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 April 7, 2012
9 @dircategory GNU organization
11 * Standards: (standards). GNU coding standards.
14 @c @setchapternewpage odd
15 @setchapternewpage off
17 @c Put everything in one index (arbitrarily chosen to be the concept index).
23 @c This is used by a cross ref in make-stds.texi
27 The GNU coding standards, last updated @value{lastupdate}.
29 Copyright @copyright{} 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
30 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010,
31 2011, 2012 Free Software Foundation, Inc.
33 Permission is granted to copy, distribute and/or modify this document
34 under the terms of the GNU Free Documentation License, Version 1.3 or
35 any later version published by the Free Software Foundation; with no
36 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
37 Texts. A copy of the license is included in the section entitled
38 ``GNU Free Documentation License''.
42 @title GNU Coding Standards
43 @author Richard Stallman, et al.
44 @author last updated @value{lastupdate}
46 @vskip 0pt plus 1filll
54 @top GNU Coding Standards
60 * Preface:: About the GNU Coding Standards.
61 * Legal Issues:: Keeping free software free.
62 * Design Advice:: General program design.
63 * Program Behavior:: Program behavior for all programs
64 * Writing C:: Making the best use of C.
65 * Documentation:: Documenting programs.
66 * Managing Releases:: The release process.
67 * References:: Mentioning non-free software or documentation.
68 * GNU Free Documentation License:: Copying and sharing this manual.
74 @chapter About the GNU Coding Standards
76 The GNU Coding Standards were written by Richard Stallman and other GNU
77 Project volunteers. Their purpose is to make the GNU system clean,
78 consistent, and easy to install. This document can also be read as a
79 guide to writing portable, robust and reliable programs. It focuses on
80 programs written in C, but many of the rules and principles are useful
81 even if you write in another programming language. The rules often
82 state reasons for writing in a certain way.
84 @cindex where to obtain @code{standards.texi}
85 @cindex downloading this manual
86 If you did not obtain this file directly from the GNU project and
87 recently, please check for a newer version. You can get the GNU
88 Coding Standards from the GNU web server in many
89 different formats, including the Texinfo source, PDF, HTML, DVI, plain
90 text, and more, at: @uref{http://www.gnu.org/prep/standards/}.
92 If you are maintaining an official GNU package, in addition to this
93 document, please read and follow the GNU maintainer information
94 (@pxref{Top, , Contents, maintain, Information for Maintainers of GNU
97 @cindex @code{gnustandards-commit@@gnu.org} mailing list
98 If you want to receive diffs for every change to these GNU documents,
99 join the mailing list @code{gnustandards-commit@@gnu.org}, via the web
101 @url{http://lists.gnu.org/mailman/listinfo/gnustandards-commit}.
102 Archives are also available there.
104 @cindex @code{bug-standards@@gnu.org} email address
105 @cindex Savannah repository for gnustandards
106 @cindex gnustandards project repository
107 Please send corrections or suggestions for this document to
108 @email{bug-standards@@gnu.org}. If you make a suggestion, please
109 include a suggested new wording for it, to help us consider the
110 suggestion efficiently. We prefer a context diff to the Texinfo
111 source, but if that's difficult for you, you can make a context diff
112 for some other version of this document, or propose it in any way that
113 makes it clear. The source repository for this document can be found
114 at @url{http://savannah.gnu.org/projects/gnustandards}.
116 These standards cover the minimum of what is important when writing a
117 GNU package. Likely, the need for additional standards will come up.
118 Sometimes, you might suggest that such standards be added to this
119 document. If you think your standards would be generally useful, please
122 You should also set standards for your package on many questions not
123 addressed or not firmly specified here. The most important point is to
124 be self-consistent---try to stick to the conventions you pick, and try
125 to document them as much as possible. That way, your program will be
126 more maintainable by others.
128 The GNU Hello program serves as an example of how to follow the GNU
129 coding standards for a trivial program.
130 @uref{http://www.gnu.org/software/hello/hello.html}.
132 This release of the GNU Coding Standards was last updated
137 @chapter Keeping Free Software Free
138 @cindex legal aspects
140 This chapter discusses how you can make sure that GNU software
141 avoids legal difficulties, and other related issues.
144 * Reading Non-Free Code:: Referring to proprietary programs.
145 * Contributions:: Accepting contributions.
146 * Trademarks:: How we deal with trademark issues.
149 @node Reading Non-Free Code
150 @section Referring to Proprietary Programs
151 @cindex proprietary programs
152 @cindex avoiding proprietary code
154 Don't in any circumstances refer to Unix source code for or during
155 your work on GNU! (Or to any other proprietary programs.)
157 If you have a vague recollection of the internals of a Unix program,
158 this does not absolutely mean you can't write an imitation of it, but
159 do try to organize the imitation internally along different lines,
160 because this is likely to make the details of the Unix version
161 irrelevant and dissimilar to your results.
163 For example, Unix utilities were generally optimized to minimize
164 memory use; if you go for speed instead, your program will be very
165 different. You could keep the entire input file in memory and scan it
166 there instead of using stdio. Use a smarter algorithm discovered more
167 recently than the Unix program. Eliminate use of temporary files. Do
168 it in one pass instead of two (we did this in the assembler).
170 Or, on the contrary, emphasize simplicity instead of speed. For some
171 applications, the speed of today's computers makes simpler algorithms
174 Or go for generality. For example, Unix programs often have static
175 tables or fixed-size strings, which make for arbitrary limits; use
176 dynamic allocation instead. Make sure your program handles NULs and
177 other funny characters in the input files. Add a programming language
178 for extensibility and write part of the program in that language.
180 Or turn some parts of the program into independently usable libraries.
181 Or use a simple garbage collector instead of tracking precisely when
182 to free memory, or use a new GNU facility such as obstacks.
186 @section Accepting Contributions
188 @cindex accepting contributions
190 If the program you are working on is copyrighted by the Free Software
191 Foundation, then when someone else sends you a piece of code to add to
192 the program, we need legal papers to use it---just as we asked you to
193 sign papers initially. @emph{Each} person who makes a nontrivial
194 contribution to a program must sign some sort of legal papers in order
195 for us to have clear title to the program; the main author alone is not
198 So, before adding in any contributions from other people, please tell
199 us, so we can arrange to get the papers. Then wait until we tell you
200 that we have received the signed papers, before you actually use the
203 This applies both before you release the program and afterward. If
204 you receive diffs to fix a bug, and they make significant changes, we
205 need legal papers for that change.
207 This also applies to comments and documentation files. For copyright
208 law, comments and code are just text. Copyright applies to all kinds of
209 text, so we need legal papers for all kinds.
211 We know it is frustrating to ask for legal papers; it's frustrating for
212 us as well. But if you don't wait, you are going out on a limb---for
213 example, what if the contributor's employer won't sign a disclaimer?
214 You might have to take that code out again!
216 You don't need papers for changes of a few lines here or there, since
217 they are not significant for copyright purposes. Also, you don't need
218 papers if all you get from the suggestion is some ideas, not actual code
219 which you use. For example, if someone sent you one implementation, but
220 you write a different implementation of the same idea, you don't need to
223 The very worst thing is if you forget to tell us about the other
224 contributor. We could be very embarrassed in court some day as a
227 We have more detailed advice for maintainers of GNU packages. If you
228 have reached the stage of maintaining a GNU program (whether released
229 or not), please take a look: @pxref{Legal Matters,,, maintain,
230 Information for GNU Maintainers}.
237 Please do not include any trademark acknowledgements in GNU software
238 packages or documentation.
240 Trademark acknowledgements are the statements that such-and-such is a
241 trademark of so-and-so. The GNU Project has no objection to the basic
242 idea of trademarks, but these acknowledgements feel like kowtowing,
243 and there is no legal requirement for them, so we don't use them.
245 What is legally required, as regards other people's trademarks, is to
246 avoid using them in ways which a reader might reasonably understand as
247 naming or labeling our own programs or activities. For example, since
248 ``Objective C'' is (or at least was) a trademark, we made sure to say
249 that we provide a ``compiler for the Objective C language'' rather
250 than an ``Objective C compiler''. The latter would have been meant as
251 a shorter way of saying the former, but it does not explicitly state
252 the relationship, so it could be misinterpreted as using ``Objective
253 C'' as a label for the compiler rather than for the language.
255 Please don't use ``win'' as an abbreviation for Microsoft Windows in
256 GNU software or documentation. In hacker terminology, calling
257 something a ``win'' is a form of praise. If you wish to praise
258 Microsoft Windows when speaking on your own, by all means do so, but
259 not in GNU software. Usually we write the name ``Windows'' in full,
260 but when brevity is very important (as in file names and sometimes
261 symbol names), we abbreviate it to ``w''. For instance, the files and
262 functions in Emacs that deal with Windows start with @samp{w32}.
265 @chapter General Program Design
266 @cindex program design
268 This chapter discusses some of the issues you should take into
269 account when designing your program.
271 @c Standard or ANSI C
273 @c In 1989 the American National Standards Institute (ANSI) standardized
274 @c C as standard X3.159-1989. In December of that year the
275 @c International Standards Organization ISO adopted the ANSI C standard
276 @c making minor changes. In 1990 ANSI then re-adopted ISO standard
277 @c C. This version of C is known as either ANSI C or Standard C.
279 @c A major revision of the C Standard appeared in 1999.
282 * Source Language:: Which languages to use.
283 * Compatibility:: Compatibility with other implementations.
284 * Using Extensions:: Using non-standard features.
285 * Standard C:: Using standard C features.
286 * Conditional Compilation:: Compiling code only if a conditional is true.
289 @node Source Language
290 @section Which Languages to Use
291 @cindex programming languages
293 When you want to use a language that gets compiled and runs at high
294 speed, the best language to use is C. Using another language is like
295 using a non-standard feature: it will cause trouble for users. Even if
296 GCC supports the other language, users may find it inconvenient to have
297 to install the compiler for that other language in order to build your
298 program. For example, if you write your program in C++, people will
299 have to install the GNU C++ compiler in order to compile your program.
301 C has one other advantage over C++ and other compiled languages: more
302 people know C, so more people will find it easy to read and modify the
303 program if it is written in C.
305 So in general it is much better to use C, rather than the
306 comparable alternatives.
308 But there are two exceptions to that conclusion:
312 It is no problem to use another language to write a tool specifically
313 intended for use with that language. That is because the only people
314 who want to build the tool will be those who have installed the other
318 If an application is of interest only to a narrow part of the community,
319 then the question of which language it is written in has less effect on
320 other people, so you may as well please yourself.
323 Many programs are designed to be extensible: they include an interpreter
324 for a language that is higher level than C. Often much of the program
325 is written in that language, too. The Emacs editor pioneered this
329 @cindex GNOME and Guile
330 The standard extensibility interpreter for GNU software is Guile
331 (@uref{http://www.gnu.org/@/software/@/guile/}), which implements the
332 language Scheme (an especially clean and simple dialect of Lisp).
333 Guile also includes bindings for GTK+/GNOME, making it practical to
334 write modern GUI functionality within Guile. We don't reject programs
335 written in other ``scripting languages'' such as Perl and Python, but
336 using Guile is very important for the overall consistency of the GNU
341 @section Compatibility with Other Implementations
342 @cindex compatibility with C and @sc{posix} standards
343 @cindex @sc{posix} compatibility
345 With occasional exceptions, utility programs and libraries for GNU
346 should be upward compatible with those in Berkeley Unix, and upward
347 compatible with Standard C if Standard C specifies their
348 behavior, and upward compatible with @sc{posix} if @sc{posix} specifies
351 When these standards conflict, it is useful to offer compatibility
352 modes for each of them.
354 @cindex options for compatibility
355 Standard C and @sc{posix} prohibit many kinds of extensions. Feel
356 free to make the extensions anyway, and include a @samp{--ansi},
357 @samp{--posix}, or @samp{--compatible} option to turn them off.
358 However, if the extension has a significant chance of breaking any real
359 programs or scripts, then it is not really upward compatible. So you
360 should try to redesign its interface to make it upward compatible.
362 @cindex @code{POSIXLY_CORRECT}, environment variable
363 Many GNU programs suppress extensions that conflict with @sc{posix} if the
364 environment variable @code{POSIXLY_CORRECT} is defined (even if it is
365 defined with a null value). Please make your program recognize this
366 variable if appropriate.
368 When a feature is used only by users (not by programs or command
369 files), and it is done poorly in Unix, feel free to replace it
370 completely with something totally different and better. (For example,
371 @code{vi} is replaced with Emacs.) But it is nice to offer a compatible
372 feature as well. (There is a free @code{vi} clone, so we offer it.)
374 Additional useful features are welcome regardless of whether
375 there is any precedent for them.
377 @node Using Extensions
378 @section Using Non-standard Features
379 @cindex non-standard extensions
381 Many GNU facilities that already exist support a number of convenient
382 extensions over the comparable Unix facilities. Whether to use these
383 extensions in implementing your program is a difficult question.
385 On the one hand, using the extensions can make a cleaner program.
386 On the other hand, people will not be able to build the program
387 unless the other GNU tools are available. This might cause the
388 program to work on fewer kinds of machines.
390 With some extensions, it might be easy to provide both alternatives.
391 For example, you can define functions with a ``keyword'' @code{INLINE}
392 and define that as a macro to expand into either @code{inline} or
393 nothing, depending on the compiler.
395 In general, perhaps it is best not to use the extensions if you can
396 straightforwardly do without them, but to use the extensions if they
397 are a big improvement.
399 An exception to this rule are the large, established programs (such as
400 Emacs) which run on a great variety of systems. Using GNU extensions in
401 such programs would make many users unhappy, so we don't do that.
403 Another exception is for programs that are used as part of compilation:
404 anything that must be compiled with other compilers in order to
405 bootstrap the GNU compilation facilities. If these require the GNU
406 compiler, then no one can compile them without having them installed
407 already. That would be extremely troublesome in certain cases.
410 @section Standard C and Pre-Standard C
411 @cindex @sc{ansi} C standard
413 1989 Standard C is widespread enough now that it is ok to use its
414 features in new programs. There is one exception: do not ever use the
415 ``trigraph'' feature of Standard C.
417 1999 Standard C is not widespread yet, so please do not require its
418 features in programs. It is ok to use its features if they are present.
420 However, it is easy to support pre-standard compilers in most programs,
421 so if you know how to do that, feel free. If a program you are
422 maintaining has such support, you should try to keep it working.
424 @cindex function prototypes
425 To support pre-standard C, instead of writing function definitions in
426 standard prototype form,
435 write the definition in pre-standard style like this,
445 and use a separate declaration to specify the argument prototype:
451 You need such a declaration anyway, in a header file, to get the benefit
452 of prototypes in all the files where the function is called. And once
453 you have the declaration, you normally lose nothing by writing the
454 function definition in the pre-standard style.
456 This technique does not work for integer types narrower than @code{int}.
457 If you think of an argument as being of a type narrower than @code{int},
458 declare it as @code{int} instead.
460 There are a few special cases where this technique is hard to use. For
461 example, if a function argument needs to hold the system type
462 @code{dev_t}, you run into trouble, because @code{dev_t} is shorter than
463 @code{int} on some machines; but you cannot use @code{int} instead,
464 because @code{dev_t} is wider than @code{int} on some machines. There
465 is no type you can safely use on all machines in a non-standard
466 definition. The only way to support non-standard C and pass such an
467 argument is to check the width of @code{dev_t} using Autoconf and choose
468 the argument type accordingly. This may not be worth the trouble.
470 In order to support pre-standard compilers that do not recognize
471 prototypes, you may want to use a preprocessor macro like this:
474 /* Declare the prototype for a general external function. */
475 #if defined (__STDC__) || defined (WINDOWSNT)
476 #define P_(proto) proto
482 @node Conditional Compilation
483 @section Conditional Compilation
485 When supporting configuration options already known when building your
486 program we prefer using @code{if (... )} over conditional compilation,
487 as in the former case the compiler is able to perform more extensive
488 checking of all possible code paths.
490 For example, please write
510 A modern compiler such as GCC will generate exactly the same code in
511 both cases, and we have been using similar techniques with good success
512 in several projects. Of course, the former method assumes that
513 @code{HAS_FOO} is defined as either 0 or 1.
515 While this is not a silver bullet solving all portability problems,
516 and is not always appropriate, following this policy would have saved
517 GCC developers many hours, or even days, per year.
519 In the case of function-like macros like @code{REVERSIBLE_CC_MODE} in
520 GCC which cannot be simply used in @code{if (...)} statements, there is
521 an easy workaround. Simply introduce another macro
522 @code{HAS_REVERSIBLE_CC_MODE} as in the following example:
525 #ifdef REVERSIBLE_CC_MODE
526 #define HAS_REVERSIBLE_CC_MODE 1
528 #define HAS_REVERSIBLE_CC_MODE 0
532 @node Program Behavior
533 @chapter Program Behavior for All Programs
535 This chapter describes conventions for writing robust
536 software. It also describes general standards for error messages, the
537 command line interface, and how libraries should behave.
540 * Non-GNU Standards:: We consider standards such as POSIX;
541 we don't "obey" them.
542 * Semantics:: Writing robust programs.
543 * Libraries:: Library behavior.
544 * Errors:: Formatting error messages.
545 * User Interfaces:: Standards about interfaces generally.
546 * Graphical Interfaces:: Standards for graphical interfaces.
547 * Command-Line Interfaces:: Standards for command line interfaces.
548 * Dynamic Plug-In Interfaces:: Standards for dynamic plug-in interfaces.
549 * Option Table:: Table of long options.
550 * OID Allocations:: Table of OID slots for GNU.
551 * Memory Usage:: When and how to care about memory needs.
552 * File Usage:: Which files to use, and where.
555 @node Non-GNU Standards
556 @section Non-GNU Standards
558 The GNU Project regards standards published by other organizations as
559 suggestions, not orders. We consider those standards, but we do not
560 ``obey'' them. In developing a GNU program, you should implement
561 an outside standard's specifications when that makes the GNU system
562 better overall in an objective sense. When it doesn't, you shouldn't.
564 In most cases, following published standards is convenient for
565 users---it means that their programs or scripts will work more
566 portably. For instance, GCC implements nearly all the features of
567 Standard C as specified by that standard. C program developers would
568 be unhappy if it did not. And GNU utilities mostly follow
569 specifications of POSIX.2; shell script writers and users would be
570 unhappy if our programs were incompatible.
572 But we do not follow either of these specifications rigidly, and there
573 are specific points on which we decided not to follow them, so as to
574 make the GNU system better for users.
576 For instance, Standard C says that nearly all extensions to C are
577 prohibited. How silly! GCC implements many extensions, some of which
578 were later adopted as part of the standard. If you want these
579 constructs to give an error message as ``required'' by the standard,
580 you must specify @samp{--pedantic}, which was implemented only so that
581 we can say ``GCC is a 100% implementation of the standard'', not
582 because there is any reason to actually use it.
584 POSIX.2 specifies that @samp{df} and @samp{du} must output sizes by
585 default in units of 512 bytes. What users want is units of 1k, so
586 that is what we do by default. If you want the ridiculous behavior
587 ``required'' by POSIX, you must set the environment variable
588 @samp{POSIXLY_CORRECT} (which was originally going to be named
589 @samp{POSIX_ME_HARDER}).
591 GNU utilities also depart from the letter of the POSIX.2 specification
592 when they support long-named command-line options, and intermixing
593 options with ordinary arguments. This minor incompatibility with
594 POSIX is never a problem in practice, and it is very useful.
596 In particular, don't reject a new feature, or remove an old one,
597 merely because a standard says it is ``forbidden'' or ``deprecated''.
601 @section Writing Robust Programs
603 @cindex arbitrary limits on data
604 Avoid arbitrary limits on the length or number of @emph{any} data
605 structure, including file names, lines, files, and symbols, by allocating
606 all data structures dynamically. In most Unix utilities, ``long lines
607 are silently truncated''. This is not acceptable in a GNU utility.
609 @cindex @code{NUL} characters
611 Utilities reading files should not drop NUL characters, or any other
612 nonprinting characters @emph{including those with codes above 0177}.
613 The only sensible exceptions would be utilities specifically intended
614 for interface to certain types of terminals or printers that can't
615 handle those characters. Whenever possible, try to make programs work
616 properly with sequences of bytes that represent multibyte characters;
617 UTF-8 is the most important.
619 @cindex error messages
620 Check every system call for an error return, unless you know you wish
621 to ignore errors. Include the system error text (from @code{perror},
622 @code{strerror}, or equivalent) in @emph{every} error message
623 resulting from a failing system call, as well as the name of the file
624 if any and the name of the utility. Just ``cannot open foo.c'' or
625 ``stat failed'' is not sufficient.
627 @cindex @code{malloc} return value
628 @cindex memory allocation failure
629 Check every call to @code{malloc} or @code{realloc} to see if it
630 returned zero. Check @code{realloc} even if you are making the block
631 smaller; in a system that rounds block sizes to a power of 2,
632 @code{realloc} may get a different block if you ask for less space.
634 In Unix, @code{realloc} can destroy the storage block if it returns
635 zero. GNU @code{realloc} does not have this bug: if it fails, the
636 original block is unchanged. Feel free to assume the bug is fixed. If
637 you wish to run your program on Unix, and wish to avoid lossage in this
638 case, you can use the GNU @code{malloc}.
640 You must expect @code{free} to alter the contents of the block that was
641 freed. Anything you want to fetch from the block, you must fetch before
644 If @code{malloc} fails in a noninteractive program, make that a fatal
645 error. In an interactive program (one that reads commands from the
646 user), it is better to abort the command and return to the command
647 reader loop. This allows the user to kill other processes to free up
648 virtual memory, and then try the command again.
650 @cindex command-line arguments, decoding
651 Use @code{getopt_long} to decode arguments, unless the argument syntax
652 makes this unreasonable.
654 When static storage is to be written in during program execution, use
655 explicit C code to initialize it. Reserve C initialized declarations
656 for data that will not be changed.
659 Try to avoid low-level interfaces to obscure Unix data structures (such
660 as file directories, utmp, or the layout of kernel memory), since these
661 are less likely to work compatibly. If you need to find all the files
662 in a directory, use @code{readdir} or some other high-level interface.
663 These are supported compatibly by GNU.
665 @cindex signal handling
666 The preferred signal handling facilities are the BSD variant of
667 @code{signal}, and the @sc{posix} @code{sigaction} function; the
668 alternative USG @code{signal} interface is an inferior design.
670 Nowadays, using the @sc{posix} signal functions may be the easiest way
671 to make a program portable. If you use @code{signal}, then on GNU/Linux
672 systems running GNU libc version 1, you should include
673 @file{bsd/signal.h} instead of @file{signal.h}, so as to get BSD
674 behavior. It is up to you whether to support systems where
675 @code{signal} has only the USG behavior, or give up on them.
677 @cindex impossible conditions
678 In error checks that detect ``impossible'' conditions, just abort.
679 There is usually no point in printing any message. These checks
680 indicate the existence of bugs. Whoever wants to fix the bugs will have
681 to read the source code and run a debugger. So explain the problem with
682 comments in the source. The relevant data will be in variables, which
683 are easy to examine with the debugger, so there is no point moving them
686 Do not use a count of errors as the exit status for a program.
687 @emph{That does not work}, because exit status values are limited to 8
688 bits (0 through 255). A single run of the program might have 256
689 errors; if you try to return 256 as the exit status, the parent process
690 will see 0 as the status, and it will appear that the program succeeded.
692 @cindex temporary files
693 @cindex @code{TMPDIR} environment variable
694 If you make temporary files, check the @code{TMPDIR} environment
695 variable; if that variable is defined, use the specified directory
696 instead of @file{/tmp}.
698 In addition, be aware that there is a possible security problem when
699 creating temporary files in world-writable directories. In C, you can
700 avoid this problem by creating temporary files in this manner:
703 fd = open (filename, O_WRONLY | O_CREAT | O_EXCL, 0600);
707 or by using the @code{mkstemps} function from Gnulib
708 (@pxref{mkstemps,,, gnulib, Gnulib}).
710 In bash, use @code{set -C} (long name @code{noclobber}) to avoid this
711 problem. In addition, the @code{mktemp} utility is a more general
712 solution for creating temporary files from shell scripts
713 (@pxref{mktemp invocation,,, coreutils, GNU Coreutils}).
717 @section Library Behavior
720 Try to make library functions reentrant. If they need to do dynamic
721 storage allocation, at least try to avoid any nonreentrancy aside from
722 that of @code{malloc} itself.
724 Here are certain name conventions for libraries, to avoid name
727 Choose a name prefix for the library, more than two characters long.
728 All external function and variable names should start with this
729 prefix. In addition, there should only be one of these in any given
730 library member. This usually means putting each one in a separate
733 An exception can be made when two external symbols are always used
734 together, so that no reasonable program could use one without the
735 other; then they can both go in the same file.
737 External symbols that are not documented entry points for the user
738 should have names beginning with @samp{_}. The @samp{_} should be
739 followed by the chosen name prefix for the library, to prevent
740 collisions with other libraries. These can go in the same files with
741 user entry points if you like.
743 Static functions and variables can be used as you like and need not
744 fit any naming convention.
747 @section Formatting Error Messages
748 @cindex formatting error messages
749 @cindex error messages, formatting
751 Error messages from compilers should look like this:
754 @var{sourcefile}:@var{lineno}: @var{message}
758 If you want to mention the column number, use one of these formats:
761 @var{sourcefile}:@var{lineno}:@var{column}: @var{message}
762 @var{sourcefile}:@var{lineno}.@var{column}: @var{message}
767 Line numbers should start from 1 at the beginning of the file, and
768 column numbers should start from 1 at the beginning of the line.
769 (Both of these conventions are chosen for compatibility.) Calculate
770 column numbers assuming that space and all ASCII printing characters
771 have equal width, and assuming tab stops every 8 columns. For
772 non-ASCII characters, Unicode character widths should be used when in
773 a UTF-8 locale; GNU libc and GNU gnulib provide suitable
774 @code{wcwidth} functions.
776 The error message can also give both the starting and ending positions
777 of the erroneous text. There are several formats so that you can
778 avoid redundant information such as a duplicate line number.
779 Here are the possible formats:
782 @var{sourcefile}:@var{line1}.@var{column1}-@var{line2}.@var{column2}: @var{message}
783 @var{sourcefile}:@var{line1}.@var{column1}-@var{column2}: @var{message}
784 @var{sourcefile}:@var{line1}-@var{line2}: @var{message}
788 When an error is spread over several files, you can use this format:
791 @var{file1}:@var{line1}.@var{column1}-@var{file2}:@var{line2}.@var{column2}: @var{message}
794 Error messages from other noninteractive programs should look like this:
797 @var{program}:@var{sourcefile}:@var{lineno}: @var{message}
801 when there is an appropriate source file, or like this:
804 @var{program}: @var{message}
808 when there is no relevant source file.
810 If you want to mention the column number, use this format:
813 @var{program}:@var{sourcefile}:@var{lineno}:@var{column}: @var{message}
816 In an interactive program (one that is reading commands from a
817 terminal), it is better not to include the program name in an error
818 message. The place to indicate which program is running is in the
819 prompt or with the screen layout. (When the same program runs with
820 input from a source other than a terminal, it is not interactive and
821 would do best to print error messages using the noninteractive style.)
823 The string @var{message} should not begin with a capital letter when
824 it follows a program name and/or file name, because that isn't the
825 beginning of a sentence. (The sentence conceptually starts at the
826 beginning of the line.) Also, it should not end with a period.
828 Error messages from interactive programs, and other messages such as
829 usage messages, should start with a capital letter. But they should not
832 @node User Interfaces
833 @section Standards for Interfaces Generally
835 @cindex program name and its behavior
836 @cindex behavior, dependent on program's name
837 Please don't make the behavior of a utility depend on the name used
838 to invoke it. It is useful sometimes to make a link to a utility
839 with a different name, and that should not change what it does.
841 Instead, use a run time option or a compilation switch or both
842 to select among the alternate behaviors.
844 @cindex output device and program's behavior
845 Likewise, please don't make the behavior of the program depend on the
846 type of output device it is used with. Device independence is an
847 important principle of the system's design; do not compromise it merely
848 to save someone from typing an option now and then. (Variation in error
849 message syntax when using a terminal is ok, because that is a side issue
850 that people do not depend on.)
852 If you think one behavior is most useful when the output is to a
853 terminal, and another is most useful when the output is a file or a
854 pipe, then it is usually best to make the default behavior the one that
855 is useful with output to a terminal, and have an option for the other
858 Compatibility requires certain programs to depend on the type of output
859 device. It would be disastrous if @code{ls} or @code{sh} did not do so
860 in the way all users expect. In some of these cases, we supplement the
861 program with a preferred alternate version that does not depend on the
862 output device type. For example, we provide a @code{dir} program much
863 like @code{ls} except that its default output format is always
867 @node Graphical Interfaces
868 @section Standards for Graphical Interfaces
869 @cindex graphical user interface
870 @cindex interface styles
871 @cindex user interface styles
874 When you write a program that provides a graphical user interface,
875 please make it work with the X Window System and the GTK+ toolkit
876 unless the functionality specifically requires some alternative (for
877 example, ``displaying jpeg images while in console mode'').
879 In addition, please provide a command-line interface to control the
880 functionality. (In many cases, the graphical user interface can be a
881 separate program which invokes the command-line program.) This is
882 so that the same jobs can be done from scripts.
887 @cindex keyboard interface
888 @cindex library interface
889 Please also consider providing a D-bus interface for use from other
890 running programs, such as within GNOME. (GNOME used to use CORBA
891 for this, but that is being phased out.) In addition, consider
892 providing a library interface (for use from C), and perhaps a
893 keyboard-driven console interface (for use by users from console
894 mode). Once you are doing the work to provide the functionality and
895 the graphical interface, these won't be much extra work.
898 @node Command-Line Interfaces
899 @section Standards for Command Line Interfaces
900 @cindex command-line interface
903 It is a good idea to follow the @sc{posix} guidelines for the
904 command-line options of a program. The easiest way to do this is to use
905 @code{getopt} to parse them. Note that the GNU version of @code{getopt}
906 will normally permit options anywhere among the arguments unless the
907 special argument @samp{--} is used. This is not what @sc{posix}
908 specifies; it is a GNU extension.
910 @cindex long-named options
911 Please define long-named options that are equivalent to the
912 single-letter Unix-style options. We hope to make GNU more user
913 friendly this way. This is easy to do with the GNU function
916 One of the advantages of long-named options is that they can be
917 consistent from program to program. For example, users should be able
918 to expect the ``verbose'' option of any GNU program which has one, to be
919 spelled precisely @samp{--verbose}. To achieve this uniformity, look at
920 the table of common long-option names when you choose the option names
921 for your program (@pxref{Option Table}).
923 It is usually a good idea for file names given as ordinary arguments to
924 be input files only; any output files would be specified using options
925 (preferably @samp{-o} or @samp{--output}). Even if you allow an output
926 file name as an ordinary argument for compatibility, try to provide an
927 option as another way to specify it. This will lead to more consistency
928 among GNU utilities, and fewer idiosyncrasies for users to remember.
930 @cindex standard command-line options
931 @cindex options, standard command-line
932 @cindex CGI programs, standard options for
933 @cindex PATH_INFO, specifying standard options as
934 All programs should support two standard options: @samp{--version}
935 and @samp{--help}. CGI programs should accept these as command-line
936 options, and also if given as the @env{PATH_INFO}; for instance,
937 visiting @url{http://example.org/p.cgi/--help} in a browser should
938 output the same information as invoking @samp{p.cgi --help} from the
942 * --version:: The standard output for --version.
943 * --help:: The standard output for --help.
947 @subsection @option{--version}
949 @cindex @samp{--version} output
951 The standard @code{--version} option should direct the program to
952 print information about its name, version, origin and legal status,
953 all on standard output, and then exit successfully. Other options and
954 arguments should be ignored once this is seen, and the program should
955 not perform its normal function.
957 @cindex canonical name of a program
958 @cindex program's canonical name
959 The first line is meant to be easy for a program to parse; the version
960 number proper starts after the last space. In addition, it contains
961 the canonical name for this program, in this format:
968 The program's name should be a constant string; @emph{don't} compute it
969 from @code{argv[0]}. The idea is to state the standard or canonical
970 name for the program, not its file name. There are other ways to find
971 out the precise file name where a command is found in @code{PATH}.
973 If the program is a subsidiary part of a larger package, mention the
974 package name in parentheses, like this:
977 emacsserver (GNU Emacs) 19.30
981 If the package has a version number which is different from this
982 program's version number, you can mention the package version number
983 just before the close-parenthesis.
985 If you @emph{need} to mention the version numbers of libraries which
986 are distributed separately from the package which contains this program,
987 you can do so by printing an additional line of version info for each
988 library you want to mention. Use the same format for these lines as for
991 Please do not mention all of the libraries that the program uses ``just
992 for completeness''---that would produce a lot of unhelpful clutter.
993 Please mention library version numbers only if you find in practice that
994 they are very important to you in debugging.
996 The following line, after the version number line or lines, should be a
997 copyright notice. If more than one copyright notice is called for, put
998 each on a separate line.
1000 Next should follow a line stating the license, preferably using one of
1001 abbreviations below, and a brief statement that the program is free
1002 software, and that users are free to copy and change it. Also mention
1003 that there is no warranty, to the extent permitted by law. See
1004 recommended wording below.
1006 It is ok to finish the output with a list of the major authors of the
1007 program, as a way of giving credit.
1009 Here's an example of output that follows these rules:
1013 Copyright (C) 2007 Free Software Foundation, Inc.
1014 License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
1015 This is free software: you are free to change and redistribute it.
1016 There is NO WARRANTY, to the extent permitted by law.
1019 You should adapt this to your program, of course, filling in the proper
1020 year, copyright holder, name of program, and the references to
1021 distribution terms, and changing the rest of the wording as necessary.
1023 This copyright notice only needs to mention the most recent year in
1024 which changes were made---there's no need to list the years for previous
1025 versions' changes. You don't have to mention the name of the program in
1026 these notices, if that is inconvenient, since it appeared in the first
1027 line. (The rules are different for copyright notices in source files;
1028 @pxref{Copyright Notices,,,maintain,Information for GNU Maintainers}.)
1030 Translations of the above lines must preserve the validity of the
1031 copyright notices (@pxref{Internationalization}). If the translation's
1032 character set supports it, the @samp{(C)} should be replaced with the
1033 copyright symbol, as follows:
1036 (the official copyright symbol, which is the letter C in a circle);
1042 Write the word ``Copyright'' exactly like that, in English. Do not
1043 translate it into another language. International treaties recognize
1044 the English word ``Copyright''; translations into other languages do not
1045 have legal significance.
1047 Finally, here is the table of our suggested license abbreviations.
1048 Any abbreviation can be followed by @samp{v@var{version}[+]}, meaning
1049 that particular version, or later versions with the @samp{+}, as shown
1052 In the case of exceptions for extra permissions with the GPL, we use
1053 @samp{/} for a separator; the version number can follow the license
1054 abbreviation as usual, as in the examples below.
1058 GNU General Public License, @url{http://www.gnu.org/@/licenses/@/gpl.html}.
1061 GNU Lesser General Public License, @url{http://www.gnu.org/@/licenses/@/lgpl.html}.
1064 GNU GPL with the exception for Ada.
1067 The Apache Software Foundation license,
1068 @url{http://www.apache.org/@/licenses}.
1071 The Artistic license used for Perl, @url{http://www.perlfoundation.org/@/legal}.
1074 The Expat license, @url{http://www.jclark.com/@/xml/@/copying.txt}.
1077 The Mozilla Public License, @url{http://www.mozilla.org/@/MPL/}.
1080 The original (4-clause) BSD license, incompatible with the GNU GPL
1081 @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#6}.
1084 The license used for PHP, @url{http://www.php.net/@/license/}.
1087 The non-license that is being in the public domain,
1088 @url{http://www.gnu.org/@/licenses/@/license-list.html#PublicDomain}.
1091 The license for Python, @url{http://www.python.org/@/2.0.1/@/license.html}.
1094 The revised (3-clause) BSD, compatible with the GNU GPL,@*
1095 @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#5}.
1098 The simple non-copyleft license used for most versions of the X Window
1099 System, @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#3}.
1102 The license for Zlib, @url{http://www.gzip.org/@/zlib/@/zlib_license.html}.
1106 More information about these licenses and many more are on the GNU
1107 licensing web pages,
1108 @url{http://www.gnu.org/@/licenses/@/license-list.html}.
1112 @subsection @option{--help}
1114 @cindex @samp{--help} output
1116 The standard @code{--help} option should output brief documentation
1117 for how to invoke the program, on standard output, then exit
1118 successfully. Other options and arguments should be ignored once this
1119 is seen, and the program should not perform its normal function.
1121 @cindex address for bug reports
1123 Near the end of the @samp{--help} option's output, please place lines
1124 giving the email address for bug reports, the package's home page
1125 (normally @indicateurl{http://www.gnu.org/software/@var{pkg}}, and the
1126 general page for help using GNU programs. The format should be like this:
1129 Report bugs to: @var{mailing-address}
1130 @var{pkg} home page: <http://www.gnu.org/software/@var{pkg}/>
1131 General help using GNU software: <http://www.gnu.org/gethelp/>
1134 It is ok to mention other appropriate mailing lists and web pages.
1137 @node Dynamic Plug-In Interfaces
1138 @section Standards for Dynamic Plug-in Interfaces
1140 @cindex dynamic plug-ins
1142 Another aspect of keeping free programs free is encouraging
1143 development of free plug-ins, and discouraging development of
1144 proprietary plug-ins. Many GNU programs will not have anything like
1145 plug-ins at all, but those that do should follow these
1148 First, the general plug-in architecture design should closely tie the
1149 plug-in to the original code, such that the plug-in and the base
1150 program are parts of one extended program. For GCC, for example,
1151 plug-ins receive and modify GCC's internal data structures, and so
1152 clearly form an extended program with the base GCC.
1154 @vindex plugin_is_GPL_compatible
1155 Second, you should require plug-in developers to affirm that their
1156 plug-ins are released under an appropriate license. This should be
1157 enforced with a simple programmatic check. For GCC, again for
1158 example, a plug-in must define the global symbol
1159 @code{plugin_is_GPL_compatible}, thus asserting that the plug-in is
1160 released under a GPL-compatible license (@pxref{Plugins,, Plugins,
1161 gccint, GCC Internals}).
1163 By adding this check to your program you are not creating a new legal
1164 requirement. The GPL itself requires plug-ins to be free software,
1165 licensed compatibly. As long as you have followed the first rule above
1166 to keep plug-ins closely tied to your original program, the GPL and AGPL
1167 already require those plug-ins to be released under a compatible
1168 license. The symbol definition in the plug-in---or whatever equivalent
1169 works best in your program---makes it harder for anyone who might
1170 distribute proprietary plug-ins to legally defend themselves. If a case
1171 about this got to court, we can point to that symbol as evidence that
1172 the plug-in developer understood that the license had this requirement.
1176 @section Table of Long Options
1177 @cindex long option names
1178 @cindex table of long options
1180 Here is a table of long options used by GNU programs. It is surely
1181 incomplete, but we aim to list all the options that a new program might
1182 want to be compatible with. If you use names not already in the table,
1183 please send @email{bug-standards@@gnu.org} a list of them, with their
1184 meanings, so we can update the table.
1186 @c Please leave newlines between items in this table; it's much easier
1187 @c to update when it isn't completely squashed together and unreadable.
1188 @c When there is more than one short option for a long option name, put
1189 @c a semicolon between the lists of the programs that use them, not a
1190 @c period. --friedman
1194 @samp{-N} in @code{tar}.
1197 @samp{-a} in @code{du}, @code{ls}, @code{nm}, @code{stty}, @code{uname},
1198 and @code{unexpand}.
1201 @samp{-a} in @code{diff}.
1204 @samp{-A} in @code{ls}.
1207 @samp{-a} in @code{etags}, @code{tee}, @code{time};
1208 @samp{-r} in @code{tar}.
1211 @samp{-a} in @code{cp}.
1214 @samp{-n} in @code{shar}.
1217 @samp{-l} in @code{m4}.
1220 @samp{-a} in @code{diff}.
1223 @samp{-v} in @code{gawk}.
1226 @samp{-W} in @code{make}.
1229 @samp{-o} in @code{make}.
1232 @samp{-a} in @code{recode}.
1235 @samp{-a} in @code{wdiff}.
1237 @item auto-reference
1238 @samp{-A} in @code{ptx}.
1241 @samp{-n} in @code{wdiff}.
1244 For server programs, run in the background.
1246 @item backward-search
1247 @samp{-B} in @code{ctags}.
1250 @samp{-f} in @code{shar}.
1259 @samp{-b} in @code{tac}.
1262 @samp{-b} in @code{cpio} and @code{diff}.
1265 @samp{-b} in @code{shar}.
1268 Used in @code{cpio} and @code{tar}.
1271 @samp{-b} in @code{head} and @code{tail}.
1274 @samp{-b} in @code{ptx}.
1277 Used in various programs to make output shorter.
1280 @samp{-c} in @code{head}, @code{split}, and @code{tail}.
1283 @samp{-C} in @code{etags}.
1286 @samp{-A} in @code{tar}.
1289 Used in various programs to specify the directory to use.
1292 @samp{-c} in @code{chgrp} and @code{chown}.
1295 @samp{-F} in @code{ls}.
1298 @samp{-c} in @code{recode}.
1301 @samp{-c} in @code{su};
1305 @samp{-d} in @code{tar}.
1308 Used in @code{gawk}.
1311 @samp{-Z} in @code{tar} and @code{shar}.
1314 @samp{-A} in @code{tar}.
1317 @samp{-w} in @code{tar}.
1320 Used in @code{diff}.
1323 @samp{-W copyleft} in @code{gawk}.
1326 @samp{-C} in @code{ptx}, @code{recode}, and @code{wdiff};
1327 @samp{-W copyright} in @code{gawk}.
1333 @samp{-q} in @code{who}.
1336 @samp{-l} in @code{du}.
1339 Used in @code{tar} and @code{cpio}.
1342 @samp{-c} in @code{shar}.
1345 @samp{-x} in @code{ctags}.
1348 @samp{-d} in @code{touch}.
1351 @samp{-d} in @code{make} and @code{m4};
1355 @samp{-D} in @code{m4}.
1358 @samp{-d} in Bison and @code{ctags}.
1361 @samp{-D} in @code{tar}.
1364 @samp{-L} in @code{chgrp}, @code{chown}, @code{cpio}, @code{du},
1365 @code{ls}, and @code{tar}.
1367 @item dereference-args
1368 @samp{-D} in @code{du}.
1371 Specify an I/O device (special file name).
1374 @samp{-d} in @code{recode}.
1376 @item dictionary-order
1377 @samp{-d} in @code{look}.
1380 @samp{-d} in @code{tar}.
1383 @samp{-n} in @code{csplit}.
1386 Specify the directory to use, in various programs. In @code{ls}, it
1387 means to show directories themselves rather than their contents. In
1388 @code{rm} and @code{ln}, it means to not treat links to directories
1392 @samp{-x} in @code{strip}.
1394 @item discard-locals
1395 @samp{-X} in @code{strip}.
1398 @samp{-n} in @code{make}.
1401 @samp{-e} in @code{diff}.
1403 @item elide-empty-files
1404 @samp{-z} in @code{csplit}.
1407 @samp{-x} in @code{wdiff}.
1410 @samp{-z} in @code{wdiff}.
1412 @item entire-new-file
1413 @samp{-N} in @code{diff}.
1415 @item environment-overrides
1416 @samp{-e} in @code{make}.
1419 @samp{-e} in @code{xargs}.
1425 Used in @code{makeinfo}.
1428 @samp{-o} in @code{m4}.
1431 @samp{-b} in @code{ls}.
1434 @samp{-X} in @code{tar}.
1440 @samp{-x} in @code{xargs}.
1443 @samp{-e} in @code{unshar}.
1446 @samp{-t} in @code{diff}.
1449 @samp{-e} in @code{sed}.
1452 @samp{-g} in @code{nm}.
1455 @samp{-i} in @code{cpio};
1456 @samp{-x} in @code{tar}.
1459 @samp{-f} in @code{finger}.
1462 @samp{-f} in @code{su}.
1464 @item fatal-warnings
1465 @samp{-E} in @code{m4}.
1468 @samp{-f} in @code{gawk}, @code{info}, @code{make}, @code{mt},
1469 @code{sed}, and @code{tar}.
1471 @item field-separator
1472 @samp{-F} in @code{gawk}.
1478 @samp{-F} in @code{ls}.
1481 @samp{-T} in @code{tar}.
1484 Used in @code{makeinfo}.
1486 @item flag-truncation
1487 @samp{-F} in @code{ptx}.
1489 @item fixed-output-files
1493 @samp{-f} in @code{tail}.
1495 @item footnote-style
1496 Used in @code{makeinfo}.
1499 @samp{-f} in @code{cp}, @code{ln}, @code{mv}, and @code{rm}.
1502 @samp{-F} in @code{shar}.
1505 For server programs, run in the foreground;
1506 in other words, don't do anything special to run the server
1510 Used in @code{ls}, @code{time}, and @code{ptx}.
1513 @samp{-F} in @code{m4}.
1519 @samp{-g} in @code{ptx}.
1522 @samp{-x} in @code{tar}.
1525 @samp{-i} in @code{ul}.
1528 @samp{-g} in @code{recode}.
1531 @samp{-g} in @code{install}.
1534 @samp{-z} in @code{tar} and @code{shar}.
1537 @samp{-H} in @code{m4}.
1540 @samp{-h} in @code{objdump} and @code{recode}
1543 @samp{-H} in @code{who}.
1546 Used to ask for brief usage information.
1548 @item here-delimiter
1549 @samp{-d} in @code{shar}.
1551 @item hide-control-chars
1552 @samp{-q} in @code{ls}.
1555 In @code{makeinfo}, output HTML.
1558 @samp{-u} in @code{who}.
1561 @samp{-D} in @code{diff}.
1564 @samp{-I} in @code{ls};
1565 @samp{-x} in @code{recode}.
1567 @item ignore-all-space
1568 @samp{-w} in @code{diff}.
1570 @item ignore-backups
1571 @samp{-B} in @code{ls}.
1573 @item ignore-blank-lines
1574 @samp{-B} in @code{diff}.
1577 @samp{-f} in @code{look} and @code{ptx};
1578 @samp{-i} in @code{diff} and @code{wdiff}.
1581 @samp{-i} in @code{make}.
1584 @samp{-i} in @code{ptx}.
1586 @item ignore-indentation
1587 @samp{-I} in @code{etags}.
1589 @item ignore-init-file
1592 @item ignore-interrupts
1593 @samp{-i} in @code{tee}.
1595 @item ignore-matching-lines
1596 @samp{-I} in @code{diff}.
1598 @item ignore-space-change
1599 @samp{-b} in @code{diff}.
1602 @samp{-i} in @code{tar}.
1605 @samp{-i} in @code{etags};
1606 @samp{-I} in @code{m4}.
1609 @samp{-I} in @code{make}.
1612 @samp{-G} in @code{tar}.
1615 @samp{-i}, @samp{-l}, and @samp{-m} in Finger.
1618 In some programs, specify the name of the file to read as the user's
1622 @samp{-i} in @code{expand}.
1625 @samp{-T} in @code{diff}.
1628 @samp{-i} in @code{ls}.
1631 @samp{-i} in @code{cp}, @code{ln}, @code{mv}, @code{rm};
1632 @samp{-e} in @code{m4};
1633 @samp{-p} in @code{xargs};
1634 @samp{-w} in @code{tar}.
1637 @samp{-p} in @code{shar}.
1643 @samp{-j} in @code{make}.
1646 @samp{-n} in @code{make}.
1649 @samp{-k} in @code{make}.
1652 @samp{-k} in @code{csplit}.
1655 @samp{-k} in @code{du} and @code{ls}.
1658 @samp{-l} in @code{etags}.
1661 @samp{-l} in @code{wdiff}.
1663 @item level-for-gzip
1664 @samp{-g} in @code{shar}.
1667 @samp{-C} in @code{split}.
1670 Used in @code{split}, @code{head}, and @code{tail}.
1673 @samp{-l} in @code{cpio}.
1677 Used in @code{gawk}.
1680 @samp{-t} in @code{cpio};
1681 @samp{-l} in @code{recode}.
1684 @samp{-t} in @code{tar}.
1687 @samp{-N} in @code{ls}.
1690 @samp{-l} in @code{make}.
1696 Used in @code{uname}.
1699 @samp{-M} in @code{ptx}.
1702 @samp{-m} in @code{hello} and @code{uname}.
1704 @item make-directories
1705 @samp{-d} in @code{cpio}.
1708 @samp{-f} in @code{make}.
1714 @samp{-n} in @code{xargs}.
1717 @samp{-n} in @code{xargs}.
1720 @samp{-l} in @code{xargs}.
1723 @samp{-l} in @code{make}.
1726 @samp{-P} in @code{xargs}.
1729 @samp{-T} in @code{who}.
1732 @samp{-T} in @code{who}.
1735 @samp{-d} in @code{diff}.
1737 @item mixed-uuencode
1738 @samp{-M} in @code{shar}.
1741 @samp{-m} in @code{install}, @code{mkdir}, and @code{mkfifo}.
1743 @item modification-time
1744 @samp{-m} in @code{tar}.
1747 @samp{-M} in @code{tar}.
1753 @samp{-L} in @code{m4}.
1756 @samp{-a} in @code{shar}.
1759 @samp{-W} in @code{make}.
1761 @item no-builtin-rules
1762 @samp{-r} in @code{make}.
1764 @item no-character-count
1765 @samp{-w} in @code{shar}.
1767 @item no-check-existing
1768 @samp{-x} in @code{shar}.
1771 @samp{-3} in @code{wdiff}.
1774 @samp{-c} in @code{touch}.
1777 @samp{-D} in @code{etags}.
1780 @samp{-1} in @code{wdiff}.
1782 @item no-dereference
1783 @samp{-d} in @code{cp}.
1786 @samp{-2} in @code{wdiff}.
1789 @samp{-S} in @code{make}.
1795 @samp{-P} in @code{shar}.
1798 @samp{-e} in @code{gprof}.
1801 @samp{-R} in @code{etags}.
1804 @samp{-p} in @code{nm}.
1807 Don't print a startup splash screen.
1810 Used in @code{makeinfo}.
1813 @samp{-a} in @code{gprof}.
1816 @samp{-E} in @code{gprof}.
1819 @samp{-m} in @code{shar}.
1822 Used in @code{makeinfo}.
1825 Used in @code{emacsclient}.
1828 Used in various programs to inhibit warnings.
1831 @samp{-n} in @code{info}.
1834 @samp{-n} in @code{uname}.
1837 @samp{-f} in @code{cpio}.
1840 @samp{-n} in @code{objdump}.
1843 @samp{-0} in @code{xargs}.
1846 @samp{-n} in @code{cat}.
1848 @item number-nonblank
1849 @samp{-b} in @code{cat}.
1852 @samp{-n} in @code{nm}.
1854 @item numeric-uid-gid
1855 @samp{-n} in @code{cpio} and @code{ls}.
1861 @samp{-o} in @code{tar}.
1864 @samp{-o} in @code{make}.
1866 @item one-file-system
1867 @samp{-l} in @code{tar}, @code{cp}, and @code{du}.
1870 @samp{-o} in @code{ptx}.
1873 @samp{-f} in @code{gprof}.
1876 @samp{-F} in @code{gprof}.
1879 @samp{-o} in @code{getopt}, @code{fdlist}, @code{fdmount},
1880 @code{fdmountd}, and @code{fdumount}.
1883 In various programs, specify the output file name.
1886 @samp{-o} in @code{shar}.
1889 @samp{-o} in @code{rm}.
1892 @samp{-c} in @code{unshar}.
1895 @samp{-o} in @code{install}.
1898 @samp{-l} in @code{diff}.
1900 @item paragraph-indent
1901 Used in @code{makeinfo}.
1904 @samp{-p} in @code{mkdir} and @code{rmdir}.
1907 @samp{-p} in @code{ul}.
1910 @samp{-p} in @code{cpio}.
1913 @samp{-P} in @code{finger}.
1916 @samp{-c} in @code{cpio} and @code{tar}.
1919 Used in @code{gawk}.
1921 @item prefix-builtins
1922 @samp{-P} in @code{m4}.
1925 @samp{-f} in @code{csplit}.
1928 Used in @code{tar} and @code{cp}.
1930 @item preserve-environment
1931 @samp{-p} in @code{su}.
1933 @item preserve-modification-time
1934 @samp{-m} in @code{cpio}.
1936 @item preserve-order
1937 @samp{-s} in @code{tar}.
1939 @item preserve-permissions
1940 @samp{-p} in @code{tar}.
1943 @samp{-l} in @code{diff}.
1946 @samp{-L} in @code{cmp}.
1948 @item print-data-base
1949 @samp{-p} in @code{make}.
1951 @item print-directory
1952 @samp{-w} in @code{make}.
1954 @item print-file-name
1955 @samp{-o} in @code{nm}.
1958 @samp{-s} in @code{nm}.
1961 @samp{-p} in @code{wdiff}.
1964 @samp{-p} in @code{ed}.
1967 Specify an HTTP proxy.
1970 @samp{-X} in @code{shar}.
1973 @samp{-q} in @code{make}.
1976 Used in many programs to inhibit the usual output. Every
1977 program accepting @samp{--quiet} should accept @samp{--silent} as a
1981 @samp{-Q} in @code{shar}
1984 @samp{-Q} in @code{ls}.
1987 @samp{-n} in @code{diff}.
1990 Used in @code{gawk}.
1992 @item read-full-blocks
1993 @samp{-B} in @code{tar}.
1999 @samp{-n} in @code{make}.
2002 @samp{-R} in @code{tar}.
2005 Used in @code{chgrp}, @code{chown}, @code{cp}, @code{ls}, @code{diff},
2009 @samp{-r} in @code{touch}.
2012 @samp{-r} in @code{ptx}.
2015 @samp{-r} in @code{tac} and @code{etags}.
2018 @samp{-r} in @code{uname}.
2021 @samp{-R} in @code{m4}.
2024 @samp{-r} in @code{objdump}.
2027 @samp{-r} in @code{cpio}.
2030 @samp{-i} in @code{xargs}.
2032 @item report-identical-files
2033 @samp{-s} in @code{diff}.
2035 @item reset-access-time
2036 @samp{-a} in @code{cpio}.
2039 @samp{-r} in @code{ls} and @code{nm}.
2042 @samp{-f} in @code{diff}.
2044 @item right-side-defs
2045 @samp{-R} in @code{ptx}.
2048 @samp{-s} in @code{tar}.
2050 @item same-permissions
2051 @samp{-p} in @code{tar}.
2054 @samp{-g} in @code{stty}.
2059 @item sentence-regexp
2060 @samp{-S} in @code{ptx}.
2063 @samp{-S} in @code{du}.
2066 @samp{-s} in @code{tac}.
2069 Used by @code{recode} to chose files or pipes for sequencing passes.
2072 @samp{-s} in @code{su}.
2075 @samp{-A} in @code{cat}.
2077 @item show-c-function
2078 @samp{-p} in @code{diff}.
2081 @samp{-E} in @code{cat}.
2083 @item show-function-line
2084 @samp{-F} in @code{diff}.
2087 @samp{-T} in @code{cat}.
2090 Used in many programs to inhibit the usual output.
2091 Every program accepting
2092 @samp{--silent} should accept @samp{--quiet} as a synonym.
2095 @samp{-s} in @code{ls}.
2098 Specify a file descriptor for a network server to use for its socket,
2099 instead of opening and binding a new socket. This provides a way to
2100 run, in a non-privileged process, a server that normally needs a
2101 reserved port number.
2107 @samp{-W source} in @code{gawk}.
2110 @samp{-S} in @code{tar}.
2112 @item speed-large-files
2113 @samp{-H} in @code{diff}.
2116 @samp{-E} in @code{unshar}.
2118 @item split-size-limit
2119 @samp{-L} in @code{shar}.
2122 @samp{-s} in @code{cat}.
2125 @samp{-w} in @code{wdiff}.
2128 @samp{-y} in @code{wdiff}.
2131 Used in @code{tar} and @code{diff} to specify which file within
2132 a directory to start processing with.
2135 @samp{-s} in @code{wdiff}.
2137 @item stdin-file-list
2138 @samp{-S} in @code{shar}.
2141 @samp{-S} in @code{make}.
2144 @samp{-s} in @code{recode}.
2147 @samp{-s} in @code{install}.
2150 @samp{-s} in @code{strip}.
2153 @samp{-S} in @code{strip}.
2156 @samp{-s} in @code{shar}.
2159 @samp{-S} in @code{cp}, @code{ln}, @code{mv}.
2162 @samp{-b} in @code{csplit}.
2165 @samp{-s} in @code{gprof}.
2168 @samp{-s} in @code{du}.
2171 @samp{-s} in @code{ln}.
2174 Used in GDB and @code{objdump}.
2177 @samp{-s} in @code{m4}.
2180 @samp{-s} in @code{uname}.
2183 @samp{-t} in @code{expand} and @code{unexpand}.
2186 @samp{-T} in @code{ls}.
2189 @samp{-T} in @code{tput} and @code{ul}.
2190 @samp{-t} in @code{wdiff}.
2193 @samp{-a} in @code{diff}.
2196 @samp{-T} in @code{shar}.
2199 Used in @code{ls} and @code{touch}.
2202 Specify how long to wait before giving up on some operation.
2205 @samp{-O} in @code{tar}.
2208 @samp{-c} in @code{du}.
2211 @samp{-t} in @code{make}, @code{ranlib}, and @code{recode}.
2214 @samp{-t} in @code{m4}.
2217 @samp{-t} in @code{hello};
2218 @samp{-W traditional} in @code{gawk};
2219 @samp{-G} in @code{ed}, @code{m4}, and @code{ptx}.
2225 @samp{-t} in @code{ctags}.
2227 @item typedefs-and-c++
2228 @samp{-T} in @code{ctags}.
2231 @samp{-t} in @code{ptx}.
2234 @samp{-z} in @code{tar}.
2237 @samp{-u} in @code{cpio}.
2240 @samp{-U} in @code{m4}.
2242 @item undefined-only
2243 @samp{-u} in @code{nm}.
2246 @samp{-u} in @code{cp}, @code{ctags}, @code{mv}, @code{tar}.
2249 Used in @code{gawk}; same as @samp{--help}.
2252 @samp{-B} in @code{shar}.
2254 @item vanilla-operation
2255 @samp{-V} in @code{shar}.
2258 Print more information about progress. Many programs support this.
2261 @samp{-W} in @code{tar}.
2264 Print the version number.
2266 @item version-control
2267 @samp{-V} in @code{cp}, @code{ln}, @code{mv}.
2270 @samp{-v} in @code{ctags}.
2273 @samp{-V} in @code{tar}.
2276 @samp{-W} in @code{make}.
2278 @item whole-size-limit
2279 @samp{-l} in @code{shar}.
2282 @samp{-w} in @code{ls} and @code{ptx}.
2285 @samp{-W} in @code{ptx}.
2288 @samp{-T} in @code{who}.
2291 @samp{-z} in @code{gprof}.
2294 @node OID Allocations
2295 @section OID Allocations
2296 @cindex OID allocations for GNU
2301 The OID (object identifier) 1.3.6.1.4.1.11591 has been assigned to the
2302 GNU Project (thanks to Werner Koch). These are used for SNMP, LDAP,
2303 X.509 certificates, and so on. The web site
2304 @url{http://www.alvestrand.no/objectid} has a (voluntary) listing of
2305 many OID assignments.
2307 If you need a new slot for your GNU package, write
2308 @email{maintainers@@gnu.org}. Here is a list of arcs currently
2312 @include gnu-oids.texi
2317 @section Memory Usage
2318 @cindex memory usage
2320 If a program typically uses just a few meg of memory, don't bother making any
2321 effort to reduce memory usage. For example, if it is impractical for
2322 other reasons to operate on files more than a few meg long, it is
2323 reasonable to read entire input files into memory to operate on them.
2325 However, for programs such as @code{cat} or @code{tail}, that can
2326 usefully operate on very large files, it is important to avoid using a
2327 technique that would artificially limit the size of files it can handle.
2328 If a program works by lines and could be applied to arbitrary
2329 user-supplied input files, it should keep only a line in memory, because
2330 this is not very hard and users will want to be able to operate on input
2331 files that are bigger than will fit in memory all at once.
2333 If your program creates complicated data structures, just make them in
2334 memory and give a fatal error if @code{malloc} returns zero.
2338 Memory analysis tools such as @command{valgrind} can be useful, but
2339 don't complicate a program merely to avoid their false alarms. For
2340 example, if memory is used until just before a process exits, don't
2341 free it simply to silence such a tool.
2348 Programs should be prepared to operate when @file{/usr} and @file{/etc}
2349 are read-only file systems. Thus, if the program manages log files,
2350 lock files, backup files, score files, or any other files which are
2351 modified for internal purposes, these files should not be stored in
2352 @file{/usr} or @file{/etc}.
2354 There are two exceptions. @file{/etc} is used to store system
2355 configuration information; it is reasonable for a program to modify
2356 files in @file{/etc} when its job is to update the system configuration.
2357 Also, if the user explicitly asks to modify one file in a directory, it
2358 is reasonable for the program to store other files in the same
2362 @chapter Making The Best Use of C
2364 This chapter provides advice on how best to use the C language
2365 when writing GNU software.
2368 * Formatting:: Formatting your source code.
2369 * Comments:: Commenting your work.
2370 * Syntactic Conventions:: Clean use of C constructs.
2371 * Names:: Naming variables, functions, and files.
2372 * System Portability:: Portability among different operating systems.
2373 * CPU Portability:: Supporting the range of CPU types.
2374 * System Functions:: Portability and ``standard'' library functions.
2375 * Internationalization:: Techniques for internationalization.
2376 * Character Set:: Use ASCII by default.
2377 * Quote Characters:: Use "..." or '...' in the C locale.
2378 * Mmap:: How you can safely use @code{mmap}.
2382 @section Formatting Your Source Code
2383 @cindex formatting source code
2386 @cindex braces, in C source
2387 @cindex function definitions, formatting
2388 It is important to put the open-brace that starts the body of a C
2389 function in column one, so that they will start a defun. Several
2390 tools look for open-braces in column one to find the beginnings of C
2391 functions. These tools will not work on code not formatted that way.
2393 Avoid putting open-brace, open-parenthesis or open-bracket in column
2394 one when they are inside a function, so that they won't start a defun.
2395 The open-brace that starts a @code{struct} body can go in column one
2396 if you find it useful to treat that definition as a defun.
2398 It is also important for function definitions to start the name of the
2399 function in column one. This helps people to search for function
2400 definitions, and may also help certain tools recognize them. Thus,
2401 using Standard C syntax, the format is this:
2405 concat (char *s1, char *s2)
2412 or, if you want to use traditional C syntax, format the definition like
2417 concat (s1, s2) /* Name starts in column one here */
2419 @{ /* Open brace in column one here */
2424 In Standard C, if the arguments don't fit nicely on one line,
2429 lots_of_args (int an_integer, long a_long, short a_short,
2430 double a_double, float a_float)
2434 @cindex @code{struct} types, formatting
2435 @cindex @code{enum} types, formatting
2436 For @code{struct} and @code{enum} types, likewise put the braces in
2437 column one, unless the whole contents fits on one line:
2445 struct foo @{ int a, b; @}
2448 The rest of this section gives our recommendations for other aspects of
2449 C formatting style, which is also the default style of the @code{indent}
2450 program in version 1.2 and newer. It corresponds to the options
2453 -nbad -bap -nbc -bbo -bl -bli2 -bls -ncdb -nce -cp1 -cs -di2
2454 -ndj -nfc1 -nfca -hnl -i2 -ip5 -lp -pcs -psl -nsc -nsob
2457 We don't think of these recommendations as requirements, because it
2458 causes no problems for users if two different programs have different
2461 But whatever style you use, please use it consistently, since a mixture
2462 of styles within one program tends to look ugly. If you are
2463 contributing changes to an existing program, please follow the style of
2466 For the body of the function, our recommended style looks like this:
2478 return ++x + bar ();
2482 @cindex spaces before open-paren
2483 We find it easier to read a program when it has spaces before the
2484 open-parentheses and after the commas. Especially after the commas.
2486 When you split an expression into multiple lines, split it
2487 before an operator, not after one. Here is the right way:
2489 @cindex expressions, splitting
2491 if (foo_this_is_long && bar > win (x, y, z)
2492 && remaining_condition)
2495 Try to avoid having two operators of different precedence at the same
2496 level of indentation. For example, don't write this:
2499 mode = (inmode[j] == VOIDmode
2500 || GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])
2501 ? outmode[j] : inmode[j]);
2504 Instead, use extra parentheses so that the indentation shows the nesting:
2507 mode = ((inmode[j] == VOIDmode
2508 || (GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])))
2509 ? outmode[j] : inmode[j]);
2512 Insert extra parentheses so that Emacs will indent the code properly.
2513 For example, the following indentation looks nice if you do it by hand,
2516 v = rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2517 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000;
2521 but Emacs would alter it. Adding a set of parentheses produces
2522 something that looks equally nice, and which Emacs will preserve:
2525 v = (rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2526 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000);
2529 Format do-while statements like this:
2541 Please use formfeed characters (control-L) to divide the program into
2542 pages at logical places (but not within a function). It does not matter
2543 just how long the pages are, since they do not have to fit on a printed
2544 page. The formfeeds should appear alone on lines by themselves.
2547 @section Commenting Your Work
2550 Every program should start with a comment saying briefly what it is for.
2551 Example: @samp{fmt - filter for simple filling of text}. This comment
2552 should be at the top of the source file containing the @samp{main}
2553 function of the program.
2555 Also, please write a brief comment at the start of each source file,
2556 with the file name and a line or two about the overall purpose of the
2559 Please write the comments in a GNU program in English, because English
2560 is the one language that nearly all programmers in all countries can
2561 read. If you do not write English well, please write comments in
2562 English as well as you can, then ask other people to help rewrite them.
2563 If you can't write comments in English, please find someone to work with
2564 you and translate your comments into English.
2566 Please put a comment on each function saying what the function does,
2567 what sorts of arguments it gets, and what the possible values of
2568 arguments mean and are used for. It is not necessary to duplicate in
2569 words the meaning of the C argument declarations, if a C type is being
2570 used in its customary fashion. If there is anything nonstandard about
2571 its use (such as an argument of type @code{char *} which is really the
2572 address of the second character of a string, not the first), or any
2573 possible values that would not work the way one would expect (such as,
2574 that strings containing newlines are not guaranteed to work), be sure
2577 Also explain the significance of the return value, if there is one.
2579 Please put two spaces after the end of a sentence in your comments, so
2580 that the Emacs sentence commands will work. Also, please write
2581 complete sentences and capitalize the first word. If a lower-case
2582 identifier comes at the beginning of a sentence, don't capitalize it!
2583 Changing the spelling makes it a different identifier. If you don't
2584 like starting a sentence with a lower case letter, write the sentence
2585 differently (e.g., ``The identifier lower-case is @dots{}'').
2587 The comment on a function is much clearer if you use the argument
2588 names to speak about the argument values. The variable name itself
2589 should be lower case, but write it in upper case when you are speaking
2590 about the value rather than the variable itself. Thus, ``the inode
2591 number NODE_NUM'' rather than ``an inode''.
2593 There is usually no purpose in restating the name of the function in
2594 the comment before it, because readers can see that for themselves.
2595 There might be an exception when the comment is so long that the function
2596 itself would be off the bottom of the screen.
2598 There should be a comment on each static variable as well, like this:
2601 /* Nonzero means truncate lines in the display;
2602 zero means continue them. */
2606 @cindex conditionals, comments for
2607 @cindex @code{#endif}, commenting
2608 Every @samp{#endif} should have a comment, except in the case of short
2609 conditionals (just a few lines) that are not nested. The comment should
2610 state the condition of the conditional that is ending, @emph{including
2611 its sense}. @samp{#else} should have a comment describing the condition
2612 @emph{and sense} of the code that follows. For example:
2620 #endif /* not foo */
2630 but, by contrast, write the comments this way for a @samp{#ifndef}:
2643 #endif /* not foo */
2647 @node Syntactic Conventions
2648 @section Clean Use of C Constructs
2649 @cindex syntactic conventions
2651 @cindex implicit @code{int}
2652 @cindex function argument, declaring
2653 Please explicitly declare the types of all objects. For example, you
2654 should explicitly declare all arguments to functions, and you should
2655 declare functions to return @code{int} rather than omitting the
2658 @cindex compiler warnings
2659 @cindex @samp{-Wall} compiler option
2660 Some programmers like to use the GCC @samp{-Wall} option, and change the
2661 code whenever it issues a warning. If you want to do this, then do.
2662 Other programmers prefer not to use @samp{-Wall}, because it gives
2663 warnings for valid and legitimate code which they do not want to change.
2664 If you want to do this, then do. The compiler should be your servant,
2669 Don't make the program ugly just to placate static analysis tools such
2670 as @command{lint}, @command{clang}, and GCC with extra warnings
2671 options such as @option{-Wconversion} and @option{-Wundef}. These
2672 tools can help find bugs and unclear code, but they can also generate
2673 so many false alarms that it hurts readability to silence them with
2674 unnecessary casts, wrappers, and other complications. For example,
2675 please don't insert casts to @code{void} or calls to do-nothing
2676 functions merely to pacify a lint checker.
2678 Declarations of external functions and functions to appear later in the
2679 source file should all go in one place near the beginning of the file
2680 (somewhere before the first function definition in the file), or else
2681 should go in a header file. Don't put @code{extern} declarations inside
2684 @cindex temporary variables
2685 It used to be common practice to use the same local variables (with
2686 names like @code{tem}) over and over for different values within one
2687 function. Instead of doing this, it is better to declare a separate local
2688 variable for each distinct purpose, and give it a name which is
2689 meaningful. This not only makes programs easier to understand, it also
2690 facilitates optimization by good compilers. You can also move the
2691 declaration of each local variable into the smallest scope that includes
2692 all its uses. This makes the program even cleaner.
2694 Don't use local variables or parameters that shadow global identifiers.
2695 GCC's @samp{-Wshadow} option can detect this problem.
2697 @cindex multiple variables in a line
2698 Don't declare multiple variables in one declaration that spans lines.
2699 Start a new declaration on each line, instead. For example, instead
2725 (If they are global variables, each should have a comment preceding it
2728 When you have an @code{if}-@code{else} statement nested in another
2729 @code{if} statement, always put braces around the @code{if}-@code{else}.
2730 Thus, never write like this:
2753 If you have an @code{if} statement nested inside of an @code{else}
2754 statement, either write @code{else if} on one line, like this,
2764 with its @code{then}-part indented like the preceding @code{then}-part,
2765 or write the nested @code{if} within braces like this:
2777 Don't declare both a structure tag and variables or typedefs in the
2778 same declaration. Instead, declare the structure tag separately
2779 and then use it to declare the variables or typedefs.
2781 Try to avoid assignments inside @code{if}-conditions (assignments
2782 inside @code{while}-conditions are ok). For example, don't write
2786 if ((foo = (char *) malloc (sizeof *foo)) == 0)
2787 fatal ("virtual memory exhausted");
2791 instead, write this:
2794 foo = (char *) malloc (sizeof *foo);
2796 fatal ("virtual memory exhausted");
2799 This example uses zero without a cast as a null pointer constant.
2800 This is perfectly fine, except that a cast is needed when calling a
2801 varargs function or when using @code{sizeof}.
2804 @section Naming Variables, Functions, and Files
2806 @cindex names of variables, functions, and files
2807 The names of global variables and functions in a program serve as
2808 comments of a sort. So don't choose terse names---instead, look for
2809 names that give useful information about the meaning of the variable or
2810 function. In a GNU program, names should be English, like other
2813 Local variable names can be shorter, because they are used only within
2814 one context, where (presumably) comments explain their purpose.
2816 Try to limit your use of abbreviations in symbol names. It is ok to
2817 make a few abbreviations, explain what they mean, and then use them
2818 frequently, but don't use lots of obscure abbreviations.
2820 Please use underscores to separate words in a name, so that the Emacs
2821 word commands can be useful within them. Stick to lower case; reserve
2822 upper case for macros and @code{enum} constants, and for name-prefixes
2823 that follow a uniform convention.
2825 For example, you should use names like @code{ignore_space_change_flag};
2826 don't use names like @code{iCantReadThis}.
2828 Variables that indicate whether command-line options have been
2829 specified should be named after the meaning of the option, not after
2830 the option-letter. A comment should state both the exact meaning of
2831 the option and its letter. For example,
2835 /* Ignore changes in horizontal whitespace (-b). */
2836 int ignore_space_change_flag;
2840 When you want to define names with constant integer values, use
2841 @code{enum} rather than @samp{#define}. GDB knows about enumeration
2844 @cindex file-name limitations
2846 You might want to make sure that none of the file names would conflict
2847 if the files were loaded onto an MS-DOS file system which shortens the
2848 names. You can use the program @code{doschk} to test for this.
2850 Some GNU programs were designed to limit themselves to file names of 14
2851 characters or less, to avoid file name conflicts if they are read into
2852 older System V systems. Please preserve this feature in the existing
2853 GNU programs that have it, but there is no need to do this in new GNU
2854 programs. @code{doschk} also reports file names longer than 14
2858 @node System Portability
2859 @section Portability between System Types
2860 @cindex portability, between system types
2862 In the Unix world, ``portability'' refers to porting to different Unix
2863 versions. For a GNU program, this kind of portability is desirable, but
2866 The primary purpose of GNU software is to run on top of the GNU kernel,
2867 compiled with the GNU C compiler, on various types of @sc{cpu}. So the
2868 kinds of portability that are absolutely necessary are quite limited.
2869 But it is important to support Linux-based GNU systems, since they
2870 are the form of GNU that is popular.
2872 Beyond that, it is good to support the other free operating systems
2873 (*BSD), and it is nice to support other Unix-like systems if you want
2874 to. Supporting a variety of Unix-like systems is desirable, although
2875 not paramount. It is usually not too hard, so you may as well do it.
2876 But you don't have to consider it an obligation, if it does turn out to
2880 The easiest way to achieve portability to most Unix-like systems is to
2881 use Autoconf. It's unlikely that your program needs to know more
2882 information about the host platform than Autoconf can provide, simply
2883 because most of the programs that need such knowledge have already been
2886 Avoid using the format of semi-internal data bases (e.g., directories)
2887 when there is a higher-level alternative (@code{readdir}).
2889 @cindex non-@sc{posix} systems, and portability
2890 As for systems that are not like Unix, such as MSDOS, Windows, VMS, MVS,
2891 and older Macintosh systems, supporting them is often a lot of work.
2892 When that is the case, it is better to spend your time adding features
2893 that will be useful on GNU and GNU/Linux, rather than on supporting
2894 other incompatible systems.
2896 If you do support Windows, please do not abbreviate it as ``win''. In
2897 hacker terminology, calling something a ``win'' is a form of praise.
2898 You're free to praise Microsoft Windows on your own if you want, but
2899 please don't do this in GNU packages. Instead of abbreviating
2900 ``Windows'' to ``win'', you can write it in full or abbreviate it to
2901 ``woe'' or ``w''. In GNU Emacs, for instance, we use @samp{w32} in
2902 file names of Windows-specific files, but the macro for Windows
2903 conditionals is called @code{WINDOWSNT}.
2905 It is a good idea to define the ``feature test macro''
2906 @code{_GNU_SOURCE} when compiling your C files. When you compile on GNU
2907 or GNU/Linux, this will enable the declarations of GNU library extension
2908 functions, and that will usually give you a compiler error message if
2909 you define the same function names in some other way in your program.
2910 (You don't have to actually @emph{use} these functions, if you prefer
2911 to make the program more portable to other systems.)
2913 But whether or not you use these GNU extensions, you should avoid
2914 using their names for any other meanings. Doing so would make it hard
2915 to move your code into other GNU programs.
2917 @node CPU Portability
2918 @section Portability between @sc{cpu}s
2920 @cindex data types, and portability
2921 @cindex portability, and data types
2922 Even GNU systems will differ because of differences among @sc{cpu}
2923 types---for example, difference in byte ordering and alignment
2924 requirements. It is absolutely essential to handle these differences.
2925 However, don't make any effort to cater to the possibility that an
2926 @code{int} will be less than 32 bits. We don't support 16-bit machines
2929 Similarly, don't make any effort to cater to the possibility that
2930 @code{long} will be smaller than predefined types like @code{size_t}.
2931 For example, the following code is ok:
2934 printf ("size = %lu\n", (unsigned long) sizeof array);
2935 printf ("diff = %ld\n", (long) (pointer2 - pointer1));
2938 1989 Standard C requires this to work, and we know of only one
2939 counterexample: 64-bit programs on Microsoft Windows. We will leave
2940 it to those who want to port GNU programs to that environment to
2941 figure out how to do it.
2943 Predefined file-size types like @code{off_t} are an exception: they are
2944 longer than @code{long} on many platforms, so code like the above won't
2945 work with them. One way to print an @code{off_t} value portably is to
2946 print its digits yourself, one by one.
2948 Don't assume that the address of an @code{int} object is also the
2949 address of its least-significant byte. This is false on big-endian
2950 machines. Thus, don't make the following mistake:
2955 while ((c = getchar ()) != EOF)
2956 write (file_descriptor, &c, 1);
2959 @noindent Instead, use @code{unsigned char} as follows. (The @code{unsigned}
2960 is for portability to unusual systems where @code{char} is signed and
2961 where there is integer overflow checking.)
2965 while ((c = getchar ()) != EOF)
2967 unsigned char u = c;
2968 write (file_descriptor, &u, 1);
2972 @cindex casting pointers to integers
2973 Avoid casting pointers to integers if you can. Such casts greatly
2974 reduce portability, and in most programs they are easy to avoid. In the
2975 cases where casting pointers to integers is essential---such as, a Lisp
2976 interpreter which stores type information as well as an address in one
2977 word---you'll have to make explicit provisions to handle different word
2978 sizes. You will also need to make provision for systems in which the
2979 normal range of addresses you can get from @code{malloc} starts far away
2983 @node System Functions
2984 @section Calling System Functions
2986 @cindex C library functions, and portability
2987 @cindex POSIX functions, and portability
2988 @cindex library functions, and portability
2989 @cindex portability, and library functions
2991 Historically, C implementations differed substantially, and many
2992 systems lacked a full implementation of ANSI/ISO C89. Nowadays,
2993 however, very few systems lack a C89 compiler and GNU C supports
2994 almost all of C99. Similarly, most systems implement POSIX.1-1993
2995 libraries and tools, and many have POSIX.1-2001.
2997 Hence, there is little reason to support old C or non-POSIX systems,
2998 and you may want to take advantage of C99 and POSIX-1.2001 to write
2999 clearer, more portable, or faster code. You should use standard
3000 interfaces where possible; but if GNU extensions make your program
3001 more maintainable, powerful, or otherwise better, don't hesitate to
3002 use them. In any case, don't make your own declaration of system
3003 functions; that's a recipe for conflict.
3005 Despite the standards, nearly every library function has some sort of
3006 portability issue on some system or another. Here are some examples:
3010 Names with trailing @code{/}'s are mishandled on many platforms.
3013 @code{long double} may be unimplemented; floating values Infinity and
3014 NaN are often mishandled; output for large precisions may be
3018 May return @code{int} instead of @code{ssize_t}.
3021 On Windows, @code{errno} is not set on failure.
3025 @uref{http://www.gnu.org/software/gnulib/, Gnulib} is a big help in
3026 this regard. Gnulib provides implementations of standard interfaces
3027 on many of the systems that lack them, including portable
3028 implementations of enhanced GNU interfaces, thereby making their use
3029 portable, and of POSIX-1.2008 interfaces, some of which are missing
3030 even on up-to-date GNU systems.
3032 @findex xmalloc, in Gnulib
3033 @findex error messages, in Gnulib
3034 @findex data structures, in Gnulib
3035 Gnulib also provides many useful non-standard interfaces; for example,
3036 C implementations of standard data structures (hash tables, binary
3037 trees), error-checking type-safe wrappers for memory allocation
3038 functions (@code{xmalloc}, @code{xrealloc}), and output of error
3041 Gnulib integrates with GNU Autoconf and Automake to remove much of the
3042 burden of writing portable code from the programmer: Gnulib makes your
3043 configure script automatically determine what features are missing and
3044 use the Gnulib code to supply the missing pieces.
3046 The Gnulib and Autoconf manuals have extensive sections on
3047 portability: @ref{Top,, Introduction, gnulib, Gnulib} and
3048 @pxref{Portable C and C++,,, autoconf, Autoconf}. Please consult them
3049 for many more details.
3052 @node Internationalization
3053 @section Internationalization
3054 @cindex internationalization
3057 GNU has a library called GNU gettext that makes it easy to translate the
3058 messages in a program into various languages. You should use this
3059 library in every program. Use English for the messages as they appear
3060 in the program, and let gettext provide the way to translate them into
3063 Using GNU gettext involves putting a call to the @code{gettext} macro
3064 around each string that might need translation---like this:
3067 printf (gettext ("Processing file '%s'..."), file);
3071 This permits GNU gettext to replace the string @code{"Processing file
3072 '%s'..."} with a translated version.
3074 Once a program uses gettext, please make a point of writing calls to
3075 @code{gettext} when you add new strings that call for translation.
3077 Using GNU gettext in a package involves specifying a @dfn{text domain
3078 name} for the package. The text domain name is used to separate the
3079 translations for this package from the translations for other packages.
3080 Normally, the text domain name should be the same as the name of the
3081 package---for example, @samp{coreutils} for the GNU core utilities.
3083 @cindex message text, and internationalization
3084 To enable gettext to work well, avoid writing code that makes
3085 assumptions about the structure of words or sentences. When you want
3086 the precise text of a sentence to vary depending on the data, use two or
3087 more alternative string constants each containing a complete sentences,
3088 rather than inserting conditionalized words or phrases into a single
3091 Here is an example of what not to do:
3094 printf ("%s is full", capacity > 5000000 ? "disk" : "floppy disk");
3097 If you apply gettext to all strings, like this,
3100 printf (gettext ("%s is full"),
3101 capacity > 5000000 ? gettext ("disk") : gettext ("floppy disk"));
3105 the translator will hardly know that "disk" and "floppy disk" are meant to
3106 be substituted in the other string. Worse, in some languages (like French)
3107 the construction will not work: the translation of the word "full" depends
3108 on the gender of the first part of the sentence; it happens to be not the
3109 same for "disk" as for "floppy disk".
3111 Complete sentences can be translated without problems:
3114 printf (capacity > 5000000 ? gettext ("disk is full")
3115 : gettext ("floppy disk is full"));
3118 A similar problem appears at the level of sentence structure with this
3122 printf ("# Implicit rule search has%s been done.\n",
3123 f->tried_implicit ? "" : " not");
3127 Adding @code{gettext} calls to this code cannot give correct results for
3128 all languages, because negation in some languages requires adding words
3129 at more than one place in the sentence. By contrast, adding
3130 @code{gettext} calls does the job straightforwardly if the code starts
3134 printf (f->tried_implicit
3135 ? "# Implicit rule search has been done.\n",
3136 : "# Implicit rule search has not been done.\n");
3139 Another example is this one:
3142 printf ("%d file%s processed", nfiles,
3143 nfiles != 1 ? "s" : "");
3147 The problem with this example is that it assumes that plurals are made
3148 by adding `s'. If you apply gettext to the format string, like this,
3151 printf (gettext ("%d file%s processed"), nfiles,
3152 nfiles != 1 ? "s" : "");
3156 the message can use different words, but it will still be forced to use
3157 `s' for the plural. Here is a better way, with gettext being applied to
3158 the two strings independently:
3161 printf ((nfiles != 1 ? gettext ("%d files processed")
3162 : gettext ("%d file processed")),
3167 But this still doesn't work for languages like Polish, which has three
3168 plural forms: one for nfiles == 1, one for nfiles == 2, 3, 4, 22, 23, 24, ...
3169 and one for the rest. The GNU @code{ngettext} function solves this problem:
3172 printf (ngettext ("%d files processed", "%d file processed", nfiles),
3178 @section Character Set
3179 @cindex character set
3181 @cindex ASCII characters
3182 @cindex non-ASCII characters
3184 Sticking to the ASCII character set (plain text, 7-bit characters) is
3185 preferred in GNU source code comments, text documents, and other
3186 contexts, unless there is good reason to do something else because of
3187 the application domain. For example, if source code deals with the
3188 French Revolutionary calendar, it is OK if its literal strings contain
3189 accented characters in month names like ``Flor@'eal''. Also, it is OK
3190 (but not required) to use non-ASCII characters to represent proper
3191 names of contributors in change logs (@pxref{Change Logs}).
3193 If you need to use non-ASCII characters, you should normally stick
3194 with one encoding, certainly within a single file. UTF-8 is likely to
3198 @node Quote Characters
3199 @section Quote Characters
3200 @cindex quote characters
3201 @cindex locale-specific quote characters
3204 @cindex opening quote
3205 @cindex single quote
3206 @cindex double quote
3207 @cindex grave accent
3208 @set txicodequoteundirected
3209 @set txicodequotebacktick
3211 In the C locale, the output of GNU programs should stick to plain
3212 ASCII for quotation characters in messages to users: preferably 0x22
3213 (@samp{"}) or 0x27 (@samp{'}) for both opening and closing quotes.
3214 Although GNU programs traditionally used 0x60 (@samp{`}) for opening
3215 and 0x27 (@samp{'}) for closing quotes, nowadays quotes @samp{`like
3216 this'} are typically rendered asymmetrically, so quoting @samp{"like
3217 this"} or @samp{'like this'} typically looks better.
3219 It is ok, but not required, for GNU programs to generate
3220 locale-specific quotes in non-C locales. For example:
3223 printf (gettext ("Processing file '%s'..."), file);
3227 Here, a French translation might cause @code{gettext} to return the
3228 string @code{"Traitement de fichier
3229 @guilsinglleft{}@tie{}%s@tie{}@guilsinglright{}..."}, yielding quotes
3230 more appropriate for a French locale.
3232 Sometimes a program may need to use opening and closing quotes
3233 directly. By convention, @code{gettext} translates the string
3234 @samp{"`"} to the opening quote and the string @samp{"'"} to the
3235 closing quote, and a program can use these translations. Generally,
3236 though, it is better to translate quote characters in the context of
3239 If the output of your program is ever likely to be parsed by another
3240 program, it is good to provide an option that makes this parsing
3241 reliable. For example, you could escape special characters using
3242 conventions from the C language or the Bourne shell. See for example
3243 the option @option{--quoting-style} of GNU @code{ls}.
3245 @clear txicodequoteundirected
3246 @clear txicodequotebacktick
3253 Don't assume that @code{mmap} either works on all files or fails
3254 for all files. It may work on some files and fail on others.
3256 The proper way to use @code{mmap} is to try it on the specific file for
3257 which you want to use it---and if @code{mmap} doesn't work, fall back on
3258 doing the job in another way using @code{read} and @code{write}.
3260 The reason this precaution is needed is that the GNU kernel (the HURD)
3261 provides a user-extensible file system, in which there can be many
3262 different kinds of ``ordinary files''. Many of them support
3263 @code{mmap}, but some do not. It is important to make programs handle
3264 all these kinds of files.
3268 @chapter Documenting Programs
3269 @cindex documentation
3271 A GNU program should ideally come with full free documentation, adequate
3272 for both reference and tutorial purposes. If the package can be
3273 programmed or extended, the documentation should cover programming or
3274 extending it, as well as just using it.
3277 * GNU Manuals:: Writing proper manuals.
3278 * Doc Strings and Manuals:: Compiling doc strings doesn't make a manual.
3279 * Manual Structure Details:: Specific structure conventions.
3280 * License for Manuals:: Writing the distribution terms for a manual.
3281 * Manual Credits:: Giving credit to documentation contributors.
3282 * Printed Manuals:: Mentioning the printed manual.
3283 * NEWS File:: NEWS files supplement manuals.
3284 * Change Logs:: Recording changes.
3285 * Man Pages:: Man pages are secondary.
3286 * Reading other Manuals:: How far you can go in learning
3291 @section GNU Manuals
3293 The preferred document format for the GNU system is the Texinfo
3294 formatting language. Every GNU package should (ideally) have
3295 documentation in Texinfo both for reference and for learners. Texinfo
3296 makes it possible to produce a good quality formatted book, using
3297 @TeX{}, and to generate an Info file. It is also possible to generate
3298 HTML output from Texinfo source. See the Texinfo manual, either the
3299 hardcopy, or the on-line version available through @code{info} or the
3300 Emacs Info subsystem (@kbd{C-h i}).
3302 Nowadays some other formats such as Docbook and Sgmltexi can be
3303 converted automatically into Texinfo. It is ok to produce the Texinfo
3304 documentation by conversion this way, as long as it gives good results.
3306 Make sure your manual is clear to a reader who knows nothing about the
3307 topic and reads it straight through. This means covering basic topics
3308 at the beginning, and advanced topics only later. This also means
3309 defining every specialized term when it is first used.
3311 Programmers tend to carry over the structure of the program as the
3312 structure for its documentation. But this structure is not
3313 necessarily good for explaining how to use the program; it may be
3314 irrelevant and confusing for a user.
3316 Instead, the right way to structure documentation is according to the
3317 concepts and questions that a user will have in mind when reading it.
3318 This principle applies at every level, from the lowest (ordering
3319 sentences in a paragraph) to the highest (ordering of chapter topics
3320 within the manual). Sometimes this structure of ideas matches the
3321 structure of the implementation of the software being documented---but
3322 often they are different. An important part of learning to write good
3323 documentation is to learn to notice when you have unthinkingly
3324 structured the documentation like the implementation, stop yourself,
3325 and look for better alternatives.
3327 For example, each program in the GNU system probably ought to be
3328 documented in one manual; but this does not mean each program should
3329 have its own manual. That would be following the structure of the
3330 implementation, rather than the structure that helps the user
3333 Instead, each manual should cover a coherent @emph{topic}. For example,
3334 instead of a manual for @code{diff} and a manual for @code{diff3}, we
3335 have one manual for ``comparison of files'' which covers both of those
3336 programs, as well as @code{cmp}. By documenting these programs
3337 together, we can make the whole subject clearer.
3339 The manual which discusses a program should certainly document all of
3340 the program's command-line options and all of its commands. It should
3341 give examples of their use. But don't organize the manual as a list
3342 of features. Instead, organize it logically, by subtopics. Address
3343 the questions that a user will ask when thinking about the job that
3344 the program does. Don't just tell the reader what each feature can
3345 do---say what jobs it is good for, and show how to use it for those
3346 jobs. Explain what is recommended usage, and what kinds of usage
3349 In general, a GNU manual should serve both as tutorial and reference.
3350 It should be set up for convenient access to each topic through Info,
3351 and for reading straight through (appendixes aside). A GNU manual
3352 should give a good introduction to a beginner reading through from the
3353 start, and should also provide all the details that hackers want.
3354 The Bison manual is a good example of this---please take a look at it
3355 to see what we mean.
3357 That is not as hard as it first sounds. Arrange each chapter as a
3358 logical breakdown of its topic, but order the sections, and write their
3359 text, so that reading the chapter straight through makes sense. Do
3360 likewise when structuring the book into chapters, and when structuring a
3361 section into paragraphs. The watchword is, @emph{at each point, address
3362 the most fundamental and important issue raised by the preceding text.}
3364 If necessary, add extra chapters at the beginning of the manual which
3365 are purely tutorial and cover the basics of the subject. These provide
3366 the framework for a beginner to understand the rest of the manual. The
3367 Bison manual provides a good example of how to do this.
3369 To serve as a reference, a manual should have an Index that list all the
3370 functions, variables, options, and important concepts that are part of
3371 the program. One combined Index should do for a short manual, but
3372 sometimes for a complex package it is better to use multiple indices.
3373 The Texinfo manual includes advice on preparing good index entries, see
3374 @ref{Index Entries, , Making Index Entries, texinfo, GNU Texinfo}, and
3375 see @ref{Indexing Commands, , Defining the Entries of an
3376 Index, texinfo, GNU Texinfo}.
3378 Don't use Unix man pages as a model for how to write GNU documentation;
3379 most of them are terse, badly structured, and give inadequate
3380 explanation of the underlying concepts. (There are, of course, some
3381 exceptions.) Also, Unix man pages use a particular format which is
3382 different from what we use in GNU manuals.
3384 Please include an email address in the manual for where to report
3385 bugs @emph{in the text of the manual}.
3387 Please do not use the term ``pathname'' that is used in Unix
3388 documentation; use ``file name'' (two words) instead. We use the term
3389 ``path'' only for search paths, which are lists of directory names.
3391 Please do not use the term ``illegal'' to refer to erroneous input to
3392 a computer program. Please use ``invalid'' for this, and reserve the
3393 term ``illegal'' for activities prohibited by law.
3395 Please do not write @samp{()} after a function name just to indicate
3396 it is a function. @code{foo ()} is not a function, it is a function
3397 call with no arguments.
3399 @node Doc Strings and Manuals
3400 @section Doc Strings and Manuals
3402 Some programming systems, such as Emacs, provide a documentation string
3403 for each function, command or variable. You may be tempted to write a
3404 reference manual by compiling the documentation strings and writing a
3405 little additional text to go around them---but you must not do it. That
3406 approach is a fundamental mistake. The text of well-written
3407 documentation strings will be entirely wrong for a manual.
3409 A documentation string needs to stand alone---when it appears on the
3410 screen, there will be no other text to introduce or explain it.
3411 Meanwhile, it can be rather informal in style.
3413 The text describing a function or variable in a manual must not stand
3414 alone; it appears in the context of a section or subsection. Other text
3415 at the beginning of the section should explain some of the concepts, and
3416 should often make some general points that apply to several functions or
3417 variables. The previous descriptions of functions and variables in the
3418 section will also have given information about the topic. A description
3419 written to stand alone would repeat some of that information; this
3420 redundancy looks bad. Meanwhile, the informality that is acceptable in
3421 a documentation string is totally unacceptable in a manual.
3423 The only good way to use documentation strings in writing a good manual
3424 is to use them as a source of information for writing good text.
3426 @node Manual Structure Details
3427 @section Manual Structure Details
3428 @cindex manual structure
3430 The title page of the manual should state the version of the programs or
3431 packages documented in the manual. The Top node of the manual should
3432 also contain this information. If the manual is changing more
3433 frequently than or independent of the program, also state a version
3434 number for the manual in both of these places.
3436 Each program documented in the manual should have a node named
3437 @samp{@var{program} Invocation} or @samp{Invoking @var{program}}. This
3438 node (together with its subnodes, if any) should describe the program's
3439 command line arguments and how to run it (the sort of information people
3440 would look for in a man page). Start with an @samp{@@example}
3441 containing a template for all the options and arguments that the program
3444 Alternatively, put a menu item in some menu whose item name fits one of
3445 the above patterns. This identifies the node which that item points to
3446 as the node for this purpose, regardless of the node's actual name.
3448 The @samp{--usage} feature of the Info reader looks for such a node
3449 or menu item in order to find the relevant text, so it is essential
3450 for every Texinfo file to have one.
3452 If one manual describes several programs, it should have such a node for
3453 each program described in the manual.
3455 @node License for Manuals
3456 @section License for Manuals
3457 @cindex license for manuals
3459 Please use the GNU Free Documentation License for all GNU manuals that
3460 are more than a few pages long. Likewise for a collection of short
3461 documents---you only need one copy of the GNU FDL for the whole
3462 collection. For a single short document, you can use a very permissive
3463 non-copyleft license, to avoid taking up space with a long license.
3465 See @uref{http://www.gnu.org/copyleft/fdl-howto.html} for more explanation
3466 of how to employ the GFDL.
3468 Note that it is not obligatory to include a copy of the GNU GPL or GNU
3469 LGPL in a manual whose license is neither the GPL nor the LGPL. It can
3470 be a good idea to include the program's license in a large manual; in a
3471 short manual, whose size would be increased considerably by including
3472 the program's license, it is probably better not to include it.
3474 @node Manual Credits
3475 @section Manual Credits
3476 @cindex credits for manuals
3478 Please credit the principal human writers of the manual as the authors,
3479 on the title page of the manual. If a company sponsored the work, thank
3480 the company in a suitable place in the manual, but do not cite the
3481 company as an author.
3483 @node Printed Manuals
3484 @section Printed Manuals
3486 The FSF publishes some GNU manuals in printed form. To encourage sales
3487 of these manuals, the on-line versions of the manual should mention at
3488 the very start that the printed manual is available and should point at
3489 information for getting it---for instance, with a link to the page
3490 @url{http://www.gnu.org/order/order.html}. This should not be included
3491 in the printed manual, though, because there it is redundant.
3493 It is also useful to explain in the on-line forms of the manual how the
3494 user can print out the manual from the sources.
3497 @section The NEWS File
3498 @cindex @file{NEWS} file
3500 In addition to its manual, the package should have a file named
3501 @file{NEWS} which contains a list of user-visible changes worth
3502 mentioning. In each new release, add items to the front of the file and
3503 identify the version they pertain to. Don't discard old items; leave
3504 them in the file after the newer items. This way, a user upgrading from
3505 any previous version can see what is new.
3507 If the @file{NEWS} file gets very long, move some of the older items
3508 into a file named @file{ONEWS} and put a note at the end referring the
3512 @section Change Logs
3515 Keep a change log to describe all the changes made to program source
3516 files. The purpose of this is so that people investigating bugs in the
3517 future will know about the changes that might have introduced the bug.
3518 Often a new bug can be found by looking at what was recently changed.
3519 More importantly, change logs can help you eliminate conceptual
3520 inconsistencies between different parts of a program, by giving you a
3521 history of how the conflicting concepts arose and who they came from.
3524 * Change Log Concepts::
3525 * Style of Change Logs::
3527 * Conditional Changes::
3528 * Indicating the Part Changed::
3531 @node Change Log Concepts
3532 @subsection Change Log Concepts
3534 You can think of the change log as a conceptual ``undo list'' which
3535 explains how earlier versions were different from the current version.
3536 People can see the current version; they don't need the change log
3537 to tell them what is in it. What they want from a change log is a
3538 clear explanation of how the earlier version differed.
3540 The change log file is normally called @file{ChangeLog} and covers an
3541 entire directory. Each directory can have its own change log, or a
3542 directory can use the change log of its parent directory---it's up to
3545 Another alternative is to record change log information with a version
3546 control system such as RCS or CVS. This can be converted automatically
3547 to a @file{ChangeLog} file using @code{rcs2log}; in Emacs, the command
3548 @kbd{C-x v a} (@code{vc-update-change-log}) does the job.
3550 There's no need to describe the full purpose of the changes or how
3551 they work together. However, sometimes it is useful to write one line
3552 to describe the overall purpose of a change or a batch of changes. If
3553 you think that a change calls for explanation, you're probably right.
3554 Please do explain it---but please put the full explanation in comments
3555 in the code, where people will see it whenever they see the code. For
3556 example, ``New function'' is enough for the change log when you add a
3557 function, because there should be a comment before the function
3558 definition to explain what it does.
3560 In the past, we recommended not mentioning changes in non-software
3561 files (manuals, help files, etc.) in change logs. However, we've been
3562 advised that it is a good idea to include them, for the sake of
3565 The easiest way to add an entry to @file{ChangeLog} is with the Emacs
3566 command @kbd{M-x add-change-log-entry}. An entry should have an
3567 asterisk, the name of the changed file, and then in parentheses the name
3568 of the changed functions, variables or whatever, followed by a colon.
3569 Then describe the changes you made to that function or variable.
3571 @node Style of Change Logs
3572 @subsection Style of Change Logs
3573 @cindex change logs, style
3575 Here are some simple examples of change log entries, starting with the
3576 header line that says who made the change and when it was installed,
3577 followed by descriptions of specific changes. (These examples are
3578 drawn from Emacs and GCC.)
3581 1998-08-17 Richard Stallman <rms@@gnu.org>
3583 * register.el (insert-register): Return nil.
3584 (jump-to-register): Likewise.
3586 * sort.el (sort-subr): Return nil.
3588 * tex-mode.el (tex-bibtex-file, tex-file, tex-region):
3589 Restart the tex shell if process is gone or stopped.
3590 (tex-shell-running): New function.
3592 * expr.c (store_one_arg): Round size up for move_block_to_reg.
3593 (expand_call): Round up when emitting USE insns.
3594 * stmt.c (assign_parms): Round size up for move_block_from_reg.
3597 It's important to name the changed function or variable in full. Don't
3598 abbreviate function or variable names, and don't combine them.
3599 Subsequent maintainers will often search for a function name to find all
3600 the change log entries that pertain to it; if you abbreviate the name,
3601 they won't find it when they search.
3603 For example, some people are tempted to abbreviate groups of function
3604 names by writing @samp{* register.el (@{insert,jump-to@}-register)};
3605 this is not a good idea, since searching for @code{jump-to-register} or
3606 @code{insert-register} would not find that entry.
3608 Separate unrelated change log entries with blank lines. When two
3609 entries represent parts of the same change, so that they work together,
3610 then don't put blank lines between them. Then you can omit the file
3611 name and the asterisk when successive entries are in the same file.
3613 Break long lists of function names by closing continued lines with
3614 @samp{)}, rather than @samp{,}, and opening the continuation with
3615 @samp{(} as in this example:
3618 * keyboard.c (menu_bar_items, tool_bar_items)
3619 (Fexecute_extended_command): Deal with 'keymap' property.
3622 When you install someone else's changes, put the contributor's name in
3623 the change log entry rather than in the text of the entry. In other
3627 2002-07-14 John Doe <jdoe@@gnu.org>
3629 * sewing.c: Make it sew.
3636 2002-07-14 Usual Maintainer <usual@@gnu.org>
3638 * sewing.c: Make it sew. Patch by jdoe@@gnu.org.
3641 As for the date, that should be the date you applied the change.
3643 @node Simple Changes
3644 @subsection Simple Changes
3646 Certain simple kinds of changes don't need much detail in the change
3649 When you change the calling sequence of a function in a simple fashion,
3650 and you change all the callers of the function to use the new calling
3651 sequence, there is no need to make individual entries for all the
3652 callers that you changed. Just write in the entry for the function
3653 being called, ``All callers changed''---like this:
3656 * keyboard.c (Fcommand_execute): New arg SPECIAL.
3657 All callers changed.
3660 When you change just comments or doc strings, it is enough to write an
3661 entry for the file, without mentioning the functions. Just ``Doc
3662 fixes'' is enough for the change log.
3664 There's no technical need to make change log entries for documentation
3665 files. This is because documentation is not susceptible to bugs that
3666 are hard to fix. Documentation does not consist of parts that must
3667 interact in a precisely engineered fashion. To correct an error, you
3668 need not know the history of the erroneous passage; it is enough to
3669 compare what the documentation says with the way the program actually
3672 However, you should keep change logs for documentation files when the
3673 project gets copyright assignments from its contributors, so as to
3674 make the records of authorship more accurate.
3676 @node Conditional Changes
3677 @subsection Conditional Changes
3678 @cindex conditional changes, and change logs
3679 @cindex change logs, conditional changes
3681 Source files can often contain code that is conditional to build-time
3682 or static conditions. For example, C programs can contain
3683 compile-time @code{#if} conditionals; programs implemented in
3684 interpreted languages can contain module imports of function
3685 definitions that are only performed for certain versions of the
3686 interpreter; and Automake @file{Makefile.am} files can contain
3687 variable definitions or target declarations that are only to be
3688 considered if a configure-time Automake conditional is true.
3690 Many changes are conditional as well: sometimes you add a new variable,
3691 or function, or even a new program or library, which is entirely
3692 dependent on a build-time condition. It is useful to indicate
3693 in the change log the conditions for which a change applies.
3695 Our convention for indicating conditional changes is to use
3696 @emph{square brackets around the name of the condition}.
3698 Conditional changes can happen in numerous scenarios and with many
3699 variations, so here are some examples to help clarify. This first
3700 example describes changes in C, Perl, and Python files which are
3701 conditional but do not have an associated function or entity name:
3704 * xterm.c [SOLARIS2]: Include <string.h>.
3705 * FilePath.pm [$^O eq 'VMS']: Import the VMS::Feature module.
3706 * framework.py [sys.version_info < (2, 6)]: Make "with" statement
3707 available by importing it from __future__,
3708 to support also python 2.5.
3711 Our other examples will for simplicity be limited to C, as the minor
3712 changes necessary to adapt them to other languages should be
3715 Next, here is an entry describing a new definition which is entirely
3716 conditional: the C macro @code{FRAME_WINDOW_P} is defined (and used)
3717 only when the macro @code{HAVE_X_WINDOWS} is defined:
3720 * frame.h [HAVE_X_WINDOWS] (FRAME_WINDOW_P): Macro defined.
3723 Next, an entry for a change within the function @code{init_display},
3724 whose definition as a whole is unconditional, but the changes
3725 themselves are contained in a @samp{#ifdef HAVE_LIBNCURSES}
3729 * dispnew.c (init_display) [HAVE_LIBNCURSES]: If X, call tgetent.
3732 Finally, here is an entry for a change that takes effect only when
3733 a certain macro is @emph{not} defined:
3736 (gethostname) [!HAVE_SOCKETS]: Replace with winsock version.
3739 @node Indicating the Part Changed
3740 @subsection Indicating the Part Changed
3742 Indicate the part of a function which changed by using angle brackets
3743 enclosing an indication of what the changed part does. Here is an entry
3744 for a change in the part of the function @code{sh-while-getopts} that
3745 deals with @code{sh} commands:
3748 * progmodes/sh-script.el (sh-while-getopts) <sh>: Handle case that
3749 user-specified option string is empty.
3757 In the GNU project, man pages are secondary. It is not necessary or
3758 expected for every GNU program to have a man page, but some of them do.
3759 It's your choice whether to include a man page in your program.
3761 When you make this decision, consider that supporting a man page
3762 requires continual effort each time the program is changed. The time
3763 you spend on the man page is time taken away from more useful work.
3765 For a simple program which changes little, updating the man page may be
3766 a small job. Then there is little reason not to include a man page, if
3769 For a large program that changes a great deal, updating a man page may
3770 be a substantial burden. If a user offers to donate a man page, you may
3771 find this gift costly to accept. It may be better to refuse the man
3772 page unless the same person agrees to take full responsibility for
3773 maintaining it---so that you can wash your hands of it entirely. If
3774 this volunteer later ceases to do the job, then don't feel obliged to
3775 pick it up yourself; it may be better to withdraw the man page from the
3776 distribution until someone else agrees to update it.
3778 When a program changes only a little, you may feel that the
3779 discrepancies are small enough that the man page remains useful without
3780 updating. If so, put a prominent note near the beginning of the man
3781 page explaining that you don't maintain it and that the Texinfo manual
3782 is more authoritative. The note should say how to access the Texinfo
3785 Be sure that man pages include a copyright statement and free license.
3786 The simple all-permissive license is appropriate for simple man pages
3787 (@pxref{License Notices for Other Files,,,maintain,Information for GNU
3790 For long man pages, with enough explanation and documentation that
3791 they can be considered true manuals, use the GFDL (@pxref{License for
3794 Finally, the GNU help2man program
3795 (@uref{http://www.gnu.org/software/help2man/}) is one way to automate
3796 generation of a man page, in this case from @option{--help} output.
3797 This is sufficient in many cases.
3799 @node Reading other Manuals
3800 @section Reading other Manuals
3802 There may be non-free books or documentation files that describe the
3803 program you are documenting.
3805 It is ok to use these documents for reference, just as the author of a
3806 new algebra textbook can read other books on algebra. A large portion
3807 of any non-fiction book consists of facts, in this case facts about how
3808 a certain program works, and these facts are necessarily the same for
3809 everyone who writes about the subject. But be careful not to copy your
3810 outline structure, wording, tables or examples from preexisting non-free
3811 documentation. Copying from free documentation may be ok; please check
3812 with the FSF about the individual case.
3814 @node Managing Releases
3815 @chapter The Release Process
3818 Making a release is more than just bundling up your source files in a
3819 tar file and putting it up for FTP. You should set up your software so
3820 that it can be configured to run on a variety of systems. Your Makefile
3821 should conform to the GNU standards described below, and your directory
3822 layout should also conform to the standards discussed below. Doing so
3823 makes it easy to include your package into the larger framework of
3827 * Configuration:: How configuration of GNU packages should work.
3828 * Makefile Conventions:: Makefile conventions.
3829 * Releases:: Making releases
3833 @section How Configuration Should Work
3834 @cindex program configuration
3837 Each GNU distribution should come with a shell script named
3838 @code{configure}. This script is given arguments which describe the
3839 kind of machine and system you want to compile the program for.
3840 The @code{configure} script must record the configuration options so
3841 that they affect compilation.
3843 The description here is the specification of the interface for the
3844 @code{configure} script in GNU packages. Many packages implement it
3845 using GNU Autoconf (@pxref{Top,, Introduction, autoconf, Autoconf})
3846 and/or GNU Automake (@pxref{Top,, Introduction, automake, Automake}),
3847 but you do not have to use these tools. You can implement it any way
3848 you like; for instance, by making @code{configure} be a wrapper around
3849 a completely different configuration system.
3851 Another way for the @code{configure} script to operate is to make a
3852 link from a standard name such as @file{config.h} to the proper
3853 configuration file for the chosen system. If you use this technique,
3854 the distribution should @emph{not} contain a file named
3855 @file{config.h}. This is so that people won't be able to build the
3856 program without configuring it first.
3858 Another thing that @code{configure} can do is to edit the Makefile. If
3859 you do this, the distribution should @emph{not} contain a file named
3860 @file{Makefile}. Instead, it should include a file @file{Makefile.in} which
3861 contains the input used for editing. Once again, this is so that people
3862 won't be able to build the program without configuring it first.
3864 If @code{configure} does write the @file{Makefile}, then @file{Makefile}
3865 should have a target named @file{Makefile} which causes @code{configure}
3866 to be rerun, setting up the same configuration that was set up last
3867 time. The files that @code{configure} reads should be listed as
3868 dependencies of @file{Makefile}.
3870 All the files which are output from the @code{configure} script should
3871 have comments at the beginning explaining that they were generated
3872 automatically using @code{configure}. This is so that users won't think
3873 of trying to edit them by hand.
3875 The @code{configure} script should write a file named @file{config.status}
3876 which describes which configuration options were specified when the
3877 program was last configured. This file should be a shell script which,
3878 if run, will recreate the same configuration.
3880 The @code{configure} script should accept an option of the form
3881 @samp{--srcdir=@var{dirname}} to specify the directory where sources are found
3882 (if it is not the current directory). This makes it possible to build
3883 the program in a separate directory, so that the actual source directory
3886 If the user does not specify @samp{--srcdir}, then @code{configure} should
3887 check both @file{.} and @file{..} to see if it can find the sources. If
3888 it finds the sources in one of these places, it should use them from
3889 there. Otherwise, it should report that it cannot find the sources, and
3890 should exit with nonzero status.
3892 Usually the easy way to support @samp{--srcdir} is by editing a
3893 definition of @code{VPATH} into the Makefile. Some rules may need to
3894 refer explicitly to the specified source directory. To make this
3895 possible, @code{configure} can add to the Makefile a variable named
3896 @code{srcdir} whose value is precisely the specified directory.
3898 In addition, the @samp{configure} script should take options
3899 corresponding to most of the standard directory variables
3900 (@pxref{Directory Variables}). Here is the list:
3903 --prefix --exec-prefix --bindir --sbindir --libexecdir --sysconfdir
3904 --sharedstatedir --localstatedir --libdir --includedir --oldincludedir
3905 --datarootdir --datadir --infodir --localedir --mandir --docdir
3906 --htmldir --dvidir --pdfdir --psdir
3909 The @code{configure} script should also take an argument which specifies the
3910 type of system to build the program for. This argument should look like
3914 @var{cpu}-@var{company}-@var{system}
3917 For example, an Athlon-based GNU/Linux system might be
3918 @samp{i686-pc-linux-gnu}.
3920 The @code{configure} script needs to be able to decode all plausible
3921 alternatives for how to describe a machine. Thus,
3922 @samp{athlon-pc-gnu/linux} would be a valid alias. There is a shell
3924 @uref{http://git.savannah.gnu.org/@/gitweb/@/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD,
3925 @file{config.sub}} that you can use as a subroutine to validate system
3926 types and canonicalize aliases.
3928 The @code{configure} script should also take the option
3929 @option{--build=@var{buildtype}}, which should be equivalent to a
3930 plain @var{buildtype} argument. For example, @samp{configure
3931 --build=i686-pc-linux-gnu} is equivalent to @samp{configure
3932 i686-pc-linux-gnu}. When the build type is not specified by an option
3933 or argument, the @code{configure} script should normally guess it using
3935 @uref{http://git.savannah.gnu.org/@/gitweb/@/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD,
3936 @file{config.guess}}.
3938 @cindex optional features, configure-time
3939 Other options are permitted to specify in more detail the software
3940 or hardware present on the machine, to include or exclude optional parts
3941 of the package, or to adjust the name of some tools or arguments to them:
3944 @item --enable-@var{feature}@r{[}=@var{parameter}@r{]}
3945 Configure the package to build and install an optional user-level
3946 facility called @var{feature}. This allows users to choose which
3947 optional features to include. Giving an optional @var{parameter} of
3948 @samp{no} should omit @var{feature}, if it is built by default.
3950 No @samp{--enable} option should @strong{ever} cause one feature to
3951 replace another. No @samp{--enable} option should ever substitute one
3952 useful behavior for another useful behavior. The only proper use for
3953 @samp{--enable} is for questions of whether to build part of the program
3956 @item --with-@var{package}
3957 @c @r{[}=@var{parameter}@r{]}
3958 The package @var{package} will be installed, so configure this package
3959 to work with @var{package}.
3961 @c Giving an optional @var{parameter} of
3962 @c @samp{no} should omit @var{package}, if it is used by default.
3964 Possible values of @var{package} include
3965 @samp{gnu-as} (or @samp{gas}), @samp{gnu-ld}, @samp{gnu-libc},
3971 Do not use a @samp{--with} option to specify the file name to use to
3972 find certain files. That is outside the scope of what @samp{--with}
3975 @item @var{variable}=@var{value}
3976 Set the value of the variable @var{variable} to @var{value}. This is
3977 used to override the default values of commands or arguments in the
3978 build process. For example, the user could issue @samp{configure
3979 CFLAGS=-g CXXFLAGS=-g} to build with debugging information and without
3980 the default optimization.
3982 Specifying variables as arguments to @code{configure}, like this:
3986 is preferable to setting them in environment variables:
3990 as it helps to recreate the same configuration later with
3991 @file{config.status}. However, both methods should be supported.
3994 All @code{configure} scripts should accept all of the ``detail''
3995 options and the variable settings, whether or not they make any
3996 difference to the particular package at hand. In particular, they
3997 should accept any option that starts with @samp{--with-} or
3998 @samp{--enable-}. This is so users will be able to configure an
3999 entire GNU source tree at once with a single set of options.
4001 You will note that the categories @samp{--with-} and @samp{--enable-}
4002 are narrow: they @strong{do not} provide a place for any sort of option
4003 you might think of. That is deliberate. We want to limit the possible
4004 configuration options in GNU software. We do not want GNU programs to
4005 have idiosyncratic configuration options.
4007 Packages that perform part of the compilation process may support
4008 cross-compilation. In such a case, the host and target machines for the
4009 program may be different.
4011 The @code{configure} script should normally treat the specified type of
4012 system as both the host and the target, thus producing a program which
4013 works for the same type of machine that it runs on.
4015 To compile a program to run on a host type that differs from the build
4016 type, use the configure option @option{--host=@var{hosttype}}, where
4017 @var{hosttype} uses the same syntax as @var{buildtype}. The host type
4018 normally defaults to the build type.
4020 To configure a cross-compiler, cross-assembler, or what have you, you
4021 should specify a target different from the host, using the configure
4022 option @samp{--target=@var{targettype}}. The syntax for
4023 @var{targettype} is the same as for the host type. So the command would
4027 ./configure --host=@var{hosttype} --target=@var{targettype}
4030 The target type normally defaults to the host type.
4031 Programs for which cross-operation is not meaningful need not accept the
4032 @samp{--target} option, because configuring an entire operating system for
4033 cross-operation is not a meaningful operation.
4035 Some programs have ways of configuring themselves automatically. If
4036 your program is set up to do this, your @code{configure} script can simply
4037 ignore most of its arguments.
4039 @comment The makefile standards are in a separate file that is also
4040 @comment included by make.texinfo. Done by roland@gnu.ai.mit.edu on 1/6/93.
4041 @comment For this document, turn chapters into sections, etc.
4043 @include make-stds.texi
4047 @section Making Releases
4050 You should identify each release with a pair of version numbers, a
4051 major version and a minor. We have no objection to using more than
4052 two numbers, but it is very unlikely that you really need them.
4054 Package the distribution of @code{Foo version 69.96} up in a gzipped tar
4055 file with the name @file{foo-69.96.tar.gz}. It should unpack into a
4056 subdirectory named @file{foo-69.96}.
4058 Building and installing the program should never modify any of the files
4059 contained in the distribution. This means that all the files that form
4060 part of the program in any way must be classified into @dfn{source
4061 files} and @dfn{non-source files}. Source files are written by humans
4062 and never changed automatically; non-source files are produced from
4063 source files by programs under the control of the Makefile.
4065 @cindex @file{README} file
4066 The distribution should contain a file named @file{README} which gives
4067 the name of the package, and a general description of what it does. It
4068 is also good to explain the purpose of each of the first-level
4069 subdirectories in the package, if there are any. The @file{README} file
4070 should either state the version number of the package, or refer to where
4071 in the package it can be found.
4073 The @file{README} file should refer to the file @file{INSTALL}, which
4074 should contain an explanation of the installation procedure.
4076 The @file{README} file should also refer to the file which contains the
4077 copying conditions. The GNU GPL, if used, should be in a file called
4078 @file{COPYING}. If the GNU LGPL is used, it should be in a file called
4079 @file{COPYING.LESSER}.
4081 Naturally, all the source files must be in the distribution. It is
4082 okay to include non-source files in the distribution along with the
4083 source files they are generated from, provided they are up-to-date
4084 with the source they are made from, and machine-independent, so that
4085 normal building of the distribution will never modify them. We
4086 commonly include non-source files produced by Autoconf, Automake,
4087 Bison, @code{lex}, @TeX{}, and @code{makeinfo}; this helps avoid
4088 unnecessary dependencies between our distributions, so that users can
4089 install whichever packages they want to install.
4091 Non-source files that might actually be modified by building and
4092 installing the program should @strong{never} be included in the
4093 distribution. So if you do distribute non-source files, always make
4094 sure they are up to date when you make a new distribution.
4096 Make sure that all the files in the distribution are world-readable, and
4097 that directories are world-readable and world-searchable (octal mode 755).
4098 We used to recommend that all directories in the distribution also be
4099 world-writable (octal mode 777), because ancient versions of @code{tar}
4100 would otherwise not cope when extracting the archive as an unprivileged
4101 user. That can easily lead to security issues when creating the archive,
4102 however, so now we recommend against that.
4104 Don't include any symbolic links in the distribution itself. If the tar
4105 file contains symbolic links, then people cannot even unpack it on
4106 systems that don't support symbolic links. Also, don't use multiple
4107 names for one file in different directories, because certain file
4108 systems cannot handle this and that prevents unpacking the
4111 Try to make sure that all the file names will be unique on MS-DOS. A
4112 name on MS-DOS consists of up to 8 characters, optionally followed by a
4113 period and up to three characters. MS-DOS will truncate extra
4114 characters both before and after the period. Thus,
4115 @file{foobarhacker.c} and @file{foobarhacker.o} are not ambiguous; they
4116 are truncated to @file{foobarha.c} and @file{foobarha.o}, which are
4119 @cindex @file{texinfo.tex}, in a distribution
4120 Include in your distribution a copy of the @file{texinfo.tex} you used
4121 to test print any @file{*.texinfo} or @file{*.texi} files.
4123 Likewise, if your program uses small GNU software packages like regex,
4124 getopt, obstack, or termcap, include them in the distribution file.
4125 Leaving them out would make the distribution file a little smaller at
4126 the expense of possible inconvenience to a user who doesn't know what
4130 @chapter References to Non-Free Software and Documentation
4131 @cindex references to non-free material
4133 A GNU program should not recommend, promote, or grant legitimacy to
4134 the use of any non-free program. Proprietary software is a social and
4135 ethical problem, and our aim is to put an end to that problem. We
4136 can't stop some people from writing proprietary programs, or stop
4137 other people from using them, but we can and should refuse to
4138 advertise them to new potential customers, or to give the public the
4139 idea that their existence is ethical.
4141 The GNU definition of free software is found on the GNU web site at
4142 @url{http://www.gnu.org/@/philosophy/@/free-sw.html}, and the definition
4143 of free documentation is found at
4144 @url{http://www.gnu.org/@/philosophy/@/free-doc.html}. The terms ``free''
4145 and ``non-free'', used in this document, refer to those definitions.
4147 A list of important licenses and whether they qualify as free is in
4148 @url{http://www.gnu.org/@/licenses/@/license-list.html}. If it is not
4149 clear whether a license qualifies as free, please ask the GNU Project
4150 by writing to @email{licensing@@gnu.org}. We will answer, and if the
4151 license is an important one, we will add it to the list.
4153 When a non-free program or system is well known, you can mention it in
4154 passing---that is harmless, since users who might want to use it
4155 probably already know about it. For instance, it is fine to explain
4156 how to build your package on top of some widely used non-free
4157 operating system, or how to use it together with some widely used
4160 However, you should give only the necessary information to help those
4161 who already use the non-free program to use your program with
4162 it---don't give, or refer to, any further information about the
4163 proprietary program, and don't imply that the proprietary program
4164 enhances your program, or that its existence is in any way a good
4165 thing. The goal should be that people already using the proprietary
4166 program will get the advice they need about how to use your free
4167 program with it, while people who don't already use the proprietary
4168 program will not see anything likely to lead them to take an interest
4171 If a non-free program or system is obscure in your program's domain,
4172 your program should not mention or support it at all, since doing so
4173 would tend to popularize the non-free program more than it popularizes
4174 your program. (You cannot hope to find many additional users for your
4175 program among the users of Foobar, if the existence of Foobar is not
4176 generally known among people who might want to use your program.)
4178 Sometimes a program is free software in itself but depends on a
4179 non-free platform in order to run. For instance, many Java programs
4180 depend on some non-free Java libraries. To recommend or promote such
4181 a program is to promote the other programs it needs. This is why we
4182 are careful about listing Java programs in the Free Software
4183 Directory: we don't want to promote the non-free Java libraries.
4185 We hope this particular problem with Java will be gone by and by, as
4186 we replace the remaining non-free standard Java libraries with free
4187 software, but the general principle will remain the same: don't
4188 recommend, promote or legitimize programs that depend on non-free
4191 Some free programs strongly encourage the use of non-free software. A
4192 typical example is @command{mplayer}. It is free software in itself,
4193 and the free code can handle some kinds of files. However,
4194 @command{mplayer} recommends use of non-free codecs for other kinds of
4195 files, and users that install @command{mplayer} are very likely to
4196 install those codecs along with it. To recommend @command{mplayer}
4197 is, in effect, to promote use of the non-free codecs.
4199 Thus, you should not recommend programs that strongly encourage the
4200 use of non-free software. This is why we do not list
4201 @command{mplayer} in the Free Software Directory.
4203 A GNU package should not refer the user to any non-free documentation
4204 for free software. Free documentation that can be included in free
4205 operating systems is essential for completing the GNU system, or any
4206 free operating system, so encouraging it is a priority; to recommend
4207 use of documentation that we are not allowed to include undermines the
4208 impetus for the community to produce documentation that we can
4209 include. So GNU packages should never recommend non-free
4212 By contrast, it is ok to refer to journal articles and textbooks in
4213 the comments of a program for explanation of how it functions, even
4214 though they are non-free. This is because we don't include such
4215 things in the GNU system even if they are free---they are outside the
4216 scope of what a software distribution needs to include.
4218 Referring to a web site that describes or recommends a non-free
4219 program is promoting that program, so please do not make links to (or
4220 mention by name) web sites that contain such material. This policy is
4221 relevant particularly for the web pages for a GNU package.
4223 Following links from nearly any web site can lead eventually to
4224 non-free software; this is inherent in the nature of the web. So it
4225 makes no sense to criticize a site for having such links. As long as
4226 the site does not itself recommend a non-free program, there is no
4227 need to consider the question of the sites that it links to for other
4230 Thus, for example, you should not refer to AT&T's web site if that
4231 recommends AT&T's non-free software packages; you should not refer to
4232 a site that links to AT&T's site presenting it as a place to get some
4233 non-free program, because that link recommends and legitimizes the
4234 non-free program. However, that a site contains a link to AT&T's web
4235 site for some other purpose (such as long-distance telephone service)
4236 is not an objection against it.
4238 @node GNU Free Documentation License
4239 @appendix GNU Free Documentation License
4241 @cindex FDL, GNU Free Documentation License
4251 eval: (add-hook 'write-file-hooks 'time-stamp)
4252 time-stamp-start: "@set lastupdate "
4254 time-stamp-format: "%:b %:d, %:y"
4255 compile-command: "cd work.s && make"