[Ffmpeg-devel-irc] ffmpeg-devel.log.20150320

burek burek021 at gmail.com
Sat Mar 21 02:05:03 CET 2015


[00:04:48 CET] <cone-068> ffmpeg 03Michael Niedermayer 07master:2c660e34cf3c: avcodec/dnxhddec: Check that the frame is interlaced before using cur_field
[00:04:49 CET] <cone-068> ffmpeg 03Michael Niedermayer 07master:d3bd943108d3: avcodec/dnxhddec: Reset is_444 if format is not 444
[00:04:50 CET] <cone-068> ffmpeg 03Michael Niedermayer 07master:77bd35d76db6: avcodec/dnxhddec: reset cur_field for non interlaced frames
[00:24:32 CET] <cone-068> ffmpeg 03Hendrik Leppkes 07master:11fe56c8bbf3: avcodec: add profile define for DTS Express
[00:52:50 CET] <cone-068> ffmpeg 03Hendrik Leppkes 07master:519868de7d9c: avcodec: add libdcadec decoder
[00:53:48 CET] <wm4> wow, did the patch just get merged
[01:15:23 CET] <nevcairiel> apparently
[01:16:38 CET] <kasper93> shouldn't it be "libdcadec/dca_context.h" in configure?
[01:17:48 CET] <nevcairiel> hm i guess it should, i wrote it when his project didnt have a install target at all
[01:46:09 CET] <cone-068> ffmpeg 03Anton Khirnov 07master:cf18e777aed9: movenc: write the display transform matrix
[01:46:10 CET] <cone-068> ffmpeg 03Michael Niedermayer 07master:750cf4e5374c: Merge commit 'cf18e777aed9f73011c7ab7cbab8003189d4b9d8'
[02:08:14 CET] <cone-068> ffmpeg 03Andreas Cadhalpun 07master:30e6abd1a8cc: webp: ensure that each transform is only used once
[02:08:15 CET] <cone-068> ffmpeg 03Michael Niedermayer 07master:85cc76050c5e: Merge commit '30e6abd1a8cc4fd5daf2e23ad2e768862c39e975'
[02:38:27 CET] <cone-068> ffmpeg 03Kacper MichajBow 07master:4f3c31df3f56: avcodec: Fix libdcadec include dir
[04:13:49 CET] <cone-068> ffmpeg 03James Almer 07master:03e9b1931b92: avcodec/libdcadec: add missing version bump and changelog entry
[04:40:35 CET] <cone-068> ffmpeg 03Martin Vignali 07master:57e8047972b4: avfilter/vf_pp7: remove unused var
[04:40:36 CET] <cone-068> ffmpeg 03Martin Vignali 07master:5f9d30ae8f6a: avcodec/dv: remove unused var dv100qstep
[04:40:37 CET] <cone-068> ffmpeg 03Martin Vignali 07master:7fb186784aad: avcodec/dv: remove unused var dvquantareas
[06:52:01 CET] <cone-068> ffmpeg 03James Almer 07master:b0fea4ad7e4c: x86/xvididct: remove obsolete function prototypes
[06:52:02 CET] <cone-068> ffmpeg 03James Almer 07master:6b940b8c998c: x86/xvididct: add some yasm guards
[10:50:44 CET] <cone-853> ffmpeg 03Stefano Sabatini 07master:e6547cce7278: lavfi/eq: factorize code in process_command()
[10:50:45 CET] <cone-853> ffmpeg 03arwa arif 07master:9015ca359f27: lavfi/eq: rework expression evaluation
[11:02:16 CET] <saste> anybody care about the segment patches?
[11:53:08 CET] <akshay_> Hey, I am Akshay
[11:53:21 CET] <akshay_> I wanted to tal to NIcolas George
[11:53:37 CET] <akshay_> Can I have his IRC nick ?
[12:12:15 CET] <wm4> cigaes, but he's not often here
[12:29:31 CET] <Daemon404> s/not often/never/
[12:48:54 CET] <ubitux> > Fix wrongly placed comment
[12:48:56 CET] <ubitux> > 237 files changed
[12:48:59 CET] <ubitux> this is disturbing
[12:49:01 CET] <ubitux> :D
[12:50:52 CET] <azk> One hell of a comment.
[14:26:19 CET] <Daemon404> dir listing api... what the shit?
[14:27:03 CET] <wm4> is the gsoc about implementing the samba protocol? that's what it sounds like in the mail...
[14:27:45 CET] <Daemon404> .......................................
[14:27:50 CET] <Daemon404> ............................................................
[14:28:01 CET] <Daemon404> what the *fuck(*
[14:28:11 CET] <wm4> it probably isn't
[14:28:24 CET] <wm4> but I can't find the project on the wiki
[14:28:40 CET] <Daemon404> yeah me neither
[14:29:11 CET] <wm4> "In a couple 
[14:29:11 CET] <wm4> of minutes I'll send these patches rebased against current master plus 
[14:29:11 CET] <wm4> my take on samba protocol."
[14:29:30 CET] <wm4> ooh, that probably just means the libsmbclient wrapper
[14:29:37 CET] <wm4> sorry for giving you a heart attack
[14:29:42 CET] <Daemon404> o ok
[14:29:53 CET] <Daemon404> still... lavf seems like the last place this should be
[14:30:09 CET] <TimNich> avio_wb32 writes Uint32, but how do I write int32?
[14:30:11 CET] <wm4> this one? https://trac.ffmpeg.org/wiki/SponsoringPrograms/GSoC/2015#Browsingcontentontheserver
[14:30:46 CET] <wm4> "Implementation for FTP, HTTP and SFTP is required."
[14:30:53 CET] <wm4> wut, it can't be implemented for http
[14:31:10 CET] <BtbN> It can, HTTP has dir listing stuff
[14:31:40 CET] <BtbN> It propably means WebDAV
[14:32:05 CET] <wm4> and also with the patches the guy sent (revived from older patches), isn't the project almost fully implemented?
[14:32:18 CET] <Daemon404> sounds sketchy in terms of funding yes
[14:36:24 CET] <av500> lavf needs a browser?
[14:36:30 CET] <av500> like Norton Commander?
[14:37:07 CET] <wm4> webkit
[14:37:42 CET] <wm4> BtbN: is webdav still used in the public?
[14:37:53 CET] <ubitux> TimNich: uint32 and int32 are the same, only the interpretation changes
[14:39:10 CET] <BtbN> wm4, yeah, it's quite popular i think.
[14:39:17 CET] <TimNich> ubitux:  but I need to write a value that I know will be interpreted as int32, currently its being misinterpreted
[14:39:34 CET] <ubitux> that's because it's read as uint32
[14:39:35 CET] <BtbN> A lot of NAS boxes also support it
[14:46:52 CET] <TimNich> Sure tis not the other way round,  other files are interpreted OK only ffmpeg ones come out wrong
[14:54:35 CET] <Daemon404> bits are bits
[14:54:39 CET] <Daemon404> theyre written the same way
[15:02:31 CET] <TimNich> the bits in a number represented as a signed int are layed out differently to the unsigned layout.
[15:04:49 CET] <Daemon404> you pass it 32 bits, it wrets 32 bits
[15:04:53 CET] <Daemon404> no transformation takes place
[15:05:32 CET] <Daemon404> example: http://pastie.org/10042038
[15:05:34 CET] <Daemon404> output:
[15:05:39 CET] <Daemon404> 0xFFFFFFE0.
[15:06:00 CET] Last message repeated 1 time(s).
[15:38:36 CET] <TimNich> Daemon404:  yes, but avio_wb32 shuffles bits assuming an unisgned int so if you give it a negative number&.
[15:42:13 CET] <Daemon404> it doesnt shuffle bits, it shuffles bytes
[15:42:23 CET] <Daemon404> which is entirely expected for wb
[15:42:27 CET] <Daemon404> (rather than wl)
[15:54:57 CET] <cone-853> ffmpeg 03Michael Niedermayer 07master:e8446a685077: configure: Silence warnings about constant unsigned overflows in MSVC
[17:35:50 CET] <nevcairiel> i wonder if the dcadec guy knew to what he would open the floodgates
[17:36:08 CET] <Daemon404> what happened?
[17:36:46 CET] <nevcairiel> ever since I posted the patch to the ML, he has been getting a whole bunch of issue reports, pull requests and general comments
[17:37:10 CET] <wm4> wouldn't he be happy about his code being used?
[17:37:17 CET] <Daemon404> not if he wants to be left alone
[17:37:23 CET] <Daemon404> or rather
[17:37:28 CET] <Daemon404> people on github are douchenozzles
[17:37:29 CET] <Daemon404> usually
[17:37:45 CET] <nevcairiel> I guess he might be happy, but the sudden spike of activity might be unfamiliar
[17:37:51 CET] <jamrial> technically, you opened the floodgates :p
[17:37:57 CET] <wm4> he posted his project in an mpv issue where I angrily closed a request for dts dll binary loader as WONTFIX
[17:38:13 CET] <wm4> (and then also angrily pointed out that his project was in D lol)
[17:38:29 CET] <nevcairiel> now its pretty-looking C
[17:38:32 CET] <nevcairiel> so all is well
[17:39:12 CET] <kasper93> people on github are douchenozzles << You Talkin' me?
[17:39:52 CET] <rcombs> maybe he should relicense it to LGPL and send it to ffmpeg upstream and then people would stop bugging him about it
[17:40:01 CET] <nevcairiel> it is LGPL
[17:40:10 CET] <rcombs> oh? I thought it was GPL?
[17:40:17 CET] <nevcairiel> he relicensed two days ago or so
[17:40:18 CET] <wm4> and there is a wrapper in ffmpeg
[17:40:22 CET] <wm4> (since yesterday)
[17:40:32 CET] <wm4> I wouldn't demand from him cleaning up ffmpeg's shitcode though
[17:40:33 CET] <kasper93> and well if he wants to be left alone he can just ignore us...
[17:40:40 CET] <wm4> some coding slave like mini can do this
[17:40:56 CET] <wm4> time to send my passive-aggressive hack removal patch again
[17:41:15 CET] <rcombs> ah, well then
[17:41:49 CET] <jamrial> github requests are the only way to share code with this project at the moment, what with no ml and such
[17:42:00 CET] <Daemon404> kasper93, in general i just find the quality of issues, patches, and people in general to be significantly lower on github
[17:42:05 CET] <Daemon404> compared to traditional means
[17:42:08 CET] <jamrial> and so far all have been for the build system
[17:42:15 CET] <rcombs> Daemon404: lower barrier to entry?
[17:42:17 CET] <nevcairiel> its a bit unfortunate that github doesnt have a simple discussion board or something
[17:42:24 CET] <nevcairiel> so you cant even ask questions
[17:42:28 CET] <wm4> (this also demonstrates why we should prefer clean code over performance hacks: potential contributors just won't bother with ffmpeg and e.g. write standalone libs from scratch, like it happened with dca)
[17:42:30 CET] <nevcairiel> ..so people abuse issues for that
[17:42:37 CET] <Daemon404> rcombs, also consider the kind of quality you get from people who dont even read your site to know you dont accept github PRs
[17:42:57 CET] <wm4> github should just allow to disable PRs
[17:43:05 CET] <Daemon404> linus asked for that years ago
[17:43:06 CET] <Daemon404> never happened
[17:43:10 CET] <wm4> yeah
[17:43:18 CET] <Daemon404> i think it' actually because github's RoR app is crap
[17:43:23 CET] <wm4> plan B: write a script that auto-closes PRs with a message where to send patches
[17:43:25 CET] <rcombs> they're kinda an integral part of github
[17:43:28 CET] <Daemon404> we use github enterprise, and i submitted a few requests
[17:43:41 CET] <Daemon404> and the reasons was always "too hard to do with our current app"
[17:43:51 CET] <Daemon404> were(
[17:43:52 CET] <Daemon404> *
[17:44:24 CET] <wm4> were the requests reasonable?
[17:44:28 CET] <Daemon404> yes
[17:44:37 CET] <Daemon404> stuff like "can i comment on lines of a gist"
[17:44:54 CET] <Daemon404> apparently gist is an entirely separate codebase and its impossible to do this.
[17:45:12 CET] <wm4> lolwut
[17:45:21 CET] <Daemon404> we use gist a lot to snowball ideas
[17:45:23 CET] <Daemon404> and get feedback
[17:45:25 CET] <Daemon404> their response
[17:45:27 CET] <Daemon404> literally was
[17:45:32 CET] <Daemon404> "youre not supposed to use it like that:
[17:45:37 CET] <Daemon404> "use a git repo and a branch for ideas"
[17:45:41 CET] <Daemon404> (which is fucking ugly imo)
[17:45:43 CET] <nevcairiel> that sounds like a apple answer
[17:45:44 CET] <wm4> I thought gist was trivial to implement for github because everything is a git repo or something
[17:45:48 CET] <Daemon404> nevcairiel, :D
[17:45:49 CET] <nevcairiel> "you are holding it wrong"
[17:45:52 CET] <Daemon404> yep.
[17:46:29 CET] <rcombs> at least it makes a nice-looking gitweb
[17:46:42 CET] <nevcairiel> i dont even like their web thing that much
[17:46:45 CET] <wm4> the repo browser is the worst part about github
[17:46:46 CET] <nevcairiel> the diff viewer is horrible
[17:46:53 CET] <nevcairiel> its so slow
[17:47:01 CET] <Daemon404> everything is slow
[17:47:07 CET] <Daemon404> it cant even list the lavc directory
[17:47:09 CET] <wm4> your browser is too slow!
[17:47:09 CET] <Daemon404> "too many files"
[17:47:16 CET] <Daemon404> itl iterally leaves out fules
[17:47:18 CET] <Daemon404> files*
[17:47:50 CET] <kasper93> I like like blam just stop loading a file in the middle
[17:47:54 CET] <kasper93> blame*
[17:47:57 CET] <nevcairiel> in an individual commit, you cannot look at diffs for individual files changed in that commit, it has to load the entire commit and then use anchors to scroll down on a mile long page .. which makes it unusable in a lot of situations
[17:48:19 CET] <Daemon404> this is why i still prefer mailing list + patches
[17:48:22 CET] <Daemon404> even if i feel liek a caveman
[17:48:29 CET] <Daemon404> and webjerks cant figure out how to do it
[17:48:32 CET] <nevcairiel> i use ordinary gitweb on my own server
[17:48:32 CET] <Daemon404> (maybe thats a bonus)
[17:48:35 CET] <nevcairiel> it does this just fine!
[17:48:44 CET] <Daemon404> nevcairiel, you dont get many 3rd party patches really
[17:48:44 CET] <Daemon404> afaict
[17:48:48 CET] <nevcairiel> nah
[17:48:53 CET] <nevcairiel> but i use it to browse ffmpeg :D
[17:48:57 CET] <Daemon404> lol
[17:49:03 CET] <Daemon404> i use videolan''s
[17:49:07 CET] <nevcairiel> since it has all blame and search function enabled
[17:49:10 CET] <nevcairiel> which videolan does not
[17:49:32 CET] <nevcairiel> for performance reasons I imagine
[17:49:57 CET] <wm4> apparently the script videolan uses is a POS
[17:50:09 CET] <wm4> most sites switched to cgit as standard
[17:50:11 CET] <wm4> even kernel.org
[17:50:11 CET] <nevcairiel> just looks like vanilla gitweb
[17:50:25 CET] <wm4> yeah
[17:50:27 CET] <nevcairiel> i dont know what should be so shit about it, it just runs for me
[17:50:34 CET] <j-b> doesn't matter
[17:50:35 CET] <nevcairiel> but maybe its performance doesnt scale up or something
[17:50:38 CET] <nevcairiel> i wouldnt know
[17:50:39 CET] <j-b> it will be removed soon
[17:50:44 CET] <j-b> and replaced by gitlab
[17:51:31 CET] <nevcairiel> gitlabs interface is so meh :(
[17:51:49 CET] <Daemon404> j-b, last time i tried to deploy gitlab it was insanely fucktarded
[17:51:56 CET] <Daemon404> if you dont use pre-built vm images it's ruby hell
[17:52:00 CET] <j-b> Daemon404: sure.
[17:52:05 CET] <jamrial> i like the current gitweb interface on videolan. i use it all the time to browse the tree
[17:52:21 CET] <j-b> Daemon404: now, it's 'dpkg -i bigfuckinggitlab.deb'
[17:52:26 CET] <jamrial> it's responsive, unlike github
[17:52:30 CET] <Daemon404> j-b, lol
[17:52:39 CET] <Daemon404> there's also gogs, which is terrible too.
[17:52:47 CET] <ubitux> there is Gogs now too
[17:52:47 CET] <nevcairiel> jamrial: more importantly, its simple, and not made all modern-looking like github (and gitlab as well)
[17:53:17 CET] <ubitux> meh, can't reply fast
[17:53:20 CET] Action: ubitux hides
[17:53:26 CET] <Daemon404> j-b, will you finally get rid of videolan svn then? :)
[17:54:35 CET] <jamrial> ubitux: you're leaving this weekend, right?
[17:54:42 CET] <ubitux> yes
[17:54:48 CET] <ubitux> i'll disconnect the laptop in a few hours
[17:55:32 CET] <ubitux> i ordered the backbone part of the x250 with the appropriate trackpad to replace the one on x240, but it didn't arrive yet :(
[17:55:40 CET] <ubitux> so i'll leave with that shitty clickpad
[17:55:40 CET] <j-b> Daemon404: probably
[17:55:42 CET] Action: ubitux sad
[17:56:15 CET] <wm4> jamrial: I just use my filesystem and gitk
[17:57:28 CET] <jamrial> ubitux: ok
[17:58:24 CET] <jamrial> tried gitk and didn't like it much. super fast to see commits from 5 years go unlike any web interface, though
[18:00:19 CET] <ubitux> alias -> lgg = log --graph --oneline --decorate
[18:00:35 CET] <ubitux> but yeah, gitk is nice
[18:00:41 CET] <wm4> gitk is kind of ugly, but its blame function is great
[18:00:57 CET] <wm4> it also shows dirty files and the staging area
[18:01:06 CET] <wm4> too bad it's written in TCL
[18:04:45 CET] <Daemon404> tk is ugly as balls...
[18:04:59 CET] <Daemon404> anyway, i find the webuis much faster to browse blames and history
[18:06:38 CET] <ubitux> you can change the colors in the UI
[18:06:40 CET] <ubitux> ;)
[18:07:28 CET] <ubitux> http://ubitux.fr/pub/pics/_gitk-recolored.png look how pretty it can be *_*
[18:07:58 CET] <wm4> Daemon404: which webui can blame in a useful way?
[18:08:18 CET] <wm4> ubitux: wins the fugly ui price
[18:08:21 CET] <wm4> *prize
[18:08:53 CET] <ubitux> :D
[18:09:04 CET] <Daemon404> wm4, i use github when possible
[18:09:28 CET] <Daemon404> because it wi generally do blame -> click rev -> find out diego moved it -> click parent -> click blame
[18:09:31 CET] <Daemon404> repeat
[18:09:34 CET] <Daemon404> for as many diegos as needed
[18:13:15 CET] <wm4> gitk lands you directly in the git, usually close to the removed line
[18:13:29 CET] <jamrial> one good thing about github is the whole network graph for a project's forks
[18:13:46 CET] <jamrial> it's slow and unresponsive, but nice to find what other people are working on
[18:13:48 CET] <Daemon404> wm4, i also dotn always have a cloned repo + term near me
[18:13:51 CET] <Daemon404> line on my phone
[18:14:25 CET] <Daemon404> jamrial, i find it very hard to make heads or tails out of
[18:14:31 CET] <Daemon404> especially with >10 forks
[18:15:21 CET] <Daemon404> https://github.com/FFmpeg/FFmpeg/pull/123
[18:15:22 CET] <jamrial> true. with a project like ffmpeg it's a nightmare to browse
[18:15:23 CET] <Daemon404> wtf is this?
[18:15:30 CET] <Daemon404> remember when i said github is low quality
[18:15:32 CET] <Daemon404> <_<
[18:15:36 CET] <Daemon404> jamrial, yea
[18:15:54 CET] <wm4> Daemon404: amazing, worse than what I expected
[18:15:57 CET] <wm4> (that PR)
[18:16:15 CET] <Daemon404> :D
[18:16:43 CET] <jamrial> haha, wtf. a "feature request" in the form of a PR that pulls a bunch of unrelated commits
[18:17:46 CET] <Daemon404> hmmm
[18:17:51 CET] <Daemon404> it should be simple to make a bot like golang has
[18:17:55 CET] <Daemon404> that autocloses PRs
[18:17:59 CET] <Daemon404> and tells them how to submit
[18:18:58 CET] <jamrial> or point them to trac for things like this PR
[18:19:23 CET] <Daemon404> i somehow think this level of stupid might be confused by a bug tracke.r
[18:19:39 CET] <Daemon404> i really dont get though, why you can disable issues
[18:19:41 CET] <Daemon404> but not PRs
[18:33:12 CET] <wm4> jamrial: also something non-standard (didn't he ask for flac in mp4?)
[18:33:48 CET] <Daemon404> thats so silly
[18:33:57 CET] <Daemon404> why not use alac then?
[18:36:19 CET] <saste> ubitux, you here?
[19:21:04 CET] <cone-853> ffmpeg 03Michael Niedermayer 07master:59c06902e838: avfilter/vf_phase: Change enum to int, which is accessed via AVOption as int
[19:21:05 CET] <cone-853> ffmpeg 03Michael Niedermayer 07master:1a79850a8fa9: avfilter/vf_overlay: Change enums to int, which are accessed via AVOption as int
[19:21:06 CET] <cone-853> ffmpeg 03Michael Niedermayer 07master:d6dba15b97dc: avfilter/vf_mcdeint: Change enums to int, which are accessed via AVOption as int
[19:39:32 CET] <cone-853> ffmpeg 03Michael Niedermayer 07master:2a8198b32f27: avcodec/libvpxenc: Set min/max quantizer to 0 for lossless mode
[19:47:02 CET] <rcombs> who wants to look at porn for an img2dec bug
[19:48:44 CET] <cone-853> ffmpeg 03James Almer 07master:d13e521cc861: libavutil: add av_mod_uintp2
[19:48:45 CET] <cone-853> ffmpeg 03James Almer 07master:60b9373dbd52: libavutil: add bmi2 optimized av_mod_uintp2
[19:51:08 CET] <wm4> rcombs: how can img2dec be buggy?
[19:51:16 CET] <wm4> I mean it doesn't do much
[19:51:59 CET] <rcombs> https://gist.github.com/1c0811ac0e1a55084ef9
[19:52:09 CET] <rcombs> something in the probing
[19:52:23 CET] <rcombs> (the image link is NSFW)
[19:53:14 CET] <rcombs> works fine if not piped
[19:53:35 CET] <wm4> then it's probably a jpeg_pipe file detection issue
[19:53:46 CET] <wm4> not piped will use the file extension
[19:58:22 CET] <rcombs> using ffmpeg for this is pretty silly of me anyway; I'd be better off with, say, imagemagick
[19:58:29 CET] <rcombs> since I'm just resizing images
[20:00:32 CET] <wm4> ffmpeg also can't handle high resolution images
[20:03:30 CET] <rcombs> heh
[20:06:50 CET] <jamrial> it can! although last time i tried to resize a huge image it ate like 2gb of ram
[20:07:04 CET] <wm4> jamrial: no, it stops somewhere around 16000x16000
[20:07:17 CET] <wm4> because it could overflow int with some pixel formats
[20:27:06 CET] <cone-853> ffmpeg 03Luca Barbato 07master:342fc46c6919: cmdutils: Add a stream specifier to map usable streams
[20:27:07 CET] <cone-853> ffmpeg 03Michael Niedermayer 07master:50255b7a865f: Merge commit '342fc46c69199b076b7c210f42208ce39a2bde1c'
[20:43:05 CET] <rcombs> welp switched to imagemagick for this
[20:43:11 CET] <rcombs> if anyone wants to play with that issue, feel free
[20:43:41 CET] <cone-853> ffmpeg 03Luca Barbato 07master:22a6d48b0d73: avconv: Print the avfilter errors
[20:43:42 CET] <cone-853> ffmpeg 03Michael Niedermayer 07master:215c2fe0ae2a: Merge commit '22a6d48b0d7367b3a45dd6db245f8ffaaa364995'
[21:26:08 CET] <wm4> is there any reason why h264dec doesn't allow byte seeks?
[21:31:47 CET] <nevcairiel> you mean the raw demuxer?
[21:32:23 CET] <wm4> yes
[21:33:14 CET] <wm4> if I ignore the flag and issue a byte seek anyway, it works pretty well with h264
[21:33:35 CET] <wm4> though hevc shows broken frames (instead of discarding them), and spams "PPS id out of range: 0"
[21:33:37 CET] <nevcairiel> i guess there is no reason it shouldnt work
[21:33:46 CET] <nevcairiel> yeah hevc doesnt have seek recovery properly yet
[21:33:57 CET] <nevcairiel> i mean, it recovers fine, but it outputs frames
[22:10:25 CET] <wm4> actually my idea that h264dec doesn't allow byte seeks was a completely brainfart
[22:48:55 CET] <cone-853> ffmpeg 03Martin Vignali 07master:276a8666d2d9: avcodec/metasound_data: remove unused variable lsp16s
[23:18:27 CET] <kierank> Damn FFmpeg misdetecting opus as aac again
[23:18:32 CET] <kierank> Its a never ending struggle
[23:37:23 CET] <ubitux> add a test
[23:37:53 CET] <ubitux> tests/fate/probe.mak
[00:00:00 CET] --- Sat Mar 21 2015


More information about the Ffmpeg-devel-irc mailing list