[Ffmpeg-devel-irc] ffmpeg-devel.log.20121118

[burek]

[06:04] <cone-229> ffmpeg.git 03Michael Niedermayer 07dab19048a1fb: mjpegdec: reset h/v_count, fix assertion failure.
[08:54] <cone-229> ffmpeg.git 03Carl Eugen Hoyos 07ab82b77822a7: Fix possible NULL-pointer dereference when decoding mov files.
[12:50] <durandal_1707> michaelni: is it ok to push mjpeg444 as it is?
[13:14] <cone-910> ffmpeg.git 03Paul B Mahol 07f5f29910dda9: img2enc: remove dead code
[13:25] <michaelni> durandal_1707, it should be ok except that iam not sure the block ordering wont bite in the future
[13:25] <durandal_1707> michaelni: how it could bite?
[13:26] <michaelni> well luma and chroma is ordered differently in it
[13:27] <michaelni> with 4 Y blocks, 4 Cb blocks, 4 Cr blocks, i would naively expect each 4 blocks to be ordered the same
[13:27] <michaelni> it would allow future encoders to use the same function and indexing to address each
[13:29] <durandal_1707> and how this could be fixed?
[13:30] <michaelni> one would have to change order in mpegvideo and do the same change in jpeg
[13:32] <durandal_1707> michaelni: the one i did?
[13:33] <cone-910> ffmpeg.git 03Paul B Mahol 07d0a503c97cc5: ast: check bit depth too
[13:39] <cone-910> ffmpeg.git 03Mans Rullgard 075af530f5a95e: parisc: work around bug in gcc 4.3 and later
[13:39] <cone-910> ffmpeg.git 03Justin Ruggles 0702738792c5cd: lavr: fix the decision for writing directly to the output buffer
[13:39] <cone-910> ffmpeg.git 03Ronald S. Bultje 077a9e65aceee6: x86: lavr: fix stack allocation for 7 and 8 channel downmixing on x86-32
[13:39] <cone-910> ffmpeg.git 03Michael Niedermayer 070c79a3d9cb0c: Merge remote-tracking branch 'qatar/master'
[13:40] <michaelni> the order for the newly added yuv444, yes
[13:55] <cone-910> ffmpeg.git 03Peter Ross 07d26eeb0dc1dd: iff: recognise more DEEP colorspaces
[13:56] <cone-910> ffmpeg.git 03Peter Ross 0705001dd7ca39: iff: process DEEP DLOC chunk to obtain image dimensions
[14:02] <cone-910> ffmpeg.git 03Marton Balint 074bee03034bf4: lavf: let av_find_best_stream use bitrate info if available
[14:34] <cone-910> ffmpeg.git 03Michael Niedermayer 074fecc3cf0998: h264: Skip odd NALs in extradata, prevent undefined behavior
[15:49] <cone-910> ffmpeg.git 03Piotr Bandurski 074413a8428a75: iff: decode HAM8 images with masking correctly.
[15:56] <cone-910> ffmpeg.git 03Nicolas George 07c437c251b8e5: lavfi/drawtext: implement more generic expansion.
[16:45] <cone-910> ffmpeg.git 03Clément BSsch 07e034b07e4056: pp: add various const for pointers.
[16:45] <cone-910> ffmpeg.git 03Clément BSsch 0759d686f10086: pp: add auto detection cpu flag.
[16:45] <cone-910> ffmpeg.git 03Clément BSsch 07f6be711a69ed: lavfi/mp/pp: use PP_CPU_CAPS_AUTO.
[16:45] <cone-910> ffmpeg.git 03Clément BSsch 074e264d1c79cf: pp: add support for SSE2.
[16:45] <cone-910> ffmpeg.git 03Clément BSsch 07f48cddfe4cf0: pp: add SSE2 deInterlaceInterpolateCubic().
[16:47] <ubitux> michaelni: thx for the review
[16:48] <cone-910> ffmpeg.git 03Michael Niedermayer 071a947dfa8347: ff_mpeg_update_thread_context: assert context unequality
[16:48] <cone-910> ffmpeg.git 03Michael Niedermayer 072d5f1addbec2: h264: fix integer overflow, assert failure
[16:48] <cone-910> ffmpeg.git 03Michael Niedermayer 07d6c184880ee2: h264: correct ref count check and limit, fix out of array accesses.
[16:54] <cone-910> ffmpeg.git 03Stefano Sabatini 076c7ae49330cc: lavfi/frei0r: extend load_path() to support arbitrarily long paths
[16:54] <cone-910> ffmpeg.git 03Stefano Sabatini 07158763312f97: lavfi/frei0r: in init() check path loop, free resources in a single point
[18:59] <cone-910> ffmpeg.git 03Paul B Mahol 079f02d4ed0f72: mjpegenc: yuvj444p support
[19:17] <cone-910> ffmpeg.git 03Michael Niedermayer 0736cf247e4302: ff_h264_direct_ref_list_init: fix B slice check.
[20:47] <cone-910> ffmpeg.git 03Paul B Mahol 07a6f4c41a7e7f: fate: add mjpeg-422 and mjpeg-444 tests
[21:34] <ubitux> durandal11707: you can move the int i into the for loop to avoid the { } after the case
[21:49] <cone-910> ffmpeg.git 03Paul B Mahol 076557c46d91d4: pcm: move i into for loop to avoid { } inside switch statement
[21:51] <durandal11707> ubitux: fixed locally
[21:58] <ubitux> you could have use a do { } while (0) here, since it will generates empty statements with the ';' after the DECODE_PLANAR() "calls"
[22:00] <durandal11707> actually i added ; for no reason there
[22:00] <ubitux> it's fine to have them after "DECODE_PLANAR()"
[22:01] <durandal11707> but not consistent with DECODE
[22:01] <ubitux> it sucks a bit for editors
[22:01] <ubitux> (for indent)
[22:01] <ubitux> and "it's weird" at reading
[22:02] <ubitux> well whatever, all these remarks are just nits, don't worry to much about it
[22:15] <cone-910> ffmpeg.git 03Paul B Mahol 070df7d95a5e53: pcmenc: remove dead increment
[23:38] <llogan> carl is curating
[00:00] --- Mon Nov 19 2012