[Ffmpeg-devel-irc] ffmpeg-devel.log.20121218

[burek]

[00:07] <cone-263> ffmpeg.git 03Stefano Sabatini 07master:992772b94757: doc: remove references to external documents
[00:07] <cone-263> ffmpeg.git 03Stefano Sabatini 07master:f14e248783f4: doc/texi2pod: add support to @ifhtml and @ifnothtml directives
[00:07] <cone-263> ffmpeg.git 03Stefano Sabatini 07master:9bba68357800: doc/Makefile: add docclean rule
[00:07] <cone-263> ffmpeg.git 03Stefano Sabatini 07master:1f9855dbc110: doc/texi2pod: correctly handle @url commands in the form @url{URL,...}
[00:07] <cone-263> ffmpeg.git 03Stefano Sabatini 07master:85e6f6d9c180: doc/indevs: update syntax for the color filter in the lavfi device examples
[00:46] <cbsrobot> michaelni: is it possible to disable skipping samples because of side data ?
[00:48] <michaelni> removing the specific side data should work
[00:50] <cbsrobot> is it stored in the container ?
[00:54] <cbsrobot> michaelni: ^ ?
[00:55] <michaelni> mp4/mov has a few ways to store initially skiped samples
[00:57] <cbsrobot> and midstream too ?
[00:58] <cbsrobot> well I guess it's a yes
[00:58] <michaelni> yes mov/mp4 edit lists allow random shuffling of parts but we dont support that (and many other players dont support it either)
[00:59] <cbsrobot> there's a user on #ffmpeg
[00:59] <cbsrobot> I will tell him to submit a bugreport
[01:02] <Compn> cbsrobot : lololol
[01:03] <cbsrobot> o/
[01:22] <cone-263> ffmpeg.git 03Michael Niedermayer 07master:28b85b1fff42: mjpegenc: fix YUV444 with w%16!=0
[01:22] <cone-263> ffmpeg.git 03Michael Niedermayer 07master:41bebda48577: mjpegenc: fix restart interval parameters for yuv444
[01:32] <ubitux> can anyone give me some hints about the "strength limitation" for gradfun to avoid the overflow?
[01:33] <ubitux> (the delta<<2 shift)
[01:33] <ubitux> i've a hard time understanding the algorithm :p
[01:42] <ubitux> so if delta is in [0x3fff-0xffff], if we want to avoid the overflow it seems we have to limit thresh to 64 maximum
[01:43] <ubitux> so the range would be between 0.51 and 64 instead of 0.51 and 255
[01:43] <ubitux> would that be appropriate?
[01:48] <michaelni> ubitux, if >64 makes sense you can also limit the optimization to the smaller range
[01:48] <michaelni> and use C for the rest
[01:49] <ubitux> ah, right ok
[01:49] <michaelni> Daemon404, i think you are the primary user of the doxygen documentation, maybe you want to officially maintain the doxygen tags in the source ?
[01:50] <michaelni> someone who uses & reads the doxy stuff would be ideal for this ...
[01:52] <Daemon404> what
[01:53] <Daemon404> 'tags' ?
[01:53] <michaelni> that stuff between /** and */ i meant
[01:53] <Daemon404> yes and that should be written and maintaiend by the person who writes the stuff they document.
[01:53] <michaelni> sure but 
[01:54] <michaelni> the generated doxy has some glitches
[01:54] <Daemon404> "fix them when i notice them" is the best i can offer
[01:55] <michaelni> ok, thanks alot
[01:55] <michaelni> first example: http://ffmpeg.org/doxygen/trunk/group__libavf.html
[01:55] <michaelni> "Passing options to (de)muxers" is there twice
[01:56] <Daemon404> need to cook dinner, will look later
[01:56] <michaelni> sure
[01:56] <michaelni> also theres a huge amount of "warning: maximum nesting level exceeded for group lavf_encoding: check for possible recursive group relation!" errors
[01:57] <michaelni> these might be a bug in doxygen, iam not sure though
[01:57] <michaelni> the maximum nesting errors come from avformat.h
[01:58] <cone-263> ffmpeg.git 03Carl Eugen Hoyos 07master:a71de9a2a2e3: Fix typo in mvc decoder.
[02:01] <cone-263> ffmpeg.git 03Stefano Sabatini 07master:1dd33d472a60: doc/ffmpeg-bitstream-filters: fix typo in title
[02:41] <Compn> carl and ami dukin it out on the tracker
[02:47] <cone-263> ffmpeg.git 03Clément BSsch 07master:8b63744f483e: Fix "bistream" typo and add a check in tools/patcheck.
[02:49] <ubitux> seems i failed at math.
[03:24] <KellenS> has anyone got the libav* librarys to compile for win8 / ARM?
[03:26] <KellenS> anyone konw what the cross compile flag would be?
[03:26] <Compn> hehe
[03:26] <KellenS> ie the equiv of this "--cross-prefix=i386-mingw32msvc"
[03:26] <Compn> thats a thing huh
[03:26] <Compn> maybe find a distro or win8 arm group ?
[03:26] <Compn> google 'how to compile win8 arm linux' and then see if anyone else figured it out ?
[03:26] <Compn> maybe someone else figured it out
[03:27] <KellenS> I'm compiling on windows
[03:27] <Compn> i dont remember anyone compiled anything for win8 yet
[03:27] <KellenS> with visual studio
[03:27] <KellenS> following these steps: http://ffmpeg.org/platform.html#Microsoft-Visual-C_002b_002b
[03:27] <KellenS> works well for x86
[03:27] <Compn> what arm cpu is it ?
[03:27] <KellenS> Tegra3
[03:28] <KellenS> not sure what version of ARM that is
[03:28] <KellenS> v7?
[03:28] <Compn> no idea
[03:28] <KellenS> yeah ARMv7 instruction set
[03:28] <KellenS> should be similar to compiling for android tablets 
[03:28] <Compn> does mingw even support arm ?
[03:28] <Compn> i dont think it does...
[03:28] <Compn> i could be wrong
[03:28] <KellenS> the compiler doesnt
[03:29] <KellenS> I'm going to take a shot at using MSYS + the new visual studio compiler
[03:29] <Compn> ok
[03:29] <Compn> i found somethin
[03:29] <KellenS> the same compiler works to create a x86 binary of it
[03:29] <KellenS> and it now supports cross compiling to ARM
[03:29] <Compn> er nevermind
[03:30] <KellenS> guess I basically need the --cross-prefix= for ARMv7 instruction set
[03:31] <Compn> vlc tried to port to windows 8 / metro
[03:31] <Compn> http://www.jbkempf.com/blog/
[03:32] <Compn> j-b probably know all about trying to compile ffmpeg for win8 ;D
[03:32] <KellenS> thats what I'm working on ;-)
[03:32] <KellenS> they're working on getting mingw cross compiling to work
[03:32] <KellenS> which is awesome
[03:32] <KellenS> but I'm wondering if in the mean time we can't just use microsoft's compiler
[03:32] <KellenS> using the c99-c89 converter it might work
[03:34] <Daemon404> its not that simple
[03:34] <Daemon404> you cant use the win32 api at all
[03:34] <KellenS> you can use some of it
[03:34] <Daemon404> your best bet is to Wait (TM)
[03:34] <KellenS> but not much
[03:34] <KellenS> I'm aware of the api restrictions though
[03:34] <KellenS> I wrote what VLC has so far
[03:34] <KellenS> I'd be waiting for myself :-p
[03:35] <Daemon404> smeone else hasb een working on win8/arm for libav&&ffmpeg
[03:35] <Daemon404> including asm support
[03:35] <KellenS> I can wait for j-b etc with the mingw .. but I thought I'd give this a shot
[03:36] <KellenS> oh really?
[03:36] <KellenS> nice
[03:36] <Compn> KellenS : why not just find the armv7 compilation instructions for ffmpeg and use those then ?
[03:36] <KellenS> Compn: yeah good idea
[03:36] <KellenS> I was looking through android compile guides
[03:37] <Compn> i dont want to insult you, but people who come in here and ask how to compile ffmpeg ... generally dont get very far into development ....
[03:37] <Daemon404> those arent going to help you for win8.
[03:37] <Daemon404> Compn, try reading what he said/is doing
[03:37] <Daemon404> before you insult
[03:37] <Daemon404> :)
[03:37] <Compn> i did, he doesnt believe you or me that it wont work
[03:37] <Compn> so i give him tools to shoot himself in foot :)
[03:37] <Compn> let him worry about foot
[03:37] <Daemon404> no you gave him somethign completely unrelated
[03:37] <Daemon404> and useless
[03:38] <Compn> also i gave him 'use vlc' :P
[03:38] <Compn> ehe
[03:38] <wm4> Compn the master troll
[03:39] <Daemon404> KellenS, ill poke the person who is working on win8/arm/winrt for libav*
[03:39] <Daemon404> but it's nighttime in europe
[03:39] <KellenS> thats cool
[03:39] <KellenS> I'll do some more reading on it
[03:39] <KellenS> this is a complete shot in the dark
[03:40] <Compn> KellenS : what programming experience do you have ?
[03:40] <Compn> i have none, thas why i ask :)
[03:40] <Daemon404> [21:34] < KellenS> I wrote what VLC has so far
[03:40] <Daemon404> perhaps you missed this.
[03:40] <Compn> i thought he was talking about blog post or so ;)
[03:40] <Daemon404> wat
[03:40] <KellenS> like you say, it's probably best to wait for the rest of the guys to get support into mingw
[03:41] <Daemon404> KellenS, mingw wont be so relevant per se
[03:41] <Compn> Daemon404 : i ask because it seems like hes developing on mingw...
[03:41] <Daemon404> youre still api-restricted
[03:41] <Daemon404> so it works out the same as using c99-c89 + msvs
[03:41] <Daemon404> with the benefit of asm, i suppose
[03:43] <KellenS> I'm not working on mingw
[03:46] <Compn> so anyways
[03:47] <Compn> what errors does msvc throw when you compile with arm7 ?
[03:47] <Daemon404> "you are not allowed to use win32 api"
[03:47] <Daemon404> compiler refuses to compile
[03:47] <Compn> msvc compiler 
[03:47] <Compn> blarg
[03:47] <Daemon404> yes
[03:47] <Daemon404> the msvc compiler.
[03:47] <Daemon404> says this.
[03:47] <KellenS> Daemon404: the compiler does not enforce api restrictions
[03:48] <Daemon404> KellenS, it spat out complains to me over api
[03:48] <Daemon404> from vs2012
[03:48] <ubitux> "BSD sockets are gone"
[03:48] <ubitux> i lol'd
[03:48] <KellenS> really?  I've never seen that
[03:48] <KellenS> I just compiled for windows 8 on x86
[03:49] <Daemon404> the x86 compiler dosnt
[03:49] <Daemon404> only the arm compiler does
[03:49] <KellenS> isnt that the job of WACK?
[03:49] <Daemon404> WACK?
[03:49] <KellenS> yeah
[03:49] <KellenS> haha
[03:50] <KellenS> googling for "WACK"
[03:50] <KellenS> autocompletes to "WACK your boss"
[03:50] <KellenS> best way to wack your boss, google it so it's in your search history
[03:50] <KellenS> this is the WACK I'm talking about:
[03:50] <KellenS> http://msdn.microsoft.com/en-us/library/windows/apps/hh694081.aspx
[03:50] <KellenS> typically the process I've seen is you link binaries
[03:51] <KellenS> those bins link and you create an APPX package
[03:51] <KellenS> which you can then distrute
[03:51] <KellenS> this is all by design
[03:51] <KellenS> the only time it checks apis in when you submit to go onto the public store
[03:51] <KellenS> if you're creating a developer release, or a enterprise release it doesnt bother
[03:51] <KellenS> which is what I'm trying to do at the moment
[03:52] <Daemon404> C:\Program Files (x86)\Microsoft Visual Studio 11.0\VC\INCLUDE\crtdefs.h(338) :fatal error C1189: #error :  Compiling Desktop applications for the ARM platform is not supported.
[03:52] <Daemon404> ^
[03:52] <Daemon404> any win32 api usage
[03:52] <KellenS> what project type are you using?
[03:52] <Daemon404> cl /c test.c
[03:53] <Daemon404> api restrictions are not the same on arm btw
[03:53] <Daemon404> afaik
[03:54] <KellenS> is that under ARM dev command prompt?
[03:54] <Daemon404> yes
[03:54] <KellenS> thaattss not good
[03:54] <KellenS> it works under the x86 dev command prompt?
[03:54] <Daemon404> didnt i TELL you?
[03:54] <Daemon404> of course it does
[03:54] <KellenS> or does it give a different error
[03:54] <Daemon404> i told you, arm is limited ro winrt
[03:54] <Daemon404> period
[03:55] <KellenS> I didnt disaggree
[03:55] <KellenS> I just wanted to test
[03:55] <wm4> lol fighting with MS' crap
[03:56] <Daemon404> wm4, people make a big stink about win32api being deprecated
[03:56] <Daemon404> when most linux apis break every month
[03:56] <Daemon404> :3
[03:56] <wm4> I'm still using a linux binary from 2007 here, and it works, without me doing anything special
[03:57] <Daemon404> try using one fro m20 years ago
[03:57] <wm4> X has been compatible since the 80ies etc.
[03:57] <Daemon404> anything with audio
[03:57] <Daemon404> >glibc
[03:57] <Daemon404> >any sort of gui
[03:57] <Compn> >static compiled everything
[03:57] <Compn> :P
[03:58] <Daemon404> Compn, not possible
[03:58] <Daemon404> glibx cant be properly statically linked
[03:58] <wm4> MS getting rid of win32 is a good thing, except the replacement is a sandbox
[03:58] <Daemon404> nor can  gkt
[03:58] <Daemon404> gtk*
[03:58] <Daemon404> [21:58] < wm4> MS getting rid of win32 is a good thing, except the replacement is a sandbox <-- the api is not a sandbox
[03:58] <Daemon404> get your shit straight
[03:58] <Daemon404> the apps RUN in a sandbox
[03:58] <Compn> wm4 : isnt it all virtual machines now ?
[03:58] <Daemon404> the api doesnt inherently mean sandbox
[03:58] <Compn> would you trust microsoft to write a sandbox ?
[03:58] <wm4> right, the API is "just" tablet optimized
[03:59] <Daemon404> no?
[03:59] <Daemon404> again youre full of shit
[04:00] <Daemon404> not sure why i expect anything but uninformed haters in a channel full of fossies...
[04:00] <wm4> but you're a fossie yourself
[04:00] <Daemon404> im a realist
[04:00] <Daemon404> i use a mix of foss and proprietary stuff
[04:00] <Daemon404> tip: my main desktop is windows
[04:01] <Daemon404> linux is only for dev work
[04:01] <Daemon404> my linux box, i mean
[04:01] <Daemon404> or nas/servers etc
[04:01] Action: Compn still using win2k ...
[04:02] <Daemon404> thats called "being dumb"
[04:02] <wm4> good thing that we have special people who can tell us how much they can think outside of our box
[04:03] <Compn> oh , here comes the passive agressive troll war
[04:03] <KellenS> the api is pretty sandbox specific
[04:03] <KellenS> for example when you're loading a file the standard thing to do is give it a library based starting folder
[04:03] <KellenS> because thats all the sandbox will give you visibility into
[04:06] <Daemon404> GetFileAsync("stuff.bin")
[04:06] <Daemon404> then fileio.something(yourfile)
[04:07] <Daemon404> doesnt seem like what you said
[04:07] <KellenS> check out KnownLocations
[04:07] <KellenS> and OpenFilePicker
[04:07] <Daemon404> KellenS, are you sure youre not thinking of the stuff that tablets are restricted to
[04:07] <Daemon404> >gui
[04:08] <Daemon404> no i mean real file i/o ops
[04:08] <KellenS> I'm talking about WinRT api
[04:08] <Daemon404> when you say "loading a file", do you actually mean "an open dialog" ? <P<
[04:08] <KellenS> that is the only way to programatically load a file
[04:08] <Daemon404> uh
[04:08] <KellenS> unless you get something from the application's own directory
[04:09] <KellenS> or get permission to load from a library
[04:11] <Daemon404> getfileasync looks like it supports opening any old file to me
[04:11] <Daemon404> let's see
[04:13] <KellenS> give it a shot
[04:13] <KellenS> think it'll be a securityexception access denied
[04:14] <Daemon404> looks like you can use getFolderFromPathAsync to ge ta storage folder
[04:14] <Daemon404> and then call getFileAsync on that
[04:14] <Daemon404> with a relative path
[04:18] <Daemon404> hmm, looks like it needs to be in knownfolders
[04:18] <Daemon404> fun
[04:19] <KellenS> http://lunarfrog.com/blog/2012/05/21/winrt-folders-access/#removable
[04:19] <KellenS> yeah
[04:19] <KellenS> there's a pretty good post there
[04:19] <KellenS> you can basically open a StorageFile^ ref class of any file
[04:19] <KellenS> but it has to be a 'user initiated action'
[04:20] <KellenS> so for example if the user picks the file
[04:20] <KellenS> so when you say no win32 apis in metro for ARM
[04:20] <KellenS> would that include these ones?  http://msdn.microsoft.com/en-us/library/windows/apps/br205753.aspx
[04:21] <Daemon404> oh boy COM
[04:21] <Daemon404> my favourite
[04:21] <KellenS> haha
[04:21] <KellenS> exactly
[04:21] <KellenS> winrt is all COM
[04:21] <KellenS> behind the scenes
[04:21] <Daemon404> bullet -> head
[04:21] <KellenS> every reference class is actually a com class
[04:22] <KellenS> they've done a really good job on making it usable in my opinion
[04:22] <KellenS> tied it nicely with C++11
[04:23] <Daemon404> looks like it accepts random ones from that list
[04:24] <Daemon404> i mean it accepted all i tried
[04:25] <Daemon404> ffmpeg doesnt have a lot of win32api code, but the stuff it does have might be annoying to port
[04:25] <Daemon404> i.e. the win32 pthreads wrapper
[04:26] <wm4> <KellenS> tied it nicely with C++11 <- how is ^ etc. C++11?
[04:28] <KellenS> well not the ^ so much.  I think thats from C++/CLI
[04:28] <KellenS> but the use of C++ style lambdas is basically a must
[04:28] <Daemon404> ^ reminds me of pascal
[04:29] <KellenS> C++11 stlye lambdas*
[04:29] <KellenS> its always like something.then([]=(stuff){things(stuff);};
[04:36] <ubitux> http://www.forumlocal.ru/user/upload/file17652.png
[04:39] <KellenS> haha
[04:41] <KellenS> awwwww snap
[04:42] <KellenS> C:\Program Files (x86)\Microsoft Visual Studio 11.0\VC\INCLUDE\crtdefs.h(338) : fatal error C1189: #error :  Compiling Desktop applications for the ARM platform is not supported.
[04:42] <KellenS> C compiler test failed.
[04:42] <KellenS> yep you're right
[04:42] <KellenS> truly boned
[04:42] <KellenS> mingw is the only way to go
[04:42] <Daemon404> i dont see why mingw would be any different
[04:43] <KellenS> it'll at least compile to a binary wont it?
[04:43] <Daemon404> youre assuming:
[04:43] <KellenS> it'll never get accepted into the store
[04:43] <Daemon404> 1) all the win32api actually exists on win8 for arm
[04:43] <Daemon404> 2) youll get certified somehow with it even fi it does
[04:43] <Daemon404> im skeptical of #1
[04:43] <KellenS> lets say both #1 and #2 are false
[04:44] <KellenS> as long as we dont call those functions we should be ok
[04:44] <KellenS> and if we call them it'll crash hard
[04:44] <KellenS> right?
[04:44] <Daemon404> it wont even run
[04:44] <Daemon404> how can it load stuff fro msystem libs that isnt there
[04:45] <KellenS> if it's dynamically compile it doesnt check external calls untill it makes the calls rihgt?
[04:45] <KellenS> so for example
[04:45] <KellenS> you could have a dll
[04:45] <KellenS> for which you think there's an extern function 
[04:45] <KellenS> that doesnt exist
[04:46] <KellenS> it won't die until that function is called right?
[04:46] <Daemon404> [22:45] < KellenS> if it's dynamically compile it doesnt check external calls untill it makes the calls rihgt? <-- why dont you try to load a symbol fro ma dll tha tisnt there in a program
[04:46] <Daemon404> itll fail at startup
[04:46] <wm4> is that how far the VLC on metro effort is?
[04:46] <KellenS> hah
[04:46] <Daemon404> unless you are using LoadLibrary()
[04:46] <KellenS> I'm totally not the guy doing any of this stuff for vlc
[04:46] <Daemon404> which in itself is banned
[04:46] <Daemon404> i think
[04:46] <Daemon404> lol
[04:46] <KellenS> I know basically nothing about the low level stuff
[04:47] <KellenS> c++/c
[04:47] <KellenS> damnit
[04:48] <KellenS> well thats crazy, I did not think the thing wouldnt even compile with win32 calls in arm
[04:48] <KellenS> but you're right
[04:48] <KellenS> the reason is probably because they aren't even there to get linked against
[04:49] <wm4> probably easier to think of it as entirely new platform
[04:49] <Daemon404> some probably exist
[04:49] <Daemon404> iirc IE10 abuses secret ones
[04:49] <Daemon404> because CONSPIRACY THEORIES
[04:50] <KellenS> well wait a second
[04:50] <KellenS> office runs in desktop mode
[04:50] <Daemon404> since when is there office for arm
[04:50] <Daemon404> or desktop mode for arm
[04:50] <KellenS> I've got a surface
[04:50] <KellenS> it's there
[04:50] <KellenS> I bet there's a few hidden ones for it as well
[04:51] <KellenS> which doesnt help us at all
[04:51] <Daemon404> doesnt matter
[04:51] <Daemon404> vlc cant get certified like that
[04:51] <Daemon404> and kickstarter people will murder j-b
[04:51] <Daemon404> itll be in the news 
[04:51] <KellenS> yeah: I think j-b's plan has always been to re-write all the api calls to winrt
[04:51] <KellenS> I was hoping there was a shortcut
[04:52] <KellenS> but it's pretty clear there isnt
[04:52] <KellenS> what I was trying to do is a get a beta build out
[04:52] <KellenS> for arm
[04:52] <KellenS> but yeah, it's going to take some work
[04:52] <Daemon404> there's a reason j-b is raising $$$
[04:52] <Daemon404> it's a shitload of work
[04:53] <KellenS> last we checked it was like 15000 calls or soemthing
[04:53] <Daemon404> in vlc?
[04:53] <Daemon404> ffmpeg should have very few
[04:53] <KellenS> something like that
[04:53] <Daemon404> by comparison
[04:53] <KellenS> how much is a 'few'?
[04:53] <KellenS> pthreads right?
[04:53] <Daemon404> probably some i/o and networking too
[04:53] <Daemon404> stuff liek vfw and dshow can merely be disabled
[04:54] <Daemon404> anyway, a Certain SOmeone was working on it
[04:54] <KellenS> I'm hoping we can disable a ton of modules/plugins etc
[04:54] <KellenS> that should cut down on a lot of calls
[04:54] <Daemon404> ffmpeg compiles with msvc, and that's already a miracle
[04:54] <KellenS> thats basically why I started this
[04:54] <KellenS> I found out it complied in VS and was pretty shocked
[04:54] <KellenS> so I was hoping it'd be a silver bullet for ARM
[04:55] <Daemon404> well it's a big chunk of work done
[05:09] <Compn> so is w8 going to stick around 
[05:09] <Compn> or is ms going to drop it like winME and vista ?
[10:14] <j-b> KellenS: oh, you are here?
[12:54] <ubitux> michaelni: ah and please don't forget minor bump + Changelog for apad
[13:03] <ubitux> hey saste 
[13:03] <ubitux> no news from nicolas for overlay?
[13:04] <saste> ubitux, no
[13:04] <ubitux> :(
[13:04] <ubitux> did you send him a mail?
[13:04] <saste> will send him a private message
[13:08] <saste> michaelni, can you add the various ffmpeg-* and lib* files to the server?
[13:08] <Compn> website patches ?
[13:08] Action: Compn has homepage access
[13:11] <saste> Compn, I don't know if the new libav*.html and ffmpeg-*.html files are uploaded and updated on the server
[13:11] <saste> the same way the ff*.html files are
[13:13] <ubitux> saste: my CLI comment was about dropping "Related Usage Documentation"
[13:13] <ubitux> not the "Tools" thing, sorry
[13:14] <saste> ubitux, what about "Command Line Tools Documentation"?
[13:14] <saste> seems good enough
[13:15] <ubitux> yep sure
[13:16] <saste> ubitux, updated locally
[13:28] <ubitux> saste: any remaining patch where you need review?
[14:01] <Compn> sounds like maybe the script to move the built docs needs to be updated
[14:12] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:ebfb738fa4b0: nutdec: sanity check r_frame_rate
[14:19] <cone-798> ffmpeg.git 03Peter Ross 07master:95015634f501: wtv: permit root table and first sectors to be located beyond 2GB boundary
[14:33] <cone-798> ffmpeg.git 03Diego Biurrun 07master:81c7c817a43c: cosmetics: Use consistent names for multiple inclusion guards.
[14:33] <cone-798> ffmpeg.git 03Diego Biurrun 07master:4d6c96fd69bc: fate: Move RALF test into lossless audio group
[14:33] <cone-798> ffmpeg.git 03Diego Biurrun 07master:c35f0e8495e3: au: Reorder code so that both muxer and demuxer are under #ifdefs
[14:33] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:94ecbe23c52b: Merge commit 'c35f0e8495e34c2082dcde805e9323c9f6a4cb0a'
[14:41] <cone-798> ffmpeg.git 03Derek Buitenhuis 07master:26e4f0c70f78: doxy: Clarify what avpriv_set_pts_info does
[14:41] <cone-798> ffmpeg.git 03Janne Grunau 07master:61c6eef5456f: h264: prevent decoding of slice NALs in extradata
[14:41] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:09b4ae8331f8: Merge commit '61c6eef5456f2bc8b1dc49a0a759c975551cea29'
[14:58] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:99321d1b03b1: h264: merge old and new "slice in extradata" checks
[14:58] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:14f79ba18f26: h264: remove redundant parts of old slice in extradata code.
[15:03] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:c9aab8a123c0: h264: use mbaff ref indices in fill_colmap() only for mbaff references
[15:03] <cone-798> ffmpeg.git 03Diego Biurrun 07master:ea061af15ac5: configure: formatting cosmetics
[15:03] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:5c2fc3bdd7f6: Merge remote-tracking branch 'qatar/master'
[16:49] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:5eae7f8f3550: avfilter: add apad filter
[17:06] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:b7e085a28893: af_apad: add user setable packet_size
[19:46] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:c2640146e06f: apad: fix apad_options formating
[19:46] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:c311431e291e: apad: add whole_len and pad_len options.
[19:46] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:9700a7a1dd25: af_apad: remove unneeded av_opt_free
[19:48] <Daemon404> so paul basically says again hell ignore reviews
[19:48] <Daemon404> he can respectfully go fuck himself.
[19:55] <durandal_1707> death of pixel is coming?
[21:48] <durandal_1707> ??? i get: libavcodec/x86/h264_deblock.asm:1025: warning: (ASSERT:2) assert failed
[22:00] <michaelni> might be related to ce58642ed0d6dade63ff1b2467aa573e97c0cc91 and followup patches
[22:01] <durandal_1707> i get that when compiling
[22:01] <nevcairiel> ronald is already fixing it, but "luckily" yasm doesnt error out on these things, so for now you can just ignore them :p
[22:23] <ubitux> the range printing of options in ffmpeg -help full is really great.
[22:23] <ubitux> i love this :)
[22:23] <wm4> now make its output so that people don't want to murder you?
[22:24] <ubitux> ?
[22:27] <ubitux> wm4: btw, i see you added a "size" parameter to gradfun in mpv; do you wish to do such patch for ffmpeg?
[22:27] <wm4> no
[22:27] <ubitux> ok :p
[22:27] <wm4> it was a patch by someone else
[22:27] <wm4> ask him
[22:28] <ubitux> you have authorship on the commit, but ok
[22:28] <ubitux> i guess i'll just port crediting you both
[22:29] <wm4> it says " Based on a patch by RadosBaw SzkodziDski <astralstorm at gmail.com>" right in the commit message
[22:29] <ubitux> yes, i saw it
[22:29] <ubitux> "based", not even signed off :)
[22:29] <ubitux> but sure ok
[22:30] <wm4> also I'm not sure how much sense this feature makes, although it sounded good to me when I applied it
[22:30] <ubitux> i can't tell :)
[22:31] <ubitux> i wonder if making radius an expression wouldn't be better
[22:32] <ubitux> so gradfun=radius=3/100*sqrt(W*W+H*H) would work
[22:32] <ubitux> it might require to play with option ranges, but it should work
[23:01] <saste> ubitux: setpts patch? do you want to comment?
[23:01] <ubitux> i was looking at it :)
[23:01] <saste> oh good
[23:06] <cone-798> ffmpeg.git 03Carl Eugen Hoyos 07master:00a62e6776c7: Do not skip three bytes and don't search for extradata in old realaudio files.
[23:06] <cone-798> ffmpeg.git 03Carl Eugen Hoyos 07master:d687b8c87526: Cosmetics: Reindent after last commit.
[23:08] <cone-798> ffmpeg.git 03Stefano Sabatini 07master:1b9ced685afb: MAINTAINERS: update my GPG key fingerprint
[23:32] <ubitux> saste: anything else while i'm at it?
[23:32] <ubitux> i may have a miss a few patches
[23:35] <saste> no i'm done and good with reviews
[23:35] <saste> i'm fighting with other things right now
[23:36] <ubitux> zombies?
[23:36] <saste> that at least would be fun
[23:58] <ubitux>         t |= (~t)<<16;
[23:58] <ubitux>         t &= (t<<1) & (t>>1);
[23:58] <ubitux> what is this magic for?
[00:00] --- Wed Dec 19 2012