[Ffmpeg-devel-irc] ffmpeg-devel.log.20140520

burek burek021 at gmail.com
Wed May 21 02:05:02 CEST 2014

[00:00] <michaelni> yes
[00:01] <iive> on both intel and amd?
[00:03] <michaelni> i dont have a working amd here
[00:21] <cone-612> ffmpeg.git 03Oliver Fromme 07master:ab6228316a59: avcodec/dvdsubenc: accept forced flag
[00:23] <llogan> i have one if you want me to test
[00:34] <michaelni> llogan, as you prefer, i dont think its very important ...
[00:35] <cone-612> ffmpeg.git 03Lukasz Marek 07master:7336e39f3cd2: lavu/fifo: add av_fifo_alloc_array function
[00:35] <cone-612> ffmpeg.git 03Lukasz Marek 07master:51bad41a13c7: lavd/jack_audio: use av_fifo_alloc_array
[00:35] <cone-612> ffmpeg.git 03Lukasz Marek 07master:4d686674b265: ffmpeg: use av_fifo_alloc_array
[00:35] <cone-612> ffmpeg.git 03Lukasz Marek 07master:d15db8e413e9: lavfi/buffersink: use av_fifo_alloc_array
[00:35] <cone-612> ffmpeg.git 03Lukasz Marek 07master:ce051ceefc70: lavfi/vf_fps: use av_fifo_alloc_array
[00:36] <cone-612> ffmpeg.git 03Lukasz Marek 07master:bc4f362c9281: lavc/frame_thread_encoder: use av_fifo_alloc_array
[00:36] <cone-612> ffmpeg.git 03Lukasz Marek 07master:8aa29880365c: lavc/flac_parser: use av_fifo_alloc_array
[00:36] <cone-612> ffmpeg.git 03Lukasz Marek 07master:74f9c5947d69: lavf/dvenc: use av_fifo_alloc_array
[00:36] <cone-612> ffmpeg.git 03Lukasz Marek 07master:c61cdefa4d79: lavf/audiointerleave: use av_fifo_alloc_array
[00:36] <cone-612> ffmpeg.git 03Lukasz Marek 07master:2ed9e17ed179: lavf/audiointerleave: check for allocation failure
[00:36] <cone-612> ffmpeg.git 03Lukasz Marek 07master:37fa4b9bb967: lavf/audiointerleave: return more meaningful error codes
[00:36] <cone-612> ffmpeg.git 03Lukasz Marek 07master:b9419b58826e: lavf/ftp: favour EPSV over PASV command
[00:36] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:6218831844fa: Merge remote-tracking branch 'lukaszmluki/master'
[00:38] <llogan> michaelni: i guess i'll pass. i don't know how to get dezicycles anyway, but feel free to ping me (w/ instructions) if you do need me to test some other time
[00:40] <michaelni> ok
[00:56] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:b2a8bc3d0482: ffmpeg/set_encoder_id: dont allocate encoder string if one is already set
[02:16] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:be2bbc5266d1: ffmpeg_opt: dont copy "encoder" metadata unless stream copy is used
[02:47] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:0739614c9c48: avcodec: remove no longer needed ff_codec_close_recursive()
[02:47] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:8408720453d7: avcodec/smvjpegdec: Dont ignore avcodec_close() return code
[02:48] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:9893e883ea21: avformat/mux: Remove localized encoder strings
[03:30] <cone-612> ffmpeg.git 03Janne Grunau 07master:132a5711a2aa: check: add libavcodec/vda_internal.h to SKIPHEADERS-$(CONFIG_VDA)
[03:30] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:447b988b5319: Merge commit '132a5711a2aaaa6c7d139a0a30046a6ad41b64aa'
[03:38] <cone-612> ffmpeg.git 03Janne Grunau 07master:afe176265480: build: export library dependencies in ${name}_FFLIBS
[03:38] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:9e6769d2dc7b: Merge commit 'afe176265480880e1f702c96a8ba434b0d88728b'
[03:51] <cone-612> ffmpeg.git 03Janne Grunau 07master:c23c96b638cb: lavf: add av_stream_get_side_data
[03:51] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:7d25af1547f8: Merge commit 'c23c96b638cbf6c489fd301e6b3d5555632fba37'
[04:34] <cone-612> ffmpeg.git 03Katerina Barone-Adesi 07master:e50ae60d4618: avcodec/fate-idct8x8: Defined behavior: eliminate negative left-shifts.
[06:24] <cone-612> ffmpeg.git 03Michael Niedermayer 07master:6310eb8010b7: avutil/cpu: force mmx on selection of higher x86 SIMD features
[14:32] <ubitux> [opengl outdev @ 0x1aa8520] SDL driver: 'x11'.
[14:32] <ubitux> [opengl outdev @ 0x1aa8520] Provided pixel format 'yuyv422' is not supported
[14:32] <ubitux> :(
[14:33] <ubitux> isn't it supposed to insert the correct pix fmt convert by itself?
[14:35] <thardin> what does it take? yuv420p?
[14:36] <ubitux> yuv420p seems to be accepted yes
[14:38] <thardin> can you insert one manually?
[14:38] <thardin> (shouldn't be necessary of course)
[14:38] <ubitux> yep sure that's wjat i did
[14:38] <ubitux> what*
[15:36] <cone-79> ffmpeg.git 03Yu Xiaolei 07master:2491579d4715: configure: arm: do not enable fast-unaligned if it is explicitly disabled via commandline
[16:21] <Keestu> hello, avcodec_decode_video2 (), does it have an impact on thread_count value on Codec Context ?
[17:22] <kurosu> BBB / michaelni: was leaving some inline asm stuff in C files intended?
[17:22] <kurosu> I recently ported the fpel part but seeing the hpeldsp part, this looks like it was intended
[17:23] <kurosu> (only the mmx implementations being left)
[21:26] <cone-711> ffmpeg.git 03Alexis Ballier 07master:c9ab8790118c: Fix build with --enable-ladspa --disable-frei0r.
[21:38] <cone-711> ffmpeg.git 03Michael Niedermayer 07master:be656c362d43: avcodec/h264_picture: Fix build without error resilience
[21:40] <Daemon404> i always see these sorts of fixed
[21:40] <Daemon404> why not just iterate over all combinations once and be done wth it
[22:24] <nevcairiel> Daemon404: i bet libav keeps breaking it
[22:32] <Daemon404> diego runs such a script
[22:34] <nevcairiel> but they dont support that buildmode, iirc
[22:34] <Daemon404> no its some script sitting on a pc somewhere
[22:34] <Daemon404> small wrapper
[23:08] <cone-711> ffmpeg.git 03Dale Curtis 07master:ca2fd25988d3: Fix compilation errors when deprecated features are disabled.
[23:20] <kurosu> what does HAVE_MMX_EXTERNAL mean? yasm supporting mmx ?
[23:21] <Daemon404> yes
[23:21] <Daemon404> im not sure why that exists...
[23:21] <Daemon404> maybe im mistaken
[23:27] <kurosu> if I'm not mistaken, yasm/nasm is for x86
[23:27] <kurosu> so yasm not supporting mmx (only from sse2 on?) is as weird as a cpu having only sse2 on 
[23:28] <kurosu> but ok
[23:28] <kurosu> I don't thing I'm going anywhere with the code in relation, anyway
[23:31] <J_Darnley> I think it must come from the idea that someone may use a gcc so old it doesn't support inline mmx instructions so there may be a yasm so old that it too doesn't support mmx instructions.
[23:48] <jamrial> those macros are mainly for when ffmpeg is configured with --disable-{opt}
[23:49] <kurosu> ok
[23:49] <kurosu> no that I think of it, most of those functions are used for when block size is 8
[23:49] <kurosu> *now
[23:49] <jamrial> the oldest yasm version we support is 0.8, which can assemble up to AVX and FMA3
[23:50] <kurosu> so, they may actually be ran even on newer cpus
[23:50] <kurosu> (because 128b-wide regs don't provide a benefit)
[00:00] --- Wed May 21 2014

More information about the Ffmpeg-devel-irc mailing list