1 \input texinfo @c -*- texinfo -*-
3 @settitle Developer Documentation
6 @center @titlefont{Developer Documentation}
11 @chapter Developers Guide
15 @item libavcodec is the library containing the codecs (both encoding and
16 decoding). Look at @file{libavcodec/apiexample.c} to see how to use it.
18 @item libavformat is the library containing the file format handling (mux and
19 demux code for several formats). Look at @file{ffplay.c} to use it in a
20 player. See @file{libavformat/output-example.c} to use it to generate
21 audio or video streams.
25 @section Integrating libavcodec or libavformat in your program
27 You can integrate all the source code of the libraries to link them
28 statically to avoid any version problem. All you need is to provide a
29 'config.mak' and a 'config.h' in the parent directory. See the defines
30 generated by ./configure to understand what is needed.
32 You can use libavcodec or libavformat in your commercial program, but
33 @emph{any patch you make must be published}. The best way to proceed is
34 to send your patches to the FFmpeg mailing list.
39 FFmpeg is programmed in the ISO C90 language with a few additional
40 features from ISO C99, namely:
43 the @samp{inline} keyword;
47 designated struct initializers (@samp{struct s x = @{ .i = 17 @};})
49 compound literals (@samp{x = (struct s) @{ 17, 23 @};})
52 These features are supported by all compilers we care about, so we will not
53 accept patches to remove their use unless they absolutely do not impair
54 clarity and performance.
56 All code must compile with GCC 2.95 and GCC 3.3. Currently, FFmpeg also
57 compiles with several other compilers, such as the Compaq ccc compiler
58 or Sun Studio 9, and we would like to keep it that way unless it would
59 be exceedingly involved. To ensure compatibility, please do not use any
60 additional C99 features or GCC extensions. Especially watch out for:
63 mixing statements and declarations;
65 @samp{long long} (use @samp{int64_t} instead);
67 @samp{__attribute__} not protected by @samp{#ifdef __GNUC__} or similar;
69 GCC statement expressions (@samp{(x = (@{ int y = 4; y; @})}).
73 The presentation is one inspired by 'indent -i4 -kr -nut'.
74 The TAB character is forbidden outside of Makefiles as is any
75 form of trailing whitespace. Commits containing either will be
76 rejected by the Subversion repository.
78 The main priority in FFmpeg is simplicity and small code size in order to
79 minimize the bug count.
81 Comments: Use the JavaDoc/Doxygen
82 format (see examples below) so that code documentation
83 can be generated automatically. All nontrivial functions should have a comment
84 above them explaining what the function does, even if it is just one sentence.
85 All structures and their member variables should be documented, too.
86 Impersonal form must be used for the function and parameter
87 descriptions, e.g. "Set the bikeshed color." is favored over "Sets the
101 typedef struct Foobar@{
102 int var1; /**< var1 description */
103 int var2; ///< var2 description
104 /** var3 description */
112 * @@param my_parameter description of my_parameter
113 * @@return return value description
115 int myfunc(int my_parameter)
119 fprintf and printf are forbidden in libavformat and libavcodec,
120 please use av_log() instead.
122 Casts should be used only when necessary. Unneeded parentheses
123 should also be avoided if they don't make the code easier to understand.
125 @section Development Policy
129 Contributions should be licensed under the LGPL 2.1, including an
130 "or any later version" clause, or the MIT license. GPL 2 including
131 an "or any later version" clause is also acceptable, but LGPL is
134 You must not commit code which breaks FFmpeg! (Meaning unfinished but
135 enabled code which breaks compilation or compiles but does not work or
136 breaks the regression tests)
137 You can commit unfinished stuff (for testing etc), but it must be disabled
138 (#ifdef etc) by default so it does not interfere with other developers'
141 You do not have to over-test things. If it works for you, and you think it
142 should work for others, then commit. If your code has problems
143 (portability, triggers compiler bugs, unusual environment etc) they will be
144 reported and eventually fixed.
146 Do not commit unrelated changes together, split them into self-contained
147 pieces. Also do not forget that if part B depends on part A, but A does not
148 depend on B, then A can and should be committed first and separate from B.
149 Keeping changes well split into self-contained parts makes reviewing and
150 understanding them on the commit log mailing list easier. This also helps
151 in case of debugging later on.
152 Also if you have doubts about splitting or not splitting, do not hesitate to
153 ask/discuss it on the developer mailing list.
155 Do not change behavior of the program (renaming options etc) without
156 first discussing it on the ffmpeg-devel mailing list. Do not remove
157 functionality from the code. Just improve!
159 Note: Redundant code can be removed.
161 Do not commit changes to the build system (Makefiles, configure script)
162 which change behavior, defaults etc, without asking first. The same
163 applies to compiler warning fixes, trivial looking fixes and to code
164 maintained by other developers. We usually have a reason for doing things
165 the way we do. Send your changes as patches to the ffmpeg-devel mailing
166 list, and if the code maintainers say OK, you may commit. This does not
167 apply to files you wrote and/or maintain.
169 We refuse source indentation and other cosmetic changes if they are mixed
170 with functional changes, such commits will be rejected and removed. Every
171 developer has his own indentation style, you should not change it. Of course
172 if you (re)write something, you can use your own style, even though we would
173 prefer if the indentation throughout FFmpeg was consistent (Many projects
174 force a given indentation style - we do not.). If you really need to make
175 indentation changes (try to avoid this), separate them strictly from real
178 NOTE: If you had to put if()@{ .. @} over a large (> 5 lines) chunk of code,
179 then either do NOT change the indentation of the inner part within (do not
180 move it to the right)! or do so in a separate commit
182 Always fill out the commit log message. Describe in a few lines what you
183 changed and why. You can refer to mailing list postings if you fix a
184 particular bug. Comments such as "fixed!" or "Changed it." are unacceptable.
186 If you apply a patch by someone else, include the name and email address in
187 the log message. Since the ffmpeg-cvslog mailing list is publicly
188 archived you should add some SPAM protection to the email address. Send an
189 answer to ffmpeg-devel (or wherever you got the patch from) saying that
190 you applied the patch.
192 When applying patches that have been discussed (at length) on the mailing
193 list, reference the thread in the log message.
195 Do NOT commit to code actively maintained by others without permission.
196 Send a patch to ffmpeg-devel instead. If no one answers within a reasonable
197 timeframe (12h for build failures and security fixes, 3 days small changes,
198 1 week for big patches) then commit your patch if you think it is OK.
199 Also note, the maintainer can simply ask for more time to review!
201 Subscribe to the ffmpeg-cvslog mailing list. The diffs of all commits
202 are sent there and reviewed by all the other developers. Bugs and possible
203 improvements or general questions regarding commits are discussed there. We
204 expect you to react if problems with your code are uncovered.
206 Update the documentation if you change behavior or add features. If you are
207 unsure how best to do this, send a patch to ffmpeg-devel, the documentation
208 maintainer(s) will review and commit your stuff.
210 Try to keep important discussions and requests (also) on the public
211 developer mailing list, so that all developers can benefit from them.
213 Never write to unallocated memory, never write over the end of arrays,
214 always check values read from some untrusted source before using them
215 as array index or other risky things.
217 Remember to check if you need to bump versions for the specific libav
218 parts (libavutil, libavcodec, libavformat) you are changing. You need
219 to change the version integer.
220 Incrementing the first component means no backward compatibility to
221 previous versions (e.g. removal of a function from the public API).
222 Incrementing the second component means backward compatible change
223 (e.g. addition of a function to the public API or extension of an
224 existing data structure).
225 Incrementing the third component means a noteworthy binary compatible
226 change (e.g. encoder bug fix that matters for the decoder).
228 Compiler warnings indicate potential bugs or code with bad style. If a type of
229 warning always points to correct and clean code, that warning should
230 be disabled, not the code changed.
231 Thus the remaining warnings can either be bugs or correct code.
232 If it is a bug, the bug has to be fixed. If it is not, the code should
233 be changed to not generate a warning unless that causes a slowdown
234 or obfuscates the code.
236 If you add a new file, give it a proper license header. Do not copy and
237 paste it from a random place, use an existing file as template.
240 We think our rules are not too hard. If you have comments, contact us.
242 Note, these rules are mostly borrowed from the MPlayer project.
244 @section Submitting patches
246 First, (@pxref{Coding Rules}) above if you did not yet.
248 When you submit your patch, try to send a unified diff (diff '-up'
249 option). We cannot read other diffs :-)
251 Also please do not submit a patch which contains several unrelated changes.
252 Split it into separate, self-contained pieces. This does not mean splitting
253 file by file. Instead, make the patch as small as possible while still
254 keeping it as a logical unit that contains an individual change, even
255 if it spans multiple files. This makes reviewing your patches much easier
256 for us and greatly increases your chances of getting your patch applied.
258 Use the patcheck tool of FFmpeg to check your patch.
259 The tool is located in the tools directory.
261 Run the regression tests before submitting a patch so that you can
262 verify that there are no big problems.
264 Patches should be posted as base64 encoded attachments (or any other
265 encoding which ensures that the patch will not be trashed during
266 transmission) to the ffmpeg-devel mailing list, see
267 @url{http://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-devel}
269 It also helps quite a bit if you tell us what the patch does (for example
270 'replaces lrint by lrintf'), and why (for example '*BSD isn't C99 compliant
273 Also please if you send several patches, send each patch as a separate mail,
274 do not attach several unrelated patches to the same mail.
276 Your patch will be reviewed on the mailing list. You will likely be asked
277 to make some changes and are expected to send in an improved version that
278 incorporates the requests from the review. This process may go through
279 several iterations. Once your patch is deemed good enough, some developer
280 will pick it up and commit it to the official FFmpeg tree.
282 Give us a few days to react. But if some time passes without reaction,
283 send a reminder by email. Your patch should eventually be dealt with.
286 @section New codecs or formats checklist
290 Did you use av_cold for codec initialization and close functions?
292 Did you add a long_name under NULL_IF_CONFIG_SMALL to the AVCodec or
293 AVInputFormat/AVOutputFormat struct?
295 Did you bump the minor version number in @file{avcodec.h} or
298 Did you register it in @file{allcodecs.c} or @file{allformats.c}?
300 Did you add the CodecID to @file{avcodec.h}?
302 If it has a fourcc, did you add it to @file{libavformat/riff.c},
303 even if it is only a decoder?
305 Did you add a rule to compile the appropriate files in the Makefile?
306 Remember to do this even if you're just adding a format to a file that is
307 already being compiled by some other rule, like a raw demuxer.
309 Did you add an entry to the table of supported formats or codecs in
310 @file{doc/general.texi}?
312 Did you add an entry in the Changelog?
314 If it depends on a parser or a library, did you add that dependency in
317 Did you "svn add" the appropriate files before commiting?
320 @section patch submission checklist
324 Do the regression tests pass with the patch applied?
326 Does @code{make checkheaders} pass with the patch applied?
328 Is the patch a unified diff?
330 Is the patch against latest FFmpeg SVN?
332 Are you subscribed to ffmpeg-dev?
333 (the list is subscribers only due to spam)
335 Have you checked that the changes are minimal, so that the same cannot be
336 achieved with a smaller patch and/or simpler final code?
338 If the change is to speed critical code, did you benchmark it?
340 If you did any benchmarks, did you provide them in the mail?
342 Have you checked that the patch does not introduce buffer overflows or
343 other security issues?
345 Did you test your decoder or demuxer against damaged data? If no, see
346 tools/trasher and the noise bitstream filter. Your decoder or demuxer
347 should not crash or end in a (near) infinite loop when fed damaged data.
349 Is the patch created from the root of the source tree, so it can be
350 applied with @code{patch -p0}?
352 Does the patch not mix functional and cosmetic changes?
354 Did you add tabs or trailing whitespace to the code? Both are forbidden.
356 Is the patch attached to the email you send?
358 Is the mime type of the patch correct? It should be text/x-diff or
359 text/x-patch or at least text/plain and not application/octet-stream.
361 If the patch fixes a bug, did you provide a verbose analysis of the bug?
363 If the patch fixes a bug, did you provide enough information, including
364 a sample, so the bug can be reproduced and the fix can be verified?
365 Note please do not attach samples >100k to mails but rather provide a
366 URL, you can upload to ftp://upload.ffmpeg.org
368 Did you provide a verbose summary about what the patch does change?
370 Did you provide a verbose explanation why it changes things like it does?
372 Did you provide a verbose summary of the user visible advantages and
373 disadvantages if the patch is applied?
375 Did you provide an example so we can verify the new feature added by the
378 If you added a new file, did you insert a license header? It should be
379 taken from FFmpeg, not randomly copied and pasted from somewhere else.
381 You should maintain alphabetical order in alphabetically ordered lists as
382 long as doing so does not break API/ABI compatibility.
384 Lines with similar content should be aligned vertically when doing so
385 improves readability.
387 Did you provide a suggestion for a clear commit log message?
390 @section Patch review process
392 All patches posted to ffmpeg-devel will be reviewed, unless they contain a
393 clear note that the patch is not for SVN.
394 Reviews and comments will be posted as replies to the patch on the
395 mailing list. The patch submitter then has to take care of every comment,
396 that can be by resubmitting a changed patch or by discussion. Resubmitted
397 patches will themselves be reviewed like any other patch. If at some point
398 a patch passes review with no comments then it is approved, that can for
399 simple and small patches happen immediately while large patches will generally
400 have to be changed and reviewed many times before they are approved.
401 After a patch is approved it will be committed to the repository.
403 We will review all submitted patches, but sometimes we are quite busy so
404 especially for large patches this can take several weeks.
406 When resubmitting patches, please do not make any significant changes
407 not related to the comments received during review. Such patches will
408 be rejected. Instead, submit significant changes or new features as
411 @section Regression tests
413 Before submitting a patch (or committing to the repository), you should at least
414 test that you did not break anything.
416 The regression tests build a synthetic video stream and a synthetic
417 audio stream. These are then encoded and decoded with all codecs or
418 formats. The CRC (or MD5) of each generated file is recorded in a
419 result file. A 'diff' is launched to compare the reference results and
420 the result file. The output is checked immediately after each test
423 The regression tests then go on to test the FFserver code with a
424 limited set of streams. It is important that this step runs correctly
427 Run 'make test' to test all the codecs and formats. Commands like
428 'make regtest-mpeg2' can be used to run a single test. By default,
429 make will abort if any test fails. To run all tests regardless,
430 use make -k. To get a more verbose output, use 'make V=1 test' or
433 Run 'make fulltest' to test all the codecs, formats and FFserver.
435 [Of course, some patches may change the results of the regression tests. In
436 this case, the reference results of the regression tests shall be modified