1 \input texinfo @c -*-texinfo-*-
3 @setfilename standards.info
4 @settitle GNU Coding Standards
5 @c This date is automagically updated when you save this file:
6 @set lastupdate August 26, 1998
12 * Standards: (standards). GNU coding standards.
17 @c @setchapternewpage odd
18 @setchapternewpage off
20 @c This is used by a cross ref in make-stds.texi
31 Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998 Free Software Foundation, Inc.
33 Permission is granted to make and distribute verbatim copies of
34 this manual provided the copyright notice and this permission notice
35 are preserved on all copies.
38 Permission is granted to process this file through TeX and print the
39 results, provided the printed document carries copying permission
40 notice identical to this one except for the removal of this paragraph
41 (this paragraph not being relevant to the printed manual).
44 Permission is granted to copy and distribute modified versions of this
45 manual under the conditions for verbatim copying, provided that the entire
46 resulting derived work is distributed under the terms of a permission
47 notice identical to this one.
49 Permission is granted to copy and distribute translations of this manual
50 into another language, under the above conditions for modified versions,
51 except that this permission notice may be stated in a translation approved
52 by the Free Software Foundation.
56 @title GNU Coding Standards
57 @author Richard Stallman
58 @author last updated @value{lastupdate}
61 @vskip 0pt plus 1filll
62 Copyright @copyright{} 1992, 1993, 1994, 1995, 1996, 1997, 1998 Free Software Foundation, Inc.
64 Permission is granted to make and distribute verbatim copies of
65 this manual provided the copyright notice and this permission notice
66 are preserved on all copies.
68 Permission is granted to copy and distribute modified versions of this
69 manual under the conditions for verbatim copying, provided that the entire
70 resulting derived work is distributed under the terms of a permission
71 notice identical to this one.
73 Permission is granted to copy and distribute translations of this manual
74 into another language, under the above conditions for modified versions,
75 except that this permission notice may be stated in a translation approved
76 by the Free Software Foundation.
80 @node Top, Preface, (dir), (dir)
83 Last updated @value{lastupdate}.
87 * Preface:: About the GNU Coding Standards
88 * Intellectual Property:: Keeping Free Software Free
89 * Design Advice:: General Program Design
90 * Program Behavior:: Program Behavior for All Programs
91 * Writing C:: Making The Best Use of C
92 * Documentation:: Documenting Programs
93 * Managing Releases:: The Release Process
97 @chapter About the GNU Coding Standards
99 The GNU Coding Standards were written by Richard Stallman and other GNU
100 Project volunteers. Their purpose is to make the GNU system clean,
101 consistent, and easy to install. This document can also be read as a
102 guide to writing portable, robust and reliable programs. It focuses on
103 programs written in C, but many of the rules and principles are useful
104 even if you write in another programming language. The rules often
105 state reasons for writing in a certain way.
107 Corrections or suggestions for this document should be sent to
108 @email{gnu@@gnu.org}. If you make a suggestion, please include a
109 suggested new wording for it; our time is limited. We prefer a context
110 diff to the @file{standards.texi} or @file{make-stds.texi} files, but if
111 you don't have those files, please mail your suggestion anyway.
113 This release of the GNU Coding Standards was last updated
116 @node Intellectual Property
117 @chapter Keeping Free Software Free
119 This @value{CHAPTER} discusses how you can make sure that GNU software
120 remains unencumbered.
123 * Reading Non-Free Code:: Referring to Proprietary Programs
124 * Contributions:: Accepting Contributions
127 @node Reading Non-Free Code
128 @section Referring to Proprietary Programs
130 Don't in any circumstances refer to Unix source code for or during
131 your work on GNU! (Or to any other proprietary programs.)
133 If you have a vague recollection of the internals of a Unix program,
134 this does not absolutely mean you can't write an imitation of it, but
135 do try to organize the imitation internally along different lines,
136 because this is likely to make the details of the Unix version
137 irrelevant and dissimilar to your results.
139 For example, Unix utilities were generally optimized to minimize
140 memory use; if you go for speed instead, your program will be very
141 different. You could keep the entire input file in core and scan it
142 there instead of using stdio. Use a smarter algorithm discovered more
143 recently than the Unix program. Eliminate use of temporary files. Do
144 it in one pass instead of two (we did this in the assembler).
146 Or, on the contrary, emphasize simplicity instead of speed. For some
147 applications, the speed of today's computers makes simpler algorithms
150 Or go for generality. For example, Unix programs often have static
151 tables or fixed-size strings, which make for arbitrary limits; use
152 dynamic allocation instead. Make sure your program handles NULs and
153 other funny characters in the input files. Add a programming language
154 for extensibility and write part of the program in that language.
156 Or turn some parts of the program into independently usable libraries.
157 Or use a simple garbage collector instead of tracking precisely when
158 to free memory, or use a new GNU facility such as obstacks.
162 @section Accepting Contributions
164 If someone else sends you a piece of code to add to the program you are
165 working on, we need legal papers to use it---the same sort of legal
166 papers we will need to get from you. @emph{Each} significant
167 contributor to a program must sign some sort of legal papers in order
168 for us to have clear title to the program. The main author alone is not
171 So, before adding in any contributions from other people, please tell
172 us, so we can arrange to get the papers. Then wait until we tell you
173 that we have received the signed papers, before you actually use the
176 This applies both before you release the program and afterward. If
177 you receive diffs to fix a bug, and they make significant changes, we
178 need legal papers for that change.
180 This also applies to comments and documentation files. For copyright
181 law, comments and code are just text. Copyright applies to all kinds of
182 text, so we need legal papers for all kinds.
184 You don't need papers for changes of a few lines here or there, since
185 they are not significant for copyright purposes. Also, you don't need
186 papers if all you get from the suggestion is some ideas, not actual code
187 which you use. For example, if you write a different solution to the
188 problem, you don't need to get papers.
190 We know this is frustrating; it's frustrating for us as well. But if
191 you don't wait, you are going out on a limb---for example, what if the
192 contributor's employer won't sign a disclaimer? You might have to take
195 The very worst thing is if you forget to tell us about the other
196 contributor. We could be very embarrassed in court some day as a
199 We have more detailed advice for maintainers of programs; if you have
200 reached the stage of actually maintaining a program for GNU (whether
201 released or not), please ask us for a copy.
204 @chapter General Program Design
206 This @value{CHAPTER} discusses some of the issues you should take into
207 account when designing your program.
210 * Compatibility:: Compatibility with other implementations
211 * Using Extensions:: Using non-standard features
212 * ANSI C:: Using ANSI C features
213 * Source Language:: Using languages other than C
217 @section Compatibility with Other Implementations
219 With occasional exceptions, utility programs and libraries for GNU
220 should be upward compatible with those in Berkeley Unix, and upward
221 compatible with @sc{ansi} C if @sc{ansi} C specifies their behavior, and
222 upward compatible with @sc{posix} if @sc{posix} specifies their
225 When these standards conflict, it is useful to offer compatibility
226 modes for each of them.
228 @sc{ansi} C and @sc{posix} prohibit many kinds of extensions. Feel free
229 to make the extensions anyway, and include a @samp{--ansi},
230 @samp{--posix}, or @samp{--compatible} option to turn them off.
231 However, if the extension has a significant chance of breaking any real
232 programs or scripts, then it is not really upward compatible. Try to
233 redesign its interface.
235 Many GNU programs suppress extensions that conflict with @sc{posix} if the
236 environment variable @code{POSIXLY_CORRECT} is defined (even if it is
237 defined with a null value). Please make your program recognize this
238 variable if appropriate.
240 When a feature is used only by users (not by programs or command
241 files), and it is done poorly in Unix, feel free to replace it
242 completely with something totally different and better. (For example,
243 @code{vi} is replaced with Emacs.) But it is nice to offer a compatible
244 feature as well. (There is a free @code{vi} clone, so we offer it.)
246 Additional useful features not in Berkeley Unix are welcome.
248 @node Using Extensions
249 @section Using Non-standard Features
251 Many GNU facilities that already exist support a number of convenient
252 extensions over the comparable Unix facilities. Whether to use these
253 extensions in implementing your program is a difficult question.
255 On the one hand, using the extensions can make a cleaner program.
256 On the other hand, people will not be able to build the program
257 unless the other GNU tools are available. This might cause the
258 program to work on fewer kinds of machines.
260 With some extensions, it might be easy to provide both alternatives.
261 For example, you can define functions with a ``keyword'' @code{INLINE}
262 and define that as a macro to expand into either @code{inline} or
263 nothing, depending on the compiler.
265 In general, perhaps it is best not to use the extensions if you can
266 straightforwardly do without them, but to use the extensions if they
267 are a big improvement.
269 An exception to this rule are the large, established programs (such as
270 Emacs) which run on a great variety of systems. Such programs would
271 be broken by use of GNU extensions.
273 Another exception is for programs that are used as part of
274 compilation: anything that must be compiled with other compilers in
275 order to bootstrap the GNU compilation facilities. If these require
276 the GNU compiler, then no one can compile them without having them
277 installed already. That would be no good.
280 @section @sc{ansi} C and pre-@sc{ansi} C
282 Do not ever use the ``trigraph'' feature of @sc{ansi} C.
284 @sc{ansi} C is widespread enough now that it is ok to write new programs
285 that use @sc{ansi} C features (and therefore will not work in
286 non-@sc{ansi} compilers). And if a program is already written in
287 @sc{ansi} C, there's no need to convert it to support non-@sc{ansi}
290 However, it is easy to support non-@sc{ansi} compilers in most programs,
291 so you might still consider doing so when you write a program. Instead
292 of writing function definitions in @sc{ansi} prototype form,
301 write the definition in pre-@sc{ansi} style like this,
311 and use a separate declaration to specify the argument prototype:
317 You need such a declaration anyway, in a header file, to get the benefit
318 of @sc{ansi} C prototypes in all the files where the function is called.
319 And once you have it, you lose nothing by writing the function
320 definition in the pre-@sc{ansi} style.
322 If you don't know non-@sc{ansi} C, there's no need to learn it; just
323 write in @sc{ansi} C.
325 @node Source Language
326 @section Using Languages Other Than C
328 Using a language other than C is like using a non-standard feature: it
329 will cause trouble for users. Even if GCC supports the other language,
330 users may find it inconvenient to have to install the compiler for that
331 other language in order to build your program. For example, if you
332 write your program in C++, people will have to install the C++ compiler
333 in order to compile your program. Thus, it is better if you write in C.
335 But there are three situations when there is no disadvantage in using
340 It is okay to use another language if your program contains an
341 interpreter for that language.
343 For example, if your program links with GUILE, it is ok to write part of
344 the program in Scheme or another language supported by GUILE.
347 It is okay to use another language in a tool specifically intended for
348 use with that language.
350 This is okay because the only people who want to build the tool will be
351 those who have installed the other language anyway.
354 If an application is of interest to a narrow community, then perhaps
355 it's not important if the application is inconvenient to install.
358 C has one other advantage over C++ and other compiled languages: more
359 people know C, so more people will find it easy to read and modify the
360 program if it is written in C.
362 @node Program Behavior
363 @chapter Program Behavior for All Programs
365 This @value{CHAPTER} describes how to write robust software. It also
366 describes general standards for error messages, the command line interface,
367 and how libraries should behave.
370 * Semantics:: Writing robust programs
371 * Libraries:: Library behavior
372 * Errors:: Formatting error messages
373 * User Interfaces:: Standards for command line interfaces
374 * Option Table:: Table of long options.
375 * Memory Usage:: When and how to care about memory needs
379 @section Writing Robust Programs
381 Avoid arbitrary limits on the length or number of @emph{any} data
382 structure, including file names, lines, files, and symbols, by allocating
383 all data structures dynamically. In most Unix utilities, ``long lines
384 are silently truncated''. This is not acceptable in a GNU utility.
386 Utilities reading files should not drop NUL characters, or any other
387 nonprinting characters @emph{including those with codes above 0177}. The
388 only sensible exceptions would be utilities specifically intended for
389 interface to certain types of printers that can't handle those characters.
391 Check every system call for an error return, unless you know you wish to
392 ignore errors. Include the system error text (from @code{perror} or
393 equivalent) in @emph{every} error message resulting from a failing
394 system call, as well as the name of the file if any and the name of the
395 utility. Just ``cannot open foo.c'' or ``stat failed'' is not
398 Check every call to @code{malloc} or @code{realloc} to see if it
399 returned zero. Check @code{realloc} even if you are making the block
400 smaller; in a system that rounds block sizes to a power of 2,
401 @code{realloc} may get a different block if you ask for less space.
403 In Unix, @code{realloc} can destroy the storage block if it returns
404 zero. GNU @code{realloc} does not have this bug: if it fails, the
405 original block is unchanged. Feel free to assume the bug is fixed. If
406 you wish to run your program on Unix, and wish to avoid lossage in this
407 case, you can use the GNU @code{malloc}.
409 You must expect @code{free} to alter the contents of the block that was
410 freed. Anything you want to fetch from the block, you must fetch before
413 If @code{malloc} fails in a noninteractive program, make that a fatal
414 error. In an interactive program (one that reads commands from the
415 user), it is better to abort the command and return to the command
416 reader loop. This allows the user to kill other processes to free up
417 virtual memory, and then try the command again.
419 Use @code{getopt_long} to decode arguments, unless the argument syntax
420 makes this unreasonable.
422 When static storage is to be written in during program execution, use
423 explicit C code to initialize it. Reserve C initialized declarations
424 for data that will not be changed.
427 Try to avoid low-level interfaces to obscure Unix data structures (such
428 as file directories, utmp, or the layout of kernel memory), since these
429 are less likely to work compatibly. If you need to find all the files
430 in a directory, use @code{readdir} or some other high-level interface.
431 These will be supported compatibly by GNU.
433 The preferred signal handling facilities are the BSD variant of
434 @code{signal}, and the @sc{posix} @code{sigaction} function; the
435 alternative USG @code{signal} interface is an inferior design.
437 Nowadays, using the @sc{posix} signal functions may be the easiest way
438 to make a program portable. If you use @code{signal}, then on GNU/Linux
439 systems running GNU libc version 1, you should include
440 @file{bsd/signal.h} instead of @file{signal.h}, so as to get BSD
441 behavior. It is up to you whether to support systems where
442 @code{signal} has only the USG behavior, or give up on them.
444 In error checks that detect ``impossible'' conditions, just abort.
445 There is usually no point in printing any message. These checks
446 indicate the existence of bugs. Whoever wants to fix the bugs will have
447 to read the source code and run a debugger. So explain the problem with
448 comments in the source. The relevant data will be in variables, which
449 are easy to examine with the debugger, so there is no point moving them
452 Do not use a count of errors as the exit status for a program.
453 @emph{That does not work}, because exit status values are limited to 8
454 bits (0 through 255). A single run of the program might have 256
455 errors; if you try to return 256 as the exit status, the parent process
456 will see 0 as the status, and it will appear that the program succeeded.
458 If you make temporary files, check the @code{TMPDIR} environment
459 variable; if that variable is defined, use the specified directory
460 instead of @file{/tmp}.
463 @section Library Behavior
465 Try to make library functions reentrant. If they need to do dynamic
466 storage allocation, at least try to avoid any nonreentrancy aside from
467 that of @code{malloc} itself.
469 Here are certain name conventions for libraries, to avoid name
472 Choose a name prefix for the library, more than two characters long.
473 All external function and variable names should start with this
474 prefix. In addition, there should only be one of these in any given
475 library member. This usually means putting each one in a separate
478 An exception can be made when two external symbols are always used
479 together, so that no reasonable program could use one without the
480 other; then they can both go in the same file.
482 External symbols that are not documented entry points for the user
483 should have names beginning with @samp{_}. They should also contain
484 the chosen name prefix for the library, to prevent collisions with
485 other libraries. These can go in the same files with user entry
488 Static functions and variables can be used as you like and need not
489 fit any naming convention.
492 @section Formatting Error Messages
494 Error messages from compilers should look like this:
497 @var{source-file-name}:@var{lineno}: @var{message}
500 Error messages from other noninteractive programs should look like this:
503 @var{program}:@var{source-file-name}:@var{lineno}: @var{message}
507 when there is an appropriate source file, or like this:
510 @var{program}: @var{message}
514 when there is no relevant source file.
516 In an interactive program (one that is reading commands from a
517 terminal), it is better not to include the program name in an error
518 message. The place to indicate which program is running is in the
519 prompt or with the screen layout. (When the same program runs with
520 input from a source other than a terminal, it is not interactive and
521 would do best to print error messages using the noninteractive style.)
523 The string @var{message} should not begin with a capital letter when
524 it follows a program name and/or file name. Also, it should not end
527 Error messages from interactive programs, and other messages such as
528 usage messages, should start with a capital letter. But they should not
531 @node User Interfaces
532 @section Standards for Command Line Interfaces
534 Please don't make the behavior of a utility depend on the name used
535 to invoke it. It is useful sometimes to make a link to a utility
536 with a different name, and that should not change what it does.
538 Instead, use a run time option or a compilation switch or both
539 to select among the alternate behaviors.
541 Likewise, please don't make the behavior of the program depend on the
542 type of output device it is used with. Device independence is an
543 important principle of the system's design; do not compromise it merely
544 to save someone from typing an option now and then. (Variation in error
545 message syntax when using a terminal is ok, because that is a side issue
546 that people do not depend on.)
548 If you think one behavior is most useful when the output is to a
549 terminal, and another is most useful when the output is a file or a
550 pipe, then it is usually best to make the default behavior the one that
551 is useful with output to a terminal, and have an option for the other
554 Compatibility requires certain programs to depend on the type of output
555 device. It would be disastrous if @code{ls} or @code{sh} did not do so
556 in the way all users expect. In some of these cases, we supplement the
557 program with a preferred alternate version that does not depend on the
558 output device type. For example, we provide a @code{dir} program much
559 like @code{ls} except that its default output format is always
562 It is a good idea to follow the @sc{posix} guidelines for the
563 command-line options of a program. The easiest way to do this is to use
564 @code{getopt} to parse them. Note that the GNU version of @code{getopt}
565 will normally permit options anywhere among the arguments unless the
566 special argument @samp{--} is used. This is not what @sc{posix}
567 specifies; it is a GNU extension.
569 Please define long-named options that are equivalent to the
570 single-letter Unix-style options. We hope to make GNU more user
571 friendly this way. This is easy to do with the GNU function
574 One of the advantages of long-named options is that they can be
575 consistent from program to program. For example, users should be able
576 to expect the ``verbose'' option of any GNU program which has one, to be
577 spelled precisely @samp{--verbose}. To achieve this uniformity, look at
578 the table of common long-option names when you choose the option names
579 for your program (@pxref{Option Table}).
581 It is usually a good idea for file names given as ordinary arguments to
582 be input files only; any output files would be specified using options
583 (preferably @samp{-o} or @samp{--output}). Even if you allow an output
584 file name as an ordinary argument for compatibility, try to provide an
585 option as another way to specify it. This will lead to more consistency
586 among GNU utilities, and fewer idiosyncracies for users to remember.
588 All programs should support two standard options: @samp{--version}
593 This option should direct the program to information about its name,
594 version, origin and legal status, all on standard output, and then exit
595 successfully. Other options and arguments should be ignored once this
596 is seen, and the program should not perform its normal function.
598 The first line is meant to be easy for a program to parse; the version
599 number proper starts after the last space. In addition, it contains
600 the canonical name for this program, in this format:
607 The program's name should be a constant string; @emph{don't} compute it
608 from @code{argv[0]}. The idea is to state the standard or canonical
609 name for the program, not its file name. There are other ways to find
610 out the precise file name where a command is found in @code{PATH}.
612 If the program is a subsidiary part of a larger package, mention the
613 package name in parentheses, like this:
616 emacsserver (GNU Emacs) 19.30
620 If the package has a version number which is different from this
621 program's version number, you can mention the package version number
622 just before the close-parenthesis.
624 If you @strong{need} to mention the version numbers of libraries which
625 are distributed separately from the package which contains this program,
626 you can do so by printing an additional line of version info for each
627 library you want to mention. Use the same format for these lines as for
630 Please do not mention all of the libraries that the program uses ``just
631 for completeness''---that would produce a lot of unhelpful clutter.
632 Please mention library version numbers only if you find in practice that
633 they are very important to you in debugging.
635 The following line, after the version number line or lines, should be a
636 copyright notice. If more than one copyright notice is called for, put
637 each on a separate line.
639 Next should follow a brief statement that the program is free software,
640 and that users are free to copy and change it on certain conditions. If
641 the program is covered by the GNU GPL, say so here. Also mention that
642 there is no warranty, to the extent permitted by law.
644 It is ok to finish the output with a list of the major authors of the
645 program, as a way of giving credit.
647 Here's an example of output that follows these rules:
651 Copyright (C) 1996 Free Software Foundation, Inc.
652 GNU Emacs comes with NO WARRANTY,
653 to the extent permitted by law.
654 You may redistribute copies of GNU Emacs
655 under the terms of the GNU General Public License.
656 For more information about these matters,
657 see the files named COPYING.
660 You should adapt this to your program, of course, filling in the proper
661 year, copyright holder, name of program, and the references to
662 distribution terms, and changing the rest of the wording as necessary.
664 This copyright notice only needs to mention the most recent year in
665 which changes were made---there's no need to list the years for previous
666 versions' changes. You don't have to mention the name of the program in
667 these notices, if that is inconvenient, since it appeared in the first
671 This option should output brief documentation for how to invoke the
672 program, on standard output, then exit successfully. Other options and
673 arguments should be ignored once this is seen, and the program should
674 not perform its normal function.
676 Near the end of the @samp{--help} option's output there should be a line
677 that says where to mail bug reports. It should have this format:
680 Report bugs to @var{mailing-address}.
685 @section Table of Long Options
687 Here is a table of long options used by GNU programs. It is surely
688 incomplete, but we aim to list all the options that a new program might
689 want to be compatible with. If you use names not already in the table,
690 please send @email{gnu@@gnu.org} a list of them, with their
691 meanings, so we can update the table.
693 @c Please leave newlines between items in this table; it's much easier
694 @c to update when it isn't completely squashed together and unreadable.
695 @c When there is more than one short option for a long option name, put
696 @c a semicolon between the lists of the programs that use them, not a
697 @c period. --friedman
701 @samp{-N} in @code{tar}.
704 @samp{-a} in @code{du}, @code{ls}, @code{nm}, @code{stty}, @code{uname},
708 @samp{-a} in @code{diff}.
711 @samp{-A} in @code{ls}.
714 @samp{-a} in @code{etags}, @code{tee}, @code{time};
715 @samp{-r} in @code{tar}.
718 @samp{-a} in @code{cp}.
721 @samp{-n} in @code{shar}.
724 @samp{-l} in @code{m4}.
727 @samp{-a} in @code{diff}.
730 @samp{-v} in @code{gawk}.
739 @samp{-a} in @code{recode}.
742 @samp{-a} in @code{wdiff}.
745 @samp{-A} in @code{ptx}.
748 @samp{-n} in @code{wdiff}.
751 For server programs, run in the background.
753 @item backward-search
754 @samp{-B} in @code{ctags}.
757 @samp{-f} in @code{shar}.
766 @samp{-b} in @code{tac}.
769 @samp{-b} in @code{cpio} and @code{diff}.
772 @samp{-b} in @code{shar}.
775 Used in @code{cpio} and @code{tar}.
778 @samp{-b} in @code{head} and @code{tail}.
781 @samp{-b} in @code{ptx}.
784 Used in various programs to make output shorter.
787 @samp{-c} in @code{head}, @code{split}, and @code{tail}.
790 @samp{-C} in @code{etags}.
793 @samp{-A} in @code{tar}.
796 Used in various programs to specify the directory to use.
799 @samp{-c} in @code{chgrp} and @code{chown}.
802 @samp{-F} in @code{ls}.
805 @samp{-c} in @code{recode}.
808 @samp{-c} in @code{su};
812 @samp{-d} in @code{tar}.
818 @samp{-Z} in @code{tar} and @code{shar}.
821 @samp{-A} in @code{tar}.
824 @samp{-w} in @code{tar}.
830 @samp{-W copyleft} in @code{gawk}.
833 @samp{-C} in @code{ptx}, @code{recode}, and @code{wdiff};
834 @samp{-W copyright} in @code{gawk}.
840 @samp{-q} in @code{who}.
843 @samp{-l} in @code{du}.
846 Used in @code{tar} and @code{cpio}.
849 @samp{-c} in @code{shar}.
852 @samp{-x} in @code{ctags}.
855 @samp{-d} in @code{touch}.
858 @samp{-d} in Make and @code{m4};
862 @samp{-D} in @code{m4}.
865 @samp{-d} in Bison and @code{ctags}.
868 @samp{-D} in @code{tar}.
871 @samp{-L} in @code{chgrp}, @code{chown}, @code{cpio}, @code{du},
872 @code{ls}, and @code{tar}.
874 @item dereference-args
875 @samp{-D} in @code{du}.
878 @samp{-d} in @code{recode}.
880 @item dictionary-order
881 @samp{-d} in @code{look}.
884 @samp{-d} in @code{tar}.
887 @samp{-n} in @code{csplit}.
890 Specify the directory to use, in various programs. In @code{ls}, it
891 means to show directories themselves rather than their contents. In
892 @code{rm} and @code{ln}, it means to not treat links to directories
896 @samp{-x} in @code{strip}.
899 @samp{-X} in @code{strip}.
905 @samp{-e} in @code{diff}.
907 @item elide-empty-files
908 @samp{-z} in @code{csplit}.
911 @samp{-x} in @code{wdiff}.
914 @samp{-z} in @code{wdiff}.
916 @item entire-new-file
917 @samp{-N} in @code{diff}.
919 @item environment-overrides
923 @samp{-e} in @code{xargs}.
929 Used in @code{makeinfo}.
932 @samp{-o} in @code{m4}.
935 @samp{-b} in @code{ls}.
938 @samp{-X} in @code{tar}.
944 @samp{-x} in @code{xargs}.
947 @samp{-e} in @code{unshar}.
950 @samp{-t} in @code{diff}.
953 @samp{-e} in @code{sed}.
956 @samp{-g} in @code{nm}.
959 @samp{-i} in @code{cpio};
960 @samp{-x} in @code{tar}.
963 @samp{-f} in @code{finger}.
966 @samp{-f} in @code{su}.
969 @samp{-E} in @code{m4}.
972 @samp{-f} in @code{info}, @code{gawk}, Make, @code{mt}, and @code{tar};
973 @samp{-n} in @code{sed};
974 @samp{-r} in @code{touch}.
976 @item field-separator
977 @samp{-F} in @code{gawk}.
983 @samp{-F} in @code{ls}.
986 @samp{-T} in @code{tar}.
989 Used in @code{makeinfo}.
991 @item flag-truncation
992 @samp{-F} in @code{ptx}.
994 @item fixed-output-files
998 @samp{-f} in @code{tail}.
1000 @item footnote-style
1001 Used in @code{makeinfo}.
1004 @samp{-f} in @code{cp}, @code{ln}, @code{mv}, and @code{rm}.
1007 @samp{-F} in @code{shar}.
1010 For server programs, run in the foreground;
1011 in other words, don't do anything special to run the server
1015 Used in @code{ls}, @code{time}, and @code{ptx}.
1018 @samp{-F} in @code{m4}.
1024 @samp{-g} in @code{ptx}.
1027 @samp{-x} in @code{tar}.
1030 @samp{-i} in @code{ul}.
1033 @samp{-g} in @code{recode}.
1036 @samp{-g} in @code{install}.
1039 @samp{-z} in @code{tar} and @code{shar}.
1042 @samp{-H} in @code{m4}.
1045 @samp{-h} in @code{objdump} and @code{recode}
1048 @samp{-H} in @code{who}.
1051 Used to ask for brief usage information.
1053 @item here-delimiter
1054 @samp{-d} in @code{shar}.
1056 @item hide-control-chars
1057 @samp{-q} in @code{ls}.
1060 @samp{-u} in @code{who}.
1063 @samp{-D} in @code{diff}.
1066 @samp{-I} in @code{ls};
1067 @samp{-x} in @code{recode}.
1069 @item ignore-all-space
1070 @samp{-w} in @code{diff}.
1072 @item ignore-backups
1073 @samp{-B} in @code{ls}.
1075 @item ignore-blank-lines
1076 @samp{-B} in @code{diff}.
1079 @samp{-f} in @code{look} and @code{ptx};
1080 @samp{-i} in @code{diff} and @code{wdiff}.
1086 @samp{-i} in @code{ptx}.
1088 @item ignore-indentation
1089 @samp{-I} in @code{etags}.
1091 @item ignore-init-file
1094 @item ignore-interrupts
1095 @samp{-i} in @code{tee}.
1097 @item ignore-matching-lines
1098 @samp{-I} in @code{diff}.
1100 @item ignore-space-change
1101 @samp{-b} in @code{diff}.
1104 @samp{-i} in @code{tar}.
1107 @samp{-i} in @code{etags};
1108 @samp{-I} in @code{m4}.
1114 @samp{-G} in @code{tar}.
1117 @samp{-i}, @samp{-l}, and @samp{-m} in Finger.
1120 @samp{-i} in @code{expand}.
1123 @samp{-T} in @code{diff}.
1126 @samp{-i} in @code{ls}.
1129 @samp{-i} in @code{cp}, @code{ln}, @code{mv}, @code{rm};
1130 @samp{-e} in @code{m4};
1131 @samp{-p} in @code{xargs};
1132 @samp{-w} in @code{tar}.
1135 @samp{-p} in @code{shar}.
1147 @samp{-k} in @code{csplit}.
1150 @samp{-k} in @code{du} and @code{ls}.
1153 @samp{-l} in @code{etags}.
1156 @samp{-l} in @code{wdiff}.
1158 @item level-for-gzip
1159 @samp{-g} in @code{shar}.
1162 @samp{-C} in @code{split}.
1165 Used in @code{split}, @code{head}, and @code{tail}.
1168 @samp{-l} in @code{cpio}.
1172 Used in @code{gawk}.
1175 @samp{-t} in @code{cpio};
1176 @samp{-l} in @code{recode}.
1179 @samp{-t} in @code{tar}.
1182 @samp{-N} in @code{ls}.
1191 No listing of which programs already use this;
1192 someone should check to
1193 see if any actually do, and tell @email{gnu@@gnu.org}.
1196 @samp{-M} in @code{ptx}.
1199 @samp{-m} in @code{hello} and @code{uname}.
1201 @item make-directories
1202 @samp{-d} in @code{cpio}.
1211 @samp{-n} in @code{xargs}.
1214 @samp{-n} in @code{xargs}.
1217 @samp{-l} in @code{xargs}.
1223 @samp{-P} in @code{xargs}.
1226 @samp{-T} in @code{who}.
1229 @samp{-T} in @code{who}.
1232 @samp{-d} in @code{diff}.
1234 @item mixed-uuencode
1235 @samp{-M} in @code{shar}.
1238 @samp{-m} in @code{install}, @code{mkdir}, and @code{mkfifo}.
1240 @item modification-time
1241 @samp{-m} in @code{tar}.
1244 @samp{-M} in @code{tar}.
1250 @samp{-L} in @code{m4}.
1253 @samp{-a} in @code{shar}.
1258 @item no-builtin-rules
1261 @item no-character-count
1262 @samp{-w} in @code{shar}.
1264 @item no-check-existing
1265 @samp{-x} in @code{shar}.
1268 @samp{-3} in @code{wdiff}.
1271 @samp{-c} in @code{touch}.
1274 @samp{-D} in @code{etags}.
1277 @samp{-1} in @code{wdiff}.
1279 @item no-dereference
1280 @samp{-d} in @code{cp}.
1283 @samp{-2} in @code{wdiff}.
1292 @samp{-P} in @code{shar}.
1295 @samp{-e} in @code{gprof}.
1298 @samp{-R} in @code{etags}.
1301 @samp{-p} in @code{nm}.
1304 Used in @code{makeinfo}.
1307 @samp{-a} in @code{gprof}.
1310 @samp{-E} in @code{gprof}.
1313 @samp{-m} in @code{shar}.
1316 Used in @code{makeinfo}.
1319 Used in @code{emacsclient}.
1322 Used in various programs to inhibit warnings.
1325 @samp{-n} in @code{info}.
1328 @samp{-n} in @code{uname}.
1331 @samp{-f} in @code{cpio}.
1334 @samp{-n} in @code{objdump}.
1337 @samp{-0} in @code{xargs}.
1340 @samp{-n} in @code{cat}.
1342 @item number-nonblank
1343 @samp{-b} in @code{cat}.
1346 @samp{-n} in @code{nm}.
1348 @item numeric-uid-gid
1349 @samp{-n} in @code{cpio} and @code{ls}.
1355 @samp{-o} in @code{tar}.
1360 @item one-file-system
1361 @samp{-l} in @code{tar}, @code{cp}, and @code{du}.
1364 @samp{-o} in @code{ptx}.
1367 @samp{-f} in @code{gprof}.
1370 @samp{-F} in @code{gprof}.
1373 In various programs, specify the output file name.
1376 @samp{-o} in @code{shar}.
1379 @samp{-o} in @code{rm}.
1382 @samp{-c} in @code{unshar}.
1385 @samp{-o} in @code{install}.
1388 @samp{-l} in @code{diff}.
1390 @item paragraph-indent
1391 Used in @code{makeinfo}.
1394 @samp{-p} in @code{mkdir} and @code{rmdir}.
1397 @samp{-p} in @code{ul}.
1400 @samp{-p} in @code{cpio}.
1403 @samp{-P} in @code{finger}.
1406 @samp{-c} in @code{cpio} and @code{tar}.
1409 Used in @code{gawk}.
1411 @item prefix-builtins
1412 @samp{-P} in @code{m4}.
1415 @samp{-f} in @code{csplit}.
1418 Used in @code{tar} and @code{cp}.
1420 @item preserve-environment
1421 @samp{-p} in @code{su}.
1423 @item preserve-modification-time
1424 @samp{-m} in @code{cpio}.
1426 @item preserve-order
1427 @samp{-s} in @code{tar}.
1429 @item preserve-permissions
1430 @samp{-p} in @code{tar}.
1433 @samp{-l} in @code{diff}.
1436 @samp{-L} in @code{cmp}.
1438 @item print-data-base
1441 @item print-directory
1444 @item print-file-name
1445 @samp{-o} in @code{nm}.
1448 @samp{-s} in @code{nm}.
1451 @samp{-p} in @code{wdiff}.
1454 @samp{-p} in @code{ed}.
1457 @samp{-X} in @code{shar}.
1463 Used in many programs to inhibit the usual output. @strong{Note:} every
1464 program accepting @samp{--quiet} should accept @samp{--silent} as a
1468 @samp{-Q} in @code{shar}
1471 @samp{-Q} in @code{ls}.
1474 @samp{-n} in @code{diff}.
1477 Used in @code{gawk}.
1479 @item read-full-blocks
1480 @samp{-B} in @code{tar}.
1489 @samp{-R} in @code{tar}.
1492 Used in @code{chgrp}, @code{chown}, @code{cp}, @code{ls}, @code{diff},
1495 @item reference-limit
1496 Used in @code{makeinfo}.
1499 @samp{-r} in @code{ptx}.
1502 @samp{-r} in @code{tac} and @code{etags}.
1505 @samp{-r} in @code{uname}.
1508 @samp{-R} in @code{m4}.
1511 @samp{-r} in @code{objdump}.
1514 @samp{-r} in @code{cpio}.
1517 @samp{-i} in @code{xargs}.
1519 @item report-identical-files
1520 @samp{-s} in @code{diff}.
1522 @item reset-access-time
1523 @samp{-a} in @code{cpio}.
1526 @samp{-r} in @code{ls} and @code{nm}.
1529 @samp{-f} in @code{diff}.
1531 @item right-side-defs
1532 @samp{-R} in @code{ptx}.
1535 @samp{-s} in @code{tar}.
1537 @item same-permissions
1538 @samp{-p} in @code{tar}.
1541 @samp{-g} in @code{stty}.
1546 @item sentence-regexp
1547 @samp{-S} in @code{ptx}.
1550 @samp{-S} in @code{du}.
1553 @samp{-s} in @code{tac}.
1556 Used by @code{recode} to chose files or pipes for sequencing passes.
1559 @samp{-s} in @code{su}.
1562 @samp{-A} in @code{cat}.
1564 @item show-c-function
1565 @samp{-p} in @code{diff}.
1568 @samp{-E} in @code{cat}.
1570 @item show-function-line
1571 @samp{-F} in @code{diff}.
1574 @samp{-T} in @code{cat}.
1577 Used in many programs to inhibit the usual output.
1578 @strong{Note:} every program accepting
1579 @samp{--silent} should accept @samp{--quiet} as a synonym.
1582 @samp{-s} in @code{ls}.
1585 Specify a file descriptor for a network server to use for its socket,
1586 instead of opening and binding a new socket. This provides a way to
1587 run, in a nonpriveledged process, a server that normally needs a
1588 reserved port number.
1594 @samp{-W source} in @code{gawk}.
1597 @samp{-S} in @code{tar}.
1599 @item speed-large-files
1600 @samp{-H} in @code{diff}.
1603 @samp{-E} in @code{unshar}.
1605 @item split-size-limit
1606 @samp{-L} in @code{shar}.
1609 @samp{-s} in @code{cat}.
1612 @samp{-w} in @code{wdiff}.
1615 @samp{-y} in @code{wdiff}.
1618 Used in @code{tar} and @code{diff} to specify which file within
1619 a directory to start processing with.
1622 @samp{-s} in @code{wdiff}.
1624 @item stdin-file-list
1625 @samp{-S} in @code{shar}.
1631 @samp{-s} in @code{recode}.
1634 @samp{-s} in @code{install}.
1637 @samp{-s} in @code{strip}.
1640 @samp{-S} in @code{strip}.
1643 @samp{-s} in @code{shar}.
1646 @samp{-S} in @code{cp}, @code{ln}, @code{mv}.
1649 @samp{-b} in @code{csplit}.
1652 @samp{-s} in @code{gprof}.
1655 @samp{-s} in @code{du}.
1658 @samp{-s} in @code{ln}.
1661 Used in GDB and @code{objdump}.
1664 @samp{-s} in @code{m4}.
1667 @samp{-s} in @code{uname}.
1670 @samp{-t} in @code{expand} and @code{unexpand}.
1673 @samp{-T} in @code{ls}.
1676 @samp{-T} in @code{tput} and @code{ul}.
1677 @samp{-t} in @code{wdiff}.
1680 @samp{-a} in @code{diff}.
1683 @samp{-T} in @code{shar}.
1686 Used in @code{ls} and @code{touch}.
1689 @samp{-O} in @code{tar}.
1692 @samp{-c} in @code{du}.
1695 @samp{-t} in Make, @code{ranlib}, and @code{recode}.
1698 @samp{-t} in @code{m4}.
1701 @samp{-t} in @code{hello};
1702 @samp{-W traditional} in @code{gawk};
1703 @samp{-G} in @code{ed}, @code{m4}, and @code{ptx}.
1709 @samp{-t} in @code{ctags}.
1711 @item typedefs-and-c++
1712 @samp{-T} in @code{ctags}.
1715 @samp{-t} in @code{ptx}.
1718 @samp{-z} in @code{tar}.
1721 @samp{-u} in @code{cpio}.
1724 @samp{-U} in @code{m4}.
1726 @item undefined-only
1727 @samp{-u} in @code{nm}.
1730 @samp{-u} in @code{cp}, @code{ctags}, @code{mv}, @code{tar}.
1733 Used in @code{gawk}; same as @samp{--help}.
1736 @samp{-B} in @code{shar}.
1738 @item vanilla-operation
1739 @samp{-V} in @code{shar}.
1742 Print more information about progress. Many programs support this.
1745 @samp{-W} in @code{tar}.
1748 Print the version number.
1750 @item version-control
1751 @samp{-V} in @code{cp}, @code{ln}, @code{mv}.
1754 @samp{-v} in @code{ctags}.
1757 @samp{-V} in @code{tar}.
1762 @item whole-size-limit
1763 @samp{-l} in @code{shar}.
1766 @samp{-w} in @code{ls} and @code{ptx}.
1769 @samp{-W} in @code{ptx}.
1772 @samp{-T} in @code{who}.
1775 @samp{-z} in @code{gprof}.
1779 @section Memory Usage
1781 If it typically uses just a few meg of memory, don't bother making any
1782 effort to reduce memory usage. For example, if it is impractical for
1783 other reasons to operate on files more than a few meg long, it is
1784 reasonable to read entire input files into core to operate on them.
1786 However, for programs such as @code{cat} or @code{tail}, that can
1787 usefully operate on very large files, it is important to avoid using a
1788 technique that would artificially limit the size of files it can handle.
1789 If a program works by lines and could be applied to arbitrary
1790 user-supplied input files, it should keep only a line in memory, because
1791 this is not very hard and users will want to be able to operate on input
1792 files that are bigger than will fit in core all at once.
1794 If your program creates complicated data structures, just make them in
1795 core and give a fatal error if @code{malloc} returns zero.
1798 @chapter Making The Best Use of C
1800 This @value{CHAPTER} provides advice on how best to use the C language
1801 when writing GNU software.
1804 * Formatting:: Formatting Your Source Code
1805 * Comments:: Commenting Your Work
1806 * Syntactic Conventions:: Clean Use of C Constructs
1807 * Names:: Naming Variables and Functions
1808 * System Portability:: Portability between different operating systems
1809 * CPU Portability:: Supporting the range of CPU types
1810 * System Functions:: Portability and ``standard'' library functions
1811 * Internationalization:: Techniques for internationalization
1812 * Mmap:: How you can safely use @code{mmap}.
1816 @section Formatting Your Source Code
1818 It is important to put the open-brace that starts the body of a C
1819 function in column zero, and avoid putting any other open-brace or
1820 open-parenthesis or open-bracket in column zero. Several tools look
1821 for open-braces in column zero to find the beginnings of C functions.
1822 These tools will not work on code not formatted that way.
1824 It is also important for function definitions to start the name of the
1825 function in column zero. This helps people to search for function
1826 definitions, and may also help certain tools recognize them. Thus,
1827 the proper format is this:
1831 concat (s1, s2) /* Name starts in column zero here */
1833 @{ /* Open brace in column zero here */
1839 or, if you want to use @sc{ansi} C, format the definition like this:
1843 concat (char *s1, char *s2)
1849 In @sc{ansi} C, if the arguments don't fit nicely on one line,
1854 lots_of_args (int an_integer, long a_long, short a_short,
1855 double a_double, float a_float)
1859 For the body of the function, we prefer code formatted like this:
1871 return ++x + bar ();
1875 We find it easier to read a program when it has spaces before the
1876 open-parentheses and after the commas. Especially after the commas.
1878 When you split an expression into multiple lines, split it
1879 before an operator, not after one. Here is the right way:
1882 if (foo_this_is_long && bar > win (x, y, z)
1883 && remaining_condition)
1886 Try to avoid having two operators of different precedence at the same
1887 level of indentation. For example, don't write this:
1890 mode = (inmode[j] == VOIDmode
1891 || GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])
1892 ? outmode[j] : inmode[j]);
1895 Instead, use extra parentheses so that the indentation shows the nesting:
1898 mode = ((inmode[j] == VOIDmode
1899 || (GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])))
1900 ? outmode[j] : inmode[j]);
1903 Insert extra parentheses so that Emacs will indent the code properly.
1904 For example, the following indentation looks nice if you do it by hand,
1905 but Emacs would mess it up:
1908 v = rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
1909 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000;
1912 But adding a set of parentheses solves the problem:
1915 v = (rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
1916 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000);
1919 Format do-while statements like this:
1929 Please use formfeed characters (control-L) to divide the program into
1930 pages at logical places (but not within a function). It does not matter
1931 just how long the pages are, since they do not have to fit on a printed
1932 page. The formfeeds should appear alone on lines by themselves.
1936 @section Commenting Your Work
1938 Every program should start with a comment saying briefly what it is for.
1939 Example: @samp{fmt - filter for simple filling of text}.
1941 Please write the comments in a GNU program in English, because English
1942 is the one language that nearly all programmers in all countries can
1943 read. If you do not write English well, please write comments in
1944 English as well as you can, then ask other people to help rewrite them.
1945 If you can't write comments in English, please find someone to work with
1946 you and translate your comments into English.
1948 Please put a comment on each function saying what the function does,
1949 what sorts of arguments it gets, and what the possible values of
1950 arguments mean and are used for. It is not necessary to duplicate in
1951 words the meaning of the C argument declarations, if a C type is being
1952 used in its customary fashion. If there is anything nonstandard about
1953 its use (such as an argument of type @code{char *} which is really the
1954 address of the second character of a string, not the first), or any
1955 possible values that would not work the way one would expect (such as,
1956 that strings containing newlines are not guaranteed to work), be sure
1959 Also explain the significance of the return value, if there is one.
1961 Please put two spaces after the end of a sentence in your comments, so
1962 that the Emacs sentence commands will work. Also, please write
1963 complete sentences and capitalize the first word. If a lower-case
1964 identifier comes at the beginning of a sentence, don't capitalize it!
1965 Changing the spelling makes it a different identifier. If you don't
1966 like starting a sentence with a lower case letter, write the sentence
1967 differently (e.g., ``The identifier lower-case is @dots{}'').
1969 The comment on a function is much clearer if you use the argument
1970 names to speak about the argument values. The variable name itself
1971 should be lower case, but write it in upper case when you are speaking
1972 about the value rather than the variable itself. Thus, ``the inode
1973 number NODE_NUM'' rather than ``an inode''.
1975 There is usually no purpose in restating the name of the function in
1976 the comment before it, because the reader can see that for himself.
1977 There might be an exception when the comment is so long that the function
1978 itself would be off the bottom of the screen.
1980 There should be a comment on each static variable as well, like this:
1983 /* Nonzero means truncate lines in the display;
1984 zero means continue them. */
1988 Every @samp{#endif} should have a comment, except in the case of short
1989 conditionals (just a few lines) that are not nested. The comment should
1990 state the condition of the conditional that is ending, @emph{including
1991 its sense}. @samp{#else} should have a comment describing the condition
1992 @emph{and sense} of the code that follows. For example:
2000 #endif /* not foo */
2010 but, by contrast, write the comments this way for a @samp{#ifndef}:
2023 #endif /* not foo */
2027 @node Syntactic Conventions
2028 @section Clean Use of C Constructs
2030 Please explicitly declare all arguments to functions.
2031 Don't omit them just because they are @code{int}s.
2033 Declarations of external functions and functions to appear later in the
2034 source file should all go in one place near the beginning of the file
2035 (somewhere before the first function definition in the file), or else
2036 should go in a header file. Don't put @code{extern} declarations inside
2039 It used to be common practice to use the same local variables (with
2040 names like @code{tem}) over and over for different values within one
2041 function. Instead of doing this, it is better declare a separate local
2042 variable for each distinct purpose, and give it a name which is
2043 meaningful. This not only makes programs easier to understand, it also
2044 facilitates optimization by good compilers. You can also move the
2045 declaration of each local variable into the smallest scope that includes
2046 all its uses. This makes the program even cleaner.
2048 Don't use local variables or parameters that shadow global identifiers.
2050 Don't declare multiple variables in one declaration that spans lines.
2051 Start a new declaration on each line, instead. For example, instead
2077 (If they are global variables, each should have a comment preceding it
2080 When you have an @code{if}-@code{else} statement nested in another
2081 @code{if} statement, always put braces around the @code{if}-@code{else}.
2082 Thus, never write like this:
2105 If you have an @code{if} statement nested inside of an @code{else}
2106 statement, either write @code{else if} on one line, like this,
2116 with its @code{then}-part indented like the preceding @code{then}-part,
2117 or write the nested @code{if} within braces like this:
2129 Don't declare both a structure tag and variables or typedefs in the
2130 same declaration. Instead, declare the structure tag separately
2131 and then use it to declare the variables or typedefs.
2133 Try to avoid assignments inside @code{if}-conditions. For example,
2137 if ((foo = (char *) malloc (sizeof *foo)) == 0)
2138 fatal ("virtual memory exhausted");
2142 instead, write this:
2145 foo = (char *) malloc (sizeof *foo);
2147 fatal ("virtual memory exhausted");
2150 Don't make the program ugly to placate @code{lint}. Please don't insert any
2151 casts to @code{void}. Zero without a cast is perfectly fine as a null
2152 pointer constant, except when calling a varargs function.
2155 @section Naming Variables and Functions
2157 The names of global variables and functions in a program serve as
2158 comments of a sort. So don't choose terse names---instead, look for
2159 names that give useful information about the meaning of the variable or
2160 function. In a GNU program, names should be English, like other
2163 Local variable names can be shorter, because they are used only within
2164 one context, where (presumably) comments explain their purpose.
2166 Please use underscores to separate words in a name, so that the Emacs
2167 word commands can be useful within them. Stick to lower case; reserve
2168 upper case for macros and @code{enum} constants, and for name-prefixes
2169 that follow a uniform convention.
2171 For example, you should use names like @code{ignore_space_change_flag};
2172 don't use names like @code{iCantReadThis}.
2174 Variables that indicate whether command-line options have been
2175 specified should be named after the meaning of the option, not after
2176 the option-letter. A comment should state both the exact meaning of
2177 the option and its letter. For example,
2181 /* Ignore changes in horizontal whitespace (-b). */
2182 int ignore_space_change_flag;
2186 When you want to define names with constant integer values, use
2187 @code{enum} rather than @samp{#define}. GDB knows about enumeration
2190 Use file names of 14 characters or less, to avoid creating gratuitous
2191 problems on older System V systems. You can use the program
2192 @code{doschk} to test for this. @code{doschk} also tests for potential
2193 name conflicts if the files were loaded onto an MS-DOS file
2194 system---something you may or may not care about.
2196 @node System Portability
2197 @section Portability between System Types
2199 In the Unix world, ``portability'' refers to porting to different Unix
2200 versions. For a GNU program, this kind of portability is desirable, but
2203 The primary purpose of GNU software is to run on top of the GNU kernel,
2204 compiled with the GNU C compiler, on various types of @sc{cpu}. The
2205 amount and kinds of variation among GNU systems on different @sc{cpu}s
2206 will be comparable to the variation among Linux-based GNU systems or
2207 among BSD systems today. So the kinds of portability that are absolutely
2208 necessary are quite limited.
2210 But many users do run GNU software on non-GNU Unix or Unix-like systems.
2211 So supporting a variety of Unix-like systems is desirable, although not
2214 The easiest way to achieve portability to most Unix-like systems is to
2215 use Autoconf. It's unlikely that your program needs to know more
2216 information about the host platform than Autoconf can provide, simply
2217 because most of the programs that need such knowledge have already been
2220 Avoid using the format of semi-internal data bases (e.g., directories)
2221 when there is a higher-level alternative (@code{readdir}).
2223 As for systems that are not like Unix, such as MSDOS, Windows, the
2224 Macintosh, VMS, and MVS, supporting them is usually so much work that it
2225 is better if you don't.
2227 The planned GNU kernel is not finished yet, but you can tell which
2228 facilities it will provide by looking at the GNU C Library Manual. The
2229 GNU kernel is based on Mach, so the features of Mach will also be
2230 available. However, if you use Mach features, you'll probably have
2231 trouble debugging your program today.
2233 @node CPU Portability
2234 @section Portability between @sc{cpu}s
2236 Even GNU systems will differ because of differences among @sc{cpu}
2237 types---for example, difference in byte ordering and alignment
2238 requirements. It is absolutely essential to handle these differences.
2239 However, don't make any effort to cater to the possibility that an
2240 @code{int} will be less than 32 bits. We don't support 16-bit machines
2243 Don't assume that the address of an @code{int} object is also the
2244 address of its least-significant byte. This is false on big-endian
2245 machines. Thus, don't make the following mistake:
2250 while ((c = getchar()) != EOF)
2251 write(file_descriptor, &c, 1);
2254 When calling functions, you need not worry about the difference between
2255 pointers of various types, or between pointers and integers. On most
2256 machines, there's no difference anyway. As for the few machines where
2257 there is a difference, all of them support @sc{ansi} C, so you can use
2258 prototypes (conditionalized to be active only in @sc{ansi} C) to make
2259 the code work on those systems.
2261 In certain cases, it is ok to pass integer and pointer arguments
2262 indiscriminately to the same function, and use no prototype on any
2263 system. For example, many GNU programs have error-reporting functions
2264 that pass their arguments along to @code{printf} and friends:
2267 error (s, a1, a2, a3)
2271 fprintf (stderr, "error: ");
2272 fprintf (stderr, s, a1, a2, a3);
2277 In practice, this works on all machines, and it is much simpler than any
2278 ``correct'' alternative. Be sure @emph{not} to use a prototype
2281 However, avoid casting pointers to integers unless you really need to.
2282 These assumptions really reduce portability, and in most programs they
2283 are easy to avoid. In the cases where casting pointers to integers is
2284 essential---such as, a Lisp interpreter which stores type information as
2285 well as an address in one word---it is ok to do so, but you'll have to
2286 make explicit provisions to handle different word sizes.
2288 @node System Functions
2289 @section Calling System Functions
2291 C implementations differ substantially. @sc{ansi} C reduces but does not
2292 eliminate the incompatibilities; meanwhile, many users wish to compile
2293 GNU software with pre-@sc{ansi} compilers. This chapter gives
2294 recommendations for how to use the more or less standard C library
2295 functions to avoid unnecessary loss of portability.
2299 Don't use the value of @code{sprintf}. It returns the number of
2300 characters written on some systems, but not on all systems.
2303 @code{main} should be declared to return type @code{int}. It should
2304 terminate either by calling @code{exit} or by returning the integer
2305 status code; make sure it cannot ever return an undefined value.
2308 Don't declare system functions explicitly.
2310 Almost any declaration for a system function is wrong on some system.
2311 To minimize conflicts, leave it to the system header files to declare
2312 system functions. If the headers don't declare a function, let it
2315 While it may seem unclean to use a function without declaring it, in
2316 practice this works fine for most system library functions on the
2317 systems where this really happens; thus, the disadvantage is only
2318 theoretical. By contrast, actual declarations have frequently caused
2322 If you must declare a system function, don't specify the argument types.
2323 Use an old-style declaration, not an @sc{ansi} prototype. The more you
2324 specify about the function, the more likely a conflict.
2327 In particular, don't unconditionally declare @code{malloc} or
2330 Most GNU programs use those functions just once, in functions
2331 conventionally named @code{xmalloc} and @code{xrealloc}. These
2332 functions call @code{malloc} and @code{realloc}, respectively, and
2335 Because @code{xmalloc} and @code{xrealloc} are defined in your program,
2336 you can declare them in other files without any risk of type conflict.
2338 On most systems, @code{int} is the same length as a pointer; thus, the
2339 calls to @code{malloc} and @code{realloc} work fine. For the few
2340 exceptional systems (mostly 64-bit machines), you can use
2341 @strong{conditionalized} declarations of @code{malloc} and
2342 @code{realloc}---or put these declarations in configuration files
2343 specific to those systems.
2346 The string functions require special treatment. Some Unix systems have
2347 a header file @file{string.h}; others have @file{strings.h}. Neither
2348 file name is portable. There are two things you can do: use Autoconf to
2349 figure out which file to include, or don't include either file.
2352 If you don't include either strings file, you can't get declarations for
2353 the string functions from the header file in the usual way.
2355 That causes less of a problem than you might think. The newer @sc{ansi}
2356 string functions should be avoided anyway because many systems still
2357 don't support them. The string functions you can use are these:
2360 strcpy strncpy strcat strncat
2361 strlen strcmp strncmp
2365 The copy and concatenate functions work fine without a declaration as
2366 long as you don't use their values. Using their values without a
2367 declaration fails on systems where the width of a pointer differs from
2368 the width of @code{int}, and perhaps in other cases. It is trivial to
2369 avoid using their values, so do that.
2371 The compare functions and @code{strlen} work fine without a declaration
2372 on most systems, possibly all the ones that GNU software runs on.
2373 You may find it necessary to declare them @strong{conditionally} on a
2376 The search functions must be declared to return @code{char *}. Luckily,
2377 there is no variation in the data type they return. But there is
2378 variation in their names. Some systems give these functions the names
2379 @code{index} and @code{rindex}; other systems use the names
2380 @code{strchr} and @code{strrchr}. Some systems support both pairs of
2381 names, but neither pair works on all systems.
2383 You should pick a single pair of names and use it throughout your
2384 program. (Nowadays, it is better to choose @code{strchr} and
2385 @code{strrchr} for new programs, since those are the standard @sc{ansi}
2386 names.) Declare both of those names as functions returning @code{char
2387 *}. On systems which don't support those names, define them as macros
2388 in terms of the other pair. For example, here is what to put at the
2389 beginning of your file (or in a header) if you want to use the names
2390 @code{strchr} and @code{strrchr} throughout:
2394 #define strchr index
2396 #ifndef HAVE_STRRCHR
2397 #define strrchr rindex
2405 Here we assume that @code{HAVE_STRCHR} and @code{HAVE_STRRCHR} are
2406 macros defined in systems where the corresponding functions exist.
2407 One way to get them properly defined is to use Autoconf.
2409 @node Internationalization
2410 @section Internationalization
2412 GNU has a library called GNU gettext that makes it easy to translate the
2413 messages in a program into various languages. You should use this
2414 library in every program. Use English for the messages as they appear
2415 in the program, and let gettext provide the way to translate them into
2418 Using GNU gettext involves putting a call to the @code{gettext} macro
2419 around each string that might need translation---like this:
2422 printf (gettext ("Processing file `%s'..."));
2426 This permits GNU gettext to replace the string @code{"Processing file
2427 `%s'..."} with a translated version.
2429 Once a program uses gettext, please make a point of writing calls to
2430 @code{gettext} when you add new strings that call for translation.
2432 Using GNU gettext in a package involves specifying a @dfn{text domain
2433 name} for the package. The text domain name is used to separate the
2434 translations for this package from the translations for other packages.
2435 Normally, the text domain name should be the same as the name of the
2436 package---for example, @samp{fileutils} for the GNU file utilities.
2438 To enable gettext to work well, avoid writing code that makes
2439 assumptions about the structure of words or sentences. When you want
2440 the precise text of a sentence to vary depending on the data, use two or
2441 more alternative string constants each containing a complete sentences,
2442 rather than inserting conditionalized words or phrases into a single
2445 Here is an example of what not to do:
2448 printf ("%d file%s processed", nfiles,
2449 nfiles != 1 ? "s" : "");
2453 The problem with that example is that it assumes that plurals are made
2454 by adding `s'. If you apply gettext to the format string, like this,
2457 printf (gettext ("%d file%s processed"), nfiles,
2458 nfiles != 1 ? "s" : "");
2462 the message can use different words, but it will still be forced to use
2463 `s' for the plural. Here is a better way:
2466 printf ((nfiles != 1 ? "%d files processed"
2467 : "%d file processed"),
2472 This way, you can apply gettext to each of the two strings
2476 printf ((nfiles != 1 ? gettext ("%d files processed")
2477 : gettext ("%d file processed")),
2482 This can be any method of forming the plural of the word for ``file'', and
2483 also handles languages that require agreement in the word for
2486 A similar problem appears at the level of sentence structure with this
2490 printf ("# Implicit rule search has%s been done.\n",
2491 f->tried_implicit ? "" : " not");
2495 Adding @code{gettext} calls to this code cannot give correct results for
2496 all languages, because negation in some languages requires adding words
2497 at more than one place in the sentence. By contrast, adding
2498 @code{gettext} calls does the job straightfowardly if the code starts
2502 printf (f->tried_implicit
2503 ? "# Implicit rule search has been done.\n",
2504 : "# Implicit rule search has not been done.\n");
2510 Don't assume that @code{mmap} either works on all files or fails
2511 for all files. It may work on some files and fail on others.
2513 The proper way to use @code{mmap} is to try it on the specific file for
2514 which you want to use it---and if @code{mmap} doesn't work, fall back on
2515 doing the job in another way using @code{read} and @code{write}.
2517 The reason this precaution is needed is that the GNU kernel (the HURD)
2518 provides a user-extensible file system, in which there can be many
2519 different kinds of ``ordinary files.'' Many of them support
2520 @code{mmap}, but some do not. It is important to make programs handle
2521 all these kinds of files.
2524 @chapter Documenting Programs
2527 * GNU Manuals:: Writing proper manuals.
2528 * Manual Structure Details:: Specific structure conventions.
2529 * NEWS File:: NEWS files supplement manuals.
2530 * Change Logs:: Recording Changes
2531 * Man Pages:: Man pages are secondary.
2532 * Reading other Manuals:: How far you can go in learning
2537 @section GNU Manuals
2539 The preferred way to document part of the GNU system is to write a
2540 manual in the Texinfo formatting language. See the Texinfo manual,
2541 either the hardcopy, or the on-line version available through
2542 @code{info} or the Emacs Info subsystem (@kbd{C-h i}).
2544 Programmers often find it most natural to structure the documentation
2545 following the structure of the implementation, which they know. But
2546 this structure is not necessarily good for explaining how to use the
2547 program; it may be irrelevant and confusing for a user.
2549 At every level, from the sentences in a paragraph to the grouping of
2550 topics into separate manuals, the right way to structure documentation
2551 is according to the concepts and questions that a user will have in mind
2552 when reading it. Sometimes this structure of ideas matches the
2553 structure of the implementation of the software being documented---but
2554 often they are different. Often the most important part of learning to
2555 write good documentation is learning to notice when you are structuring
2556 the documentation like the implementation, and think about better
2559 For example, each program in the GNU system probably ought to be
2560 documented in one manual; but this does not mean each program should
2561 have its own manual. That would be following the structure of the
2562 implementation, rather than the structure that helps the user
2565 Instead, each manual should cover a coherent @emph{topic}. For example,
2566 instead of a manual for @code{diff} and a manual for @code{diff3}, we
2567 have one manual for ``comparison of files'' which covers both of those
2568 programs, as well as @code{cmp}. By documenting these programs
2569 together, we can make the whole subject clearer.
2571 The manual which discusses a program should document all of the
2572 program's command-line options and all of its commands. It should give
2573 examples of their use. But don't organize the manual as a list of
2574 features. Instead, organize it logically, by subtopics. Address the
2575 questions that a user will ask when thinking about the job that the
2578 In general, a GNU manual should serve both as tutorial and reference.
2579 It should be set up for convenient access to each topic through Info,
2580 and for reading straight through (appendixes aside). A GNU manual
2581 should give a good introduction to a beginner reading through from the
2582 start, and should also provide all the details that hackers want.
2584 That is not as hard as it first sounds. Arrange each chapter as a
2585 logical breakdown of its topic, but order the sections, and write their
2586 text, so that reading the chapter straight through makes sense. Do
2587 likewise when structuring the book into chapters, and when structuring a
2588 section into paragraphs. The watchword is, @emph{at each point, address
2589 the most fundamental and important issue raised by the preceding text.}
2591 If necessary, add extra chapters at the beginning of the manual which
2592 are purely tutorial and cover the basics of the subject. These provide
2593 the framework for a beginner to understand the rest of the manual. The
2594 Bison manual provides a good example of how to do this.
2596 Don't use Unix man pages as a model for how to write GNU documentation;
2597 most of them are terse, badly structured, and give inadequate
2598 explanation of the underlying concepts. (There are, of course
2599 exceptions.) Also Unix man pages use a particular format which is
2600 different from what we use in GNU manuals.
2602 Please do not use the term ``pathname'' that is used in Unix
2603 documentation; use ``file name'' (two words) instead. We use the term
2604 ``path'' only for search paths, which are lists of file names.
2606 Please do not use the term ``illegal'' to refer to erroneous input to a
2607 computer program. Please use ``invalid'' for this, and reserve the term
2608 ``illegal'' for violations of law.
2610 @node Manual Structure Details
2611 @section Manual Structure Details
2613 The title page of the manual should state the version of the programs or
2614 packages documented in the manual. The Top node of the manual should
2615 also contain this information. If the manual is changing more
2616 frequently than or independent of the program, also state a version
2617 number for the manual in both of these places.
2619 Each program documented in the manual should should have a node named
2620 @samp{@var{program} Invocation} or @samp{Invoking @var{program}}. This
2621 node (together with its subnodes, if any) should describe the program's
2622 command line arguments and how to run it (the sort of information people
2623 would look in a man page for). Start with an @samp{@@example}
2624 containing a template for all the options and arguments that the program
2627 Alternatively, put a menu item in some menu whose item name fits one of
2628 the above patterns. This identifies the node which that item points to
2629 as the node for this purpose, regardless of the node's actual name.
2631 There will be automatic features for specifying a program name and
2632 quickly reading just this part of its manual.
2634 If one manual describes several programs, it should have such a node for
2635 each program described.
2638 @section The NEWS File
2640 In addition to its manual, the package should have a file named
2641 @file{NEWS} which contains a list of user-visible changes worth
2642 mentioning. In each new release, add items to the front of the file and
2643 identify the version they pertain to. Don't discard old items; leave
2644 them in the file after the newer items. This way, a user upgrading from
2645 any previous version can see what is new.
2647 If the @file{NEWS} file gets very long, move some of the older items
2648 into a file named @file{ONEWS} and put a note at the end referring the
2652 @section Change Logs
2654 Keep a change log to describe all the changes made to program source
2655 files. The purpose of this is so that people investigating bugs in the
2656 future will know about the changes that might have introduced the bug.
2657 Often a new bug can be found by looking at what was recently changed.
2658 More importantly, change logs can help you eliminate conceptual
2659 inconsistencies between different parts of a program, by giving you a
2660 history of how the conflicting concepts arose and who they came from.
2663 * Change Log Concepts::
2664 * Style of Change Logs::
2666 * Conditional Changes::
2669 @node Change Log Concepts
2670 @subsection Change Log Concepts
2672 You can think of the change log as a conceptual ``undo list'' which
2673 explains how earlier versions were different from the current version.
2674 People can see the current version; they don't need the change log
2675 to tell them what is in it. What they want from a change log is a
2676 clear explanation of how the earlier version differed.
2678 The change log file is normally called @file{ChangeLog} and covers an
2679 entire directory. Each directory can have its own change log, or a
2680 directory can use the change log of its parent directory--it's up to
2683 Another alternative is to record change log information with a version
2684 control system such as RCS or CVS. This can be converted automatically
2685 to a @file{ChangeLog} file.
2687 There's no need to describe the full purpose of the changes or how they
2688 work together. If you think that a change calls for explanation, you're
2689 probably right. Please do explain it---but please put the explanation
2690 in comments in the code, where people will see it whenever they see the
2691 code. For example, ``New function'' is enough for the change log when
2692 you add a function, because there should be a comment before the
2693 function definition to explain what it does.
2695 However, sometimes it is useful to write one line to describe the
2696 overall purpose of a batch of changes.
2698 The easiest way to add an entry to @file{ChangeLog} is with the Emacs
2699 command @kbd{M-x add-change-log-entry}. An entry should have an
2700 asterisk, the name of the changed file, and then in parentheses the name
2701 of the changed functions, variables or whatever, followed by a colon.
2702 Then describe the changes you made to that function or variable.
2704 @node Style of Change Logs
2705 @subsection Style of Change Logs
2707 Here are some examples of change log entries:
2710 * register.el (insert-register): Return nil.
2711 (jump-to-register): Likewise.
2713 * sort.el (sort-subr): Return nil.
2715 * tex-mode.el (tex-bibtex-file, tex-file, tex-region):
2716 Restart the tex shell if process is gone or stopped.
2717 (tex-shell-running): New function.
2719 * expr.c (store_one_arg): Round size up for move_block_to_reg.
2720 (expand_call): Round up when emitting USE insns.
2721 * stmt.c (assign_parms): Round size up for move_block_from_reg.
2724 It's important to name the changed function or variable in full. Don't
2725 abbreviate function or variable names, and don't combine them.
2726 Subsequent maintainers will often search for a function name to find all
2727 the change log entries that pertain to it; if you abbreviate the name,
2728 they won't find it when they search.
2730 For example, some people are tempted to abbreviate groups of function
2731 names by writing @samp{* register.el (@{insert,jump-to@}-register)};
2732 this is not a good idea, since searching for @code{jump-to-register} or
2733 @code{insert-register} would not find that entry.
2735 Separate unrelated change log entries with blank lines. When two
2736 entries represent parts of the same change, so that they work together,
2737 then don't put blank lines between them. Then you can omit the file
2738 name and the asterisk when successive entries are in the same file.
2740 @node Simple Changes
2741 @subsection Simple Changes
2743 Certain simple kinds of changes don't need much detail in the change
2746 When you change the calling sequence of a function in a simple fashion,
2747 and you change all the callers of the function, there is no need to make
2748 individual entries for all the callers that you changed. Just write in
2749 the entry for the function being called, ``All callers changed.''
2752 * keyboard.c (Fcommand_execute): New arg SPECIAL.
2753 All callers changed.
2756 When you change just comments or doc strings, it is enough to write an
2757 entry for the file, without mentioning the functions. Just ``Doc
2758 fixes'' is enough for the change log.
2760 There's no need to make change log entries for documentation files.
2761 This is because documentation is not susceptible to bugs that are hard
2762 to fix. Documentation does not consist of parts that must interact in a
2763 precisely engineered fashion. To correct an error, you need not know
2764 the history of the erroneous passage; it is enough to compare what the
2765 documentation says with the way the program actually works.
2767 @node Conditional Changes
2768 @subsection Conditional Changes
2770 C programs often contain compile-time @code{#if} conditionals. Many
2771 changes are conditional; sometimes you add a new definition which is
2772 entirely contained in a conditional. It is very useful to indicate in
2773 the change log the conditions for which the change applies.
2775 Our convention for indicating conditional changes is to use square
2776 brackets around the name of the condition.
2778 Here is a simple example, describing a change which is conditional but
2779 does not have a function or entity name associated with it:
2782 * xterm.c [SOLARIS2]: Include string.h.
2785 Here is an entry describing a new definition which is entirely
2786 conditional. This new definition for the macro @code{FRAME_WINDOW_P} is
2787 used only when @code{HAVE_X_WINDOWS} is defined:
2790 * frame.h [HAVE_X_WINDOWS] (FRAME_WINDOW_P): Macro defined.
2793 Here is an entry for a change within the function @code{init_display},
2794 whose definition as a whole is unconditional, but the changes themselves
2795 are contained in a @samp{#ifdef HAVE_LIBNCURSES} conditional:
2798 * dispnew.c (init_display) [HAVE_LIBNCURSES]: If X, call tgetent.
2801 Here is an entry for a change that takes affect only when
2802 a certain macro is @emph{not} defined:
2805 (gethostname) [!HAVE_SOCKETS]: Replace with winsock version.
2811 In the GNU project, man pages are secondary. It is not necessary or
2812 expected for every GNU program to have a man page, but some of them do.
2813 It's your choice whether to include a man page in your program.
2815 When you make this decision, consider that supporting a man page
2816 requires continual effort each time the program is changed. The time
2817 you spend on the man page is time taken away from more useful work.
2819 For a simple program which changes little, updating the man page may be
2820 a small job. Then there is little reason not to include a man page, if
2823 For a large program that changes a great deal, updating a man page may
2824 be a substantial burden. If a user offers to donate a man page, you may
2825 find this gift costly to accept. It may be better to refuse the man
2826 page unless the same person agrees to take full responsibility for
2827 maintaining it---so that you can wash your hands of it entirely. If
2828 this volunteer later ceases to do the job, then don't feel obliged to
2829 pick it up yourself; it may be better to withdraw the man page from the
2830 distribution until someone else agrees to update it.
2832 When a program changes only a little, you may feel that the
2833 discrepancies are small enough that the man page remains useful without
2834 updating. If so, put a prominent note near the beginning of the man
2835 page explaining that you don't maintain it and that the Texinfo manual
2836 is more authoritative. The note should say how to access the Texinfo
2839 @node Reading other Manuals
2840 @section Reading other Manuals
2842 There may be non-free books or documentation files that describe the
2843 program you are documenting.
2845 It is ok to use these documents for reference, just as the author of a
2846 new algebra textbook can read other books on algebra. A large portion
2847 of any non-fiction book consists of facts, in this case facts about how
2848 a certain program works, and these facts are necessarily the same for
2849 everyone who writes about the subject. But be careful not to copy your
2850 outline structure, wording, tables or examples from preexisting non-free
2851 documentation. Copying from free documentation may be ok; please check
2852 with the FSF about the individual case.
2854 @node Managing Releases
2855 @chapter The Release Process
2857 Making a release is more than just bundling up your source files in a
2858 tar file and putting it up for FTP. You should set up your software so
2859 that it can be configured to run on a variety of systems. Your Makefile
2860 should conform to the GNU standards described below, and your directory
2861 layout should also conform to the standards discussed below. Doing so
2862 makes it easy to include your package into the larger framework of
2866 * Configuration:: How Configuration Should Work
2867 * Makefile Conventions:: Makefile Conventions
2868 * Releases:: Making Releases
2872 @section How Configuration Should Work
2874 Each GNU distribution should come with a shell script named
2875 @code{configure}. This script is given arguments which describe the
2876 kind of machine and system you want to compile the program for.
2878 The @code{configure} script must record the configuration options so
2879 that they affect compilation.
2881 One way to do this is to make a link from a standard name such as
2882 @file{config.h} to the proper configuration file for the chosen system.
2883 If you use this technique, the distribution should @emph{not} contain a
2884 file named @file{config.h}. This is so that people won't be able to
2885 build the program without configuring it first.
2887 Another thing that @code{configure} can do is to edit the Makefile. If
2888 you do this, the distribution should @emph{not} contain a file named
2889 @file{Makefile}. Instead, it should include a file @file{Makefile.in} which
2890 contains the input used for editing. Once again, this is so that people
2891 won't be able to build the program without configuring it first.
2893 If @code{configure} does write the @file{Makefile}, then @file{Makefile}
2894 should have a target named @file{Makefile} which causes @code{configure}
2895 to be rerun, setting up the same configuration that was set up last
2896 time. The files that @code{configure} reads should be listed as
2897 dependencies of @file{Makefile}.
2899 All the files which are output from the @code{configure} script should
2900 have comments at the beginning explaining that they were generated
2901 automatically using @code{configure}. This is so that users won't think
2902 of trying to edit them by hand.
2904 The @code{configure} script should write a file named @file{config.status}
2905 which describes which configuration options were specified when the
2906 program was last configured. This file should be a shell script which,
2907 if run, will recreate the same configuration.
2909 The @code{configure} script should accept an option of the form
2910 @samp{--srcdir=@var{dirname}} to specify the directory where sources are found
2911 (if it is not the current directory). This makes it possible to build
2912 the program in a separate directory, so that the actual source directory
2915 If the user does not specify @samp{--srcdir}, then @code{configure} should
2916 check both @file{.} and @file{..} to see if it can find the sources. If
2917 it finds the sources in one of these places, it should use them from
2918 there. Otherwise, it should report that it cannot find the sources, and
2919 should exit with nonzero status.
2921 Usually the easy way to support @samp{--srcdir} is by editing a
2922 definition of @code{VPATH} into the Makefile. Some rules may need to
2923 refer explicitly to the specified source directory. To make this
2924 possible, @code{configure} can add to the Makefile a variable named
2925 @code{srcdir} whose value is precisely the specified directory.
2927 The @code{configure} script should also take an argument which specifies the
2928 type of system to build the program for. This argument should look like
2932 @var{cpu}-@var{company}-@var{system}
2935 For example, a Sun 3 might be @samp{m68k-sun-sunos4.1}.
2937 The @code{configure} script needs to be able to decode all plausible
2938 alternatives for how to describe a machine. Thus, @samp{sun3-sunos4.1}
2939 would be a valid alias. For many programs, @samp{vax-dec-ultrix} would
2940 be an alias for @samp{vax-dec-bsd}, simply because the differences
2941 between Ultrix and @sc{BSD} are rarely noticeable, but a few programs
2942 might need to distinguish them.
2943 @c Real 4.4BSD now runs on some Suns.
2945 There is a shell script called @file{config.sub} that you can use
2946 as a subroutine to validate system types and canonicalize aliases.
2948 Other options are permitted to specify in more detail the software
2949 or hardware present on the machine, and include or exclude optional
2950 parts of the package:
2953 @item --enable-@var{feature}@r{[}=@var{parameter}@r{]}
2954 Configure the package to build and install an optional user-level
2955 facility called @var{feature}. This allows users to choose which
2956 optional features to include. Giving an optional @var{parameter} of
2957 @samp{no} should omit @var{feature}, if it is built by default.
2959 No @samp{--enable} option should @strong{ever} cause one feature to
2960 replace another. No @samp{--enable} option should ever substitute one
2961 useful behavior for another useful behavior. The only proper use for
2962 @samp{--enable} is for questions of whether to build part of the program
2965 @item --with-@var{package}
2966 @c @r{[}=@var{parameter}@r{]}
2967 The package @var{package} will be installed, so configure this package
2968 to work with @var{package}.
2970 @c Giving an optional @var{parameter} of
2971 @c @samp{no} should omit @var{package}, if it is used by default.
2973 Possible values of @var{package} include
2974 @samp{gnu-as} (or @samp{gas}), @samp{gnu-ld}, @samp{gnu-libc},
2980 Do not use a @samp{--with} option to specify the file name to use to
2981 find certain files. That is outside the scope of what @samp{--with}
2985 The target machine has no floating point processor.
2988 The target machine assembler is GAS, the GNU assembler.
2989 This is obsolete; users should use @samp{--with-gnu-as} instead.
2992 The target machine has the X Window System installed.
2993 This is obsolete; users should use @samp{--with-x} instead.
2996 All @code{configure} scripts should accept all of these ``detail''
2997 options, whether or not they make any difference to the particular
2998 package at hand. In particular, they should accept any option that
2999 starts with @samp{--with-} or @samp{--enable-}. This is so users will
3000 be able to configure an entire GNU source tree at once with a single set
3003 You will note that the categories @samp{--with-} and @samp{--enable-}
3004 are narrow: they @strong{do not} provide a place for any sort of option
3005 you might think of. That is deliberate. We want to limit the possible
3006 configuration options in GNU software. We do not want GNU programs to
3007 have idiosyncratic configuration options.
3009 Packages that perform part of the compilation process may support cross-compilation.
3010 In such a case, the host and target machines for the program may be
3011 different. The @code{configure} script should normally treat the
3012 specified type of system as both the host and the target, thus producing
3013 a program which works for the same type of machine that it runs on.
3015 The way to build a cross-compiler, cross-assembler, or what have you, is
3016 to specify the option @samp{--host=@var{hosttype}} when running
3017 @code{configure}. This specifies the host system without changing the
3018 type of target system. The syntax for @var{hosttype} is the same as
3021 Bootstrapping a cross-compiler requires compiling it on a machine other
3022 than the host it will run on. Compilation packages accept a
3023 configuration option @samp{--build=@var{hosttype}} for specifying the
3024 configuration on which you will compile them, in case that is different
3027 Programs for which cross-operation is not meaningful need not accept the
3028 @samp{--host} option, because configuring an entire operating system for
3029 cross-operation is not a meaningful thing.
3031 Some programs have ways of configuring themselves automatically. If
3032 your program is set up to do this, your @code{configure} script can simply
3033 ignore most of its arguments.
3035 @comment The makefile standards are in a separate file that is also
3036 @comment included by make.texinfo. Done by roland@gnu.ai.mit.edu on 1/6/93.
3037 @comment For this document, turn chapters into sections, etc.
3039 @include make-stds.texi
3043 @section Making Releases
3045 Package the distribution of @code{Foo version 69.96} up in a gzipped tar
3046 file with the name @file{foo-69.96.tar.gz}. It should unpack into a
3047 subdirectory named @file{foo-69.96}.
3049 Building and installing the program should never modify any of the files
3050 contained in the distribution. This means that all the files that form
3051 part of the program in any way must be classified into @dfn{source
3052 files} and @dfn{non-source files}. Source files are written by humans
3053 and never changed automatically; non-source files are produced from
3054 source files by programs under the control of the Makefile.
3056 Naturally, all the source files must be in the distribution. It is okay
3057 to include non-source files in the distribution, provided they are
3058 up-to-date and machine-independent, so that building the distribution
3059 normally will never modify them. We commonly include non-source files
3060 produced by Bison, @code{lex}, @TeX{}, and @code{makeinfo}; this helps avoid
3061 unnecessary dependencies between our distributions, so that users can
3062 install whichever packages they want to install.
3064 Non-source files that might actually be modified by building and
3065 installing the program should @strong{never} be included in the
3066 distribution. So if you do distribute non-source files, always make
3067 sure they are up to date when you make a new distribution.
3069 Make sure that the directory into which the distribution unpacks (as
3070 well as any subdirectories) are all world-writable (octal mode 777).
3071 This is so that old versions of @code{tar} which preserve the
3072 ownership and permissions of the files from the tar archive will be
3073 able to extract all the files even if the user is unprivileged.
3075 Make sure that all the files in the distribution are world-readable.
3077 Make sure that no file name in the distribution is more than 14
3078 characters long. Likewise, no file created by building the program
3079 should have a name longer than 14 characters. The reason for this is
3080 that some systems adhere to a foolish interpretation of the @sc{posix}
3081 standard, and refuse to open a longer name, rather than truncating as
3082 they did in the past.
3084 Don't include any symbolic links in the distribution itself. If the tar
3085 file contains symbolic links, then people cannot even unpack it on
3086 systems that don't support symbolic links. Also, don't use multiple
3087 names for one file in different directories, because certain file
3088 systems cannot handle this and that prevents unpacking the
3091 Try to make sure that all the file names will be unique on MS-DOS. A
3092 name on MS-DOS consists of up to 8 characters, optionally followed by a
3093 period and up to three characters. MS-DOS will truncate extra
3094 characters both before and after the period. Thus,
3095 @file{foobarhacker.c} and @file{foobarhacker.o} are not ambiguous; they
3096 are truncated to @file{foobarha.c} and @file{foobarha.o}, which are
3099 Include in your distribution a copy of the @file{texinfo.tex} you used
3100 to test print any @file{*.texinfo} or @file{*.texi} files.
3102 Likewise, if your program uses small GNU software packages like regex,
3103 getopt, obstack, or termcap, include them in the distribution file.
3104 Leaving them out would make the distribution file a little smaller at
3105 the expense of possible inconvenience to a user who doesn't know what
3112 update-date-leading-regexp: "@c This date is automagically updated when you save this file:\n@set lastupdate "
3113 update-date-trailing-regexp: ""
3114 eval: (load "/gd/gnuorg/update-date.el")
3115 eval: (add-hook 'write-file-hooks 'update-date)