1 \input texinfo @c -*- texinfo -*-
3 @settitle Developer Documentation
5 @center @titlefont{Developer Documentation}
12 @chapter Developers Guide
14 @section Notes for external developers
16 This document is mostly useful for internal FFmpeg developers.
17 External developers who need to use the API in their application should
18 refer to the API doxygen documentation in the public headers, and
19 check the examples in @file{doc/examples} and in the source code to
20 see how the public API is employed.
22 You can use the FFmpeg libraries in your commercial program, but you
23 are encouraged to @emph{publish any patch you make}. In this case the
24 best way to proceed is to send your patches to the ffmpeg-devel
25 mailing list following the guidelines illustrated in the remainder of
28 For more detailed legal information about the use of FFmpeg in
29 external programs read the @file{LICENSE} file in the source tree and
30 consult @url{http://ffmpeg.org/legal.html}.
34 There are 3 ways by which code gets into ffmpeg.
36 @item Submitting Patches to the main developer mailing list
37 see @ref{Submitting patches} for details.
38 @item Directly committing changes to the main tree.
39 @item Committing changes to a git clone, for example on github.com or
40 gitorious.org. And asking us to merge these changes.
43 Whichever way, changes should be reviewed by the maintainer of the code
44 before they are committed. And they should follow the @ref{Coding Rules}.
45 The developer making the commit and the author are responsible for their changes
46 and should try to fix issues their commit causes.
51 @subsection Code formatting conventions
53 There are the following guidelines regarding the indentation in files:
60 The TAB character is forbidden outside of Makefiles as is any
61 form of trailing whitespace. Commits containing either will be
62 rejected by the git repository.
65 You should try to limit your code lines to 80 characters; however, do so if
66 and only if this improves readability.
68 The presentation is one inspired by 'indent -i4 -kr -nut'.
70 The main priority in FFmpeg is simplicity and small code size in order to
71 minimize the bug count.
74 Use the JavaDoc/Doxygen format (see examples below) so that code documentation
75 can be generated automatically. All nontrivial functions should have a comment
76 above them explaining what the function does, even if it is just one sentence.
77 All structures and their member variables should be documented, too.
79 Avoid Qt-style and similar Doxygen syntax with @code{!} in it, i.e. replace
80 @code{//!} with @code{///} and similar. Also @@ syntax should be employed
81 for markup commands, i.e. use @code{@@param} and not @code{\param}.
95 typedef struct Foobar @{
96 int var1; /**< var1 description */
97 int var2; ///< var2 description
98 /** var3 description */
106 * @@param my_parameter description of my_parameter
107 * @@return return value description
109 int myfunc(int my_parameter)
113 @subsection C language features
115 FFmpeg is programmed in the ISO C90 language with a few additional
116 features from ISO C99, namely:
120 the @samp{inline} keyword;
126 designated struct initializers (@samp{struct s x = @{ .i = 17 @};})
129 compound literals (@samp{x = (struct s) @{ 17, 23 @};})
132 These features are supported by all compilers we care about, so we will not
133 accept patches to remove their use unless they absolutely do not impair
134 clarity and performance.
136 All code must compile with recent versions of GCC and a number of other
137 currently supported compilers. To ensure compatibility, please do not use
138 additional C99 features or GCC extensions. Especially watch out for:
142 mixing statements and declarations;
145 @samp{long long} (use @samp{int64_t} instead);
148 @samp{__attribute__} not protected by @samp{#ifdef __GNUC__} or similar;
151 GCC statement expressions (@samp{(x = (@{ int y = 4; y; @})}).
154 @subsection Naming conventions
155 All names should be composed with underscores (_), not CamelCase. For example,
156 @samp{avfilter_get_video_buffer} is an acceptable function name and
157 @samp{AVFilterGetVideo} is not. The exception from this are type names, like
158 for example structs and enums; they should always be in the CamelCase
160 There are the following conventions for naming variables and functions:
164 For local variables no prefix is required.
167 For file-scope variables and functions declared as @code{static}, no prefix
171 For variables and functions visible outside of file scope, but only used
172 internally by a library, an @code{ff_} prefix should be used,
173 e.g. @samp{ff_w64_demuxer}.
176 For variables and functions visible outside of file scope, used internally
177 across multiple libraries, use @code{avpriv_} as prefix, for example,
178 @samp{avpriv_aac_parse_header}.
181 Each library has its own prefix for public symbols, in addition to the
182 commonly used @code{av_} (@code{avformat_} for libavformat,
183 @code{avcodec_} for libavcodec, @code{swr_} for libswresample, etc).
184 Check the existing code and choose names accordingly.
185 Note that some symbols without these prefixes are also exported for
186 retro-compatibility reasons. These exceptions are declared in the
187 @code{lib<name>/lib<name>.v} files.
190 Furthermore, name space reserved for the system should not be invaded.
191 Identifiers ending in @code{_t} are reserved by
192 @url{http://pubs.opengroup.org/onlinepubs/007904975/functions/xsh_chap02_02.html#tag_02_02_02, POSIX}.
193 Also avoid names starting with @code{__} or @code{_} followed by an uppercase
194 letter as they are reserved by the C standard. Names starting with @code{_}
195 are reserved at the file level and may not be used for externally visible
196 symbols. If in doubt, just avoid names starting with @code{_} altogether.
198 @subsection Miscellaneous conventions
202 fprintf and printf are forbidden in libavformat and libavcodec,
203 please use av_log() instead.
206 Casts should be used only when necessary. Unneeded parentheses
207 should also be avoided if they don't make the code easier to understand.
210 @subsection Editor configuration
211 In order to configure Vim to follow FFmpeg formatting conventions, paste
212 the following snippet into your @file{.vimrc}:
214 " indentation rules for FFmpeg: 4 spaces, no tabs
220 " Allow tabs in Makefiles.
221 autocmd FileType make,automake set noexpandtab shiftwidth=8 softtabstop=8
222 " Trailing whitespace and tabs are forbidden, so highlight them.
223 highlight ForbiddenWhitespace ctermbg=red guibg=red
224 match ForbiddenWhitespace /\s\+$\|\t/
225 " Do not highlight spaces at the end of line while typing on that line.
226 autocmd InsertEnter * match ForbiddenWhitespace /\t\|\s\+\%#\@@<!$/
229 For Emacs, add these roughly equivalent lines to your @file{.emacs.d/init.el}:
231 (c-add-style "ffmpeg"
234 (indent-tabs-mode . nil)
235 (show-trailing-whitespace . t)
237 (statement-cont . (c-lineup-assignments +)))
240 (setq c-default-style "ffmpeg")
243 @section Development Policy
247 Contributions should be licensed under the
248 @uref{http://www.gnu.org/licenses/lgpl-2.1.html, LGPL 2.1},
249 including an "or any later version" clause, or, if you prefer
250 a gift-style license, the
251 @uref{http://opensource.org/licenses/isc-license.txt, ISC} or
252 @uref{http://mit-license.org/, MIT} license.
253 @uref{http://www.gnu.org/licenses/gpl-2.0.html, GPL 2} including
254 an "or any later version" clause is also acceptable, but LGPL is
256 If you add a new file, give it a proper license header. Do not copy and
257 paste it from a random place, use an existing file as template.
260 You must not commit code which breaks FFmpeg! (Meaning unfinished but
261 enabled code which breaks compilation or compiles but does not work or
262 breaks the regression tests)
263 You can commit unfinished stuff (for testing etc), but it must be disabled
264 (#ifdef etc) by default so it does not interfere with other developers'
268 The commit message should have a short first line in the form of
269 a @samp{topic: short description} as a header, separated by a newline
270 from the body consisting of an explanation of why the change is necessary.
271 If the commit fixes a known bug on the bug tracker, the commit message
272 should include its bug ID. Referring to the issue on the bug tracker does
273 not exempt you from writing an excerpt of the bug in the commit message.
276 You do not have to over-test things. If it works for you, and you think it
277 should work for others, then commit. If your code has problems
278 (portability, triggers compiler bugs, unusual environment etc) they will be
279 reported and eventually fixed.
282 Do not commit unrelated changes together, split them into self-contained
283 pieces. Also do not forget that if part B depends on part A, but A does not
284 depend on B, then A can and should be committed first and separate from B.
285 Keeping changes well split into self-contained parts makes reviewing and
286 understanding them on the commit log mailing list easier. This also helps
287 in case of debugging later on.
288 Also if you have doubts about splitting or not splitting, do not hesitate to
289 ask/discuss it on the developer mailing list.
292 Do not change behavior of the programs (renaming options etc) or public
293 API or ABI without first discussing it on the ffmpeg-devel mailing list.
294 Do not remove functionality from the code. Just improve!
296 Note: Redundant code can be removed.
299 Do not commit changes to the build system (Makefiles, configure script)
300 which change behavior, defaults etc, without asking first. The same
301 applies to compiler warning fixes, trivial looking fixes and to code
302 maintained by other developers. We usually have a reason for doing things
303 the way we do. Send your changes as patches to the ffmpeg-devel mailing
304 list, and if the code maintainers say OK, you may commit. This does not
305 apply to files you wrote and/or maintain.
308 We refuse source indentation and other cosmetic changes if they are mixed
309 with functional changes, such commits will be rejected and removed. Every
310 developer has his own indentation style, you should not change it. Of course
311 if you (re)write something, you can use your own style, even though we would
312 prefer if the indentation throughout FFmpeg was consistent (Many projects
313 force a given indentation style - we do not.). If you really need to make
314 indentation changes (try to avoid this), separate them strictly from real
317 NOTE: If you had to put if()@{ .. @} over a large (> 5 lines) chunk of code,
318 then either do NOT change the indentation of the inner part within (do not
319 move it to the right)! or do so in a separate commit
322 Always fill out the commit log message. Describe in a few lines what you
323 changed and why. You can refer to mailing list postings if you fix a
324 particular bug. Comments such as "fixed!" or "Changed it." are unacceptable.
326 area changed: Short 1 line description
328 details describing what and why and giving references.
331 Make sure the author of the commit is set correctly. (see git commit --author)
332 If you apply a patch, send an
333 answer to ffmpeg-devel (or wherever you got the patch from) saying that
334 you applied the patch.
337 When applying patches that have been discussed (at length) on the mailing
338 list, reference the thread in the log message.
341 Do NOT commit to code actively maintained by others without permission.
342 Send a patch to ffmpeg-devel instead. If no one answers within a reasonable
343 timeframe (12h for build failures and security fixes, 3 days small changes,
344 1 week for big patches) then commit your patch if you think it is OK.
345 Also note, the maintainer can simply ask for more time to review!
348 Subscribe to the ffmpeg-cvslog mailing list. The diffs of all commits
349 are sent there and reviewed by all the other developers. Bugs and possible
350 improvements or general questions regarding commits are discussed there. We
351 expect you to react if problems with your code are uncovered.
354 Update the documentation if you change behavior or add features. If you are
355 unsure how best to do this, send a patch to ffmpeg-devel, the documentation
356 maintainer(s) will review and commit your stuff.
359 Try to keep important discussions and requests (also) on the public
360 developer mailing list, so that all developers can benefit from them.
363 Never write to unallocated memory, never write over the end of arrays,
364 always check values read from some untrusted source before using them
365 as array index or other risky things.
368 Remember to check if you need to bump versions for the specific libav*
369 parts (libavutil, libavcodec, libavformat) you are changing. You need
370 to change the version integer.
371 Incrementing the first component means no backward compatibility to
372 previous versions (e.g. removal of a function from the public API).
373 Incrementing the second component means backward compatible change
374 (e.g. addition of a function to the public API or extension of an
375 existing data structure).
376 Incrementing the third component means a noteworthy binary compatible
377 change (e.g. encoder bug fix that matters for the decoder). The third
378 component always starts at 100 to distinguish FFmpeg from Libav.
381 Compiler warnings indicate potential bugs or code with bad style. If a type of
382 warning always points to correct and clean code, that warning should
383 be disabled, not the code changed.
384 Thus the remaining warnings can either be bugs or correct code.
385 If it is a bug, the bug has to be fixed. If it is not, the code should
386 be changed to not generate a warning unless that causes a slowdown
387 or obfuscates the code.
390 Make sure that no parts of the codebase that you maintain are missing from the
391 @file{MAINTAINERS} file. If something that you want to maintain is missing add it with
393 If at some point you no longer want to maintain some code, then please help
394 finding a new maintainer and also don't forget updating the @file{MAINTAINERS} file.
397 We think our rules are not too hard. If you have comments, contact us.
399 @anchor{Submitting patches}
400 @section Submitting patches
402 First, read the @ref{Coding Rules} above if you did not yet, in particular
403 the rules regarding patch submission.
405 When you submit your patch, please use @code{git format-patch} or
406 @code{git send-email}. We cannot read other diffs :-)
408 Also please do not submit a patch which contains several unrelated changes.
409 Split it into separate, self-contained pieces. This does not mean splitting
410 file by file. Instead, make the patch as small as possible while still
411 keeping it as a logical unit that contains an individual change, even
412 if it spans multiple files. This makes reviewing your patches much easier
413 for us and greatly increases your chances of getting your patch applied.
415 Use the patcheck tool of FFmpeg to check your patch.
416 The tool is located in the tools directory.
418 Run the @ref{Regression tests} before submitting a patch in order to verify
419 it does not cause unexpected problems.
421 It also helps quite a bit if you tell us what the patch does (for example
422 'replaces lrint by lrintf'), and why (for example '*BSD isn't C99 compliant
425 Also please if you send several patches, send each patch as a separate mail,
426 do not attach several unrelated patches to the same mail.
428 Patches should be posted to the
429 @uref{http://lists.ffmpeg.org/mailman/listinfo/ffmpeg-devel, ffmpeg-devel}
430 mailing list. Use @code{git send-email} when possible since it will properly
431 send patches without requiring extra care. If you cannot, then send patches
432 as base64-encoded attachments, so your patch is not trashed during
435 Your patch will be reviewed on the mailing list. You will likely be asked
436 to make some changes and are expected to send in an improved version that
437 incorporates the requests from the review. This process may go through
438 several iterations. Once your patch is deemed good enough, some developer
439 will pick it up and commit it to the official FFmpeg tree.
441 Give us a few days to react. But if some time passes without reaction,
442 send a reminder by email. Your patch should eventually be dealt with.
445 @section New codecs or formats checklist
449 Did you use av_cold for codec initialization and close functions?
452 Did you add a long_name under NULL_IF_CONFIG_SMALL to the AVCodec or
453 AVInputFormat/AVOutputFormat struct?
456 Did you bump the minor version number (and reset the micro version
457 number) in @file{libavcodec/version.h} or @file{libavformat/version.h}?
460 Did you register it in @file{allcodecs.c} or @file{allformats.c}?
463 Did you add the AVCodecID to @file{avcodec.h}?
464 When adding new codec IDs, also add an entry to the codec descriptor
465 list in @file{libavcodec/codec_desc.c}.
468 If it has a FourCC, did you add it to @file{libavformat/riff.c},
469 even if it is only a decoder?
472 Did you add a rule to compile the appropriate files in the Makefile?
473 Remember to do this even if you're just adding a format to a file that is
474 already being compiled by some other rule, like a raw demuxer.
477 Did you add an entry to the table of supported formats or codecs in
478 @file{doc/general.texi}?
481 Did you add an entry in the Changelog?
484 If it depends on a parser or a library, did you add that dependency in
488 Did you @code{git add} the appropriate files before committing?
491 Did you make sure it compiles standalone, i.e. with
492 @code{configure --disable-everything --enable-decoder=foo}
493 (or @code{--enable-demuxer} or whatever your component is)?
497 @section patch submission checklist
501 Does @code{make fate} pass with the patch applied?
504 Was the patch generated with git format-patch or send-email?
507 Did you sign off your patch? (git commit -s)
508 See @url{http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=blob_plain;f=Documentation/SubmittingPatches} for the meaning
512 Did you provide a clear git commit log message?
515 Is the patch against latest FFmpeg git master branch?
518 Are you subscribed to ffmpeg-devel?
519 (the list is subscribers only due to spam)
522 Have you checked that the changes are minimal, so that the same cannot be
523 achieved with a smaller patch and/or simpler final code?
526 If the change is to speed critical code, did you benchmark it?
529 If you did any benchmarks, did you provide them in the mail?
532 Have you checked that the patch does not introduce buffer overflows or
533 other security issues?
536 Did you test your decoder or demuxer against damaged data? If no, see
537 tools/trasher, the noise bitstream filter, and
538 @uref{http://caca.zoy.org/wiki/zzuf, zzuf}. Your decoder or demuxer
539 should not crash, end in a (near) infinite loop, or allocate ridiculous
540 amounts of memory when fed damaged data.
543 Does the patch not mix functional and cosmetic changes?
546 Did you add tabs or trailing whitespace to the code? Both are forbidden.
549 Is the patch attached to the email you send?
552 Is the mime type of the patch correct? It should be text/x-diff or
553 text/x-patch or at least text/plain and not application/octet-stream.
556 If the patch fixes a bug, did you provide a verbose analysis of the bug?
559 If the patch fixes a bug, did you provide enough information, including
560 a sample, so the bug can be reproduced and the fix can be verified?
561 Note please do not attach samples >100k to mails but rather provide a
562 URL, you can upload to ftp://upload.ffmpeg.org
565 Did you provide a verbose summary about what the patch does change?
568 Did you provide a verbose explanation why it changes things like it does?
571 Did you provide a verbose summary of the user visible advantages and
572 disadvantages if the patch is applied?
575 Did you provide an example so we can verify the new feature added by the
579 If you added a new file, did you insert a license header? It should be
580 taken from FFmpeg, not randomly copied and pasted from somewhere else.
583 You should maintain alphabetical order in alphabetically ordered lists as
584 long as doing so does not break API/ABI compatibility.
587 Lines with similar content should be aligned vertically when doing so
588 improves readability.
591 Consider to add a regression test for your code.
594 If you added YASM code please check that things still work with --disable-yasm
597 Make sure you check the return values of function and return appropriate
598 error codes. Especially memory allocation functions like @code{av_malloc()}
599 are notoriously left unchecked, which is a serious problem.
602 Test your code with valgrind and or Address Sanitizer to ensure it's free
603 of leaks, out of array accesses, etc.
606 @section Patch review process
608 All patches posted to ffmpeg-devel will be reviewed, unless they contain a
609 clear note that the patch is not for the git master branch.
610 Reviews and comments will be posted as replies to the patch on the
611 mailing list. The patch submitter then has to take care of every comment,
612 that can be by resubmitting a changed patch or by discussion. Resubmitted
613 patches will themselves be reviewed like any other patch. If at some point
614 a patch passes review with no comments then it is approved, that can for
615 simple and small patches happen immediately while large patches will generally
616 have to be changed and reviewed many times before they are approved.
617 After a patch is approved it will be committed to the repository.
619 We will review all submitted patches, but sometimes we are quite busy so
620 especially for large patches this can take several weeks.
622 If you feel that the review process is too slow and you are willing to try to
623 take over maintainership of the area of code you change then just clone
624 git master and maintain the area of code there. We will merge each area from
625 where its best maintained.
627 When resubmitting patches, please do not make any significant changes
628 not related to the comments received during review. Such patches will
629 be rejected. Instead, submit significant changes or new features as
632 @anchor{Regression tests}
633 @section Regression tests
635 Before submitting a patch (or committing to the repository), you should at least
636 test that you did not break anything.
638 Running 'make fate' accomplishes this, please see @url{fate.html} for details.
640 [Of course, some patches may change the results of the regression tests. In
641 this case, the reference results of the regression tests shall be modified
644 @subsection Adding files to the fate-suite dataset
646 When there is no muxer or encoder available to generate test media for a
647 specific test then the media has to be inlcuded in the fate-suite.
648 First please make sure that the sample file is as small as possible to test the
649 respective decoder or demuxer sufficiently. Large files increase network
650 bandwidth and disk space requirements.
651 Once you have a working fate test and fate sample, provide in the commit
652 message or introductionary message for the patch series that you post to
653 the ffmpeg-devel mailing list, a direct link to download the sample media.
656 @subsection Visualizing Test Coverage
658 The FFmpeg build system allows visualizing the test coverage in an easy
659 manner with the coverage tools @code{gcov}/@code{lcov}. This involves
664 Configure to compile with instrumentation enabled:
665 @code{configure --toolchain=gcov}.
668 Run your test case, either manually or via FATE. This can be either
669 the full FATE regression suite, or any arbitrary invocation of any
670 front-end tool provided by FFmpeg, in any combination.
673 Run @code{make lcov} to generate coverage data in HTML format.
676 View @code{lcov/index.html} in your preferred HTML viewer.
679 You can use the command @code{make lcov-reset} to reset the coverage
680 measurements. You will need to rerun @code{make lcov} after running a
683 @subsection Using Valgrind
685 The configure script provides a shortcut for using valgrind to spot bugs
686 related to memory handling. Just add the option
687 @code{--toolchain=valgrind-memcheck} or @code{--toolchain=valgrind-massif}
688 to your configure line, and reasonable defaults will be set for running
689 FATE under the supervision of either the @strong{memcheck} or the
690 @strong{massif} tool of the valgrind suite.
692 In case you need finer control over how valgrind is invoked, use the
693 @code{--target-exec='valgrind <your_custom_valgrind_options>} option in
694 your configure line instead.
696 @anchor{Release process}
697 @section Release process
699 FFmpeg maintains a set of @strong{release branches}, which are the
700 recommended deliverable for system integrators and distributors (such as
701 Linux distributions, etc.). At regular times, a @strong{release
702 manager} prepares, tests and publishes tarballs on the
703 @url{http://ffmpeg.org} website.
705 There are two kinds of releases:
709 @strong{Major releases} always include the latest and greatest
710 features and functionality.
713 @strong{Point releases} are cut from @strong{release} branches,
714 which are named @code{release/X}, with @code{X} being the release
718 Note that we promise to our users that shared libraries from any FFmpeg
719 release never break programs that have been @strong{compiled} against
720 previous versions of @strong{the same release series} in any case!
722 However, from time to time, we do make API changes that require adaptations
723 in applications. Such changes are only allowed in (new) major releases and
724 require further steps such as bumping library version numbers and/or
725 adjustments to the symbol versioning file. Please discuss such changes
726 on the @strong{ffmpeg-devel} mailing list in time to allow forward planning.
728 @anchor{Criteria for Point Releases}
729 @subsection Criteria for Point Releases
731 Changes that match the following criteria are valid candidates for
732 inclusion into a point release:
736 Fixes a security issue, preferably identified by a @strong{CVE
737 number} issued by @url{http://cve.mitre.org/}.
740 Fixes a documented bug in @url{https://trac.ffmpeg.org}.
743 Improves the included documentation.
746 Retains both source code and binary compatibility with previous
747 point releases of the same release branch.
750 The order for checking the rules is (1 OR 2 OR 3) AND 4.
753 @subsection Release Checklist
755 The release process involves the following steps:
759 Ensure that the @file{RELEASE} file contains the version number for
760 the upcoming release.
763 Add the release at @url{https://trac.ffmpeg.org/admin/ticket/versions}.
766 Announce the intent to do a release to the mailing list.
769 Make sure all relevant security fixes have been backported. See
770 @url{https://ffmpeg.org/security.html}.
773 Ensure that the FATE regression suite still passes in the release
774 branch on at least @strong{i386} and @strong{amd64}
775 (cf. @ref{Regression tests}).
778 Prepare the release tarballs in @code{bz2} and @code{gz} formats, and
779 supplementing files that contain @code{gpg} signatures
782 Publish the tarballs at @url{http://ffmpeg.org/releases}. Create and
783 push an annotated tag in the form @code{nX}, with @code{X}
784 containing the version number.
787 Propose and send a patch to the @strong{ffmpeg-devel} mailing list
788 with a news entry for the website.
791 Publish the news entry.
794 Send announcement to the mailing list.