[Ffmpeg-devel-irc] ffmpeg-devel.log.20150617

Thu Jun 18 02:05:03 CEST 2015

[00:02:25 CEST] <jamrial> wm4: you mean it should use avpriv_atomic_ptr_cas()?
[00:03:06 CEST] <wm4> jamrial: it should use pthread_once
[00:04:29 CEST] <jamrial> why? the atomic compare and swap function exists exactly for this purpose
[00:07:16 CEST] <wm4> jamrial: I suppose doing ~260 atomic ops on init isn't that expensive
[00:07:22 CEST] <jamrial> there's also apparently no wrapper for pthread_once using w32threads
[00:07:32 CEST] <wm4> but it sure as hell is roundabout as hell
[00:07:49 CEST] <wm4> yes, because fuck everything including ffmpeg and microsoft
[00:07:57 CEST] <wm4> let's just use racy code
[00:09:03 CEST] <wm4> michaelni: ok, ffio_ensure_seekback() and checksum_ptr (whatever the ridiculous crap this is) interact badly
[00:09:15 CEST] <wm4> why is aviobuf doing checksums
[00:09:28 CEST] <wm4> and in such a fragile way
[00:12:02 CEST] <nevcairiel> it would be possible to emulate pthread_once for windows without too much trouble, using the atomics, but I find pthread_once in itself an equally ugly solution
[00:12:50 CEST] <jamrial> wm4, nevcairiel: there seems to be this https://msdn.microsoft.com/en-us/library/windows/desktop/ms683493%28v=vs.85%29.aspx but it's Vista+ only
[00:12:57 CEST] <wm4> jamrial: yep
[00:13:10 CEST] <wm4> I'm confident it could be emulated on xp too
[00:13:17 CEST] <wm4> but what about OS/2?
[00:13:23 CEST] <jamrial> using atomics as nev said, yeah
[00:15:09 CEST] <jamrial> does anyone care about OS/2?
[00:15:37 CEST] <wm4> so we just break it and that-one-guy has to deal with it?
[00:15:40 CEST] <nevcairiel> apparently at least one person that keeps fixing it when we break it
[00:16:02 CEST] <nevcairiel> if we build a pre-vista layer, he could probably copy that without too much pain
[00:35:14 CEST] <wm4> rc0mbs: what the heck is bfstm anyway?
[00:35:31 CEST] <rc0mbs> (hi I'm here under my usual nick)
[00:35:35 CEST] <rc0mbs> oh wait I'm not
[00:35:44 CEST] <rcombs> welp
[00:35:55 CEST] <rcombs> wm4: WiiU sound format
[00:36:41 CEST] <wm4> ah
[00:37:38 CEST] <nevcairiel> why cant modern game consoles just use normal ordinary formats
[00:38:01 CEST] <nevcairiel> PC games usually manage to these days
[00:42:47 CEST] <BtbN> To make it harder to use their assets
[00:43:07 CEST] <wm4> nevcairiel: ordinary formats like bink?
[00:43:22 CEST] <nevcairiel> is that really used by new games anymore?
[00:43:31 CEST] <wm4> (I hope not)
[00:43:56 CEST] <wm4> but last time I talked with a gamedev about bink, he defended it
[00:46:19 CEST] <cone-928> ffmpeg 03Michael Niedermayer 07master:1125c71fd761: avcodec/jpeg2000dwt: assert that mod == 0 for encoding
[00:46:20 CEST] <cone-928> ffmpeg 03Michael Niedermayer 07master:f140a99f8b31: swscale/output: Factorize rgb48 and 64bit code
[00:46:21 CEST] <cone-928> ffmpeg 03Michael Niedermayer 07master:e29d99614969: swscale/output: Add rgba64/rgb48/bgra64/bgr48 output functions with full chroma interpolation
[00:47:04 CEST] <wm4> michaelni: that was quick
[01:26:06 CEST] <cone-928> ffmpeg 03wm4 07master:dc87758775e2: avio: fix potential crashes when combining ffio_ensure_seekback + crc
[01:58:07 CEST] <cehoyos> -Fyr-: No, I would have set the status to "open" and "reproduced" if I had.
[04:01:30 CEST] <fatter> hello 
[04:01:53 CEST] <fatter> WINDRES	libswresample/swresampleres.o
[04:01:55 CEST] <fatter> LD	libswresample/swresample-1.dll
[04:01:56 CEST] <fatter> sed -e "s/ @[^ ]*//" libswresample/swresample-1.orig.def > libswresample/swresample-1.def; lib.exe /machine:i386 /def:libswresample/swresample-1.def /out:libswresample/swresample.lib
[04:01:58 CEST] <fatter> Microsoft (R) Library Manager Version 10.00.40219.01
[04:01:59 CEST] <fatter> Copyright (C) Microsoft Corporation.  All rights reserved.
[04:02:01 CEST] <fatter>    c(ú“ libswresample/swresample.lib Œùa libswresample/swresample.exp
[04:02:02 CEST] <fatter> LD	libavcodec/avcodec-56.dll
[04:02:04 CEST] <fatter> gcc.exe: error: libavcodec/moti: No such file or directory
[04:02:05 CEST] <fatter> library.mak:106: recipe for target 'libavcodec/avcodec-56.dll' failed
[04:02:07 CEST] <fatter> make: *** [libavcodec/avcodec-56.dll] Error 1
[04:02:08 CEST] <fatter> 09:53:50 Build Finished (took 10m:26s.193ms)
[04:02:24 CEST] <fatter> what's up
[04:02:56 CEST] <fatter> i used msys2 and mingw32 and eclipse
[04:56:37 CEST] <Compn> fatter : what ver make ?
[04:57:00 CEST] Action: Compn sleeps
[05:00:06 CEST] <fatter> sorry , i have removed msys2 and reinstall it  and  retry it 
[05:36:09 CEST] <philipl> wm4: Re bink: I assumed it was because the consoles didn't include any codec licenses for devs. And bink was cheaper/easier than getting an h.264 licence.
[05:37:18 CEST] <rcombs> Nintendo outright bought a (shitty) video codec company
[05:38:05 CEST] <rcombs> for audio I guess ADPCM is light on the baby CPU and has very predictable performance
[05:39:20 CEST] <philipl> Good ol' Nintendo.
[05:39:47 CEST] <philipl> I assume the xbone and ps4 have AMD's decoder hardware and API support, but maybe I'm too optimistic.
[05:39:52 CEST] <jamrial> most games nowadays use pcm. light on the CPU and blu-rays have plenty of space anyway
[05:40:17 CEST] <philipl> unreal engine pushes people towards vorbis it seems.
[05:40:32 CEST] <rcombs> jamrial: Nintendo games (and the system menu and such) use their own special-snowflake ADPCM, but close enough
[05:41:11 CEST] <jamrial> i'm talking about ps4/xbox/PC (when they are multiplatform)
[05:41:40 CEST] <jamrial> since both the ps4 and xbox cpus are shit, those cycles they save by not decoding some lossy or lossless codec seem to be very valuable
[05:42:14 CEST] <rcombs> yup
[05:42:48 CEST] <rcombs> the WiiU seems to actually have an H.264 decoder, judging by it playing web video
[05:43:04 CEST] <rcombs> and probably a hardware encoder as well, judging by Mario Kart 8 being able to upload to YouTube
[05:46:01 CEST] <philipl> https://github.com/hcs64/ww2ogg
[05:46:11 CEST] <philipl> Why would they do that?
[05:46:33 CEST] <philipl> Vorbis but not vorbis.
[08:49:48 CEST] <j-b> 'matin
[09:11:00 CEST] <kierank> philipl: iirc vorbis codebooks are quite large
[09:29:04 CEST] <durandal_1707> j-b: ffmpeg supports sevc/evrc
[09:40:52 CEST] <Fyr> guys, if you click "Open+New Tickets", you will see 1093 bugs, part of them was opened 3-2 years and modified 3-2 years ago. is ffmpeg obsolete?
[09:42:39 CEST] <wm4> Fyr: yes it is
[09:42:54 CEST] <wm4> we're just working on it for the lack of something better tto do
[09:43:06 CEST] <Fyr> ='(
[09:43:32 CEST] <rcombs> who do I throw new fate samples at
[09:43:55 CEST] <kierank> Fyr: interesting troll
[09:44:25 CEST] <Fyr> guys, you can check that.
[09:44:44 CEST] <Fyr> the majority of the bugs was opened 3-2 years ago.
[09:46:43 CEST] <thardin> bugs or feature requests?
[09:47:02 CEST] <Fyr> tickets
[09:47:06 CEST] <thardin> bugs or feature requests?
[09:47:19 CEST] <Fyr> defect
[09:47:43 CEST] <Fyr> only a few of enhancement
[09:51:15 CEST] <thardin> seems to mostly be "ffmpeg doesn't work exactly like $TOOL"
[09:51:28 CEST] <thardin> heh, opus tickets
[09:52:12 CEST] <BtbN> Some automatic ticket closer and bumper would indeed be nice
[09:52:28 CEST] <BtbN> i'm sure most of these tickets have become pointless
[09:53:01 CEST] <Fyr> I posted a bug and see no movement for two days. =))
[09:53:10 CEST] <Fyr> even no confirmation.=/
[09:55:19 CEST] <j-b> durandal_1707: lol. Seriously?
[10:03:53 CEST] <durandal_1707> Fyr: what bug?
[10:04:07 CEST] <Fyr> https://trac.ffmpeg.org/ticket/4622#ticket
[10:04:29 CEST] <Fyr> ffmpeg doesn't add ReplayGain into m4a container.
[10:04:48 CEST] <durandal_1707> j-b: seriously, vlc have 2 years old bug report
[10:04:57 CEST] <kierank> all big projects have old bug reports
[10:05:03 CEST] <kierank> fyr
[10:05:10 CEST] <Fyr> ={
[10:06:18 CEST] <durandal_1707> Fyr: Perhaps it could be done via lavfi filter 
[10:06:28 CEST] <Fyr> maybe
[10:06:54 CEST] <Fyr> but why does ffmpeg forgets to write ReplayGain?
[10:07:11 CEST] <Fyr> it shows RG in output.
[10:07:26 CEST] <Fyr> but doesn't add it to metadata.
[10:07:35 CEST] <BtbN> Because someone forgot to implement that/there is a bug breaking it.
[10:07:37 CEST] <durandal_1707> rcombs: ask michaelni
[10:07:57 CEST] <rcombs> durandal_1707: I looked it up; turns out the correct answer is "provide a link in your commit message"
[10:08:30 CEST] <rcombs> BtbN: may as well say "because it doesn't"
[10:08:40 CEST] <rcombs> BtbN: which& is the only real answer to that question :P
[10:08:53 CEST] <durandal_1707> you could also send it 
[10:09:14 CEST] <BtbN> Isn't m4a just renamed mov/mp4?
[10:09:41 CEST] <Fyr> it is
[10:09:59 CEST] <BtbN> Yeah, not surprised if random stuff breaks with that container.
[10:10:40 CEST] <durandal_1707> perhaps if you donate some money first
[10:11:03 CEST] <Fyr> perhaps after fixing the bug.
[10:11:04 CEST] <Fyr> =))
[10:11:45 CEST] <durandal_1707> in my experiance people then just dissapear
[10:13:00 CEST] <Fyr> appreciation and gratefulness of the users are not just enough to do a pleasing to God business?
[10:13:39 CEST] <Fyr> xD
[10:14:44 CEST] <BtbN> putting a bounty on a bug definitely helps ;)
[10:14:59 CEST] <durandal_1707> there are people that steal from us
[10:15:10 CEST] <Fyr> it sounds like a bribe. =)
[10:15:10 CEST] <BtbN> There even is a service for that, and i have seen it used a few times in the ffmpeg trac
[10:15:38 CEST] <BtbN> You want some people to do something for you, in their free time. And you want them to do it faster than usual.
[11:23:31 CEST] <ubitux> wm4: it seems vda actually does the expected convert
[11:23:50 CEST] <ubitux> at least requesting rgb32 seems to work (didn't check output yet though)
[11:24:20 CEST] <wm4> interesting
[11:26:40 CEST] <ubitux> btw, any idea how much frames can be kept cached?
[11:58:24 CEST] <wm4> ubitux: what do you mean?
[12:19:40 CEST] <cbsrobot> michaelni: thanks a lot for the full "chroma interpolation" !
[13:32:05 CEST] <cone-315> ffmpeg 03Rodger Combs 07master:bb42a7d4d401: lavf/brstm: add support for BFSTM files
[14:13:41 CEST] <cone-315> ffmpeg 03Shivraj Patil 07master:63eaf529bcfa: avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for H263 lpf functions
[14:21:53 CEST] <ubitux> wm4: like, how much CVPixelBufferRef can be stored?
[14:25:39 CEST] <wm4> ubitux: I'd just assume as much as there's GPU memory
[14:25:51 CEST] <ubitux> ok
[14:25:51 CEST] <wm4> I don't know, I didn't ever have trouble with this
[14:26:05 CEST] <ubitux> your frame queue is large?
[14:26:24 CEST] <wm4> no, maybe 2 or 3 frames
[14:26:30 CEST] <ubitux> ok
[14:44:32 CEST] <JEEBsv> hmm, was the default audio bit rate also 200kbps?
[15:15:46 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07master:9841024402e8: avcodec/jpeg2000dwt: More special cases to handle singularities in 9/7 decode
[15:29:22 CEST] <JEEBsv> do I understand correctly that setting -maxrate doesn't try to collerate it between streams, random encoders and muxers just take the value of rc_max_rate and use it?
[15:29:35 CEST] <JEEBsv> at least that's how it looks like with a quick git grep
[15:33:47 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07master:7c3f7e6ed185: avcodec/jpeg2000dwt: Support 9/7 singularity cases on the encoder side
[20:50:59 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07master:2687a51a3fcf: avcodec/jpeg2000dec: Drop unneeded rlevel variable
[20:51:00 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07master:e6190045b3a1: avcodec/mqcenc: Add ff_mqc_flush_to()
[20:51:01 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07master:4624656797b6: avcodec/j2kenc: Properly flush the end of the truncated AC stream
[20:59:58 CEST] <cone-315> ffmpeg 03Vittorio Giovara 07master:d8d124eebcf2: aic: Improve error reporting
[20:59:59 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07master:98b95e746668: Merge commit 'd8d124eebcf2ec1f6a1936b12a4f00a48f08e85c'
[21:12:14 CEST] <cone-315> ffmpeg 03James Almer 07master:1e79d5c6e73a: libwebpenc: use WebPMemoryWriterClear()
[21:12:15 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07master:23c22d37d201: Merge commit '1e79d5c6e73ad131f9395f337b58a2b59ee04c1b'
[22:09:52 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:47936e1a5c87: swresample: Check the return value of resampler->init()
[22:09:53 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:b71657324d0c: libopenjpegenc: add NULL check for img before accessing it
[22:09:54 CEST] <cone-315> ffmpeg 03Ganesh Ajjanagadde 07release/2.2:0fee46e2dbc0: swresample/dither: check memory allocation
[22:09:55 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:f33adf1b5345: avformat/mxfenc: Accept MXF D-10 with 49.999840 Mbit/sec
[22:09:56 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:41e3d5bc3aef: Revert "avformat/rtpenc: check av_packet_get_side_data() return, fix null ptr dereference"
[22:09:57 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:b97fb80db9e6: swresample/swresample: Cleanup on init failure.
[22:09:58 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:eb021638a116: avcodec/atrac3plusdec: consume only as many bytes as available
[22:09:59 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:c03c2f23b229: avcodec/alsdec: Check for overread
[22:10:00 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:57c99f8fe4d6: avcodec/adpcm: Check for overreads
[22:10:01 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:78ef1be8e12e: avformat/ffmdec: Check ffio_set_buf_size() return value
[22:10:02 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:4cb077793d00: avcodec/jpeg2000dec: Check that coords match before applying ICT
[22:10:03 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:4348a241ed53: avcodec/x86/h264_weight: handle weight1=128
[22:10:04 CEST] <cone-315> ffmpeg 03Simon Thelen 07release/2.2:b2d7a0840534: libavutil/channel_layout: Correctly return layout when channel specification ends with a trailing 'c'.
[22:10:05 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:4a5a6fe6ba1a: avdevice/lavfi: do not rescale AV_NOPTS_VALUE in lavfi_read_packet()
[22:10:06 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:d6bde7be5f34: takdec: ensure chan2 is a valid channel index
[22:10:07 CEST] <cone-315> ffmpeg 03Deliang Fu 07release/2.2:557aa4c6b32a: avformat: Fix bug in parse_rps for HEVC.
[22:10:08 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:8bb21c031074: ffmpeg_opt: Check for localtime() failure
[22:10:09 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:de8e63ca8215: mov: abort on EOF in ff_mov_read_chan
[22:10:10 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:ebf0a8fbda1f: vp9: change type of tile_size from unsigned to int64_t
[22:10:11 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:fabb394a2dd0: vp8: change mv_{min,max}.{x,y} type to int
[22:10:12 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:f87d76e659e0: sonic: set avctx->channels in sonic_decode_init
[22:10:13 CEST] <cone-315> ffmpeg 03wm4 07release/2.2:539603e8770f: avio: fix potential crashes when combining ffio_ensure_seekback + crc
[22:10:14 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:d29153591299: diracdec: check that block length is valid
[22:10:15 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:3a0c421b5669: aacsbr: break infinite loop in sbr_hf_calc_npatches
[22:10:16 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:9f9af1dd04fb: libavutil/mem: use size_t for the length in av_strdup()
[22:10:17 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:0b6c8bb26c62: avutil/dict: Use size_t for appending strings
[22:10:18 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:d27b9d3bd1d2: avformat/vorbiscomment: Check entry length in ff_vorbiscomment_write()
[22:10:19 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:f76b54fba781: avutil/avstring: Use size_t in av_strlcatf()
[22:10:20 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:17ad9778a9a3: avformat/url: Use size_t for len from strlen()
[22:10:21 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:404db2b99eca: avformat/subtitles: Use size_t for len
[22:10:22 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:6bb62d64d240: tools/graph2dot: use larger data types than int for array/string sizes
[22:10:23 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:0452e3e41cd0: avformat/rtpdec_xiph: Check upper bound on len in xiph_handle_packet()
[22:10:25 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:b8060bc217e4: avformat/rtpenc_jpeg: Check remaining buffer size for SOS
[22:10:25 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:3062689f3766: avformat/nutdec: Fix use of uinitialized value
[22:10:26 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:ed578c91bfdf: avformat/matroskadec: Cleanup error handling for bz2 & zlib
[22:10:27 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:55a9a373590b: avformat/hevc: Check cpb_cnt_minus1
[22:10:28 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:06fda5bef349: avformat/hevc: Check num_negative_pics and num_positive_pics
[22:10:29 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:14d69655b782: aacdec: don't return frames without data
[22:10:30 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:e1b98555b9aa: avcodec/vqavideo: Check chunk size
[22:10:31 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:6d6acef9fb83: avcodec/hevc_sei: Check num_sps_ids_minus1 value
[22:10:32 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:eed8ccd9eb3e: avcodec/hevc: Check offset_len
[22:10:33 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:87d2c6105df2: avcodec/hevc: Check num_entry_point_offsets
[22:10:34 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:a2e183aa0552: avcodec/hevc_ps: Check vps_num_hrd_parameters
[22:10:35 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:003b006c8564: avcodec/hevc_ps: Explicitly check num_tile_* for negative values
[22:10:36 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:046439d6084e: avcodec/jpeg2000dec: fix boolean operator
[22:10:37 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:611791e77e94: avcodec/mjpegdec: fix len computation in ff_mjpeg_decode_dqt()
[22:10:38 CEST] <cone-315> ffmpeg 03Rainer Hochecker 07release/2.2:02c51e05640c: swr: fix alignment issue caused by 8ch sse functions
[22:10:39 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:05a3a3e5b0d9: avcodec/mpeg4audio: add some padding/alignment to MAX_PCE_SIZE
[22:10:40 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:e52d66c647e6: avcodec/cavsdec: Check esc_code
[22:10:41 CEST] <cone-315> ffmpeg 03wm4 07release/2.2:f06168f7dcd2: hevc: make avcodec_decode_video2() fail if get_format() fails
[22:10:42 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:d9194389049b: avcodec/dcadec: Check nchans
[22:10:43 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:b259c7e89101: avcodec/dcadec: Check subsubframes
[22:10:44 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:f8f24d5dbdf9: ffmpeg_opt: Fix -timestamp parsing
[22:10:45 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:cd489b8f6469: avcodec/proresdec2: Reset slice_count on deallocation
[22:10:46 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:9f885733d6a9: avcodec/shorten: Fix code depending on signed overflow behavior
[22:10:47 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:16af12a807ce: avcodec/shorten: Check skip_bytes()
[22:10:48 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:47db54cef679: avcodec/shorten: More complete pred_order check
[22:10:49 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:49a759fb3fe9: avcodec/smvjpegdec: check avcodec_decode_video2() return code
[22:10:50 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:3a43c712aebc: avcodec/sonic: check memory allocations
[22:10:51 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:3344a28787a7: avcodec/sonic: More completely check sample_rate_index and channels
[22:10:52 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:3edb74db3bc9: avcodec/dcadec: Check scale table index
[22:10:53 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:0122e38cb19b: avcodec/dcadec: Check active_bands
[22:10:54 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:ed6d7696db31: avcodec/libtheoraenc: Check for av_malloc failure
[22:10:55 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:9cee1c4ac7cd: avcodec/hevc: Fix typo in num_entry_point_offsets check
[22:10:56 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:c8676f9daf9d: avcodec/golomb: get_ur_golomb_jpegls: Fix reading huge k values
[22:10:57 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:6ad365467097: avformat/nutdec: Return error on EOF from get_str()
[22:10:58 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:8cfc0fdb23e1: avformat/nutdec: Fix recovery when immedeately after seeking a failure happens
[22:10:59 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:fe1e6b919d8d: avformat/nutdec: Check X in 2nd branch of index reading
[22:11:00 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:9414358f1586: nutdec: fix infinite resync loops
[22:11:01 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:4d625f270a59: nutdec: stop skipping bytes at EOF
[22:11:02 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:dfd1d487301b: nutdec: abort if EOF is reached in decode_info_header/read_sm_data
[22:11:03 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:9d6d4eb42542: avcodec/put_bits: Update size_in_bits in set_put_bits_buffer_size()
[22:11:04 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:4f09968865a0: avcodec/mpegvideo_enc: Update the buffer size as more slices are merged
[22:11:05 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:2b499b694002: avcodec/put_bits: Assert that there is enough space left in skip_put_bytes()
[22:11:06 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:249021eb2336: avcodec/bitstream: Assert that there is enough space left in avpriv_copy_bits()
[22:11:07 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:f494b906e344: avformat/utils: Ensure that AVFMT_FLAG_CUSTOM_IO is set before use
[22:11:08 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:e96e0e1333c8: mpeg4videodec: only allow a positive length
[22:11:09 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:940b9484dcde: bink: check vst->index_entries before using it
[22:11:10 CEST] <cone-315> ffmpeg 03Vittorio Giovara 07release/2.2:cd5a832b0aab: lavf: Reset global flag on deinit
[22:11:11 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:fe8a763bc706: tests/fate-run: do not attempt to parse tiny_psnrs output if it failed
[22:11:12 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:d9ff9b8ab0b5: apedec: set s->samples only when init_frame_decoder succeeded
[22:11:13 CEST] <cone-315> ffmpeg 03Maneesh Gupta 07release/2.2:25daf27484ea: OpenCL: Avoid potential buffer overflow in cmdutils_opencl.c
[22:11:14 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:02b434044caa: ffmpeg: remove incorrect network deinit
[22:11:15 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:b548a6208eed: nutdec: check for negative frame rate in decode_info_header
[22:11:16 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:ffdd93fe79b7: nutdec: fix illegal count check in decode_main_header
[22:11:17 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:45ea5ba0686c: avcodec/wavpack: Check L/R values before use to avoid harmless integer overflow and undefined behavior in fate
[22:11:18 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:1f5eeed43ba4: matroskadec: use uint64_t instead of int for index_scale
[22:11:19 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:80cd5abf7fdb: avcodec/ffv1dec: Check chroma shift parameters
[22:11:20 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:7d9725187d1a: matroskadec: check s->streams[k] before using it
[22:11:21 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:4f0d1af68ad3: avformat/matroskadec: Use tracks[k]->stream instead of s->streams[k]
[22:11:22 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:a64102e25fb9: diracdec: prevent overflow in data_unit_size check
[22:11:23 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:b946f849bb92: diracdec: avoid overflow of bytes*8 in decode_lowdelay
[22:11:24 CEST] <cone-315> ffmpeg 03Andreas Cadhalpun 07release/2.2:c9c5bd8c89b6: diracdec: check if reference could not be allocated
[22:11:25 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:cea2106fb2e1: avcodec/h264_slice: More complete cleanup in h264_slice_header_init()
[22:20:13 CEST] <durandal_1707> michaelni: backport exr fixes?
[22:21:46 CEST] <michaelni> durandal_1707, which git commit hash ?
[22:22:19 CEST] <cone-315> ffmpeg 03Michael Niedermayer 07release/2.2:051cd7dc5f42: Update for 2.2.16
[22:23:07 CEST] <michaelni> btw you can cherry pick commits to release branches too if you see anythng else that should be backported
[22:28:12 CEST] <durandal_1707> michaelni: last commit to exr, I'm on phone
[22:41:50 CEST] <michaelni> durandal_1707, that one says its a regression since 95582b5c, 95582b5c is in 2.3 but not 2.2
[22:44:20 CEST] <durandal_1707> michaelni: ok, backport to where it is sane
[22:44:36 CEST] <michaelni> ok
[23:52:17 CEST] <cone-315> ffmpeg 03George Boyle 07master:47cadf4d1fa0: avcodec/flacenc: Regression test for ticket #4628
[00:00:00 CEST] --- Thu Jun 18 2015