[Ffmpeg-devel-irc] ffmpeg-devel.log.20150717

[burek]

[00:29:53 CEST] <ubitux> rcombs: iirc it was made sure not to be dropped to not loose information because of a syntax error
[00:30:22 CEST] <ubitux> rcombs: like typically a > not closed and causing the whole sentence and the next one to disappear
[00:30:46 CEST] <ubitux> i think it's better to print "<font Hi I'm a subtitle" than nothing
[00:31:05 CEST] <ubitux> maybe in practice it's better to drop, i don't know
[00:31:09 CEST] <ubitux> your call
[00:31:26 CEST] <ubitux> might be interesting to get a bunch of samples with the most common errors and check what to do
[00:31:44 CEST] <rcombs> ubitux: I can see either way there, but webvttenc definitely needs escaping code then
[00:31:44 CEST] <ubitux> i'd love to have a huge db of subtitles, like a dump of all the shit on opensubtitles or whatever
[00:32:15 CEST] <ubitux> philipl: 4k gif is actually pretty smooth for the cpu ;)
[00:32:19 CEST] <ubitux> (decoding at least)
[00:32:20 CEST] <rcombs> e.g. <fonz> instead of <fonz>
[00:32:33 CEST] <ubitux> rcombs: ah, yeah, right
[00:32:46 CEST] <ubitux> anyway, i need to go, it's very late
[00:32:55 CEST] <rcombs> OK
[00:33:00 CEST] <rcombs> think that should go in lavu/bprintf?
[00:33:18 CEST] <rcombs> erm, lavu/bprint
[00:53:26 CEST] <cone-649> ffmpeg 03Marton Balint 07master:53f2ef2c4afb: mxfdec: calculate the index in display order
[00:53:27 CEST] <cone-649> ffmpeg 03Marton Balint 07master:7ff0137a1f3b: concatdec: add support for specifying inpoint of files
[00:53:28 CEST] <cone-649> ffmpeg 03Marton Balint 07master:12d82004c51f: concatdec: store eof condition in context
[00:53:29 CEST] <cone-649> ffmpeg 03Marton Balint 07master:25a6711c2548: concatdec: add support for specifying outpoint of files
[00:53:30 CEST] <cone-649> ffmpeg 03Marton Balint 07master:5117b5e9aaa0: concatdec: add support for injecting packet metadata
[00:53:31 CEST] <cone-649> ffmpeg 03Marton Balint 07master:c8ec2109f9b3: avformat: bump micro version after adding concatdec features
[01:00:46 CEST] <philipl> ubitux: hehe.
[01:23:30 CEST] <J_Darnley> jamrial durandal_1707: I fixed the avx2 functions but my test was too short and gave inconsistant benchmark results.  I will bench them again tomorrow.
[01:24:45 CEST] <J_Darnley> I've also correctly limited the number of pixels each function will process.
[01:31:28 CEST] <J_Darnley> https://gitlab.com/J_Darnley/ffmpeg/commit/2c769642d34beb9c60c3593329d79d36b0011a95
[01:32:06 CEST] <J_Darnley> It is messy but I thought it would be cleaner than setting it for each function.
[01:42:25 CEST] <jamrial> J_Darnley: you can pmovzxbw with avx2 using an unaligned memory operand
[01:42:29 CEST] <jamrial> no need for the movu
[01:42:49 CEST] <J_Darnley> oh
[01:43:13 CEST] <jamrial> it will load 16 bytes as well: http://www.felixcloutier.com/x86/PMOVZX.html
[01:44:00 CEST] <J_Darnley> Wow an actual page I can search in!
[01:44:19 CEST] <J_Darnley> goodbye intel pdf manuals
[01:44:50 CEST] <jamrial> haha
[02:11:15 CEST] <cone-649> ffmpeg 03Carl Eugen Hoyos 07master:7f0d5401885b: lavc: Allow 64 channels internally.
[02:11:16 CEST] <cone-649> ffmpeg 03Carl Eugen Hoyos 07master:a77401e1f712: lswr: Allow 64 channels internally.
[02:11:17 CEST] <cone-649> ffmpeg 03Carl Eugen Hoyos 07master:ed5d62e509bf: lavfi/af_pan: Support a maximum of 64 channels.
[02:11:18 CEST] <cone-649> ffmpeg 03Michael Niedermayer 07master:8250943febe9: Merge remote-tracking branch 'cehoyos/master'
[02:27:58 CEST] <J_Darnley> Have you tried the classic HTML only interface?
[02:28:18 CEST] <J_Darnley> oops, that's not right.
[02:29:07 CEST] <rcombs> what level of lavu version bump is necessary after adding a function, and what's necessary after adding an AV_ESCAPE_FLAG or AV_ESCAPE_MODE?
[02:49:23 CEST] <jamrial> rcombs: minor, probably
[03:36:14 CEST] <jamrial> oh nice, gcc 5.2 was released today
[11:36:10 CEST] <cone-545> ffmpeg 03Paul B Mahol 07master:3b365dda5cd5: avfilter/af_astats: measure minimal and mean difference between two consecutive samples
[13:31:26 CEST] <cone-545> ffmpeg 03LoRd_MuldeR 07master:21436b95dc96: avfilter: add Dynamic Audio Normalizer filter
[13:31:27 CEST] <cone-545> ffmpeg 03Paul B Mahol 07master:297df52f3035: avfilter/af_compand: fix clipping with default options
[13:42:55 CEST] <cone-545> ffmpeg 03Paul B Mahol 07master:ec6bef68f14f: avfilter: bump minor and add Changelog entry
[14:39:44 CEST] <cone-545> ffmpeg 03Kieran Kunhya 07master:b0017579b6e8: h264: Add support for Closed Caption export
[14:39:45 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:441d5a7084b5: Merge commit 'b0017579b6e8de3d0d13dcfe0a4dd848dc6baa37'
[14:47:48 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:f1840b070dd4: Revert "mov: Double-check that alias path is not an absolute path"
[14:47:49 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:694f9ab1ef14: Merge commit 'f1840b070dd449ad25e1ebbab463f10863a5e0f1'
[14:58:43 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:a53540840d26: APIchanges: Mention lavfi and lavd identification symbol addition
[14:58:44 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:187f4278e81c: Merge commit 'a53540840d26beb57a5e53b7f488c23ddf86e193'
[15:19:42 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:0cf5588d6992: hq_hqa: Fix decoding when INFO section is absent
[15:19:44 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:1cdfae0b3e3e: Merge commit '0cf5588d69922aa3e063bac6304c09c543a9ef52'
[15:50:12 CEST] <D404|Ghetto> apparently msys2's mirror also mirrored the 404 package db...
[15:50:18 CEST] <D404|Ghetto> thus making it useless 
[18:29:56 CEST] <Compn> D404|Ghetto : i'm half surprised there isnt just a big msys2 zip on mega.co.nz ...
[18:30:16 CEST] <Compn> does msys have torrent mirror ?
[18:30:57 CEST] <BtbN> The Cygwin installer is also a quite easy way to get a working msys/mingw64 development environment on windows.
[18:43:47 CEST] <durandal_1707> J_Darnley: what is actual speed improvement with removegrain?
[18:56:32 CEST] <BBB> j-b: can we pay you to take over ffmpeg hosting (everything)?
[18:57:50 CEST] <j-b> Why do you need to pay me?
[18:58:08 CEST] Action: Compn has worries about being hosted by another open source project (especially one that is not mplayer)
[18:58:33 CEST] <j-b> Compn: rofl
[18:58:36 CEST] <Compn> ehe
[18:58:52 CEST] <Compn> j-b : ffmpeg needs a new host. web, mailing lists, trac, etc.
[18:58:55 CEST] <j-b> Compn: VideoLAN non-profit goal is not about VLC
[18:59:06 CEST] <BBB> I would be the biggest fanboy of j-b hosting all of ffmpeg
[18:59:12 CEST] <j-b> http://people.videolan.org/~jb/Status-VideoLAN.pdf
[18:59:19 CEST] <BBB> I would even learn how to use git correctly and push my own patches
[18:59:26 CEST] <j-b> Article 2
[18:59:43 CEST] <jamrial> we're still looking for hosting?
[19:00:50 CEST] <Compn> jamrial : yes, there are several offers so far. but non have offered long-term solution
[19:02:07 CEST] <Compn> j-b : all i can think of is "ffmtech" , so i'll be quiet :P
[19:04:20 CEST] <Compn> oh and i forgot, its a combined thing, so ffmpeg/mplayer and rtmpdump all need hosts :P
[19:04:25 CEST] <Compn> vlc hosting mplayer ;D
[19:05:05 CEST] <D404|Ghetto> recycling bin has a good offer for mplayer
[19:05:50 CEST] <durandal_1707> :)
[19:09:25 CEST] <philipl> mplayer is more of a museum piece now. I think historical preservation could be a VideoLAN goal.
[19:13:51 CEST] <D404|Ghetto> philipl: as a warning for future generations?
[19:13:55 CEST] <Compn> lol
[19:15:41 CEST] <j-b> mplayer > *
[19:18:04 CEST] <philipl> heh
[19:35:36 CEST] <BBB> mplayer | (void)
[19:36:04 CEST] <BBB> but it might indeed be a good tutorial to future generations of programmers-to-be on how not to do things
[19:36:25 CEST] <BBB> Compn: hey, ffmtech was a good idea originally
[19:36:49 CEST] <BBB> I have no idea what happened to it (the domain is gone), but it did bring in money and stuff
[19:37:16 CEST] <kierank> It paid for good cosmetics
[19:37:25 CEST] <durandal_1707> Lol
[19:37:25 CEST] <kierank> Don't discount that
[19:37:35 CEST] Action: kierank trolls
[19:38:03 CEST] Action: j-b counter-trolls kierank 
[19:39:33 CEST] <atomnuker> https://0x0.st/sXy.png
[19:39:36 CEST] <atomnuker> https://0x0.st/sXt.png
[19:39:48 CEST] <BBB> your https is broken
[19:39:48 CEST] <atomnuker> whoever guesses where the aac fixed point decoder was used wins nothing
[19:45:22 CEST] <BBB> atomnuker: is that overzelous clipping?
[19:45:39 CEST] <J_Darnley> durandal_1707: with sse2 it was about 8-20
[19:45:41 CEST] <BBB> or is the headroom in the fixed point range too small?
[19:47:39 CEST] <atomnuker> @BBB: no idea, but the fixed point decoder has a long long way to go to
[19:47:40 CEST] <D404|Ghetto> kierank: it also paid for all that subtitle work tha twas done on Libav... er, wait
[20:03:43 CEST] <jamrial> atomnuker: there's a couple aac fixed patches that haven't been committed yet, so that may be why
[20:12:06 CEST] <atomnuker> those are for AAC SBR I think, that won't change much
[20:31:45 CEST] <rcombs> BBB: it's a 8192-bit RSA key pair, which some browsers dislike
[20:32:09 CEST] <rcombs> you can probably tweak a config setting somewhere to fix that; on OS X it's in `defaults`
[20:32:32 CEST] <wm4> <BBB> but it might indeed be a good tutorial to future generations of programmers-to-be on how not to do things <- ffmpeg is still alive to fulfill this mission
[20:33:08 CEST] <BBB> ffmpeg can still be fixed
[20:33:13 CEST] <BBB> mplayer is mortally doomed
[20:34:33 CEST] <rcombs> BBB: oh, except also the config is outright broken
[20:35:05 CEST] <j-b> 8192-bit RSA key pair, for TLS? wow.
[20:35:26 CEST] Action: rcombs bugs lachs0r
[20:35:40 CEST] <rcombs> now you have 2 problems
[20:36:17 CEST] Action: BBB is very confused
[20:36:18 CEST] <J_Darnley> 0x0 - nice domain name!
[20:41:20 CEST] <rcombs> oh hurr, I forgot OpenSSL doesn't enable SNI by default
[21:14:24 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:2ea8a480832a: avfilter/af_aresample: Check ff_all_* for allocation failures
[21:14:25 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:4c7c0d37e5c2: avfilter/af_dynaudnorm: Fix "ISO C90 forbids mixed declarations and code" warnings
[21:31:04 CEST] <cone-545> ffmpeg 03Luca Barbato 07master:e1319aa1c1be: libx264: Add support for the MPEG2 encoder
[21:31:05 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:e56f14659f58: Merge commit 'e1319aa1c1be9b64117c19170344fb78841dd67c'
[21:38:59 CEST] <cone-545> ffmpeg 03Dave Yeo 07master:4da585ae478f: OS/2: Cleanup slib_create_def_cmd
[21:39:00 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:fd280de7eeb8: Merge commit '4da585ae478f8b6f865ab13779af399aee20d899'
[21:48:44 CEST] <rcombs> e1319aa1c1be looks wrong to me
[21:49:07 CEST] <rcombs> https://gist.github.com/54ec499a9c10bb43abd3 this in particular
[21:49:32 CEST] <rcombs> looks like that #else should be an #endif and the #endif shouldn't be there
[21:50:35 CEST] <BtbN> won't that change in configure also cause it to fail checking for libx264 when X264_MPEG2 isn't defined?
[21:51:07 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:b075869bc1e1: configure: Factor out flacdsp module
[21:51:08 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:948a15195c69: Merge commit 'b075869bc1e1aadea0a8dc819ebfb758adb9e3d0'
[21:51:41 CEST] <rcombs> BtbN: that looks fine to me?
[21:52:13 CEST] <BtbN> If check_cpp_condition x264.h "X264_MPEG2" fails, it will fail the entire conditional?
[21:52:33 CEST] <rcombs> no, that all runs after the main x264 conditional
[21:52:43 CEST] <rcombs> has no effect on it
[21:52:51 CEST] <BtbN> Oh, it's after the die
[21:52:53 CEST] <BtbN> nvm
[21:52:56 CEST] <rcombs> just determines whether or not the `enable libx262` gets short-circuited
[21:53:46 CEST] <BtbN> but yes, that #if/#else/#endif is definitely wrong.
[21:55:50 CEST] <rcombs> poked #libav-devel; michaelni: ^
[21:59:04 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:2edc77dc7be5: configure: Factor out ividsp module
[21:59:05 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:ccd3edf81fc2: Merge commit '2edc77dc7be5f4a006c6295f4c827e5471f85262'
[22:10:19 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:575ec60e5421: configure: Factor out mss34dsp module
[22:10:20 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:77cd5193a5a5: Merge commit '575ec60e542114b2ef5b2f88b28eeb72389dda45'
[22:15:35 CEST] <michaelni> rcombs, ill push a fix in a moment
[22:15:39 CEST] <michaelni> thx
[22:15:42 CEST] <rcombs> welcome
[22:18:21 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:189f2ed902a6: avcodec/libx264: fix MPEG2 support
[22:31:41 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:5cb4bdb2a03c: configure: Factor out rv34dsp module
[22:31:42 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:fd29dd432c56: Merge commit '5cb4bdb2a03c3643f8f1e7d21d7094e61e0a4418'
[22:44:01 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:d4aea1aa4060: configure: Factor out vp56dsp module
[22:44:01 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:3500eccb8e7f: Merge commit 'd4aea1aa4060fb5ba7f5f426823b697d2a5af08b'
[22:55:37 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:d42191c78bef: configure: Factor out vp8dsp module
[22:55:38 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:115a9b5091cd: Merge commit 'd42191c78befc1983f23b1899b2dda513b72f1ed'
[23:03:20 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:a623aa0069df: configure: Factor out wmv2dsp module
[23:03:21 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:65b0f7ce0817: Merge commit 'a623aa0069dff586241ad2ffedcd45ed840e2553'
[23:12:24 CEST] <cone-545> ffmpeg 03Vittorio Giovara 07master:f5ee23004d11: configure: Factor out g722dsp module
[23:12:25 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:c7c33ab721cf: Merge commit 'f5ee23004d1177ca6dd99b92cb4ff4b94b2eae09'
[23:36:42 CEST] <cone-545> ffmpeg 03Henrik Gramner 07master:a344e5d094eb: x86: bswapdsp: Don't treat 32-bit integers as 64-bit
[23:36:43 CEST] <cone-545> ffmpeg 03Henrik Gramner 07master:d37f23263584: checkasm: Add unit tests for bswapdsp
[23:36:44 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:52b6d9626898: Merge commit 'a344e5d094ebcf9a23acf3a27c56cbbbc829db42'
[23:36:45 CEST] <cone-545> ffmpeg 03Michael Niedermayer 07master:72d1409e2380: Merge commit 'd37f23263584774e1798e9ac909a398304a05091'
[00:00:00 CEST] --- Sat Jul 18 2015