 This is a three step process:
20278 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20281 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20284 Write a function to print both lines, the @code{print-X-axis} function,
20285 using @code{print-X-axis-tic-line} and
20286 @code{print-X-axis-numbered-line}.
20289 @node X Axis Tic Marks
20290 @appendixsubsec X Axis Tic Marks
20292 The first function should print the X axis tic marks. We must specify
20293 the tic marks themselves and their spacing:
20297 (defvar X-axis-label-spacing
20298 (if (boundp 'graph-blank)
20299 (* 5 (length graph-blank)) 5)
20300 "Number of units from one X axis label to next.")
20305 (Note that the value of @code{graph-blank} is set by another
20306 @code{defvar}. The @code{boundp} predicate checks whether it has
20307 already been set; @code{boundp} returns @code{nil} if it has not. If
20308 @code{graph-blank} were unbound and we did not use this conditional
20309 construction, in a recent GNU Emacs, we would enter the debugger and
20310 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20311 @w{(void-variable graph-blank)}}.)
20314 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20318 (defvar X-axis-tic-symbol "|"
20319 "String to insert to point to a column in X axis.")
20324 The goal is to make a line that looks like this:
20330 The first tic is indented so that it is under the first column, which is
20331 indented to provide space for the Y axis labels.
20333 A tic element consists of the blank spaces that stretch from one tic to
20334 the next plus a tic symbol. The number of blanks is determined by the
20335 width of the tic symbol and the @code{X-axis-label-spacing}.
20338 The code looks like this:
20342 ;;; X-axis-tic-element
20346 ;; @r{Make a string of blanks.}
20347 (- (* symbol-width X-axis-label-spacing)
20348 (length X-axis-tic-symbol))
20350 ;; @r{Concatenate blanks with tic symbol.}
20356 Next, we determine how many blanks are needed to indent the first tic
20357 mark to the first column of the graph. This uses the value of
20358 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20361 The code to make @code{X-axis-leading-spaces}
20366 ;; X-axis-leading-spaces
20368 (make-string full-Y-label-width ? )
20373 We also need to determine the length of the horizontal axis, which is
20374 the length of the numbers list, and the number of ticks in the horizontal
20381 (length numbers-list)
20387 (* symbol-width X-axis-label-spacing)
20391 ;; number-of-X-ticks
20392 (if (zerop (% (X-length tic-width)))
20393 (/ (X-length tic-width))
20394 (1+ (/ (X-length tic-width))))
20399 All this leads us directly to the function for printing the X axis tic line:
20401 @findex print-X-axis-tic-line
20404 (defun print-X-axis-tic-line
20405 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20406 "Print ticks for X axis."
20407 (insert X-axis-leading-spaces)
20408 (insert X-axis-tic-symbol) ; @r{Under first column.}
20411 ;; @r{Insert second tic in the right spot.}
20414 (- (* symbol-width X-axis-label-spacing)
20415 ;; @r{Insert white space up to second tic symbol.}
20416 (* 2 (length X-axis-tic-symbol)))
20418 X-axis-tic-symbol))
20421 ;; @r{Insert remaining ticks.}
20422 (while (> number-of-X-tics 1)
20423 (insert X-axis-tic-element)
20424 (setq number-of-X-tics (1- number-of-X-tics))))
20428 The line of numbers is equally straightforward:
20431 First, we create a numbered element with blank spaces before each number:
20433 @findex X-axis-element
20436 (defun X-axis-element (number)
20437 "Construct a numbered X axis element."
20438 (let ((leading-spaces
20439 (- (* symbol-width X-axis-label-spacing)
20440 (length (number-to-string number)))))
20441 (concat (make-string leading-spaces ? )
20442 (number-to-string number))))
20446 Next, we create the function to print the numbered line, starting with
20447 the number ``1'' under the first column:
20449 @findex print-X-axis-numbered-line
20452 (defun print-X-axis-numbered-line
20453 (number-of-X-tics X-axis-leading-spaces)
20454 "Print line of X-axis numbers"
20455 (let ((number X-axis-label-spacing))
20456 (insert X-axis-leading-spaces)
20462 ;; @r{Insert white space up to next number.}
20463 (- (* symbol-width X-axis-label-spacing) 2)
20465 (number-to-string number)))
20468 ;; @r{Insert remaining numbers.}
20469 (setq number (+ number X-axis-label-spacing))
20470 (while (> number-of-X-tics 1)
20471 (insert (X-axis-element number))
20472 (setq number (+ number X-axis-label-spacing))
20473 (setq number-of-X-tics (1- number-of-X-tics)))))
20477 Finally, we need to write the @code{print-X-axis} that uses
20478 @code{print-X-axis-tic-line} and
20479 @code{print-X-axis-numbered-line}.
20481 The function must determine the local values of the variables used by both
20482 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20483 then it must call them. Also, it must print the carriage return that
20484 separates the two lines.
20486 The function consists of a varlist that specifies five local variables,
20487 and calls to each of the two line printing functions:
20489 @findex print-X-axis
20492 (defun print-X-axis (numbers-list)
20493 "Print X axis labels to length of NUMBERS-LIST."
20494 (let* ((leading-spaces
20495 (make-string full-Y-label-width ? ))
20498 ;; symbol-width @r{is provided by} graph-body-print
20499 (tic-width (* symbol-width X-axis-label-spacing))
20500 (X-length (length numbers-list))
20508 ;; @r{Make a string of blanks.}
20509 (- (* symbol-width X-axis-label-spacing)
20510 (length X-axis-tic-symbol))
20514 ;; @r{Concatenate blanks with tic symbol.}
20515 X-axis-tic-symbol))
20519 (if (zerop (% X-length tic-width))
20520 (/ X-length tic-width)
20521 (1+ (/ X-length tic-width)))))
20524 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20526 (print-X-axis-numbered-line tic-number leading-spaces)))
20531 You can test @code{print-X-axis}:
20535 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20536 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20537 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20540 Copy the following expression:
20545 (let ((full-Y-label-width 5)
20548 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20553 Switch to the @file{*scratch*} buffer and place the cursor where you
20554 want the axis labels to start.
20557 Type @kbd{M-:} (@code{eval-expression}).
20560 Yank the test expression into the minibuffer
20561 with @kbd{C-y} (@code{yank)}.
20564 Press @key{RET} to evaluate the expression.
20568 Emacs will print the horizontal axis like this:
20578 @node Print Whole Graph
20579 @appendixsec Printing the Whole Graph
20580 @cindex Printing the whole graph
20581 @cindex Whole graph printing
20582 @cindex Graph, printing all
20584 Now we are nearly ready to print the whole graph.
20586 The function to print the graph with the proper labels follows the
20587 outline we created earlier (@pxref{Full Graph, , A Graph with Labeled
20588 Axes}), but with additions.
20591 Here is the outline:
20595 (defun print-graph (numbers-list)
20596 "@var{documentation}@dots{}"
20597 (let ((height @dots{}
20601 (print-Y-axis height @dots{} )
20602 (graph-body-print numbers-list)
20603 (print-X-axis @dots{} )))
20608 * The final version:: A few changes.
20609 * Test print-graph:: Run a short test.
20610 * Graphing words in defuns:: Executing the final code.
20611 * lambda:: How to write an anonymous function.
20612 * mapcar:: Apply a function to elements of a list.
20613 * Another Bug:: Yet another bug @dots{} most insidious.
20614 * Final printed graph:: The graph itself!
20618 @node The final version
20619 @unnumberedsubsec Changes for the Final Version
20622 The final version is different from what we planned in two ways:
20623 first, it contains additional values calculated once in the varlist;
20624 second, it carries an option to specify the labels' increment per row.
20625 This latter feature turns out to be essential; otherwise, a graph may
20626 have more rows than fit on a display or on a sheet of paper.
20629 This new feature requires a change to the @code{Y-axis-column}
20630 function, to add @code{vertical-step} to it. The function looks like
20633 @findex Y-axis-column @r{Final version.}
20636 ;;; @r{Final version.}
20637 (defun Y-axis-column
20638 (height width-of-label &optional vertical-step)
20639 "Construct list of labels for Y axis.
20640 HEIGHT is maximum height of graph.
20641 WIDTH-OF-LABEL is maximum width of label.
20642 VERTICAL-STEP, an option, is a positive integer
20643 that specifies how much a Y axis label increments
20644 for each line. For example, a step of 5 means
20645 that each line is five units of the graph."
20649 (number-per-line (or vertical-step 1)))
20650 (while (> height 1)
20651 (if (zerop (% height Y-axis-label-spacing))
20654 ;; @r{Insert label.}
20658 (* height number-per-line)
20663 ;; @r{Else, insert blanks.}
20666 (make-string width-of-label ? )
20668 (setq height (1- height)))
20671 ;; @r{Insert base line.}
20672 (setq Y-axis (cons (Y-axis-element
20673 (or vertical-step 1)
20676 (nreverse Y-axis)))
20680 The values for the maximum height of graph and the width of a symbol
20681 are computed by @code{print-graph} in its @code{let} expression; so
20682 @code{graph-body-print} must be changed to accept them.
20684 @findex graph-body-print @r{Final version.}
20687 ;;; @r{Final version.}
20688 (defun graph-body-print (numbers-list height symbol-width)
20689 "Print a bar graph of the NUMBERS-LIST.
20690 The numbers-list consists of the Y-axis values.
20691 HEIGHT is maximum height of graph.
20692 SYMBOL-WIDTH is number of each column."
20695 (let (from-position)
20696 (while numbers-list
20697 (setq from-position (point))
20699 (column-of-graph height (car numbers-list)))
20700 (goto-char from-position)
20701 (forward-char symbol-width)
20704 ;; @r{Draw graph column by column.}
20706 (setq numbers-list (cdr numbers-list)))
20707 ;; @r{Place point for X axis labels.}
20708 (forward-line height)
20714 Finally, the code for the @code{print-graph} function:
20716 @findex print-graph @r{Final version.}
20719 ;;; @r{Final version.}
20721 (numbers-list &optional vertical-step)
20722 "Print labeled bar graph of the NUMBERS-LIST.
20723 The numbers-list consists of the Y-axis values.
20727 Optionally, VERTICAL-STEP, a positive integer,
20728 specifies how much a Y axis label increments for
20729 each line. For example, a step of 5 means that
20730 each row is five units."
20733 (let* ((symbol-width (length graph-blank))
20734 ;; @code{height} @r{is both the largest number}
20735 ;; @r{and the number with the most digits.}
20736 (height (apply 'max numbers-list))
20739 (height-of-top-line
20740 (if (zerop (% height Y-axis-label-spacing))
20743 (* (1+ (/ height Y-axis-label-spacing))
20744 Y-axis-label-spacing)))
20747 (vertical-step (or vertical-step 1))
20748 (full-Y-label-width
20754 (* height-of-top-line vertical-step))
20760 height-of-top-line full-Y-label-width vertical-step)
20764 numbers-list height-of-top-line symbol-width)
20765 (print-X-axis numbers-list)))
20769 @node Test print-graph
20770 @appendixsubsec Testing @code{print-graph}
20773 We can test the @code{print-graph} function with a short list of numbers:
20777 Install the final versions of @code{Y-axis-column},
20778 @code{graph-body-print}, and @code{print-graph} (in addition to the
20782 Copy the following expression:
20785 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20789 Switch to the @file{*scratch*} buffer and place the cursor where you
20790 want the axis labels to start.
20793 Type @kbd{M-:} (@code{eval-expression}).
20796 Yank the test expression into the minibuffer
20797 with @kbd{C-y} (@code{yank)}.
20800 Press @key{RET} to evaluate the expression.
20804 Emacs will print a graph that looks like this:
20825 On the other hand, if you pass @code{print-graph} a
20826 @code{vertical-step} value of 2, by evaluating this expression:
20829 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20834 The graph looks like this:
20855 (A question: is the `2' on the bottom of the vertical axis a bug or a
20856 feature? If you think it is a bug, and should be a `1' instead, (or
20857 even a `0'), you can modify the sources.)
20859 @node Graphing words in defuns
20860 @appendixsubsec Graphing Numbers of Words and Symbols
20862 Now for the graph for which all this code was written: a graph that
20863 shows how many function definitions contain fewer than 10 words and
20864 symbols, how many contain between 10 and 19 words and symbols, how
20865 many contain between 20 and 29 words and symbols, and so on.
20867 This is a multi-step process. First make sure you have loaded all the
20871 It is a good idea to reset the value of @code{top-of-ranges} in case
20872 you have set it to some different value. You can evaluate the
20877 (setq top-of-ranges
20880 110 120 130 140 150
20881 160 170 180 190 200
20882 210 220 230 240 250
20883 260 270 280 290 300)
20888 Next create a list of the number of words and symbols in each range.
20892 Evaluate the following:
20896 (setq list-for-graph
20899 (recursive-lengths-list-many-files
20900 (directory-files "/usr/local/emacs/lisp"
20908 On my old machine, this took about an hour. It looked though 303 Lisp
20909 files in my copy of Emacs version 19.23. After all that computing,
20910 the @code{list-for-graph} had this value:
20914 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
20915 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
20920 This means that my copy of Emacs had 537 function definitions with
20921 fewer than 10 words or symbols in them, 1,027 function definitions
20922 with 10 to 19 words or symbols in them, 955 function definitions with
20923 20 to 29 words or symbols in them, and so on.
20925 Clearly, just by looking at this list we can see that most function
20926 definitions contain ten to thirty words and symbols.
20928 Now for printing. We do @emph{not} want to print a graph that is
20929 1,030 lines high @dots{} Instead, we should print a graph that is
20930 fewer than twenty-five lines high. A graph that height can be
20931 displayed on almost any monitor, and easily printed on a sheet of paper.
20933 This means that each value in @code{list-for-graph} must be reduced to
20934 one-fiftieth its present value.
20936 Here is a short function to do just that, using two functions we have
20937 not yet seen, @code{mapcar} and @code{lambda}.
20941 (defun one-fiftieth (full-range)
20942 "Return list, each number one-fiftieth of previous."
20943 (mapcar (lambda (arg) (/ arg 50)) full-range))
20948 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
20949 @cindex Anonymous function
20952 @code{lambda} is the symbol for an anonymous function, a function
20953 without a name. Every time you use an anonymous function, you need to
20954 include its whole body.
20961 (lambda (arg) (/ arg 50))
20965 is a function definition that says `return the value resulting from
20966 dividing whatever is passed to me as @code{arg} by 50'.
20969 Earlier, for example, we had a function @code{multiply-by-seven}; it
20970 multiplied its argument by 7. This function is similar, except it
20971 divides its argument by 50; and, it has no name. The anonymous
20972 equivalent of @code{multiply-by-seven} is:
20975 (lambda (number) (* 7 number))
20979 (@xref{defun, , The @code{defun} Macro}.)
20983 If we want to multiply 3 by 7, we can write:
20985 @c !!! Clear print-postscript-figures if the computer formatting this
20986 @c document is too small and cannot handle all the diagrams and figures.
20987 @c clear print-postscript-figures
20988 @c set print-postscript-figures
20989 @c lambda example diagram #1
20993 (multiply-by-seven 3)
20994 \_______________/ ^
21000 @ifset print-postscript-figures
21003 @center @image{lambda-1}
21004 %%%% old method of including an image
21005 % \input /usr/local/lib/tex/inputs/psfig.tex
21006 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21011 @ifclear print-postscript-figures
21015 (multiply-by-seven 3)
21016 \_______________/ ^
21025 This expression returns 21.
21029 Similarly, we can write:
21031 @c lambda example diagram #2
21035 ((lambda (number) (* 7 number)) 3)
21036 \____________________________/ ^
21038 anonymous function argument
21042 @ifset print-postscript-figures
21045 @center @image{lambda-2}
21046 %%%% old method of including an image
21047 % \input /usr/local/lib/tex/inputs/psfig.tex
21048 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21053 @ifclear print-postscript-figures
21057 ((lambda (number) (* 7 number)) 3)
21058 \____________________________/ ^
21060 anonymous function argument
21068 If we want to divide 100 by 50, we can write:
21070 @c lambda example diagram #3
21074 ((lambda (arg) (/ arg 50)) 100)
21075 \______________________/ \_/
21077 anonymous function argument
21081 @ifset print-postscript-figures
21084 @center @image{lambda-3}
21085 %%%% old method of including an image
21086 % \input /usr/local/lib/tex/inputs/psfig.tex
21087 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21092 @ifclear print-postscript-figures
21096 ((lambda (arg) (/ arg 50)) 100)
21097 \______________________/ \_/
21099 anonymous function argument
21106 This expression returns 2. The 100 is passed to the function, which
21107 divides that number by 50.
21109 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21110 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21111 expressions derive from the Lambda Calculus.
21114 @appendixsubsec The @code{mapcar} Function
21117 @code{mapcar} is a function that calls its first argument with each
21118 element of its second argument, in turn. The second argument must be
21121 The @samp{map} part of the name comes from the mathematical phrase,
21122 `mapping over a domain', meaning to apply a function to each of the
21123 elements in a domain. The mathematical phrase is based on the
21124 metaphor of a surveyor walking, one step at a time, over an area he is
21125 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21134 (mapcar '1+ '(2 4 6))
21140 The function @code{1+} which adds one to its argument, is executed on
21141 @emph{each} element of the list, and a new list is returned.
21143 Contrast this with @code{apply}, which applies its first argument to
21145 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21149 In the definition of @code{one-fiftieth}, the first argument is the
21150 anonymous function:
21153 (lambda (arg) (/ arg 50))
21157 and the second argument is @code{full-range}, which will be bound to
21158 @code{list-for-graph}.
21161 The whole expression looks like this:
21164 (mapcar (lambda (arg) (/ arg 50)) full-range))
21167 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21168 Lisp Reference Manual}, for more about @code{mapcar}.
21170 Using the @code{one-fiftieth} function, we can generate a list in
21171 which each element is one-fiftieth the size of the corresponding
21172 element in @code{list-for-graph}.
21176 (setq fiftieth-list-for-graph
21177 (one-fiftieth list-for-graph))
21182 The resulting list looks like this:
21186 (10 20 19 15 11 9 6 5 4 3 3 2 2
21187 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21192 This, we are almost ready to print! (We also notice the loss of
21193 information: many of the higher ranges are 0, meaning that fewer than
21194 50 defuns had that many words or symbols---but not necessarily meaning
21195 that none had that many words or symbols.)
21198 @appendixsubsec Another Bug @dots{} Most Insidious
21199 @cindex Bug, most insidious type
21200 @cindex Insidious type of bug
21202 I said `almost ready to print'! Of course, there is a bug in the
21203 @code{print-graph} function @dots{} It has a @code{vertical-step}
21204 option, but not a @code{horizontal-step} option. The
21205 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21206 @code{print-graph} function will print only by ones.
21208 This is a classic example of what some consider the most insidious
21209 type of bug, the bug of omission. This is not the kind of bug you can
21210 find by studying the code, for it is not in the code; it is an omitted
21211 feature. Your best actions are to try your program early and often;
21212 and try to arrange, as much as you can, to write code that is easy to
21213 understand and easy to change. Try to be aware, whenever you can,
21214 that whatever you have written, @emph{will} be rewritten, if not soon,
21215 eventually. A hard maxim to follow.
21217 It is the @code{print-X-axis-numbered-line} function that needs the
21218 work; and then the @code{print-X-axis} and the @code{print-graph}
21219 functions need to be adapted. Not much needs to be done; there is one
21220 nicety: the numbers ought to line up under the tic marks. This takes
21224 Here is the corrected @code{print-X-axis-numbered-line}:
21228 (defun print-X-axis-numbered-line
21229 (number-of-X-tics X-axis-leading-spaces
21230 &optional horizontal-step)
21231 "Print line of X-axis numbers"
21232 (let ((number X-axis-label-spacing)
21233 (horizontal-step (or horizontal-step 1)))
21236 (insert X-axis-leading-spaces)
21237 ;; @r{Delete extra leading spaces.}
21240 (length (number-to-string horizontal-step)))))
21245 ;; @r{Insert white space.}
21247 X-axis-label-spacing)
21250 (number-to-string horizontal-step)))
21254 (* number horizontal-step))))
21257 ;; @r{Insert remaining numbers.}
21258 (setq number (+ number X-axis-label-spacing))
21259 (while (> number-of-X-tics 1)
21260 (insert (X-axis-element
21261 (* number horizontal-step)))
21262 (setq number (+ number X-axis-label-spacing))
21263 (setq number-of-X-tics (1- number-of-X-tics)))))
21268 If you are reading this in Info, you can see the new versions of
21269 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21270 reading this in a printed book, you can see the changed lines here
21271 (the full text is too much to print).
21276 (defun print-X-axis (numbers-list horizontal-step)
21278 (print-X-axis-numbered-line
21279 tic-number leading-spaces horizontal-step))
21287 &optional vertical-step horizontal-step)
21289 (print-X-axis numbers-list horizontal-step))
21297 (defun print-X-axis (numbers-list horizontal-step)
21298 "Print X axis labels to length of NUMBERS-LIST.
21299 Optionally, HORIZONTAL-STEP, a positive integer,
21300 specifies how much an X axis label increments for
21304 ;; Value of symbol-width and full-Y-label-width
21305 ;; are passed by `print-graph'.
21306 (let* ((leading-spaces
21307 (make-string full-Y-label-width ? ))
21308 ;; symbol-width @r{is provided by} graph-body-print
21309 (tic-width (* symbol-width X-axis-label-spacing))
21310 (X-length (length numbers-list))
21316 ;; @r{Make a string of blanks.}
21317 (- (* symbol-width X-axis-label-spacing)
21318 (length X-axis-tic-symbol))
21322 ;; @r{Concatenate blanks with tic symbol.}
21323 X-axis-tic-symbol))
21325 (if (zerop (% X-length tic-width))
21326 (/ X-length tic-width)
21327 (1+ (/ X-length tic-width)))))
21331 (print-X-axis-tic-line
21332 tic-number leading-spaces X-tic)
21334 (print-X-axis-numbered-line
21335 tic-number leading-spaces horizontal-step)))
21342 (numbers-list &optional vertical-step horizontal-step)
21343 "Print labeled bar graph of the NUMBERS-LIST.
21344 The numbers-list consists of the Y-axis values.
21348 Optionally, VERTICAL-STEP, a positive integer,
21349 specifies how much a Y axis label increments for
21350 each line. For example, a step of 5 means that
21351 each row is five units.
21355 Optionally, HORIZONTAL-STEP, a positive integer,
21356 specifies how much an X axis label increments for
21358 (let* ((symbol-width (length graph-blank))
21359 ;; @code{height} @r{is both the largest number}
21360 ;; @r{and the number with the most digits.}
21361 (height (apply 'max numbers-list))
21364 (height-of-top-line
21365 (if (zerop (% height Y-axis-label-spacing))
21368 (* (1+ (/ height Y-axis-label-spacing))
21369 Y-axis-label-spacing)))
21372 (vertical-step (or vertical-step 1))
21373 (full-Y-label-width
21377 (* height-of-top-line vertical-step))
21382 height-of-top-line full-Y-label-width vertical-step)
21384 numbers-list height-of-top-line symbol-width)
21385 (print-X-axis numbers-list horizontal-step)))
21392 Graphing Definitions Re-listed
21395 Here are all the graphing definitions in their final form:
21399 (defvar top-of-ranges
21402 110 120 130 140 150
21403 160 170 180 190 200
21404 210 220 230 240 250)
21405 "List specifying ranges for `defuns-per-range'.")
21409 (defvar graph-symbol "*"
21410 "String used as symbol in graph, usually an asterisk.")
21414 (defvar graph-blank " "
21415 "String used as blank in graph, usually a blank space.
21416 graph-blank must be the same number of columns wide
21421 (defvar Y-axis-tic " - "
21422 "String that follows number in a Y axis label.")
21426 (defvar Y-axis-label-spacing 5
21427 "Number of lines from one Y axis label to next.")
21431 (defvar X-axis-tic-symbol "|"
21432 "String to insert to point to a column in X axis.")
21436 (defvar X-axis-label-spacing
21437 (if (boundp 'graph-blank)
21438 (* 5 (length graph-blank)) 5)
21439 "Number of units from one X axis label to next.")
21445 (defun count-words-in-defun ()
21446 "Return the number of words and symbols in a defun."
21447 (beginning-of-defun)
21449 (end (save-excursion (end-of-defun) (point))))
21454 (and (< (point) end)
21456 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21458 (setq count (1+ count)))
21465 (defun lengths-list-file (filename)
21466 "Return list of definitions' lengths within FILE.
21467 The returned list is a list of numbers.
21468 Each number is the number of words or
21469 symbols in one function definition."
21473 (message "Working on `%s' ... " filename)
21475 (let ((buffer (find-file-noselect filename))
21477 (set-buffer buffer)
21478 (setq buffer-read-only t)
21480 (goto-char (point-min))
21484 (while (re-search-forward "^(defun" nil t)
21486 (cons (count-words-in-defun) lengths-list)))
21487 (kill-buffer buffer)
21494 (defun lengths-list-many-files (list-of-files)
21495 "Return list of lengths of defuns in LIST-OF-FILES."
21496 (let (lengths-list)
21497 ;;; @r{true-or-false-test}
21498 (while list-of-files
21504 ;;; @r{Generate a lengths' list.}
21506 (expand-file-name (car list-of-files)))))
21507 ;;; @r{Make files' list shorter.}
21508 (setq list-of-files (cdr list-of-files)))
21509 ;;; @r{Return final value of lengths' list.}
21516 (defun defuns-per-range (sorted-lengths top-of-ranges)
21517 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21518 (let ((top-of-range (car top-of-ranges))
21519 (number-within-range 0)
21520 defuns-per-range-list)
21525 (while top-of-ranges
21529 ;; @r{Need number for numeric test.}
21530 (car sorted-lengths)
21531 (< (car sorted-lengths) top-of-range))
21533 ;; @r{Count number of definitions within current range.}
21534 (setq number-within-range (1+ number-within-range))
21535 (setq sorted-lengths (cdr sorted-lengths)))
21539 ;; @r{Exit inner loop but remain within outer loop.}
21541 (setq defuns-per-range-list
21542 (cons number-within-range defuns-per-range-list))
21543 (setq number-within-range 0) ; @r{Reset count to zero.}
21545 ;; @r{Move to next range.}
21546 (setq top-of-ranges (cdr top-of-ranges))
21547 ;; @r{Specify next top of range value.}
21548 (setq top-of-range (car top-of-ranges)))
21552 ;; @r{Exit outer loop and count the number of defuns larger than}
21553 ;; @r{ the largest top-of-range value.}
21554 (setq defuns-per-range-list
21556 (length sorted-lengths)
21557 defuns-per-range-list))
21559 ;; @r{Return a list of the number of definitions within each range,}
21560 ;; @r{ smallest to largest.}
21561 (nreverse defuns-per-range-list)))
21567 (defun column-of-graph (max-graph-height actual-height)
21568 "Return list of MAX-GRAPH-HEIGHT strings;
21569 ACTUAL-HEIGHT are graph-symbols.
21570 The graph-symbols are contiguous entries at the end
21572 The list will be inserted as one column of a graph.
21573 The strings are either graph-blank or graph-symbol."
21577 (let ((insert-list nil)
21578 (number-of-top-blanks
21579 (- max-graph-height actual-height)))
21581 ;; @r{Fill in @code{graph-symbols}.}
21582 (while (> actual-height 0)
21583 (setq insert-list (cons graph-symbol insert-list))
21584 (setq actual-height (1- actual-height)))
21588 ;; @r{Fill in @code{graph-blanks}.}
21589 (while (> number-of-top-blanks 0)
21590 (setq insert-list (cons graph-blank insert-list))
21591 (setq number-of-top-blanks
21592 (1- number-of-top-blanks)))
21594 ;; @r{Return whole list.}
21601 (defun Y-axis-element (number full-Y-label-width)
21602 "Construct a NUMBERed label element.
21603 A numbered element looks like this ` 5 - ',
21604 and is padded as needed so all line up with
21605 the element for the largest number."
21608 (let* ((leading-spaces
21609 (- full-Y-label-width
21611 (concat (number-to-string number)
21616 (make-string leading-spaces ? )
21617 (number-to-string number)
21624 (defun print-Y-axis
21625 (height full-Y-label-width &optional vertical-step)
21626 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21627 Height must be the maximum height of the graph.
21628 Full width is the width of the highest label element.
21629 Optionally, print according to VERTICAL-STEP."
21632 ;; Value of height and full-Y-label-width
21633 ;; are passed by `print-graph'.
21634 (let ((start (point)))
21636 (Y-axis-column height full-Y-label-width vertical-step))
21639 ;; @r{Place point ready for inserting graph.}
21641 ;; @r{Move point forward by value of} full-Y-label-width
21642 (forward-char full-Y-label-width)))
21648 (defun print-X-axis-tic-line
21649 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21650 "Print ticks for X axis."
21651 (insert X-axis-leading-spaces)
21652 (insert X-axis-tic-symbol) ; @r{Under first column.}
21655 ;; @r{Insert second tic in the right spot.}
21658 (- (* symbol-width X-axis-label-spacing)
21659 ;; @r{Insert white space up to second tic symbol.}
21660 (* 2 (length X-axis-tic-symbol)))
21662 X-axis-tic-symbol))
21665 ;; @r{Insert remaining ticks.}
21666 (while (> number-of-X-tics 1)
21667 (insert X-axis-tic-element)
21668 (setq number-of-X-tics (1- number-of-X-tics))))
21674 (defun X-axis-element (number)
21675 "Construct a numbered X axis element."
21676 (let ((leading-spaces
21677 (- (* symbol-width X-axis-label-spacing)
21678 (length (number-to-string number)))))
21679 (concat (make-string leading-spaces ? )
21680 (number-to-string number))))
21686 (defun graph-body-print (numbers-list height symbol-width)
21687 "Print a bar graph of the NUMBERS-LIST.
21688 The numbers-list consists of the Y-axis values.
21689 HEIGHT is maximum height of graph.
21690 SYMBOL-WIDTH is number of each column."
21693 (let (from-position)
21694 (while numbers-list
21695 (setq from-position (point))
21697 (column-of-graph height (car numbers-list)))
21698 (goto-char from-position)
21699 (forward-char symbol-width)
21702 ;; @r{Draw graph column by column.}
21704 (setq numbers-list (cdr numbers-list)))
21705 ;; @r{Place point for X axis labels.}
21706 (forward-line height)
21713 (defun Y-axis-column
21714 (height width-of-label &optional vertical-step)
21715 "Construct list of labels for Y axis.
21716 HEIGHT is maximum height of graph.
21717 WIDTH-OF-LABEL is maximum width of label.
21720 VERTICAL-STEP, an option, is a positive integer
21721 that specifies how much a Y axis label increments
21722 for each line. For example, a step of 5 means
21723 that each line is five units of the graph."
21725 (number-per-line (or vertical-step 1)))
21728 (while (> height 1)
21729 (if (zerop (% height Y-axis-label-spacing))
21730 ;; @r{Insert label.}
21734 (* height number-per-line)
21739 ;; @r{Else, insert blanks.}
21742 (make-string width-of-label ? )
21744 (setq height (1- height)))
21747 ;; @r{Insert base line.}
21748 (setq Y-axis (cons (Y-axis-element
21749 (or vertical-step 1)
21752 (nreverse Y-axis)))
21758 (defun print-X-axis-numbered-line
21759 (number-of-X-tics X-axis-leading-spaces
21760 &optional horizontal-step)
21761 "Print line of X-axis numbers"
21762 (let ((number X-axis-label-spacing)
21763 (horizontal-step (or horizontal-step 1)))
21766 (insert X-axis-leading-spaces)
21768 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21771 ;; @r{Insert white space up to next number.}
21772 (- (* symbol-width X-axis-label-spacing)
21773 (1- (length (number-to-string horizontal-step)))
21776 (number-to-string (* number horizontal-step))))
21779 ;; @r{Insert remaining numbers.}
21780 (setq number (+ number X-axis-label-spacing))
21781 (while (> number-of-X-tics 1)
21782 (insert (X-axis-element (* number horizontal-step)))
21783 (setq number (+ number X-axis-label-spacing))
21784 (setq number-of-X-tics (1- number-of-X-tics)))))
21790 (defun print-X-axis (numbers-list horizontal-step)
21791 "Print X axis labels to length of NUMBERS-LIST.
21792 Optionally, HORIZONTAL-STEP, a positive integer,
21793 specifies how much an X axis label increments for
21797 ;; Value of symbol-width and full-Y-label-width
21798 ;; are passed by `print-graph'.
21799 (let* ((leading-spaces
21800 (make-string full-Y-label-width ? ))
21801 ;; symbol-width @r{is provided by} graph-body-print
21802 (tic-width (* symbol-width X-axis-label-spacing))
21803 (X-length (length numbers-list))
21809 ;; @r{Make a string of blanks.}
21810 (- (* symbol-width X-axis-label-spacing)
21811 (length X-axis-tic-symbol))
21815 ;; @r{Concatenate blanks with tic symbol.}
21816 X-axis-tic-symbol))
21818 (if (zerop (% X-length tic-width))
21819 (/ X-length tic-width)
21820 (1+ (/ X-length tic-width)))))
21824 (print-X-axis-tic-line
21825 tic-number leading-spaces X-tic)
21827 (print-X-axis-numbered-line
21828 tic-number leading-spaces horizontal-step)))
21834 (defun one-fiftieth (full-range)
21835 "Return list, each number of which is 1/50th previous."
21836 (mapcar (lambda (arg) (/ arg 50)) full-range))
21843 (numbers-list &optional vertical-step horizontal-step)
21844 "Print labeled bar graph of the NUMBERS-LIST.
21845 The numbers-list consists of the Y-axis values.
21849 Optionally, VERTICAL-STEP, a positive integer,
21850 specifies how much a Y axis label increments for
21851 each line. For example, a step of 5 means that
21852 each row is five units.
21856 Optionally, HORIZONTAL-STEP, a positive integer,
21857 specifies how much an X axis label increments for
21859 (let* ((symbol-width (length graph-blank))
21860 ;; @code{height} @r{is both the largest number}
21861 ;; @r{and the number with the most digits.}
21862 (height (apply 'max numbers-list))
21865 (height-of-top-line
21866 (if (zerop (% height Y-axis-label-spacing))
21869 (* (1+ (/ height Y-axis-label-spacing))
21870 Y-axis-label-spacing)))
21873 (vertical-step (or vertical-step 1))
21874 (full-Y-label-width
21878 (* height-of-top-line vertical-step))
21884 height-of-top-line full-Y-label-width vertical-step)
21886 numbers-list height-of-top-line symbol-width)
21887 (print-X-axis numbers-list horizontal-step)))
21894 @node Final printed graph
21895 @appendixsubsec The Printed Graph
21897 When made and installed, you can call the @code{print-graph} command
21903 (print-graph fiftieth-list-for-graph 50 10)
21933 50 - ***************** * *
21935 10 50 100 150 200 250 300 350
21942 The largest group of functions contain 10--19 words and symbols each.
21944 @node Free Software and Free Manuals
21945 @appendix Free Software and Free Manuals
21947 @strong{by Richard M. Stallman}
21950 The biggest deficiency in free operating systems is not in the
21951 software---it is the lack of good free manuals that we can include in
21952 these systems. Many of our most important programs do not come with
21953 full manuals. Documentation is an essential part of any software
21954 package; when an important free software package does not come with a
21955 free manual, that is a major gap. We have many such gaps today.
21957 Once upon a time, many years ago, I thought I would learn Perl. I got
21958 a copy of a free manual, but I found it hard to read. When I asked
21959 Perl users about alternatives, they told me that there were better
21960 introductory manuals---but those were not free.
21962 Why was this? The authors of the good manuals had written them for
21963 O'Reilly Associates, which published them with restrictive terms---no
21964 copying, no modification, source files not available---which exclude
21965 them from the free software community.
21967 That wasn't the first time this sort of thing has happened, and (to
21968 our community's great loss) it was far from the last. Proprietary
21969 manual publishers have enticed a great many authors to restrict their
21970 manuals since then. Many times I have heard a GNU user eagerly tell me
21971 about a manual that he is writing, with which he expects to help the
21972 GNU project---and then had my hopes dashed, as he proceeded to explain
21973 that he had signed a contract with a publisher that would restrict it
21974 so that we cannot use it.
21976 Given that writing good English is a rare skill among programmers, we
21977 can ill afford to lose manuals this way.
21979 Free documentation, like free software, is a matter of freedom, not
21980 price. The problem with these manuals was not that O'Reilly Associates
21981 charged a price for printed copies---that in itself is fine. The Free
21982 Software Foundation @uref{http://shop.fsf.org, sells printed copies} of
21983 free @uref{http://www.gnu.org/doc/doc.html, GNU manuals}, too.
21984 But GNU manuals are available in source code form, while these manuals
21985 are available only on paper. GNU manuals come with permission to copy
21986 and modify; the Perl manuals do not. These restrictions are the
21989 The criterion for a free manual is pretty much the same as for free
21990 software: it is a matter of giving all users certain
21991 freedoms. Redistribution (including commercial redistribution) must be
21992 permitted, so that the manual can accompany every copy of the program,
21993 on-line or on paper. Permission for modification is crucial too.
21995 As a general rule, I don't believe that it is essential for people to
21996 have permission to modify all sorts of articles and books. The issues
21997 for writings are not necessarily the same as those for software. For
21998 example, I don't think you or I are obliged to give permission to
21999 modify articles like this one, which describe our actions and our
22002 But there is a particular reason why the freedom to modify is crucial
22003 for documentation for free software. When people exercise their right
22004 to modify the software, and add or change its features, if they are
22005 conscientious they will change the manual too---so they can provide
22006 accurate and usable documentation with the modified program. A manual
22007 which forbids programmers to be conscientious and finish the job, or
22008 more precisely requires them to write a new manual from scratch if
22009 they change the program, does not fill our community's needs.
22011 While a blanket prohibition on modification is unacceptable, some
22012 kinds of limits on the method of modification pose no problem. For
22013 example, requirements to preserve the original author's copyright
22014 notice, the distribution terms, or the list of authors, are ok. It is
22015 also no problem to require modified versions to include notice that
22016 they were modified, even to have entire sections that may not be
22017 deleted or changed, as long as these sections deal with nontechnical
22018 topics. (Some GNU manuals have them.)
22020 These kinds of restrictions are not a problem because, as a practical
22021 matter, they don't stop the conscientious programmer from adapting the
22022 manual to fit the modified program. In other words, they don't block
22023 the free software community from making full use of the manual.
22025 However, it must be possible to modify all the technical content of
22026 the manual, and then distribute the result in all the usual media,
22027 through all the usual channels; otherwise, the restrictions do block
22028 the community, the manual is not free, and so we need another manual.
22030 Unfortunately, it is often hard to find someone to write another
22031 manual when a proprietary manual exists. The obstacle is that many
22032 users think that a proprietary manual is good enough---so they don't
22033 see the need to write a free manual. They do not see that the free
22034 operating system has a gap that needs filling.
22036 Why do users think that proprietary manuals are good enough? Some have
22037 not considered the issue. I hope this article will do something to
22040 Other users consider proprietary manuals acceptable for the same
22041 reason so many people consider proprietary software acceptable: they
22042 judge in purely practical terms, not using freedom as a
22043 criterion. These people are entitled to their opinions, but since
22044 those opinions spring from values which do not include freedom, they
22045 are no guide for those of us who do value freedom.
22047 Please spread the word about this issue. We continue to lose manuals
22048 to proprietary publishing. If we spread the word that proprietary
22049 manuals are not sufficient, perhaps the next person who wants to help
22050 GNU by writing documentation will realize, before it is too late, that
22051 he must above all make it free.
22053 We can also encourage commercial publishers to sell free, copylefted
22054 manuals instead of proprietary ones. One way you can help this is to
22055 check the distribution terms of a manual before you buy it, and prefer
22056 copylefted manuals to non-copylefted ones.
22060 Note: The Free Software Foundation maintains a page on its Web site
22061 that lists free books available from other publishers:@*
22062 @uref{http://www.gnu.org/doc/other-free-books.html}
22064 @node GNU Free Documentation License
22065 @appendix GNU Free Documentation License
22067 @cindex FDL, GNU Free Documentation License
22068 @include doclicense.texi
22074 MENU ENTRY: NODE NAME.
22080 @c Place biographical information on right-hand (verso) page
22083 \par\vfill\supereject
22085 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22086 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22089 % \par\vfill\supereject
22090 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22091 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22092 %\page\hbox{}%\page
22093 %\page\hbox{}%\page
22100 @c ================ Biographical information ================
22104 @center About the Author
22109 @node About the Author
22110 @unnumbered About the Author
22114 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22115 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22116 world on software freedom. Chassell was a founding Director and
22117 Treasurer of the Free Software Foundation, Inc. He is co-author of
22118 the @cite{Texinfo} manual, and has edited more than a dozen other
22119 books. He graduated from Cambridge University, in England. He has an
22120 abiding interest in social and economic history and flies his own
22127 @c @c Prevent page number on blank verso, so eject it first.
22129 @c \par\vfill\supereject
22134 @c @evenheading @thispage @| @| @thistitle
22135 @c @oddheading @| @| @thispage