[FFmpeg-devel-irc] IRC log for 2010-09-07

Wed Sep 8 02:00:36 CEST 2010

[02:19:46] <pengvado> lgpl x264_add8x4_idct_sse2 ok
[02:21:05] <Dark_Shikari> BBB: ^
[02:22:30] <BBB> awesome
[02:22:33] <BBB> thanks pengvado
[02:23:12] <BBB> that's the end of yasm h264dsp then, I'll run a few quick tests to see if I can fix win64 fate and then xvp8 will be done
[02:23:26] <BBB> Dark_Shikari: you sure you don't want to help yasmifying some of our asm?
[02:24:43] <Dark_Shikari> wait, so h264dsp_mmx.c is now gone?
[02:25:04] <BBB> deblock strength prediction is still ther
[02:25:10] <BBB> and all init crap
[02:25:12] <Dark_Shikari> Oh.  we'll want to deal with that eventually
[02:25:17] <Dark_Shikari> but I guess not important for now
[02:25:25] <BBB> indeed
[02:39:06] <astrange> do the yasm functions (the ones that don't call other functions themselves, at least) break frame-pointer backtraces?
[02:40:27] <Dark_Shikari> yes
[02:40:32] <Dark_Shikari> they don't save ebp
[02:40:37] <Dark_Shikari> well, I mean, they save it
[02:40:40] <Dark_Shikari> but they don't do that whole backup thing
[02:40:50] <Dark_Shikari> or, actually, I think they only partially break it
[02:40:53] <Dark_Shikari> not quite sure
[02:44:22] <pengvado> iirc gdb gets confused by functions that modify esp but don't use ebp as a frame pointer. most asm functions don't use the stack, so they're ok.
[02:50:03] * BBB sleep...
[02:59:15] <astrange> mru: what's the problem with intptr_t on ppc64?
[03:11:54] <Compn> heh
[04:54:33] * _av500_ is travelling with the beast
[04:56:38] <kshishkov> with or without such insignificant part as displays?
[05:07:00] <thresh> moroning
[05:11:45] <_av500_> kshishkov: with of course
[05:12:09] <_av500_> without it would not be a true beast
[05:22:56] <pJok> god morgon
[05:25:37] <kshishkov> goda morgnar
[05:25:48] <Tjoppen> morrn
[05:25:55] <Tjoppen> kaffe \o/
[05:26:47] <kshishkov> Tjoppen: some people manage to live without it
[05:28:36] <Tjoppen> :)
[05:29:08] <kshishkov> but these days I drink a liter of filmjÃ¶lk daily
[05:29:20] <kshishkov> anyway, hjortronsylt time
[05:30:18] <Tjoppen> protip: hjortronsylt goes well with waffles
[05:31:47] <kshishkov> or Ã¤lgkÃ¶ttbullar :)
[05:32:41] <Tjoppen> well, with that I'd prefer lingonsylt
[05:33:04] <Tjoppen> and perhaps saltgurka
[05:34:39] <kshishkov> maybe pressad gurka?
[05:35:12] * kshishkov likes lingonsylt
[05:55:13] <pJok> mmmmh, älgköttbullar
[05:55:47] <kshishkov> there's definitely no such thing in Denmark
[05:55:58] <pJok> nop
[05:56:07] <pJok> you can't even get älgkött in denmark
[05:56:13] <kshishkov> loosers
[05:56:25] <kshishkov> at least you have high mountain(s)
[05:57:45] <vipw> someone likes the taste of moose?
[05:58:37] <pJok> denmark is flat as a pancake, kshishkov
[05:58:46] <pJok> vipw, moose is good
[06:00:00] <vipw> i wonder if it tastes different in europe than in america
[06:00:10] <vipw> i never ate it when i lived in sweden
[06:01:41] <funman> http://daniel.haxx.se/blog/2010/09/01/nominations-for-nordic-free-software-award-2010/
[06:01:56] <pJok> wouldn't know, never tried canadian moose
[06:04:18] <_av500_> funman: can we nominate kshishkov while he is still up north?
[06:05:37] <funman> _av500_: i guess he qualifies as "people in the nordic countries"
[06:05:37] <kshishkov> _av500_: nope, he's hearing you again
[06:25:02] <benoit-> moin
[06:25:11] <kshishkov> bonjour
[06:28:15] <_av500_> salut
[06:32:37] <mru> morning
[06:32:45] <kshishkov> god morgon
[07:05:17] <_av500_> lol, get h264 in firefox via new divx player
[07:05:30] <_av500_> problem solved
[07:05:47] <kshishkov> divx solves any problem?
[07:09:04] <mru> there are few companies more annoying than divx
[07:10:17] <kshishkov> well, wait till that company that bought it renames itself to SCO^W divx corp
[07:10:17] <_av500_> true
[07:10:36] <mru> http://adangerousbusiness.com/2010/01/05/the-museum-of-soviet-video-games/
[07:11:57] <kshishkov> mru: why such interest to it?
[07:18:46] <pJok> mru, buffering inc. ?
[07:19:55] * pJok is surprised that Real is still alive
[07:20:28] <kshishkov> "it is not dead which can eternal buffer..."
[07:27:32] <merbzt> kshishkov/andoma/superdump: dinner at 1830 somewhere in the city center
[07:27:48] <superdump> sounds ok to me
[07:27:55] <kshishkov> merbzt: today?
[07:27:58] <merbzt> yes
[07:28:08] <superdump> andoma: ^?
[07:28:13] <kshishkov> fine, I'm rather free till Sunday anyway
[07:40:38] <andoma> superdump: Pontus have a cold so i'm home taking care of him, Lisa wokring long hours today so she can be home tomorrow
[07:40:44] <andoma> so I guess i'm out
[07:44:03] <kshishkov> looks like it's harder to get you than, say, MÃ¥ns
[07:50:15] <superdump> andoma: hope he gets better soon
[07:50:50] <KotH> moin
[07:50:56] <andoma> superdump: yeah, me too!
[09:11:45] <CIA-11> ffmpeg: stefano * r25055 /trunk/doc/ffprobe-doc.texi: Prefer "foo" over ``foo'' in the ffprobe docs.
[09:11:48] <CIA-11> ffmpeg: stefano * r25056 /trunk/doc/ffprobe-doc.texi: Amend documention for the option -prefix.
[11:03:01] <mru> bah, silly arm engineer trying to deny there's a bug
[11:03:37] <spaam> bug in the cpu? :)
[11:03:41] <mru> compiler
[11:24:41] <KotH> mru: what does he say? feature?
[11:25:20] <mru> he claimed the generated code was correct
[11:25:47] <cartman> in another dimension maybe
[11:26:29] <mru> oh well, this is how the game is played
[11:27:10] <cartman> mru: whats next?
[11:27:35] <av500> shotgun
[11:27:36] <mru> I reply and explain why he's wrong
[11:27:54] <mru> then he apologises for the mistake and forwards the bug to the real engineers
[11:28:00] <cartman> LOL
[11:28:31] <cartman> in the intel c++ compiler's case first you are greeted by indian "engineers" who claims there is no bug :P
[11:28:43] <cartman> then you get to a real engineer
[11:30:26] <mru> and then there's the TI compiler
[11:30:37] <mru> I have a one-line test case that makes it segfault
[11:31:46] <mru> and the best part, it doesn't even contain any code
[11:31:53] <mru> only a data declaration
[11:32:07] <cartman> mru: ibm xlc will crash if you feed it with an empty file
[11:32:14] <av500> #pragma((crash))
[11:32:34] <mru> cartman: ibm xlc doesn't even come close to compiling ffmpeg
[11:32:53] <cartman> compiles lotsa shit though :)
[11:36:00] * KotH waits for ffcc
[11:36:38] <cartman> no need there is clang
[11:36:56] <mru> hah, the second bug I reported last week was accepted without that dance
[11:37:02] <mru> still waiting for the third
[11:37:18] <cartman> mru: how pricey is the ARM compilers?
[11:37:24] <mru> no idea
[11:37:34] <cartman> ah company license?
[11:37:47] <mru> no, they gave me a free licence
[11:38:03] <cartman> ah, nice :)
[11:38:11] <cartman> so it didn't cost an ARM and a leg :P
[11:38:19] <spaam> mru: what did you do to get that?:)
[11:38:36] <mru> asked nicely
[11:38:38] <cartman> there was an ARM engineer in ffmpeg committers too
[11:38:42] <cartman> Ian something
[11:38:52] <mru> caulfield or something
[11:38:57] <cartman> yeah! :)
[11:40:19] <av500> cartman: round 5k
[11:40:27] <cartman> av500: ouch
[11:40:34] <av500> yep
[11:40:47] <thresh> 5k for a company? nothing
[11:40:47] <cartman> should be free for non-commercial :/
[11:40:57] <av500> cartman: tell that to arm
[11:41:06] <cartman> av500: I am sure they'll listen
[11:41:15] <mru> gcc is usually good enough
[11:41:28] <cartman> mru: its not gcc based is it?
[11:41:29] <mru> armcc doesn't usually beat it by more than 5-10%
[11:41:33] <av500> cartman: they didnt listen to us
[11:41:34] <cartman> gcc is GPL nm
[11:41:40] <mru> cartman: no, nothing to do with gcc
[11:41:45] <cartman> av500: they will respect my authorita!
[11:43:28] <mru> what armcc has over gcc is much finer control over certain aspects
[11:43:45] <mru> things like code layout and calling conventions
[11:44:00] <mru> which don't matter at all on a linux system
[11:46:03] <cartman> mru: product name is RVDS right?
[11:46:38] <mru> that's the full suite
[11:47:05] * cartman looks for a seperate download
[11:47:05] <mru> with jtag debugger, ide, simulator, etc
[11:47:24] <mru> there's a 30-day eval for the whole thing
[11:47:31] <cartman> if I wanted to do some development with this
[11:47:39] <cartman> I would need a board with ARM cpu right?
[11:47:47] <mru> not much point otherwise
[11:48:07] <cartman> something like beagleboard or is there a cheap board you would suggest for starters
[11:48:22] <mru> beagle is a good choice
[11:48:47] <mru> anything cheaper will have a much slower cpu
[11:48:53] <cartman> ok, would be neat to test this stuff :-)
[11:49:08] <mru> it's just another compiler
[11:49:24] <cartman> no I mean ARM asm and such
[11:49:46] <mru> there's nothing you can't do with gcc
[11:50:23] <cartman> mru: I mean to learn ARM :)
[11:50:28] <cartman> I'll use gcc ok :P
[11:50:31] <mru> you can do that with gcc too
[11:50:33] <cartman> beagleboard even has a DSP
[11:50:42] <mru> that's another world
[11:50:43] <cartman> mru: where do I run my stuff? simulator?
[11:50:49] <mru> on the beagle of course
[11:51:07] <cartman> no free beer /me goes to read the details of the board
[11:53:34] <cartman> is gumstix an option too ?
[11:54:11] <mru> same cpu as beagle
[11:54:14] <mru> slightly more expensive
[11:54:29] <mru> the cpu module is the same price iirc
[11:54:38] <mru> but you need the breakout board with connectors too
[11:55:34] <av500> unless you have a steady hand
[11:55:51] <cartman> cheaper the better
[11:56:21] <av500> cartman: I can also hook up a BB and sell you machine hours :)
[11:56:44] <cartman> av500: internet is too slow here, wouldn't be much fun
[11:57:04] <mru> av500: rent him your spare room too
[11:57:09] <av500> you could submit batch jobs via email :)
[11:57:30] <cartman> or I could get my company to pay for a BB :P
[11:57:36] <cartman> just need a compelling reason :P
[11:57:41] <av500> outrageous proposal
[11:57:51] <av500> "Its'....because!"
[11:58:00] <cartman> goes like that :D
[11:58:05] <cartman> to accelerate the business value of
[11:58:14] <cartman> What would wally do? anyway
[11:58:19] <av500> synergy effects!
[11:58:29] <mru> paradigm shifts
[11:58:38] <mru> best in class
[11:58:44] <av500> cross brand identifoo
[11:58:53] <mru> vertical markets
[11:59:03] <av500> horizontal business
[11:59:09] <cartman> lol
[11:59:15] <cartman> go with that av500
[12:04:52] <av500> mru: http://git.m2x.eu/git/cgi-bin/gitweb.cgi?p=vlc.git;a=commitdiff;h=6c2feaea88aa668e4f021fc5b3e54c5f05f3feab
[12:05:19] <mru> yes?
[12:05:25] <av500> it will run faster?
[12:05:59] <kierank> yes
[12:06:11] <mru> the init function will run about 1ns faster if you have a stupid compiler, yes
[12:07:00] <vipw> there are compilers that bad?
[12:07:13] <mru> actually it depends on the type of that value
[12:07:24] <mru> if it's signed the compiler can't do that optimisation
[13:09:40] <cartman__> seems like clang miscompiles ffmpeg
[13:09:42] <cartman__> errr
[13:09:46] <cartman__> s/ffmpeg/mplayer
[13:10:21] <mru> and the sky is blue
[13:11:03] <lu_zero> mru: here is grey
[13:11:51] <mru> no, that's just the clouds obscuring the sky
[13:11:59] <mru> the actual sky is still blue
[13:14:10] <funman> it's black here
[13:17:29] <cartman> ok no miscompilation, I was wrong, sky is white
[13:17:31] <cartman> :P
[13:20:28] <cartman> just one asm error and commenting out that file works
[13:21:10] <mru> which file?
[13:21:17] <mru> something in ffmpeg?
[13:21:24] <cartman> nope
[13:21:30] <mru> 32-bit?
[13:21:44] <cartman> http://llvm.org/bugs/show_bug.cgi?id=4504
[13:22:23] <mru> wtf is the o constraint?
[13:22:56] <cartman> no idea
[13:28:27] <funman> `o' A memory operand is allowed, but only if the address is "offsettable".
[13:35:31] <pengvado> I never figured out what "o" is supposed to mean, seeing as gcc x86 asm doesn't actually have any syntax for an offsetted memory arg
[13:35:59] <mru> isn't that one of the constraints available on all arches?
[13:36:08] <mru> whether or not they make sense
[13:36:32] <pengvado> I guess so
[13:39:28] <funman> yes
[13:39:51] <pengvado> though I question whether gcc's raw textual substitution of args is sufficient for offsetted memory args on any other arch either
[13:40:56] <mru> arm guy is back: "I apologise.  You are correct."
[13:41:00] <cartman> mru: lol
[13:41:12] <lu_zero> which was the subject?
[13:41:21] <mru> armcc bugs
[13:41:25] <lu_zero> ah
[13:41:47] <lu_zero> that reminds me that I should find some time to look at the pathscale bug
[13:42:00] <mru> they have a guy looking at it
[13:42:08] <mru> but he's on holiday
[13:42:12] <mru> back next week
[13:42:13] <lu_zero> I see ^^
[13:42:25] <lu_zero> and I'm again overbusy =E
[13:49:14] <merbzt> what is the omap4 beagleboard lookalike called ?
[13:49:25] <av500> panda
[13:49:29] <av500> board
[13:49:38] <av500> add # for irc channel
[13:50:40] <cartman> looks too recent to be available
[13:50:48] <av500> its not yet
[13:50:52] <av500> only few ppl have one
[13:51:03] <av500> like all canonical employees
[13:51:13] <mru> ~curse canonical
[13:51:15] <cartman> av500: are you an ubuntu developer?
[13:51:25] <av500> cartman: why insult me?
[13:51:33] <cartman> yeah treats you right
[13:54:12] <cartman> ubuntu on arm
[13:54:14] <cartman> uhm
[13:54:51] <kierank> armbuntu
[13:55:00] <av500> buarm
[13:55:17] <kierank> ubuntarm
[14:35:43] <kshishkov> merbzt: please see http://shishkov.homeunix.net/vb.jpg and advise what to try first
[14:36:23] * thresh almost broke his neck
[14:36:33] <cartman> thresh: oops
[14:36:49] <av500> kshishkov: try to fix gravity 1st
[14:36:53] <cartman> kshishkov: Kuba
[14:36:53] <merbzt> well try cuba cola
[14:37:01] <cartman> I knew it :D
[14:37:22] <av500> kshishkov: try none of these
[14:37:29] <av500> they all look wierd
[14:37:46] <thresh> like from 80s, right?
[14:37:50] <kshishkov> av500: you haven't seen Vira BlÃ¥tira then
[14:37:52] <av500> get something natural, from real fruit, like hops and barley
[14:38:03] <mru> +1
[14:38:11] <kshishkov> thresh: 60s or so
[14:38:28] <kshishkov> av500: I'm no horse
[14:38:40] <mru> that's oats
[14:39:14] <mru> hops, barley, and a dash of yeast... wonderful recipe
[14:39:17] <mru> and water of course
[14:40:00] <kshishkov> mru: well, only water sounds tempting for me
[14:40:22] <mru> speaking of, anyone in sw .de up for some trolling this weekend?
[14:40:33] <av500> as in?
[14:40:50] <mru> you know what I mean
[14:40:58] <kshishkov> av500: to keep mru a company, you know where he is
[14:41:37] <av500> i phear i am restricted from venturing too far from home the next days
[14:42:05] <kshishkov> av500: kids, work or worse?
[14:42:13] <mru> wife?
[14:42:17] <av500> yup
[14:43:52] <kshishkov> definitely worse then
[14:44:44] <cartman> damn wives
[14:48:21] <av500> cartman: how many do you have?
[14:48:32] <kierank> lol
[14:48:42] <cartman> av500: 1 for now
[14:48:45] <cartman> after September, 12
[14:48:49] <cartman> up to 4 hopefully
[14:48:59] <cartman> Islam baby
[14:49:00] <cartman> :P
[14:49:46] <mru> islamic polygamy works in a society where men routinely kill one another
[14:49:54] <mru> thus creating a surplus of women
[14:50:04] <kshishkov> cartman: I'd say it's one too many but Russian tradition is more concerned about mother-in-law
[14:50:18] <mru> in the civilised world, we no longer live like that
[14:50:44] <mru> at least not so much
[14:50:47] <cartman> Åžeriat is coming baby
[14:50:49] <cartman> :P
[14:51:10] <cartman> sadly I am only kidding :P
[14:51:26] <cartman> will possibly have to leave the country if that happens
[14:51:35] <av500> with 4 wives?
[14:51:41] <av500> domt come here
[14:51:46] <mru> go to utah
[14:51:53] <av500> work for novell
[14:52:03] <kshishkov> cartman: just invite another general Mustafa ;)
[14:52:17] <cartman> kshishkov: maybe one will come, possibly not soon
[14:52:23] <av500> or let your 4 wives work for novell while you lean back and drink kshishkov-drinks
[14:52:28] <cartman> av500: 1 wife is enough pain
[14:52:29] <mru> iirc polygamy is illegal but not prosecuted in utah
[14:52:50] <kshishkov> av500: you can't find them to the south of Stockholm (including it)
[14:53:02] <cartman> kshishkov: you know too much about turkey :P
[14:53:03] <cartman> damn you
[14:53:38] <kshishkov> cartman: no, I don't. And I don't know a thing about Ottoman empire or seljouks
[14:53:47] <av500> kshishkov: I am quite sure one can find fruit flavored colored sugar water all over the world
[14:53:51] <cartman> kshishkov: Mustafa Kemal is enough :P
[14:55:09] <cartman> kshishkov: what would be your opinion on Lenin vs Kemal. Are they similar ?
[14:55:19] <kshishkov> av500: that depends on your taste as well. Personally I don't think Germany has anything better than Apfelschorle
[14:56:04] <mru> that's because you refuse to try the beer
[14:56:09] <kshishkov> cartman: similar time, some similar reforms too but different approach. And Lenin would not tolerate even toy opposition
[14:56:39] <cartman> kshishkov: opposition can be deadly in the early times of a revolution
[14:56:40] <kshishkov> mru: yes, and I'm not going to change my decision
[14:57:05] <mru> you don't know what you're missing
[14:57:19] <cartman> cola > beer
[14:57:22] <cartman> kshishkov++
[14:57:25] <kierank> lol
[14:57:39] <kierank> sucks to be you cartman
[14:57:41] <kshishkov> cartman: what about after? And I would say that Lenin wasn't trying to make Russia closer to Europe
[14:57:50] <cartman> kierank: I drink vodka too, sometimes
[14:58:30] <cartman> kshishkov: what would Europe want anyway? They say our current prime minister is very very good, democratic etc.
[14:58:42] <cartman> and if you badmouth government, they put you in jail
[14:58:47] <kshishkov> kierank: I've managed to live with sober outlook on life in Ukraine, so I definitely don't need alcohol while in Europe
[14:58:51] <cartman> Europe is lying with a straight face atm.
[14:59:12] <kshishkov> prime minister != government
[14:59:18] <cartman> it is so here
[14:59:29] <cartman> deep government even
[14:59:50] <cartman> Mustafa Kemal would say, observe Europe and try to take good parts
[15:00:56] <kshishkov> and av500 would point out you've taken the worst one - Serbia, Greece, whatever
[15:01:07] <cartman> anyhow I don't trust Europe at all, maybe Lenin thought the same
[15:01:08] <cartman> dunnow
[15:01:29] <cartman> if they want a second Iran fine, thats whats happening down here
[15:01:30] <kierank> difficult to survive in england without alcohol
[15:01:58] <kshishkov> cartman: Turkey doesn't have oil, who cares?
[15:02:05] <cartman> kshishkov: Iraq does
[15:02:10] <cartman> you go to Iraq through us
[15:02:10] <mru> greece has olive oil
[15:02:16] <cartman> thats the point
[15:02:45] <cartman> look what they did to Persians
[15:02:52] <kshishkov> kierank: looks like in Scandinavia they prefer coffee to alcohol
[15:03:14] <funman> you can of course merge the two
[15:04:05] <mru> irish coffee ftw
[15:04:10] <Tjoppen> kask
[15:04:33] <funman> http://static.skynetblogs.be/media/2260/dyn001_original_316_445_jpeg_12527_00864b676e35bf44b25a7f7582ab8947.2.jpg < that makes a nice start of the day with coffee
[15:04:56] <cartman> ok off to home
[15:04:58] <cartman> see you!
[15:09:40] <BBB> pengvado, Dark_Shikari: carl eugen asks if you guys can commit the license changes to h264_idct_sse2.asm yourself (or email the list or so), so it's a little bit more public that I'm not just screwing around
[15:12:20] <superdump> merbzt, kshishkov : see you in a bit :)
[15:31:49] * kshishkov goes away too
[16:00:28] <lu_zero> eh
[16:00:43] <lu_zero> fun the top-ix lost part of is networking
[16:01:08] <lu_zero> ^^;
[18:36:32] <lu_zero> hi j0sh
[18:39:44] <bcoudurier> hi guys
[18:40:07] <kierank> hello mr bot
[18:40:08] <lu_zero> hi bcoudurier
[18:40:20] <bcoudurier> :)
[19:04:29] <merbanan> elo Baptiste
[19:06:05] <bcoudurier> hey merbanan
[19:16:15] <CIA-11> ffmpeg: stefano * r25057 /trunk/libavcore/ (imgutils.c imgutils.h avcore.h):
[19:16:15] <CIA-11> ffmpeg: Adopt a hierarchical name scheme for the imgutils.h API. Simplify
[19:16:16] <CIA-11> ffmpeg: grepping and somewhat more consistent with the scheme adopted by other
[19:16:16] <CIA-11> ffmpeg: FFmpeg modules API.
[19:16:16] <CIA-11> ffmpeg: stefano * r25058 /trunk/ (51 files in 3 dirs): Use new imgutils.h API names, fix deprecation warnings.
[19:20:26] <CIA-11> ffmpeg: stefano * r25059 /trunk/doc/APIchanges: Add APIchanges entry after the libavcore/imgutils.h renames of r25057.
[19:55:43] <CIA-11> ffmpeg: mru * r25060 /trunk/libavcodec/mpegvideo_enc.c:
[19:55:43] <CIA-11> ffmpeg: Check rc_buffer_size value using integer arithmetic
[19:55:43] <CIA-11> ffmpeg: Using floating-point here can cause erroneous rejection of
[19:55:43] <CIA-11> ffmpeg: parameters due to rounding errors leading to a slightly too
[19:55:43] <CIA-11> ffmpeg: large result.
[19:55:44] <CIA-11> ffmpeg: This fixes the mxf regression test with gcc 4.5 on x86_32.
[19:56:43] <mru> I wonder if that will fix it with sunc too
[19:58:04] <merbanan> mru: time will tell
[19:58:17] <mru> yes, it does that
[20:03:18] <kierank> bcoudurier: with 337m do you know if a guard band is meant to come after the preamble or if the preamble follows a guard bane
[20:03:21] <kierank> band*
[20:24:29] <bcoudurier> kierank, sorry don't know about that
[20:45:41] <CIA-11> ffmpeg: vitor * r25061 /trunk/libavcodec/ (lsp.h amrnbdec.c lsp.c):
[20:45:41] <CIA-11> ffmpeg: Move AMRNB lsf2lsp() function to common code for using in future AMRWB decoder.
[20:45:41] <CIA-11> ffmpeg: Patch by Marcelo Galv?o P?voa
[20:47:32] <CIA-11> ffmpeg: vitor * r25062 /trunk/libavcodec/ (lsp.h sipr.c lsp.c):
[20:47:32] <CIA-11> ffmpeg: Move lsp2lpc_sipr() function to common code so it can be reused in a
[20:47:32] <CIA-11> ffmpeg: AMRWB decoder.
[20:47:32] <CIA-11> ffmpeg: Patch by Marcelo Galv?o P?voa.
[20:59:54] <mru> so now only nut is failing with gcc 4.5
[21:00:07] <mru> and a bunch of rgb-related things fail with suncc
[21:07:11] <CIA-11> ffmpeg: cehoyos * r25063 /trunk/libavformat/utils.c:
[21:07:11] <CIA-11> ffmpeg: Fix formatting for negative start times (issue 2139).
[21:07:11] <CIA-11> ffmpeg: Patch by Christian d'Heureuse, chdh inventec ch
[21:24:43] <CIA-11> ffmpeg: stefano * r25064 /trunk/ (7 files in 2 dirs):
[21:24:43] <CIA-11> ffmpeg: Reimplement ff_img_copy_plane() as av_image_copy_plane() in libavcore,
[21:24:43] <CIA-11> ffmpeg: and deprecate the old function.
[21:24:44] <CIA-11> ffmpeg: stefano * r25065 /trunk/libavcodec/imgconvert.c: Reindent after the last commit.
[21:24:54] <CIA-11> ffmpeg: stefano * r25066 /trunk/libavcodec/imgconvert.c:
[21:24:54] <CIA-11> ffmpeg: Reimplement av_picture_data_copy() avoiding the use of PixFmtInfo
[21:24:54] <CIA-11> ffmpeg: information.
[21:24:54] <CIA-11> ffmpeg: Required for moving the function to libavcore.
[21:24:55] <CIA-11> ffmpeg: stefano * r25067 /trunk/ (8 files in 4 dirs):
[21:24:55] <CIA-11> ffmpeg: Move av_picture_data_copy() to libavcore, and rename it
[21:24:56] <CIA-11> ffmpeg: av_image_copy().
[21:24:56] <CIA-11> ffmpeg: stefano * r25068 /trunk/ (ffplay.c libavcodec/imgconvert.c libavfilter/vsrc_buffer.c): Reindent.
[21:24:57] <CIA-11> ffmpeg: stefano * r25069 /trunk/libavcodec/utils.c:
[21:24:58] <CIA-11> ffmpeg: Remove use of deprecated functions av_image_fill_pointers/linesizes in
[21:24:58] <CIA-11> ffmpeg: libavcodec/utils.c, fix warnings.
[21:24:59] <CIA-11> ffmpeg: stefano * r25070 /trunk/libavcore/ (imgutils.c imgutils.h):
[21:24:59] <CIA-11> ffmpeg: Cosmetics: use plural form for *linesizes arguments of
[21:25:00] <CIA-11> ffmpeg: av_image_copy(), more consistent with the rest of the API and more
[21:25:01] <CIA-11> ffmpeg: semantically correct.
[21:30:30] <CIA-11> ffmpeg: stefano * r25071 /trunk/doc/APIchanges:
[21:30:30] <CIA-11> ffmpeg: Add APIchanges entry after the recent changes to imgutils.h API of
[21:30:30] <CIA-11> ffmpeg: r25064 and r25067.
[23:42:37] <CIA-11> ffmpeg: stefano * r25072 /trunk/libavcore/imgutils.c:
[23:42:37] <CIA-11> ffmpeg: Add missing checks on invalid pixel format in
[23:42:37] <CIA-11> ffmpeg: av_image_fill_linesizes/pointers.