[Ffmpeg-devel-irc] ffmpeg-devel.log.20170402

[burek]

[00:03:27 CEST] <ashk43712> Hi, I was thinking to work on the issue https://trac.ffmpeg.org/ticket/2112. It's for adding support for NL filter of gimp.
[00:03:36 CEST] <ashk43712> Basically it's a non linear filter that uses a hexagonal block whose size can be set with the Radius option.
[00:03:47 CEST] <ashk43712> Are thery any similar filters in ffmpeg that I can use for reference purpose?
[00:16:56 CEST] <RiCON> ashk43712: https://ffmpeg.org/ffmpeg-filters.html#nlmeans ?
[00:18:35 CEST] <ashk43712> yeah, seems similar thanks
[01:00:50 CEST] <cone-163> ffmpeg 03James Almer 07master:cbd250293998: avcodec/extract_extradata_bsf: make sure all parameter set NAL units were found for h264/hevc
[01:39:32 CEST] <Dorjepy> I'm looking for detailed  info on developing filters? 
[03:14:09 CEST] <b0bby__> hey
[03:14:09 CEST] <b0bby__> trying to learn how to use the ffmpeg libraries (libav) and Im getting errors([vp8 @ 0x19492a0] Invalid profile 5.
[03:14:09 CEST] <b0bby__> [vp8 @ 0x19492a0] Header size larger than data provided) with the code https://gist.github.com/anonymous/2cd6151c667146dceb0b7b08762de903(taken for the examples page). It worked fine with mpeg formates but I tried changing the codec to vp8 and giving the program a webm only to get errors.
[03:43:26 CEST] <DHE> b0bby__: libav questions still go to #ffmpeg
[05:11:07 CEST] <stevenliu> Hi Clément BSsch?
[05:28:56 CEST] <jamrial> stevenliu: he's ubitux, but he's probably sleeping right now
[05:29:53 CEST] <stevenliu> Do we merge the code from libav or just merge options and functions?
[06:00:17 CEST] <cone-510> ffmpeg 03James Almer 07master:7942907878dd: compat/atomics: fix atomic_fetch_xor
[08:10:07 CEST] <kkanungo17> atomnuker : please check the shared proposal
[10:51:14 CEST] <ubitux> stevenliu: we want to have the same features
[10:51:22 CEST] <ubitux> but it doesn't need to follow exactly the same option names
[10:51:40 CEST] <ubitux> ah. meh.
[12:04:12 CEST] <ubitux> so
[12:04:33 CEST] <ubitux> http://sprunge.us/hReS this is the first batch of the bitstream api
[12:04:41 CEST] <ubitux> which of these is slower sometimes?
[12:05:22 CEST] <ubitux> i'd like to reproduce any slowdown
[12:05:40 CEST] <ubitux> if we don't do anything, it's going to stall the merge process for months again
[12:05:49 CEST] <ubitux> and i really don't want that
[12:05:59 CEST] <ubitux> we already have 5 to 6 months delay in the merge
[12:06:28 CEST] <ubitux> we can just skip any codec that is slower
[12:09:48 CEST] <nevcairiel> you would probably have to benchmark each and every codec on a bunch of arches that matter to find those
[12:12:36 CEST] <ubitux> well, if people are saying that there is a slowdown, there sure is already existing benchmarks, no?
[12:13:22 CEST] <nevcairiel> there are benchmarks of codecs that matter, that libav hasnt pushed yet, because there are slowdowns, not sure if they benchmarked every single fringe codec
[12:13:34 CEST] <nevcairiel> the benchmarking effort alone has been keeping them busy, nevermind finding out why its slower
[12:21:42 CEST] <ubitux> so far it seems they have ported 89 decoders
[12:21:52 CEST] <ubitux> (pushed)
[12:22:06 CEST] <ubitux> is there any important slowdown in those?
[12:26:47 CEST] <wbs> afaik, the intent has been so far that the pushed ones either had no significant slowdown, or were deemed to be a fringe codec where a small slowdown on some arches didn't matter
[12:29:03 CEST] <nevcairiel> imho figuring out the important cases first would've been wiser, that way you still have room to change the reader fundamentally if that needs to happen, but the entire thing was rushed
[12:29:37 CEST] <PaoloP> Hello. I did not receive any feedback for the API usage snippet I posted, according to the modifications that ubitux, michaelni, wm4 and others suggested. Http://ffmpeg.org/pipermail/ffmpeg-devel/2017-March/209494.html  Does it mean that the snippet has not been accepted or it has not been examined yet? Please tell me a word! I think the snippet is useful for learning the API and I can provide other examples as 
[12:29:39 CEST] <PaoloP> well (now that I learnt all the coding rules and the "template", it would be very easy, for me, to make them). But, at least, let me know if the snippet has been rejected or it has to be examined.
[12:31:18 CEST] <atomnuker> ubitux: my offer to keep the API and for me to merge whatever conflicts in non-obvious ways still stands
[12:31:57 CEST] <nevcairiel> will you also provide the speedups the new reader provides? :)
[12:33:01 CEST] <nevcairiel> its not only sometimes slower and sometimes equal, but also quite a bit faster on certain things
[12:33:36 CEST] <ubitux> we should probably macroify libavcodec/bitstream.h to support a 32-bit variant
[12:33:39 CEST] <atomnuker> I did say that we can have adjustable window sizes depending on an ifdef, didn't I
[12:34:01 CEST] <ubitux> well if you're doing this, can't we do this on the new api?
[12:34:14 CEST] <atomnuker> no, it sucks
[12:34:18 CEST] <nevcairiel> the problem is the api contract, you cant read 32-bits in one call when you only have a 32-bit cache
[12:35:09 CEST] <nevcairiel> which is why get_bits only read 25 bits
[12:35:33 CEST] <atomnuker> get bits isn't required to only read 25 bits
[12:36:44 CEST] <atomnuker> we can have it read any arbitrary number of bits an all get_bitsX functions just inline that version
[12:36:57 CEST] <nevcairiel> that doesnt sound very optimized
[12:36:57 CEST] <ubitux> wbs: do you know if anyone is actually working on this? (that is, working on the implementation & api to solve the problem, not just benchmarks)
[12:37:15 CEST] <wbs> ubitux: afaik it's only benchmarking so far
[12:37:21 CEST] <ubitux> :/
[12:38:06 CEST] <atomnuker> as I said before I just think the API is a steaming pile of shit and I'd like it to stay away
[12:40:22 CEST] <ubitux> still, it's faster on 64-bit and the api is simpler (a lot of macros seems to go away and replaced by one call) 
[12:40:37 CEST] <ubitux> so while it's not the best it still provides improvements over the current situation
[12:40:56 CEST] <ubitux> can you make the current API on par with these 2 points?
[12:41:46 CEST] <ubitux> i'd rather improve the new api, but we may walk on other developers feet (even though it seems they're not actually working on it)
[12:52:41 CEST] <atomnuker> ubitux: remove all _le functions, remove get_bitsz, make the reader do 32 bits by default unless something else has been defined and there's a good API
[12:54:36 CEST] <nevcairiel> at the same time it shall not get slower
[12:54:56 CEST] <atomnuker> it won't if we keep the 32 bit window
[12:55:14 CEST] <ubitux> the _le() you mean the thing only used in 2 places?
[12:55:16 CEST] <nevcairiel> you say that now
[12:55:36 CEST] <nevcairiel> but there is a reason get_bitsz exists to avoid that check in the optimized reader
[12:56:57 CEST] <nevcairiel> or why it can only read 25 bits
[12:57:25 CEST] <atomnuker> wait, nevermind, that's a dumb idea, the get_bits file is meant to contain useful functions for reading various bits in multimedia
[12:57:34 CEST] <atomnuker> and not a unified API for bit reading
[12:58:46 CEST] <nevcairiel> so you want it to  be a cess pool of random functions instead of a proper API? :D
[12:59:49 CEST] <atomnuker> well, yeah
[13:00:18 CEST] <atomnuker> if something's useful from there we use it, else we do it in codecs, if too many codecs use it we add it there
[13:01:29 CEST] <atomnuker> what's the point of having a single API if we have to do everything in the codecs anyway?
[13:02:48 CEST] <ubitux> atomnuker: tbh, i'm fine with keeping our API if and only if it's faster
[13:02:53 CEST] <ubitux> but currently it isn't
[13:03:01 CEST] <atomnuker> where?
[13:03:07 CEST] <ubitux> on 64-bit
[13:03:25 CEST] <atomnuker> theirs lost on some codecs on 64 bits too
[13:03:40 CEST] <ubitux> overall it's faster
[13:03:44 CEST] <nevcairiel> basically take a high-bitrate video codec like prores or something, it gets significant speedups there, especially on 64-bit
[13:04:17 CEST] <atomnuker> then clearly what we should do is do manual reading for those codecs
[13:04:32 CEST] <atomnuker> (like we do in dnxhr)
[13:04:49 CEST] <atomnuker> (which is faster with our reader's macros)
[13:07:21 CEST] <nevcairiel> the manual macro way is only faster because the get_bits reader lacks a persistent cache, so you make it kinda-persistent by using the macros directly (which use a local variable as a cache)
[13:07:54 CEST] <nevcairiel> but it makes for really terrible reading api
[13:17:09 CEST] <atomnuker> how about this: we merge their API but we rename all functions to bs_get_X?
[13:19:12 CEST] <atomnuker> I'd agree to that since we can change the put_bits API to bs_put_X (and they can't because they dug themselves into this hole where bitstream is only for reading)
[13:22:50 CEST] <durandal_1707> bs?
[13:24:41 CEST] <atomnuker> short for bitstream
[14:13:33 CEST] <wm4> so why does the mjpeg encoder not allow yuv420p, only the J variant?
[14:13:40 CEST] <wm4> is this a stupid ffmpeg.c thing
[14:15:26 CEST] <nevcairiel> encoding a limited range jpeg would probably be rather  broken, and encoders cant signal otherwise what they want?
[14:16:59 CEST] <wm4> no, but it requires users to use a deprecated pixfmt
[14:19:07 CEST] <nevcairiel> its been depracted for so long without being removed because some things are just not fixed to take the  range  in account otherwise, unfortunately
[14:19:30 CEST] <durandal_1707> lazy developers 
[14:20:04 CEST] <durandal_1707> and ugly swscale
[17:50:21 CEST] <wm4> Assertion index < q->nb_elements failed at libavfilter/af_dynaudnorm.c:206
[17:50:24 CEST] <wm4> is that new or my fault
[18:01:48 CEST] <ubitux> wm4: did you swtich to assert level 2 recently?
[18:01:57 CEST] <ubitux> for, let's say, debug the threading issue?
[18:02:18 CEST] <wm4> oh yes
[18:02:19 CEST] <wm4> I did
[18:52:26 CEST] <durandal_1707> wm4: how to reproduce?
[18:52:45 CEST] <wm4> durandal_1707: enable dynaudnorm during runtime in mpv
[19:00:02 CEST] <ubitux> lol why did i wrote "filter"
[19:00:05 CEST] <ubitux> i meant API, obviously
[19:10:07 CEST] <jamrial> ubitux: probably because the previous merge blocker was the new bitstream filter API :P
[19:10:21 CEST] <jamrial> also, this one would be bitstream reader API i think
[19:10:37 CEST] <ubitux> did they write a bitstream writer?
[19:10:44 CEST] <durandal_1707> wm4: i cant repro that
[19:11:34 CEST] <jamrial> a new one? i don't think they did. still put_bits.h
[19:11:45 CEST] <durandal_1707> wm4: but i can repro sync issues if one of chapters is different sample rate
[19:15:11 CEST] <atomnuker> ubitux: with the way they treat writing rawbits you'd almost think they'll remove all encoding support
[19:21:45 CEST] <ubitux> jamrial: i'll skip and continue the merge tomorrow
[19:24:34 CEST] <jamrial> do we cut 3.3 here then, or try to get to the major bump asap and cut there instead?
[19:24:47 CEST] <jamrial> right before the bump, that is
[19:25:34 CEST] <jamrial> also, by skipping the bitstream commits the backlog is ~100 commits shorter :D
[19:34:51 CEST] <ubitux> i don't really care about the release
[19:36:04 CEST] <durandal_1707> :(
[19:42:16 CEST] <jamrial> michaelni: up to you then i guess. re cut 3.3 at current git head, use the existing branch in its current cut point, or wait a few weeks until we get to the major bump
[19:52:59 CEST] <michaelni> ok, ill then do 2 ver bumps and update release/3.3 to master  unless there are objections
[19:53:06 CEST] <michaelni> ill also do some tests
[19:53:11 CEST] <nevcairiel> sounds fine to me
[20:17:39 CEST] <kierank> slice threads seems ok now but I am not fuzzing at google scale
[20:19:29 CEST] <durandal_1707> bribe google to fuzz slices
[20:20:02 CEST] <kierank> i need to email the guy at fosdem
[20:20:06 CEST] <kierank> and send him our corpus
[20:20:09 CEST] <kierank> fate is not a good test corpus
[20:20:34 CEST] <kierank> and make them use fffuzz that actually tests sliced threads
[20:22:12 CEST] <durandal_1707> whats point of audiotoolbox support and mediafoundation?
[20:22:35 CEST] <durandal_1707> are we ditching native decoders?
[20:22:53 CEST] <JEEB> it's mostly for encoding on specific things as far as I can recall
[20:23:03 CEST] <JEEB> although the wrappers do come with decoders as well
[20:24:39 CEST] <ubitux> durandal_1707: it's useful on embedded system where you can have licensing/fee issues
[20:24:51 CEST] <ubitux> (+ hw accelerated)
[20:25:28 CEST] <durandal_1707> so what, this hw accel feels like cancer
[20:25:49 CEST] <kierank> durandal_1707: it's not getting in the way imo
[20:25:54 CEST] <ubitux> durandal_1707: embedded system again
[20:26:04 CEST] <ubitux> power consumption etc
[20:26:14 CEST] <ubitux> think: iphones
[20:28:12 CEST] <durandal_1707> thats all blackbox
[20:33:49 CEST] <kierank> durandal_1707: ofc
[20:50:29 CEST] <cone-166> ffmpeg 03Michael Niedermayer 07master:e1cc7f83df00: Bump minor for 3.3
[20:50:29 CEST] <cone-166> ffmpeg 03Michael Niedermayer 07master:22b0daa1b3f0: Bump versions for master after 3.3
[20:55:48 CEST] <jamrial> wonder if the branch merge will get spammed here by the bot
[21:01:26 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:fe7bc1f16aba: configure: Do not unconditionally check for (and enable) xlib
[21:01:27 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:831005b2302c: configure: Log correct test name and use correct filter when testing objective C flags
[21:01:28 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:8a6e7a67cb29: configure: Use check_cpp in CPP flags tests
[21:01:29 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:c78495d1cdac: configure: Log name and parameters of all helper functions where it makes sense
[21:01:30 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:2dd464868c64: configure: Move license checks directly after command line parsing
[21:01:31 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:72a19f4013ec: mpegaudiodsp: aarch64: Adjust function prototype after 2caa93b813adc5dbb7771dfe615da826a2947d18
[21:01:32 CEST] <cone-166> ffmpeg 03Anton Khirnov 07release/3.3:84f225684cd3: pthread_frame: properly propagate the hw frame context across frame threads
[21:01:33 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:f4bf236338f6: matroskaenc: fix muxing AAC streams when using aac_adtstoasc bsf
[21:01:34 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:98cae966c778: matroskaenc: write updated STREAMINFO metadata for FLAC streams if available
[21:01:35 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:a4cfcddcb0f7: vp9: Make the subpel filters non-static
[21:01:36 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:c44a8a3eabcd: aarch64: Add an offset parameter to the movrel macro
[21:01:37 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:383d96aa2229: aarch64: vp9: Add NEON optimizations of VP9 MC functions
[21:01:38 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:557c1675cf0e: arm: vp9mc: Minor adjustments from review of the aarch64 version
[21:01:39 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:6a62795d4051: aarch64: h264idct: Use the offset parameter to movrel
[21:01:40 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:905cdcaa9d08: examples/decode_audio: Add missing header for av_free()
[21:01:41 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:824e8c284054: arm: Clear the gp register alias at the end of functions
[21:01:42 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:11623217e3c9: arm: vp9mc: Use a different helper register for PIC loads
[21:01:43 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:fd0fae60372c: pthread_frame: Unreference hw_frames_ctx on per-thread codec contexts
[21:01:44 CEST] <cone-166> ffmpeg 03Ronald S. Bultje 07release/3.3:0b37cd09a67c: checkasm: add vp9dsp.itxfm_add tests.
[21:01:45 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:a67ae6708315: arm: vp9: Add NEON itxfm routines
[21:01:46 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:12db2832e41a: libxvid: Require availability of mkstemp()
[21:01:47 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:e5e8a26dcf6d: libxvid: Use proper context in av_log() calls
[21:01:48 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:f7d183f08472: libxvid: Check return value of write() call
[21:01:49 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:dd299a2d6d4d: arm: vp9: Add NEON loop filters
[21:01:50 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:3b50dbc51fb0: ratecontrol: Use correct function pointer casts instead of void*
[21:01:51 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:fcbdd605b540: nut: Use correct function pointer casts instead of void*
[21:01:52 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:8ddfa5ae5ef6: vf_drawtext: Drop wrong void* cast
[21:01:53 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:d316f9cefcd8: aac: Drop pointless cast
[21:01:54 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:800d91d348c8: Drop pointless void* casts
[21:01:55 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:01348e411f96: avconv_opt: Consistently iterate through hwaccels array in all cases
[21:01:56 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:52d196fb30fb: arm: vp9itxfm: Simplify txfm string comparisons
[21:01:57 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:9d2afd1eb8c5: aarch64: vp9: Implement NEON loop filters
[21:01:58 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:3c9546dfafcd: aarch64: vp9: Add NEON itxfm routines
[21:01:59 CEST] <cone-166> ffmpeg 03Janne Grunau 07release/3.3:31756abe29eb: aarch64: vp9: loop_filter: fix typo in skip flatout8 check
[21:02:00 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:cd1047f3911f: qsvdec: Pass the correct profile to libmfx
[21:02:01 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:3297577f3eac: mpegvideo: Return correct coded frame sizes from parser
[21:02:02 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:030d84fa2e35: qsvdec: Pass field order information to libmfx
[21:02:03 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:0940b748bdba: qsvdec: Only warn about unconsumed data if it happens more than once
[21:02:04 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:fea4dc05b41f: vc1: Return stream format information from parser
[21:02:05 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:b6582b29277e: qsv: Add VC-1 decoder
[21:02:06 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:182cf170a544: vp8: Return stream format information from parser
[21:02:07 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:e0b164576f74: qsv: Add VP8 decoder
[21:02:08 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:fbd1f7639d01: af_asyncts: Use llabs instead of labs for 64-bit variable
[21:02:09 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:f7407f56cbf8: golomb: Replace __PRETTY_FUNCTION__ with __func__ for tracing
[21:02:10 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:ffe89e1edb02: configure: Move mjpeg_vaapi_decoder dependency declarations to the right place
[21:02:11 CEST] <cone-166> ffmpeg 03Gianluigi Tiesi 07release/3.3:e17567a831de: libilbc: support for latest git of libilbc
[21:02:12 CEST] <cone-166> ffmpeg 03Janne Grunau 07release/3.3:d7595de0b25e: aarch64: vp9: use alternative returns in the core loop filter function
[21:02:13 CEST] <cone-166> ffmpeg 03Janne Grunau 07release/3.3:e7ae8f7a7158: aarch64: vp9: loop filter: replace 'orr; cbn?z' with 'adds; b.{eq,ne};
[21:02:14 CEST] <cone-166> ffmpeg 03Martin Storsjö 07release/3.3:81d7f0bbca83: checkasm: vp9dsp: Benchmark the dc-only version of idct_idct separately
[21:02:15 CEST] <cone-166> ffmpeg 03Vittorio Giovara 07release/3.3:d5d62ce6d643: mov: Fix identity matrix boolean logic
[21:02:16 CEST] <cone-166> ffmpeg 03Vittorio Giovara 07release/3.3:481ff3cf0188: fate: Add h264 and hevc extradata reload tests
[21:02:17 CEST] <cone-166> ffmpeg 03Marton Balint 07release/3.3:bcefafa226dc: avisynth: Fix setting stream timebase
[21:02:18 CEST] <cone-166> ffmpeg 03Stephen Hutchinson 07release/3.3:aaae59700f7f: avisynth: Simplify the pix_fmt check for the newer AviSynth API
[21:02:19 CEST] <cone-166> ffmpeg 03Stephen Hutchinson 07release/3.3:3cc3463f306f: avisynth: Support pix_fmts added to AviSynth+
[21:02:20 CEST] <cone-166> ffmpeg 03Stephen Hutchinson 07release/3.3:bf8646274b3e: doc: Add note about recent regression in AviSynth+
[21:02:21 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:bfe92dfe60f6: Ignore all generated example binaries
[21:02:22 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:bb265b764a05: examples/transcode_aac: Drop pointless return value const qualifier
[21:02:23 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:715b82434608: qsv: Drop some unused variables
[21:02:24 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:76167140a91c: qsvdec: Drop stray extra braces around initializer
[21:02:25 CEST] <cone-166> ffmpeg 03Christian Suloway 07release/3.3:d860a3cc0a12: crypto: Add encryption support
[21:02:26 CEST] <cone-166> ffmpeg 03Luca Barbato 07release/3.3:0a4b9d0ccd10: hlsenc: Add encryption support
[21:02:27 CEST] <cone-166> ffmpeg 03Luca Barbato 07release/3.3:adb0e941c329: avpacket: Mark src pointer as constant
[21:02:28 CEST] <cone-166> ffmpeg 03Anton Khirnov 07release/3.3:0d3176e32f35: hwcontext_dxva2: do not assume the destination format during mapping is always the right one
[21:02:29 CEST] <cone-166> ffmpeg 03Anton Khirnov 07release/3.3:9d7026574bbb: hwcontext_dxva2: fix handling of the mapping flags
[21:02:30 CEST] <cone-166> ffmpeg 03Anton Khirnov 07release/3.3:5a1d605ceae4: hwcontext_dxva2: split transfer_data() into upload/download functions
[21:02:31 CEST] <cone-166> ffmpeg 03Anton Khirnov 07release/3.3:e18ba2dfd2d1: hwcontext_dxva2: make sure the sw frame format is the right one during transfer
[21:02:32 CEST] <cone-166> ffmpeg 03Michael Niedermayer 07release/3.3:58b867a7cfea: Bump minor versions for master after release/3.3 branchpoint
[21:02:33 CEST] <cone-166> ffmpeg 03Sasi Inguva 07release/3.3:ef71dc794832: lavf/mov.c: Add -advanced_editlist option for mov format.
[21:02:34 CEST] <cone-166> ffmpeg 03Ronald S. Bultje 07release/3.3:f800d6508d7e: dnxhd: initialize DNXHDContext::avctx to each thread's respective one.
[21:02:35 CEST] <cone-166> ffmpeg 03Ronald S. Bultje 07release/3.3:9e2050b698b2: codec_desc: mark fraps as an intra-only codec.
[21:02:36 CEST] <cone-166> ffmpeg 03Ronald S. Bultje 07release/3.3:73f863d751df: fic: set pict_type/key_frame after (instead of during) slice decoding.
[21:02:37 CEST] <cone-166> ffmpeg 03Ronald S. Bultje 07release/3.3:081c21ca55d7: lagarith: assign correct per-thread value to LagarithContext::avctx.
[21:02:38 CEST] <cone-166> ffmpeg 03Ronald S. Bultje 07release/3.3:b5300c8ad8c5: h264: don't write to source picture object in ff_h264_ref_picture().
[21:02:39 CEST] <cone-166> ffmpeg 03Ronald S. Bultje 07release/3.3:1ddc37051f11: h264: only assign H264Picture::mbaff for first slice.
[21:02:40 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:0505a1d9c4be: doc/libav-merge: remove line about AC3 fixed decoder speedup
[21:02:41 CEST] <cone-166> ffmpeg 03Carl Eugen Hoyos 07release/3.3:76dd87c92969: lavf/amr: Return AVERROR_EOF on EOF.
[21:02:42 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:9033e8723c86: avutil/spherical: add av_spherical_projection_name()
[21:02:43 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:2efb70c37992: avformat/dump: use av_spherical_projection_name() to print spherical projection names
[21:02:44 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:2a2854f57842: ffprobe: use av_spherical_projection_name() to print spherical projection names
[21:02:45 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:f0df60d392d6: Merge commit 'fe7bc1f16abaefe66d8a20f734ca3eb8a4ce4d43'
[21:02:46 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:e5177e8f8db0: Merge commit '831005b2302cbeb377e3f00fd18c78928bcec185'
[21:02:47 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:3795899978cf: Merge commit '8a6e7a67cb2943f552569801539829a304971302'
[21:02:48 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:8d50dd976d29: Merge commit 'c78495d1cdac6dd13786a7e5571b606604a360bd'
[21:02:49 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:0ad9aff0227a: Merge commit '2dd464868c64fa21a6e3bd63ad364ff12999c7d0'
[21:02:50 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:5694427dc384: Merge commit '72a19f4013ec2c7f8581416f8ad4bf81df163fb6'
[21:02:51 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:e2f6e1c4a825: Merge commit '84f225684cd389747907381122c073aa1c8b6bf1'
[21:02:52 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:13a211e6320d: Merge commit 'f4bf236338f6001736a4784b9c23de863057a583'
[21:02:53 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:51d95914e00c: Merge commit '98cae966c77875e26c5958206a6cfe7eba6269e8'
[21:02:54 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:28bace5c0a75: Merge commit 'a4cfcddcb0f76e837d5abc06840c2b26c0e8aefc'
[21:02:55 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:afac31d2b2cf: Merge commit 'c44a8a3eabcd6acd2ba79f32ec8a432e6ebe552c'
[21:02:56 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:183635b9e96b: doc/libav-merge: mention aac_adtstoasc extradata update fix for matroska
[21:02:56 CEST] <durandal_1707> SPAM
[21:02:57 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:2348eb171455: Merge commit '383d96aa2229f644d9bd77b821ed3a309da5e9fc'
[21:02:58 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:8408c2efe4bf: Merge commit '557c1675cf0e803b2fee43b4c8b58433842c84d0'
[21:02:59 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:67eda469ab34: Merge commit '6a62795d4051f435a9a2c59395d96913693922f8'
[21:03:00 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:ebe9808aaad6: Merge commit '905cdcaa9d081d3d945ce555b27b43a75c3af57b'
[21:03:01 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:f6e37665a9dd: Merge commit '824e8c284054f323f854892d1b4739239ed1fdc7'
[21:03:02 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:bf624e993d6f: Merge commit '11623217e3c9b859daee544e31acdd0821b61039'
[21:03:03 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:a9134fa7135b: Merge commit 'fd0fae60372cddbe0bec8830d07e760195f80bad'
[21:03:04 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:5e7288480f02: Merge commit '0b37cd09a67c3ba4db044404b99c65a32b4ad932'
[21:03:05 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:037522a75ade: Merge commit 'a67ae67083151f2f9595a1f2d17b601da19b939e'
[21:03:06 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:ef516efad445: Merge commit '12db2832e41aa71b5903ef7fa5c59c5473ded2c5'
[21:03:07 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:804ae6e30fc7: Merge commit 'e5e8a26dcf6d572e841a7a191e4c96524367e3f9'
[21:03:08 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:a1afcf8b8c30: Merge commit 'f7d183f08472e566a2e6b62a80e200a12670ed0e'
[21:03:09 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:66e7b421fa04: Merge commit 'dd299a2d6d4d1af9528ed35a8131c35946be5973'
[21:03:10 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:c4fd1e7b0116: Merge commit '3b50dbc51fb0978d09c1a5b83d4bf5a59d170e1e'
[21:03:11 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:8e7497e774af: Merge commit 'fcbdd605b5409103c3f4bfa063ea270f2229b125'
[21:03:12 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:a4ee9551138c: Merge commit '8ddfa5ae5ef64a25dd087d74954ebdb9081f0d67'
[21:03:13 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:fc2a94219df7: Merge commit 'd316f9cefcd854071985c6f524a9a15348240264'
[21:03:14 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:b725b482c6af: Merge commit '800d91d348c89fc8ca3fbec7696ab1ec8787acc6'
[21:03:15 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:10eb3259a65d: Merge commit '01348e411f962f5e4605d649fc9a47a54587ba8e'
[21:03:16 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:b27dd80255a0: doc/decode_audio: use <> to include libav* headers
[21:03:17 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:74927355fc7f: Merge commit '3c9546dfafcdfe8e7860aff9ebbf609318220f29'
[21:03:18 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:c6f06876e108: Merge commit '31756abe29eb039a11c59a42cb12e0cc2aef3b97'
[21:03:19 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:b00f44e513e9: Merge commit '3297577f3eac1c87d48dedd527942de2bd28e7a5'
[21:03:20 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:1fb2c697d5d1: Merge commit '030d84fa2e35af0e77516735de35bf1a52371c86'
[21:03:21 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:4fe9d6964830: Merge commit '0940b748bdba36c4894fc8ea6be631d821fdf578'
[21:03:22 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:99fa2fc5db06: Merge commit 'fea4dc05b41f5465bedc786b67966f204ec6150c'
[21:03:23 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:678ab33861d1: Merge commit 'b6582b29277e00e5d49f400e58beefa5a21d83b8'
[21:03:24 CEST] <cone-166> ffmpeg 03Martin Vignali 07release/3.3:6426f74272c3: fate/exr : add test for uint32 data
[21:03:25 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:b62a87591ebe: doc/libav-merge: link to the relevant ml thread in the extract_extradata line
[21:03:26 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:6171f178e70e: x86/hevc_add_res: merge last remaining changes from 3d6535983282bea542dac2e568ae50da5796be34
[21:03:27 CEST] <cone-166> ffmpeg 03Steven Liu 07release/3.3:99e5d81ef997: avutil/avstring: add av_strreplace API into avstring
[21:03:28 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:f56a5fee374c: Merge commit '182cf170a544bce069c8690c90b49381150a1f10'
[21:03:29 CEST] <cone-166> ffmpeg 03Carl Eugen Hoyos 07release/3.3:fa0a6ce34d00: configure: Fix GPL dependency for avisynth / avxsynth.
[21:03:30 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:e06b8b07d599: Merge commit 'e0b164576f7467b7b1127c18175e215dc1df011f'
[21:03:31 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:6516c62ec0e4: Merge commit 'fbd1f7639d0142c391bec85d1d840c835210843f'
[21:03:32 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:52e6fb9c597f: Merge commit 'f7407f56cbf820a147bd77d728ac9a72c587cc56'
[21:03:33 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:ab04d9bf6383: Merge commit 'ffe89e1edb0281ff65d1bda88253784e9283b717'
[21:03:34 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:c1476219cad0: Merge commit 'e17567a831dede1f24e3a1a4c305a93012d7a8ce'
[21:03:35 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:33b80adaab0f: Merge commit 'd7595de0b25e7064fd9e06dea5d0425536cef6dc'
[21:03:36 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:cdcdb7d32dd2: Merge commit 'e7ae8f7a715843a5089d18e033afb3ee19ab3057'
[21:03:37 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:edfa7ac8ec4a: Merge commit '81d7f0bbca837afda1f7e60d3ae52ab1360ab44b'
[21:03:38 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:1ef8da77a5ab: Merge commit 'd5d62ce6d643de704e7bd62a2375e6391c0ffb9a'
[21:03:39 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:e83a1c5b7ce0: Merge commit '481ff3cf018811ba3235f1c236e970f32a6300b9'
[21:03:40 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:fb1a2cbae10a: Merge commit 'bcefafa226dcda23d4d9af9601d19389cb918a5b'
[21:03:41 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:2738b4cee01c: Merge commit 'aaae59700f7fc10fd80cb93b38c5d109900872d9'
[21:03:42 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:f047da4ebe41: Merge commit '3cc3463f306f425f76bd962755df1132eeac6dfa'
[21:03:43 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:2b1a6b1ae31a: Merge commit 'bf8646274b3ea2d9a193edfdbd70a3bb5c6dd74f'
[21:03:44 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:147c5ff11ce8: Merge commit 'bfe92dfe60f601b3f20a918ffcc0acdf40a5955c'
[21:03:45 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:e181e2909b21: doc/examples/transcode_aac: replace local get_error_text with av_err2str
[21:03:46 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:6200bf89a807: Merge commit 'bb265b764a055f2dc576b9aec62460d9580868f4'
[21:03:47 CEST] <cone-166> ffmpeg 03Takayuki 'January June' Suwa 07release/3.3:e3d8963c3cb5: avcodec/dsddec: correct for DSD silence bit-ordering
[21:03:48 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:a032c5224d40: Merge commit '715b8243460836fb7dd15bf7e41668e773beb276'
[21:03:49 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:c7173e50980c: Merge commit '76167140a91c081a0cf9d0abcaa4da18d1bacadb'
[21:03:50 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:e66563aedeec: Merge commit 'd860a3cc0a12360a92b9ffd179a0c34413beaf88'
[21:03:51 CEST] <cone-166> ffmpeg 03Mark Thompson 07release/3.3:75afad774b2b: vf_deinterlace_vaapi: Mark as hwframe-aware
[21:03:52 CEST] <cone-166> ffmpeg 03Dzung Hoang 07release/3.3:68e3c9a3b680: avcodec/exr: add support for scanline file where offsets are set to zero
[21:03:53 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:edc43c571d66: Merge commit '0a4b9d0ccd10b3c39105f99bd320f696f69a75a2'
[21:03:54 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:a3fc5f535adc: doc/libav-merge: document hlsenc encryption state
[21:03:55 CEST] <cone-166> ffmpeg 03Clément BSsch 07release/3.3:507a85b93c93: Merge commit 'adb0e941c329a4778ade6dd0a326274472992f54'
[21:03:56 CEST] <cone-166> ffmpeg 03Paul B Mahol 07release/3.3:f078bc4c5e66: avcodec/dnxhdenc: DNxHR 444 and HQX support
[21:03:57 CEST] <cone-166> ffmpeg 03Paul B Mahol 07release/3.3:358d4524cc8f: avcodec/dnxhdenc: fix indentation issue
[21:03:58 CEST] <cone-166> ffmpeg 03Hendrik Leppkes 07release/3.3:d91e7aac122f: Merge commit '0d3176e32f351d18d6174d8b05796829a75a4c6b'
[21:03:59 CEST] <cone-166> ffmpeg 03Hendrik Leppkes 07release/3.3:fbfa72916c89: Merge commit '9d7026574bbbe67d004a1c32911da75375692967'
[21:04:00 CEST] <cone-166> ffmpeg 03Hendrik Leppkes 07release/3.3:0f9ce9c5fcdb: Merge commit '5a1d605ceae448b476a525f7368ec452000d1f26'
[21:04:01 CEST] <cone-166> ffmpeg 03Hendrik Leppkes 07release/3.3:5c612c5ff8c4: Merge commit 'e18ba2dfd2d19aedc8afccf011d5fd0833352423'
[21:04:02 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:5f23d8b405fa: fate: add bitexact sws_flags to hevc-extradata-reload
[21:04:03 CEST] <cone-166> ffmpeg 03Michael Niedermayer 07release/3.3:679a315424e6: avformat/oggparsedaala: Check duration for AV_NOPTS_VALUE
[21:04:04 CEST] <cone-166> ffmpeg 03Michael Niedermayer 07release/3.3:23ae3cc82291: avformat/oggparsedaala: Do not leave an invalid value in gpshift
[21:04:05 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:cbd250293998: avcodec/extract_extradata_bsf: make sure all parameter set NAL units were found for h264/hevc
[21:04:06 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:7942907878dd: compat/atomics: fix atomic_fetch_xor
[21:04:07 CEST] <cone-166> ffmpeg 03Michael Niedermayer 07release/3.3:e1cc7f83df00: Bump minor for 3.3
[21:04:11 CEST] <jamrial> we made making poor cone work
[21:11:07 CEST] <cone-166> ffmpeg 03Diego Biurrun 07master:f8e29a371622: build: Add missing object dependency for extract_extradata bitstream filter
[21:13:34 CEST] <cone-166> ffmpeg 03Diego Biurrun 07release/3.3:e0064df4ffac: build: Add missing object dependency for extract_extradata bitstream filter
[21:28:02 CEST] <J_Darnley> <jamrial> wonder if the branch merge will get spammed here by the bot  -->  I wonder :)
[23:15:40 CEST] <jamrial> 3.3 changelog seems oddly short for five months worth of development
[23:17:09 CEST] <durandal_1707> blame yourself for being lazy :)
[23:17:27 CEST] <jamrial> no, looks like some things didn't get an entry when they should have :p
[23:18:08 CEST] <durandal_1707> like ?
[23:18:48 CEST] <ubitux> we don't want the threading fixes in 3.3 btw?
[23:19:09 CEST] <ubitux> BBB: what's the state of those? isn't the 3-days timeout out?
[23:19:28 CEST] <jamrial> durandal_1707: qsv vp8 decoder
[23:19:48 CEST] <jamrial> ubitux: we do. they can be backported
[23:20:27 CEST] <BBB> I can commit a couple, some havent been reviewed yet
[23:20:35 CEST] <BBB> some reviews are making me somewhat irritated
[23:21:09 CEST] <jamrial> :(
[23:21:53 CEST] <BBB> Ill look at the status Monday
[23:22:52 CEST] <BBB> dont be sad, Im honestly trying to improve tsan reporting
[23:23:16 CEST] <BBB> and I feel most of the bugs have been ironed out and its more or less a matter on agreeing on the correct hack around it, or something like that
[23:23:24 CEST] <BBB> at this point its more politics than anything else
[23:43:10 CEST] <cone-166> ffmpeg 03James Almer 07master:50e594956651: Changelog: add 3.3
[23:43:11 CEST] <cone-166> ffmpeg 03James Almer 07master:7ab9d3f34101: Changelog: add missing entry for VP8 QSV decoder
[23:43:26 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:b1377b2d281c: Changelog: add 3.3
[23:43:27 CEST] <cone-166> ffmpeg 03James Almer 07release/3.3:2ff93effb375: Changelog: add missing entry for VP8 QSV decoder
[23:44:05 CEST] <jamrial> 400 merged commits and barely any changelog padding :p
[23:50:26 CEST] <jkqxz> VAAPI MPEG-2/VP8 changelog was merged in the list for 3.1.
[23:56:54 CEST] <jkqxz> <http://sprunge.us/OMBa>?
[00:00:00 CEST] --- Mon Apr  3 2017