[Ffmpeg-devel-irc] ffmpeg-devel.log.20150208

[burek]

[02:30] <cone-53> ffmpeg.git 03Michael Niedermayer 07master:ecfd48dc065b: avcodec/h264: Move COPY_PICTURE() to h264.h so it can be used from other parts of the h264 decoder
[02:30] <cone-53> ffmpeg.git 03Michael Niedermayer 07master:de6df4612036: avcodec/h264: Setup decoder to have matching reference to the EC code
[02:30] <cone-53> ffmpeg.git 03Michael Niedermayer 07master:2048126ea7a5: avcodec/utils: Assert that dst->progress is clear before ff_thread_ref_frame()
[02:45] <cone-53> ffmpeg.git 03Carl Eugen Hoyos 07master:ea329b60f5d3: Do not use -mdynamic-no-pic on OSX x86_64.
[02:45] <cone-53> ffmpeg.git 03Carl Eugen Hoyos 07master:2efaaf9476c8: Support decoding yuv dpx images.
[02:45] <cone-53> ffmpeg.git 03Michael Niedermayer 07master:7c148f30f364: Merge remote-tracking branch 'cehoyos/master'
[04:18] <cone-53> ffmpeg.git 03Mark Reid 07master:6a8a3bfb8737: libavformat/mxfenc.c: added mxf opatom muxer
[04:32] <cone-53> ffmpeg.git 03Georg Lippitsch 07master:97a27065c88c: avdevice/decklink: 10 Bit support for Decklink input device
[04:44] <cone-53> ffmpeg.git 03Georg Lippitsch 07master:62ea5ae2b567: avdevice/decklink_common: Fix Decklink for Mac
[04:51] <cone-53> ffmpeg.git 03James Almer 07master:92d903afaa22: x86/vp9dsp: fix clobbering of xmm6 on IDCT sse2 functions
[05:01] <cone-53> ffmpeg.git 03Michael Niedermayer 07master:5ae140c04f02: avdevice/decklink_common: Use defined(__APPLE__) instead of __APPLE__
[11:38] <ubitux> make is randomly recompiling a bunch of files in ffmpeg sometimes, even though i didn't do any modification
[11:38] <ubitux> wonder if that's an issue on my side
[11:39] <ubitux> or if it's watching the wrong files
[11:39] <cone-129> ffmpeg.git 03Eejya Singh 07master:40b198e98736: lavfi/subtitles: add force_style option
[13:39] <cone-129> ffmpeg.git 03Michael Niedermayer 07master:009a57b4792e: avformat/mxfenc: Remove unused variables
[13:45] <compn> ubitux : wasnt there some trick to get ffmpeg to only rebuild modified files ?
[13:45] <ubitux> well the "trick" is make
[13:47] <ubitux> but yeah i don't know, maybe git is touching random files or something
[13:50] <kurosu> does it inclue the git revision, ie if you commit omething, the source doesn't change but some files are recompiled?
[13:51] <nevcairiel> i havent really noticed such problems
[13:53] <wm4> I don't think git attempts to restore file timestamps on checkouts
[13:54] <compn> michaelni had a patch for it locally i think
[13:54] <compn> back when it was changed 
[13:55] <compn> it = build system
[14:31] <cone-129> ffmpeg.git 03Clément BSsch 07master:eb7efaa92447: avfilter/dctdnoiz: fix slice_h computation
[15:52] <aetasx> question: if you're specifying -ss for seeking in front of an output file, is there no combination of flags that allow you to keep the original PTS or is it going to shift all the timestamps unless you put it before the input file?
[16:01] <rcombs> aetasx: -copyts
[16:01] <rcombs> erm wait, I swapped input and output there
[16:01] <rcombs> uh, not sure then, actually
[16:04] <aetasx> yeah I tried copyts, vsync passthrough, disabling avoid_negative_ts, etc on both the input and output and it does nothing if you pass in -ss on the output video
[16:05] <aetasx> I did find -start_time 0 will do it, but out of all the options I haven't a clue why that one works while all the others don't
[16:41] <kierank> wow "Our team is working on AltiVec
[16:41] <kierank> functions for HEVC decoder."
[16:42] <JEEB> altivec, eh
[16:42] <JEEB> some people love them powerpc I guess
[17:05] <ubitux> did we agree that "AVERROR_INVALIDATA" is relative to the input data and not the user input?
[17:06] <nevcairiel> einval should be user input, imho
[17:07] <ubitux> yes, i think that's what we agreed on
[17:07] <nevcairiel> then we agree!
[17:07] <ubitux> :)
[17:09] <cone-129> ffmpeg.git 03Clément BSsch 07master:4b60bd1aacf0: avformat/internal: fix av_interleaved_write_frame typo reference
[17:09] <anshul_mahe> michaelni: you told me some days back, to update the Apichange, but the generic part you pushed are not changing any APi
[17:13] <anshul_mahe> I am taliking about support of transmuxing of data path
[17:23] <cone-129> ffmpeg.git 03Clément BSsch 07master:55feff57cec8: avfilter/hqdn3d: an invalid bit depth means a bug, not invalid read data
[17:30] <michaelni> anshul_, a47c9331ef13c4af01de582f159ac6e59fe990ed added things to avformat.h, they should be documented in APIChanges
[17:36] <anshul_> ok, thanks for pointing I was looking only at bb0344dc5582829253cc300a94fab29fb1f63f40
[17:40] <anshul_> there is one log in APIChange with 2015-01-xx - xxxxxxx, does it mean date and commit are not finalize yet
[17:55] <cone-129> ffmpeg.git 03Carl Eugen Hoyos 07master:a2bc8c56016d: Fix standalone compilation of the mov muxer with --disable-optimizations.
[17:55] <cone-129> ffmpeg.git 03Michael Niedermayer 07master:162f34497114: Merge remote-tracking branch 'cehoyos/master'
[18:29] <michaelni> if you see entries with xx / xxxxxx in APIChanges feel free to replace them by the relevant git hashes and dates, the hashes cannot be commited in the commit that adds the change and entry so they will always be xxxxxxx initially
[18:35] <anshul_mahe> ok
[18:40] <cone-129> ffmpeg.git 03Anshul Maheshwari 07master:1df841f154f8: doc/APIchanges: updated Data stream changes in APIChange
[18:42] <kierank> these patches are complete crap
[18:44] <wm4> which?
[18:45] <kierank> the whole scte35 data stream patches
[18:45] <wm4> I don't give a shit about scte35, but I hate this "data stream" thing too
[18:45] <wm4> more weird API artifacts
[18:46] <wm4> which FUCKING NOBODY knows how to use, too
[18:46] <kierank> there we go
[18:48] <kierank> the guy is just doing something for his employer who needed outsourced ffmpeg work
[18:48] <kierank> and in 4 months he'll be g one
[18:49] <kierank> and nobody will know wtf any of this stuff does
[18:53] <kierank> j-b: ^ :(
[18:55] <wm4> kierank: and "we" will refuse to remove it anyway
[18:55] <wm4> because "somebody may still use it"
[18:55] <kierank> of course
[18:55] <kierank> like mplayer filters that segfaulted
[18:55] <kierank> someone might still use it
[18:55] <wm4> that said, I wonder how long Libav needs to apply that patch that removes libavformat/seek.c, which is all 100% dead code (and has been for 6 to 3 years)
[18:57] <nevcairiel> they still havent pushed my truehd patches from  september, so .. :p
[18:58] <nevcairiel> i wont ping them anymore
[22:45] <cone-710> ffmpeg.git 03Christophe Gisquet 07master:4919b384214e: x86: hevc_mc: fewer gpr autoloads for _v filters
[22:45] <cone-710> ffmpeg.git 03Christophe Gisquet 07master:b61b9e4919dc: x86: hevc_mc: remove lea in EPEL_LOAD
[23:03] <jamrial> fionag: are you ok with me backporting (and thus relicensing) http://git.videolan.org/?p=x264.git;a=blob;f=common/x86/x86util.asm;h=b18a1d10c3b9115e10ca0673143ff6f1564ee57e;hb=HEAD#l58 (the dqqq part) to ffmpeg's x86util?
[00:00] --- Mon Feb  9 2015