[Ffmpeg-devel-irc] ffmpeg-devel.log.20140505

burek burek021 at gmail.com
Tue May 6 02:05:02 CEST 2014


[00:09] <cone-944> ffmpeg.git 03Anton Khirnov 07release/1.1:696a74e81560: matroska: add the Opus mapping
[00:09] <cone-944> ffmpeg.git 03Luca Barbato 07release/1.1:9a2387cfb1a9: swscale: Fix an undefined behaviour
[00:09] <cone-944> ffmpeg.git 03Reinhard Tartler 07release/1.1:1f5678c5e404: Update Changelog for v9.13
[00:09] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/1.1:227ac71a6093: Merge commit '696a74e81560889ef2fde76b045c5158a325a7a5' into release/1.1
[00:09] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/1.1:0597f052372b: Merge commit '9a2387cfb1a99e4b00593c22074929cdc73410f0' into release/1.1
[00:09] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/1.1:ebd74c42a259: Merge commit '1f5678c5e404ebd326f6eae1b527beadaa85a528' into release/1.1
[00:35] <ubitux> JEEBsv: http://coverage.ffmpeg.org/
[00:36] <JEEB> :o
[00:36] <JEEB> nice
[00:42] <cone-944> ffmpeg.git 03Stephan Soller 07release/2.2:e3b08b3ad4da: avformat/rtmpproto: Added handling of an initial RTMP chunk size packet.
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:366cdd35484c: avcodec/x86/idct_sse2_xvid: fix non C99 inline function
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:26becbcd2a15: swscale/swscale: fix srcStride/srcSlice typo
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:1131e7a1a4c9: avcodec/mjpegen: Fix declared argument size
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:5191b00155b5: avcodec/dcadec: fix error message suppression code
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:8ab849cddca1: avcodec/fic: Fix return value check
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:27a3a59428e7: avformat/mpegts: Remove redundant check
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:10e023c4fa73: avcodec/diracdec: fix undefined behavior with shifts
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:fc5b32877a47: avcodec/g723_1: add assert to help static code analyzers
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:83fb31a76da1: avfilter/f_select: fix loss of precission in SAD calculation
[00:42] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:16a9c5ea9e68: avfilter/vf_deshake: fix loss of precission with odd resolutions
[00:43] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:afd1f6194447: iavcodec/vc1dec: Fix missing {}
[00:43] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:c588316555f9: avfilter/filtfmts: Support dynamically allocated in/outputs
[00:43] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:f1b583018208: avformat/h263dec: Fix h263 probe
[00:43] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:2d18e7f3ef05: avcodec/mjpegdec: Fix undefined shift
[00:43] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:56436683085b: ffmpeg_filter: fix pointer to local outside scope
[00:43] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:02b7b125b522: avfilter/graphdump: Fix pointer to local outside scope
[00:43] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:9d0ff6436ef7: ffmpeg: print an error at the end if conversion failed
[00:43] <cone-944> ffmpeg.git 03Lukasz Marek 07release/2.2:c9c223ba0028: lavd/opengl_enc: fix pixel data alignment
[00:49] <cone-944> ffmpeg.git 03Michael Niedermayer 07release/2.2:c2eb66861755: Update for 2.2.2
[01:49] <BBB> michaelni: let's not be silly, the maintainers file is kinda retarded, you don't need a file that lists me as blocker for any file to know that if you have issues with vp8/9, you can ask me and I'll try to help
[01:50] <BBB> michaelni: I'll leave it up to you to consider what the best way forward for mdvc support is - I think derek hosts the fate instances so you have a fair clue of who's the best person to ask
[01:51] <michaelni> i dont need the file to know it for the vp8/9 case but other people wont know it so easily whom to contact
[01:53] <BBB> h264: would list 20 people?
[01:53] <BBB> (since there's about 20 people with relevant knowledge of that file)
[01:56] <Skyler_> you could list separately some main people and secondary people?
[01:56] <michaelni> id say h264 should list whoever is willing to actually sit down and fix an issue/bug in h264 if one is reported or who is able and willing to review related patches
[01:56] <Skyler_> yeah, that's a point
[01:56] <Skyler_> people shouldn't be put on mainters list if they're not willing to look at things/help
[01:56] <Skyler_> *maintainers
[01:58] <Compnn> theres 20 people who know h264
[01:58] <Compnn> but only a few who would say they want to maintain it
[01:58] <Compnn> i think Daemon404 was offended when patches against his code were committed before asking him :p
[01:59] <Compnn> BBB : if you dont use/like maintainers file, it wont bug you too much, right? 
[01:59] <Skyler_> the three levels of maintainers: "don't poke me, ask me if you need help, I'll-stab-you-if-you-touch-my-code" ;-)
[02:00] <Compnn> lol
[02:02] <BBB> true opensource spirit ;)
[02:02] <BBB> fine I'll stop whining
[02:07] <Daemon404> Skyler_, my level is "please let me review it"
[02:07] <Daemon404> or
[02:08] <Daemon404> "at least send it to the mailing list instead of pushing right away"
[02:43] <iive> Daemon404: no, your level is "I'll stab you if you touch my code"
[02:56] <Daemon404> evidence where?
[02:57] <Daemon404> pretty sure ive only herped a derp when unreviewed changes have been pushed to it.
[03:00] <iive> that's it.
[03:00] <Daemon404> thats certanly not the same thing
[03:01] <iive> if that is not the same thing, what would it be then?
[03:01] <Daemon404> dont touch my code != dont push withut review
[03:01] <iive> without YOUR review
[03:01] <Daemon404> where did i say that
[03:02] <Daemon404> the stuff i bitch about isnt even sent to the ML
[03:02] <Daemon404> for ANY review
[03:02] <Daemon404> thats my point.
[03:02] <iive> and this is not libav
[03:02] <Daemon404> ....
[03:02] <Daemon404> what kind of retarded argument is that
[03:02] <iive> well, libav forked over michaelni committing stuff directly.
[03:02] <iive> if you have problem with that...
[03:03] <Daemon404> well, ok
[03:16] <Compn> mean ole iive :p
[03:24] <iive> :E
[03:24] <Skyler_> geez   touchy
[03:38] <iive> n8 ppl.
[05:01] <cone-973> ffmpeg.git 03Hendrik Leppkes 07master:ae8e2d7634e0: avisynth: set duration of audio streams
[06:03] <cone-973> ffmpeg.git 03Michael Niedermayer 07master:ebf2c2c3a818: avcodec/lossless_videodsp: fix incompatible pointer type warning
[06:03] <cone-973> ffmpeg.git 03Michael Niedermayer 07master:6bd1741231d3: avfilter/f_select: fix incompatible pointer type warning
[06:54] <cone-973> ffmpeg.git 03Michael Niedermayer 07fatal: ambiguous argument 'refs/tags/n2.2.2': unknown revision or path not in the working tree.
[06:54] <cone-973> Use '--' to separate paths from revisions
[06:54] <cone-973> refs/tags/n2.2.2:HEAD: avfilter/f_select: fix incompatible pointer type warning
[09:14] <cone-973> ffmpeg.git 03Carl Eugen Hoyos 07master:db01de4c0af8: Fix standalone compilation of sonic lossless encoder.
[14:44] <cone-387> ffmpeg.git 03Anton Khirnov 07master:73ed8558a61c: avconv: simplify exit_program() by using more local vars
[14:44] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:6e397074380e: Merge commit '73ed8558a61c4f1d0ce8d1f35183ac1a0a4e31e3'
[14:54] <cone-387> ffmpeg.git 03Anton Khirnov 07master:f3a6ad22c977: avconv: rename InputStream.opts to InputStream.decoder_opts
[14:55] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:42b7bda41a4f: Merge commit 'f3a6ad22c977c8fea87c4c5e4535061db6324d66'
[15:01] <cone-387> ffmpeg.git 03Anton Khirnov 07master:650a5206a944: avconv: rename 'icodec' to 'dec_ctx'
[15:01] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:1f40536ccaea: Merge commit '650a5206a944fd36a42d92f98e745972056b1f7c'
[15:10] <cone-387> ffmpeg.git 03Anton Khirnov 07master:65ed9e7585e8: avconv: rename 'codec' to 'enc_ctx'
[15:10] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:1efc08921868: Merge commit '65ed9e7585e86e6c728bc762ac9944b157e39f8d'
[15:42] <cone-387> ffmpeg.git 03Anton Khirnov 07master:01947f07b81a: avconv: remove a useless line
[15:42] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:22ffa5390f6b: Merge commit '01947f07b81a57fb1e15114b710f6d5ae4f70a11'
[15:52] <Meths> I'm getting build failures with 2.2.1 which seem to be caused by incorrect build dependencies.  It is trying to build tools/uncoded_frame before it's built libavdevice on which it depends.
[15:52] <Meths> Have others seen this issue?
[15:53] <Daemon404> sounds like a bug
[15:53] <Daemon404> not sure why that tool needs avdevice though
[15:55] <Meths> It includes this:  #include "libavdevice/avdevice.h" avdevice_register_all();
[15:55] <Daemon404> ah
[15:56] <Daemon404> looks like its in teh wrong part of the makefile
[15:56] <Daemon404> its part of TOOLS for some reason
[15:56] <Daemon404> which is only supposed to have standalone tools
[15:57] <Daemon404> Meths, ill send a fix.
[15:57] <cone-387> ffmpeg.git 03Anton Khirnov 07master:9580818c5d93: avconv: rename OutputStream.opts to OutputStream.encoder_opts
[15:57] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:d5ec1ddf7bb0: Merge commit '9580818c5d934bde65a95efd2cee61c174721092'
[15:59] <Meths> That location seems to have been discussed on the ML in Feb.
[15:59] <Daemon404> link?
[15:59] <Daemon404> its a pretty simple fix
[15:59] <Meths> One sec...
[16:00] <Daemon404> it can still be in TOOLS
[16:00] <Daemon404> but it has to have a dep on avdevice 
[16:00] <Daemon404> (if thats possible)
[16:01] <Meths> http://lists.ffmpeg.org/pipermail/ffmpeg-devel/2014-February/154210.html
[16:02] <Daemon404> ah
[16:02] <Daemon404> that isnt related to what i mean :P
[16:02] <Meths> Ah, okay.
[16:03] <Meths> Where is the dep to be added? Lines 64 and 65 of Makefile?
[16:03] <Daemon404> i need to make sure i do it right
[16:03] <Daemon404> at the very least it needs to depend on avdevice even being built
[16:03] <Daemon404> it doesnt even do that right now
[16:04] <Daemon404> as in enabled in configure
[16:07] <Daemon404> Meths, deps seem to work in HEAD
[16:07] <Meths> hmmm
[16:07] <Daemon404> configure && make tools/uncoded_frame builds the libs first
[16:07] <Daemon404> before linking
[16:11] <cone-387> ffmpeg.git 03Jean First 07master:15fa856b82ea: libfdk-aacenc: Enable 7.1 channel encoding
[16:11] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:db9c138616c9: Merge commit '15fa856b82ea7dbbf69edc1074eb13dd00b0fa84'
[16:12] <Daemon404> michaelni, fyi, for the fdk relicensing, wbs also contact all the ffmpeg-only contributers
[16:13] <Daemon404> so it should be fine to merge.
[16:14] <Meths> Daemon404: Sorry, looks like it was a configure flags issue.  Someone explicitly disabled avdevice.  Rerunning now to check if avdevice builds on our platform.
[16:14] <Daemon404> Meths, it should not even allow you to try and build it if it is disabled
[16:15] <Meths> Agreed, but it helps you if I get the bug report right!
[16:16] <Meths> That built fine.
[16:17] <Daemon404> yeah its a bug though
[16:17] <Daemon404> confirmed it
[16:18] <michaelni> wbs, thx
[16:20] <cone-387> ffmpeg.git 03Martin Storsjö 07master:3828eb8519c1: libfdk-aac: Relicense the library wrappers to 2-clause BSD
[16:20] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:ba040d4c8c1e: Merge commit '3828eb8519c1530392ab8820b1aa784602128afc'
[16:21] <Daemon404> now fraunhofer can submit their 4 line patch
[16:21] <Daemon404> rejoice!
[16:22] <wm4> so is distributing ffmpeg with fdk now suddenly legal or something?
[16:22] <nevcairiel> why would that change
[16:22] <Daemon404> it doesnt
[16:23] <Daemon404> the point of the relicensing was so their engineers could submit patches to the wrapper
[16:23] <Daemon404> because their legal dpeartment is fucktarded
[16:23] <wm4> lol
[16:23] <kierank> to be honest credit must go to them and wbs for actually doing this
[16:23] <Daemon404> yes
[16:23] <Daemon404> well, wbs
[16:24] <Daemon404> he did it all.
[16:24] <kierank> because i can only imagine how much of a pita in fraunhofer it must have been to get permission to add to an open source
[16:24] <nevcairiel> should've forced them to change their license of the lib to make it actually more useful :p
[16:24] <Daemon404> unless you menat he decoder
[16:24] <Daemon404> the*
[16:24] <Daemon404> kierank, wouldnt that have been negotiated by google
[16:24] <kierank> dunno
[16:25] <J_Darnley> That sounds like the most pointless change.  You can't separate that file from the rest of libavcodec.
[16:26] <Daemon404> J_Darnley, ?
[16:26] <Daemon404> it was solely to make their legal dept happy
[16:26] <Daemon404> you cannot expect them to have any logic
[16:27] <J_Darnley> Well that's ture
[16:27] <J_Darnley> *true
[16:27] <nevcairiel> different files have different licenses in ffmpeg, thats just how its always been
[16:27] <nevcairiel> and it does kinda make a bit of difference if someone wants to take that code and do something else with it
[16:27] <J_Darnley> Re-implement the rest of libavcodec?
[16:28] <nevcairiel> why would you need to do that
[16:28] <J_Darnley> To be able to use that file
[16:28] <nevcairiel> you're not supposed to use it as-is, of course
[16:28] <nevcairiel> but its easy enough to recycle code
[16:29] <cone-387> ffmpeg.git 03Rafaël Carré 07release/1.1:2340ce6954db: update Changelog
[16:29] <cone-387> ffmpeg.git 03Michael Niedermayer 07release/1.1:714dd84f204f: Merge commit '2340ce6954db2c0a79c57fcba7f45e2fcc20cdc9' into release/1.1
[16:31] <Compn> i'm really surprised no one has tried to sue ffmpeg devels for codecs yet
[16:31] <Compn> :P
[16:34] <nevcairiel> didnt viacom or someone send some papers sometime about aac
[16:38] <Skyler_> Compn: j-b has apparently received quite a lot of things, like letters from mpeg-la
[16:38] <Skyler_> from what I've heard
[16:39] <Skyler_> though I guess nobody has ever actually sued?
[16:39] <Compn> j-b got one from dolby 
[16:41] <Daemon404> ac-3 patents almost expired (next year!)
[16:41] <Daemon404> \o/
[16:41] <Compn> hooray!
[16:41] <wm4> good thing ac-4 is in the works
[16:41] <Daemon404> of course
[16:41] <kierank> Daemon404: they won't say that
[16:41] <kierank> they'll just say their IPR is still being violated
[16:41] <nevcairiel> i'm sure they'll find a way to make them last longer, at least in the US
[16:43] <Daemon404> kierank, so then you can tell them to put up to shutu[
[16:43] <kierank> they'll find a way by saying their e-ac3 patents still apply
[16:50] <ubitux> http://ubitux.fr/pub/pics/_gcc-color-insanity.png someone needs to hint the gcc devs about isatty()
[16:55] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:47550e62bf71: avformat/astenc: avoid using AVCodecContext.frame_number
[17:41] <ubitux> ok i was stupid, it's because it's on stderr
[17:54] <J_Darnley> gcc has colour now?
[17:54] <J_Darnley> colour me impressed
[17:54] <ubitux> yeah, GCC_COLOR=1 in your env to enable it though
[17:55] <ubitux> i wonder why it's not the default but well
[18:08] <cone-387> ffmpeg.git 03Lukasz Marek 07master:78d3453c4a2e: lavu/fifo: add const to arguments
[18:12] <wm4> does it have good error messages yet?
[18:32] <smarter> they improved them in 4.8: http://gcc.gnu.org/wiki/ClangDiagnosticsComparison
[18:57] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:a69e16a97e40: tests/tiny_ssim: check dimensions
[18:57] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:b395167d56d5: tools/ismindex: use av_mallocz_array()
[20:11] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:2ded57216eae: avutil/audio_fifo: Use av_mallocz_array()
[20:11] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:e77ffaa0ac26: tests/tiny_ssim: more correct error message
[20:11] <cone-387> ffmpeg.git 03Michael Niedermayer 07master:4aa7848eb865: avutil/file_open: Use av_mallocz_array()
[21:23] <Daemon404> who is anshul
[21:23] <Daemon404> and why is he emailing me privately
[21:23] <Daemon404> "Sorry for personal mail but i didnt want it public, your mail is too harsh.
[21:23] <Daemon404> There should not be a problem in selecting default setting explicitly, 
[21:24] <Daemon404> Things that are considered as default now may not be default in future.Some things were not default in past.
[21:24] <Daemon404> Its your and anyone right to ask the developer to test the patch in normal co"
[21:25] <JEEB> wat :D
[21:26] <J_Darnley> I think he's sent me a reply off-list too.  Let me look...
[21:28] <J_Darnley> Maybe not.
[21:30] <Daemon404> either way it's a bit wtf
[21:31] <Daemon404> (it was in response to http://ffmpeg.org/pipermail/ffmpeg-devel/2014-May/157302.html)
[21:32] <J_Darnley> Well, you are strict in that mail but I don't see a problem with it.
[21:33] <Daemon404> i was blunt, sure
[21:33] <Daemon404> but thats different than harsh.
[21:33] <nevcairiel> i wonder what you have to do to make that fail
[21:33] <nevcairiel> might be related to one of the disable things
[21:33] <nevcairiel> but the patch is still wrong
[21:34] <Daemon404> nevcairiel, well also, passing mingw's include folder to msvc cant end well.
[21:35] <nevcairiel> does it even use /usr/local
[21:35] <nevcairiel> thats usually for your own stuff
[21:35] <Daemon404> some setups do
[21:36] <Daemon404> one of the 50 different msys zips out there
[21:36] <Daemon404> some map mingw to /usr/local
[21:36] <Daemon404> (evil)
[23:41] <cone-764> ffmpeg.git 03Michael Niedermayer 07master:1394e0a9e03e: avformat/segment: Use av_malloc_array()
[23:41] <cone-764> ffmpeg.git 03Michael Niedermayer 07master:89d2fc62db98: avcodec/svq3: Fix "incompatible pointer type" warnings
[00:00] --- Tue May  6 2014


More information about the Ffmpeg-devel-irc mailing list