[Ffmpeg-devel-irc] ffmpeg-devel.log.20180131

[burek]

[01:20:35 CET] <cone-612> ffmpeg 03Jun Zhao 07release/3.4:9aa0ed850b77: avfilter/formats: fix wrong function name in error message
[01:20:36 CET] <cone-612> ffmpeg 03Kelly Ledford 07release/3.4:a3832486e4f1: libavfilter/af_dcshift.c: Fixed repeated spelling error
[01:20:37 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:0f0a2ff5a09d: avcodec/vp9: mark frame as finished on decode_tiles() failure
[01:20:38 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:d6a13f031ced: avcodec/h264_parse: Treat escaped and unescaped decoding error equal in decode_extradata_ps_mp4()
[01:20:39 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:d147e2d55d29: avcodec/hevcdsp_template: Fix undefined shift in put_hevc_qpel_bi_w_hv()
[01:20:40 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:2e426fae43f3: avcodec/hevc_sei: Fix integer overflows in decode_nal_sei_message()
[01:20:41 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:43c03866b23a: tests/audiomatch: Add missing return code at the end of main()
[01:20:42 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:0288d15cdded: avcodec/hevc_cabac: Fix integer overflow in ff_hevc_cu_qp_delta_abs()
[01:20:43 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:e55a6c5f055c: avcodec/dirac_dwt: Fix integer overflow in COMPOSE_DD97iH0() and COMPOSE_DD137iL0()
[01:20:44 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:0e7d8ce37c2f: avcodec/hevcdsp_template.c: Fix undefined shift in FUNC(dequant)
[01:20:45 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:fb9560b366da: avcodec/flacdec: avoid undefined shift
[01:20:46 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:7e402c31efd8: avcodec/hevcdsp_template: Fix Invalid shifts in put_hevc_qpel_bi_w_h() and put_hevc_qpel_bi_w_w()
[01:20:47 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:91f5a2b7b88a: avcodec/flacdec: Fix overflow in multiplication in decode_subframe_fixed()
[01:20:48 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:6abe1e06f592: avcodec/exr: Check buf_size more completely
[01:20:49 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:b1af55778b00: avcodec/dnxhddec: Check dc vlc
[01:20:50 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:62024c127798: avcodec/h264_slice: Do not attempt to render into frames already output
[01:20:51 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:5365904e9642: avcodec/jpeg2000dsp: Fix integer overflows in ict_int()
[01:20:52 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:a3add1924095: avcodec/opus_parser: Check payload_len in parse_opus_ts_header()
[01:20:53 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:097bc4d32d59: avcodec/diracdec: Fix integer overflow with quant
[01:20:54 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:8263246ba8f6: avcodec/dirac_dwt: Fix overflows in COMPOSE_HAARiH0/COMPOSE_HAARiL0
[01:20:55 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:4715ef27a068: avcodec/h264addpx_template: Fixes integer overflows
[01:20:56 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:ece787999249: avcodec/arm/sbrdsp_neon: Use a free register instead of putting 2 things in one
[01:20:57 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:04949cc08ece: avcodec/utils: Avoid hardcoding duplicated types in sizeof()
[01:20:58 CET] <cone-612> ffmpeg 03Carl Eugen Hoyos 07release/3.4:092febb2add6: configure: bump year
[01:20:59 CET] <cone-612> ffmpeg 03Jun Zhao 07release/3.4:7b56d6584c46: lavfi/deinterlace_vaapi: fix can't show full option information.
[01:21:00 CET] <cone-612> ffmpeg 03Nikolas Bowe 07release/3.4:facd0521e440: avformat/matroskadec: Fix float-cast-overflow undefined behavior in matroska_parse_tracks()
[01:21:01 CET] <cone-612> ffmpeg 03Nikolas Bowe 07release/3.4:e755482d367a: avformat/lrcdec: Fix memory leak in lrc_read_header()
[01:21:02 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:56b0179b6a03: avcodec/ac3dec_fixed: Fix integer overflow in scale_coefs()
[01:21:03 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:f56215d3ff63: avcodec/jpeg2000: Check sum of sizes of band->prec before allocating
[01:21:04 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:bae4d39437fe: avcodec/wavpack: Fix integer overflows in wv_unpack_stereo / mono
[01:21:05 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:540f4467c825: avcodec/ulti: Check number of blocks at init
[01:21:06 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:aed915b8a62c: avcodec/snowdec: Fix integer overflow before htaps check
[01:21:07 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:6ed5e44998ed: avcodec/truemotion2: Fix integer overflow in TM2_RECALC_BLOCK()
[01:21:08 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:edf200e2bc9a: avcodec/hevc_cabac: Move prefix check in coeff_abs_level_remaining_decode() down
[01:21:09 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:c1b74d608c6e: avcodec/dxtory: Fix bits left checks
[01:21:10 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:2fdb27b5123d: avcodec/mjpegdec: Fix integer overflow in DC dequantization
[01:21:11 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:11498c22a0db: avcodec/hevc_cabac: Check prefix so as to avoid invalid shifts in coeff_abs_level_remaining_decode()
[01:21:12 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:2980b95fafb3: avfilter/vf_transpose: Fix used plane count.
[01:21:13 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:6723a436095f: avcodec/mpeg4videodec: Check mb_num also against 0
[01:21:14 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:cd478122b0a0: avcodec/get_bits: Document the return code of get_vlc2()
[01:21:15 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:d07f78ae726b: avcodec/mpeg4videodec: Avoid possibly aliasing violating casts
[01:21:16 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:93437a18d878: avcodec/hevc_ps: Check log2_sao_offset_scale_*
[01:21:17 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:d06972535e48: avcodec/indeo5: Do not leave frame_type set to an invalid value
[01:21:18 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:c1c50fc4a754: avcodec/dirac_dwt: Fix several integer overflows
[01:21:19 CET] <cone-612> ffmpeg 03Michael Niedermayer 07release/3.4:dd93df46a618: Update for 3.4.2
[02:11:24 CET] <cone-612> ffmpeg 03James Almer 07release/3.4:64f0fd599845: avcodec/hevc_ps: add a function to uninitialize parameter set buffers
[02:11:25 CET] <cone-612> ffmpeg 03James Almer 07release/3.4:d7d5a3379dfe: avcodec/hevcdec: use ff_hevc_uninit_parameter_sets()
[02:11:26 CET] <cone-612> ffmpeg 03James Almer 07release/3.4:e5bbb5219441: avcodec/hevc_parser: use ff_hevc_uninit_parameter_sets()
[02:11:27 CET] <cone-612> ffmpeg 03James Almer 07release/3.4:af54886de8ab: avcodec/mediacodecdec: use ff_hevc_ps_uninit()
[02:14:02 CET] <cone-612> ffmpeg 03James Almer 07release/3.4:9b97afe7ad06: Changelog: update for the previous four commits
[02:46:10 CET] <SortaCore> wait, you can get an nvidia primary with intel secondary?
[02:46:24 CET] <SortaCore> isn't primary decided by motherboard?
[02:46:41 CET] <SortaCore> re that ticket #6996 just posted
[05:00:39 CET] <philipl> BtbN: I can get -cq to do something. It *looks* like it's trying to do something like crf.
[05:02:03 CET] <philipl> The bitrate (-b:v) seems to influence it to, and acts as some sort of upper bound (although not literally an upper bound) so if the b:v is low, then setting -cq to a 10 (for example) doesn't do anything; it's constrained by the bitrate. but if you up the bitrate, then it makes a difference. If you set cq to 51, then you get appropriate q values reported back.
[05:02:16 CET] <philipl> But obviously, these interactions are not documented.
[06:39:43 CET] <wm4> av_opt_serialize() was used by ffserver only
[15:31:12 CET] <jdarnley> atomnuker: I want to ask you about the following two commits and applying them to ffmpeg upstream.
[15:31:18 CET] <jdarnley> https://github.com/OpenBroadcastSystems/ffmpeg/commit/e6a8de2ea62030bedc187d85e3ccbd7a8c2c392c
[15:31:24 CET] <jdarnley> https://github.com/OpenBroadcastSystems/ffmpeg/commit/eefdd82193b140b830709d6421a326ab13dbdab1
[15:32:29 CET] <jdarnley> I think I do understand the first but I am less clear about what the second one does and why.
[15:34:01 CET] <jdarnley> I thought that the packet size allocation was roughly equivalent before and after the change
[15:34:52 CET] <atomnuker> its a bloody hack because kierank was too lazy to write proper parsing code a whole few months before IBC 2016 so I had to make them equal
[15:35:14 CET] <jdarnley> Oh
[15:35:46 CET] <jdarnley> Does upstream not really want that one then?
[15:35:55 CET] <atomnuker> No.
[15:36:43 CET] <jdarnley> But the first one can be useful to others too, right?
[15:39:13 CET] <atomnuker> yes
[15:54:33 CET] <sfan5> " If using GCC, consider adding -ftree-vectorize to --extra-cflags. Most recent versions of GCC do not miscompile FFmpeg with the auto-vectorizer enabled, [...] Therefore, configure by default disables the auto-vectorizer on GCC, and it must be enabled by the user explicitly if desired, such as via the method outlined above. " ~https://trac.ffmpeg.org/wiki/CompilationGuide#PerformanceTips
[15:55:27 CET] <sfan5> this seems inaccurate, if I configure with --extra-cflags='-ftree-vectorize', config.mak ends up with CFLAGS=  -ftree-vectorize [...] -fno-tree-vectorize [...]
[15:55:36 CET] <sfan5> am I missing something or is the wikipage just outdated?
[15:58:08 CET] <jdarnley> It sounds outdated or just wrong.
[15:58:54 CET] <jdarnley> I'm sure everytime people have tried to enable vectorisation it gets disabled a few commits later when it screws something up
[16:04:56 CET] <JEEB> yea
[16:06:18 CET] <JEEB> every time people vectorization is enabled, some not-so-exotic compiler is found to muck around with things badly
[16:22:38 CET] <sfan5> no fate failures here with -ftree-vectorize, Linux x86_64 + gcc 7.2.1 20180116
[16:27:09 CET] <jdarnley> Then happily use it on your own machines
[16:28:07 CET] <sfan5> sure, I wasn't advocating for official support for this option
[16:35:06 CET] <jamrial> fate might pass, but since only about 60% of the code is actually tested by it, it doesn't mean much
[16:35:50 CET] <jamrial> last time vectorization was applied a bunch of issues were found by downstreams
[17:12:27 CET] <atomnuker> jdarnley: no one's going to review it, you should just push it
[17:18:43 CET] <jdarnley> okay
[17:42:12 CET] <cone-796> ffmpeg 03Rostislav Pehlivanov 07master:61a4ee8ab482: avcodec/vc2enc: prevent bitrate overshoots
[21:21:13 CET] <Chloe> git: 'send-email' is not a git command. See 'git --help'.
[21:21:14 CET] <Chloe> thanks gfit
[21:21:43 CET] <Chloe> why is this in a separate package
[21:24:43 CET] <JEEB> because it's just an additional perl script :D
[23:06:06 CET] <Zeranoe> Check out this message sent to me, subject line was "Do any of you idiots know how to fucking code": https://i.imgur.com/slA6QUQ.png
[23:06:43 CET] <sfan5> /g/-tier bait
[23:13:27 CET] <klaxa> lol
[23:14:42 CET] <klaxa> aren't there dozens of front-ends for ffmpeg?
[23:42:05 CET] <tmm1> lol
[23:45:31 CET] <chance83> =)
[23:47:22 CET] <chance83> ffmpeg ncurses
[23:49:29 CET] <thardin> someone doesn't know about handbrake
[23:49:52 CET] <thardin> or that you can wrap CLIs in GUIs
[23:54:36 CET] <Chloe> or that vlc and kodi use ffmpeg
[23:54:45 CET] <chance83> Chloe +1
[00:00:00 CET] --- Thu Feb  1 2018