[FFmpeg-devel-irc] IRC log for 2011-01-28

[irc at mansr.com]

[00:05:20] <CIA-38> ffmpeg: Justin Ruggles <justin.ruggles at gmail.com> master * r37cb3eb534 ffmpeg/libavcodec/iirfilter.c:
[00:05:20] <CIA-38> ffmpeg: Add special case for 2nd-order IIR filter.
[00:05:20] <CIA-38> ffmpeg: 40% faster in ff_iir_filter_flt() when c->order == 2.
[00:07:06] <CIA-38> ffmpeg: Janne Grunau <janne-ffmpeg at jannau.net> master * r795ed278e6 ffmpeg/libavformat/movenc.c:
[00:07:06] <CIA-38> ffmpeg: movenc: byteswap codec_tag in mov_write_ms_tag
[00:07:06] <CIA-38> ffmpeg: based on Alex Converse's "Fix ADPCM MS in mov muxing" patch
[00:07:17] <CIA-38> ffmpeg: John Stebbins <stebbins at jetheaddev.com> master * r97b04f5ed3 ffmpeg/libavformat/mov.c:
[00:07:17] <CIA-38> ffmpeg: mov: add support for little-endian utf16 chapter names
[00:07:17] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[00:07:18] <CIA-38> ffmpeg: Baptiste Coudurier <baptiste.coudurier at gmail.com> master * rf258964217 ffmpeg/libavformat/ (mov.c movenc.c):
[00:07:18] <CIA-38> ffmpeg: In mov muxer, mux adpcm_ms and adpcm_ima_wav the way quicktime supports it
[00:07:18] <CIA-38> ffmpeg: In mov demuxer, set adpcm_ms and adpcm_ima_wav frame size to stsd samples per packet.
[00:07:18] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[00:22:18] <CIA-38> ffmpeg: Justin Ruggles <justin.ruggles at gmail.com> master * rb5ec638343 ffmpeg/libavcodec/ (aacdec.c ac3dec.c dca.c nellymoserdec.c wmadec.c): cosmetics: indentation and spacing
[00:22:29] <CIA-38> ffmpeg: Justin Ruggles <justin.ruggles at gmail.com> master * r9d06d7bce3 ffmpeg/libavcodec/ (12 files): Remove the add bias hack for the C version of DSPContext.float_to_int16_*().
[00:22:55] <mru> hmm, cia got the order wrong
[00:24:22] <ruggles> i've seen that on the ML too when the email timestamps are identical
[00:24:35] <mru> timestamps are not the issue
[00:24:45] <mru> emails sometimes simply get reordered along the way
[00:25:17] <ruggles> ah
[00:26:32] <mru> nothing in the mail transport looks at timestamps at least
[00:27:49] <mru> every server has some kind of queue system, which doesn't need to be a strict fifo
[00:29:08] <saste> good night people
[00:29:16] <jannau> good night saste
[00:32:27] <mru> argh, libvpx is still slightly faster on two clips on omap4
[00:33:20] <mru> ffvp8 is faster on all my clips on omap3
[00:33:23] <jannau> but using both cores instead of just one?
[00:33:42] <mru> single-threaded
[00:34:10] <mru> at least that's what I asked for
[00:34:37] <mru> and numbers are consistent between omap3 and omap4 in that regard
[00:35:17] <mru> but I'm running out of functions to optimise
[00:35:34] <mru> I'll probably have to write some of the mb parsing in asm
[00:35:43] <mru> make that some more
[00:40:39] <Dark_Shikari> maybe the asm is better optimized for a9 as opposed to a8?
[00:40:43] <Dark_Shikari> check top to make sure it's only using one core
[00:41:08] <mru> it's roughtly twice as fast on 1GHz a9 as 600MHz a8
[00:41:23] <mru> same ratios for ffmpeg
[00:41:37] <mru> I think the difference is in non-asm code
[00:42:04] <mru> the a9 neon unit is in some ways slower than a8
[00:42:23] <Dark_Shikari> well, iirc the ffvp8 code is rather better structured at least
[00:42:26] <Dark_Shikari> in terms of memory layout, etc
[00:42:29] <Dark_Shikari> with stuff like the ringbuffer
[00:43:18] <Dark_Shikari> btw, structural improvements are welcome as well
[00:43:27] <Dark_Shikari> did you do neon versions of the idct_dc functions?
[00:43:31] <Dark_Shikari> i.e. the uv4 and all that
[00:43:35] <mru> yes
[00:44:20] <mru> it's all in my git repo if you're curious
[00:44:32] <Dark_Shikari> ok
[00:45:34] <mru> I'm going to mess some more with the rac functions and see what happens
[00:45:56] <Dark_Shikari> BBB:
[00:45:58] <Dark_Shikari>                 if (copy) {
[00:45:58] <Dark_Shikari>                     * (uint32_t *) (ptr+4*x)               = * (uint32_t *) (copy_dst + 12);
[00:46:01] <Dark_Shikari>                     * (uint32_t *) (ptr+4*x+s->linesize)   = * (uint32_t *) (copy_dst + 20);
[00:46:05] <Dark_Shikari>                     * (uint32_t *) (ptr+4*x+s->linesize*2) = * (uint32_t *) (copy_dst + 28);
[00:46:08] <mru> perhaps write the entire decode_mb_coeffs in asm
[00:46:08] <Dark_Shikari>                     * (uint32_t *) (ptr+4*x+s->linesize*3) = * (uint32_t *) (copy_dst + 36);
[00:46:12] <Dark_Shikari>                 }
[00:46:14] <Dark_Shikari> Stop with your aliasing ivolations please
[00:46:17] <Dark_Shikari> and go fix that
[00:46:26] <BBB> they are aligned!
[00:46:28] <BBB> isn't it ok then?
[00:46:31] <mru> no
[00:46:40] <mru> that's not what aliasing means
[00:46:40] <Dark_Shikari> No
[00:46:44] <Dark_Shikari> it's not ok
[00:46:45] <Dark_Shikari> ever
[00:46:48] <Dark_Shikari> *(x*) is not ok
[00:46:50] <mru> use the AV_RN*A macros
[00:46:56] <mru> and WN
[00:46:59] <Dark_Shikari> and please inline if(copy) into the EMULATED_EDGE section
[00:47:03] <mru> or COPY
[00:47:04] <Dark_Shikari> don't leave it hanging outside
[00:47:18] <Dark_Shikari> I don't want it slowing down my non-stupid code
[00:47:37] <Dark_Shikari> if ((y == 0 || x == 3) && mb_y == 0 && avctx->flags & CODEC_FLAG_EMU_EDGE) { topright = tr_top;
[00:47:40] <Dark_Shikari> merge the EMU_EDGE ifs please
[00:48:05] <Dark_Shikari> and while you're at it, add a comment or two to explain your mess
[00:58:52] <Dark_Shikari> BBB: you going to do that or should I
[01:28:30] <Dark_Shikari> BBB: ?
[01:39:51] <BBB> Dark_Shikari: sorry, dinner
[01:39:54] <BBB> Dark_Shikari: will fix
[01:39:58] <BBB> Dark_Shikari: what shall I use instead?
[01:40:06] <BBB> Dark_Shikari: I thought it was fine if I used alignment
[01:40:09] <BBB> sorry if it wasn't
[01:40:15] <BBB> is AV_[WR]N32A ok?
[01:40:41] <mru> there are COPY macros too
[01:40:45] <mru> they'll save you some typing
[01:49:21] <CIA-38> ffmpeg: Diego Elio Pettenò <flameeyes at gmail.com> master * re628864033 ffmpeg/libavformat/libavformat.v:
[01:49:21] <CIA-38> ffmpeg: Hide demuxers', muxers' and protocols' objects via the ld version script.
[01:49:21] <CIA-38> ffmpeg: This reduces the symbols exported by libavformat from 699 to 451.
[01:49:21] <CIA-38> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[01:49:30] <CIA-38> ffmpeg: Diego Elio Pettenò <flameeyes at gmail.com> master * rac28ce5fac ffmpeg/libavcodec/libavcodec.v:
[01:49:30] <CIA-38> ffmpeg: Hide the now-prefixed decoders, encoders, parsers, bsf, hwaccel objects.
[01:49:30] <CIA-38> ffmpeg: This significantly reduces the size of the symbol table in the generated ELF
[01:49:30] <CIA-38> ffmpeg: shared object (as well as the other linked tables).
[01:49:31] <CIA-38> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[01:57:48] <BBB> mru: ok, will use those
[02:18:42] <j0sh> is nb_rtsp_streams always 1 under tcp? (reviewing lu_zero's patch)
[02:19:29] <Dark_Shikari> BBB: do what I said
[02:19:31] <Dark_Shikari> get rid of all type puns
[02:19:39] <Dark_Shikari> merge if(copy) into the emu edge section
[02:19:44] <Dark_Shikari> get rid of the redundant emu_edge checks
[02:21:37] <lu_zero> shouldn't
[02:39:42] * Compn wonders what carl is so angry at
[02:39:43] <Compn> lol
[03:08:55] <BBB> Dark_Shikari: yes, I will - have a visitor right now but will send a patch tonight
[03:22:35] <Dark_Shikari> ok
[03:51:42] <j0sh> http://imgur.com/jacoj
[04:33:34] <peloverde> to flame or not to flame: that is the question
[04:35:35] <saintdev> fffire extinguisher
[04:43:35] <pross-au> so thats what the other f stands for
[04:43:48] <pross-au> LS
[05:05:18] <in3xes> anyone has specification sheet for Tivo format?
[05:06:56] <jai> read demux_ty.c from mplayer's source tree
[05:08:07] <in3xes> No other way? I mean, anything official?
[05:08:12] <jai> no
[05:08:33] <Sean_McG> Tivo doesn't want to write themselves out of business.
[05:09:55] <in3xes> reading this may be easier http://goo.gl/ulfLb
[05:11:07] <jai> i don't see anything useful on that page
[05:11:08] <in3xes> or not?
[05:11:18] <in3xes> source
[05:11:26] <jai> good luck with that then
[05:11:36] <in3xes> I was just asking
[05:31:21] <Compn> in3xes : btw there are devs to help reading mplayer code, if you need that
[05:31:35] <Compn> but as for specs, its a private company like apple, no specs
[05:31:59] <Compn> and even if there were specs (like on2) , the code is the spec in many cases
[05:32:13] <Compn> bwahahahaha
[05:32:26] <Compn> now you see what ffmpeg and mplayer had to put up with all of these years
[05:32:37] <jai> reimar said he would help anyone who wanted to port it, i guess you'll have to ask on the ML
[05:32:43] <Compn> in3xes : btw there might be something written up on http://wiki.multimedia.cx but i doubt it
[05:32:56] <Compn> there were some tivo projects too, maybe one of them has better docs
[05:33:19] <in3xes> I went to that page because of I was scared of mplayer codebase
[05:33:43] <jai> 23:06:40          @jai | read demux_ty.c from mplayer's source tree
[05:33:48] <jai> i stand by what i said :)
[05:34:02] <Compn> people still have tivos ?
[05:34:03] <Compn> :P
[05:34:03] <jai> it's probably the easiest way to port it
[05:34:04] <in3xes> I figured that long before
[05:34:07] * Compn trooooolls
[05:34:12] * Compn sleeps
[05:34:14] <Compn> night
[05:34:29] <Compn> now to find the comfy spot under the bridge...
[05:35:56] <in3xes> Compn: Thanks for the info
[06:02:42] <thresh> moroning
[06:27:58] <siretart> good moroning!
[06:28:10] <thresh> who's Gabu?
[06:28:30] * thresh is young and doesnt know the flamers
[06:28:42] <siretart> thresh: you really don't want to know
[06:29:24] <thresh> siretart: haha :)
[06:29:45] <thresh> it mustve been good I was ignoring mplayer
[06:29:51] <peloverde> I kind of lost it at Arpi
[06:30:03] <siretart> KotH: thank you very much for your very clear statement. I found it very enlighting!
[06:30:03] * kshishkov thinks thresh might be somehow related to VLC
[06:30:42] <thresh> kshishkov: yes, the rule of ignoring mplayer is one of the first ones in our Figh^WVideoLAN club
[06:32:11] <kshishkov> thresh: lucky you. There were at least three ex-MPlayer guys nobody have heard about for a long time coming here.
[06:32:31] <kshishkov> thresh: feels a bit like undead rising
[06:32:34] <thresh> well I remember arpi
[06:37:16] <kshishkov> as an active developer?
[06:38:54] <thresh> hey I'm not in my 40s to remember *that*
[06:41:29] <peloverde> Is there some sort of attribute that can be used to trick gcc about get_bits() range?
[07:17:34] * elenril yawns
[07:18:19] <_av500_> gm
[07:20:49] <elenril> yay, more fflames
[07:22:24] <saintdev> elenril: you should fffan the ffflames
[07:24:04] <elenril> what's up with all the ghosts from the past
[07:30:50] <Tjoppen> god morgon
[07:46:22] * elenril reads about KotH hating us all and Dark_Shikari feeding the trolls :/
[07:46:43] * wbs doesn't dare start reading the mails from last night yet
[07:48:11] <elenril> Sean_McG: chapters?
[07:53:39] <spaam> wbs: awww
[08:12:25] <peloverde> Can we make an mplayer-grumpy-old men list and reserve FFmpeg-devel for FFmpeg development?
[08:13:24] <DonDiego> go ahead and suggest it
[08:14:01] <DonDiego> btw, i woke up with a mailbox full of flames this morning, i don't intend to read it for now - anything important i need to respond to?
[08:19:17] <superdump> perhaps if you haven't read KotH's mail you wish to read that
[08:20:50] <peloverde> The MPlayer emeriti have got to go
[08:24:45] <kshishkov> that all makes me want to switch to VLC finally. Or at least to mplayer2(uau)
[08:25:04] <spaam> kshishkov: not mplayerG2 ?
[08:25:45] <kshishkov> spaam: it's like GNU Hurd but even less developed
[08:26:07] * elenril wonders where do those trolls keep coming from
[08:26:25] <spaam> kshishkov: haha :)
[08:26:37] <kshishkov> elenril: mostly Hungary. And MPlayer graveyard
[08:27:36] <elenril> just ban them all to nine hells :/
[08:27:56] <elenril> peloverde: don't bother feeding them
[08:28:41] <MultimediaMike> Sort of a troll reunion ; makes me sentimental
[08:28:52] <elenril> wow, Mike is here
[08:28:55] <elenril> hi MultimediaMike
[08:28:58] <superdump> wow, we have mr melanson too? :)
[08:29:00] <superdump> hey MultimediaMike
[08:29:09] <kshishkov> MultimediaMike: and I've just started working on Xxan decoder again :(
[08:29:49] <superdump> MultimediaMike: did you guys at adobe locate that memcpy instead of memmove yet? :)
[08:30:26] <DonDiego> peloverde: if you feel that MPlayer issues have to stay out of this, say it
[08:30:50] <Dark_Shikari> holy crap it's MultimediaMike
[08:31:01] <Dark_Shikari> michaelni AND MultimediaMike
[08:31:04] <DonDiego> MultimediaMike: mike? hey! what are you up to...
[08:31:04] <Dark_Shikari> the mike pair
[08:31:33] <DonDiego> Dark_Shikari: michael is not mike, although we always use the english instead of the german pronounciation, dunno why really...
[08:31:38] <kshishkov> now all we need is Michel from France who's also not active these days
[08:33:20] <peloverde> DonDiego: MPlayer is not FFmpeg. MPlayer developers who happen to develop FFmpeg are welcome. Former MPlayer developers who are not current FFmpeg contributors are not welcome to participate in organizational discussion. They are welcome to do code review or contribute new code.
[08:33:58] <kshishkov> peloverde: only the latter - look at MPlayer code and tell how well it was reviewed
[08:34:51] <DonDiego> peloverde: i absolutely agree, but you have to tell it to those people, not to me, please *do* express your feelings
[08:36:19] <MultimediaMike> Me? I just decided to put an IRC client on my smartphone and keep upper with the drama
[08:36:32] <Dark_Shikari> COME BACK AND DO ACTUAL WORK MIKE
[08:36:35] <Dark_Shikari> and stop whoring all day at adobe
[08:36:48] <Dark_Shikari> also you could apply for a job at gaikai, all the cool kids are doing that
[08:37:01] <MultimediaMike> First time I have seen IRC in 2 years
[08:37:04] <kshishkov> Dark_Shikari: he's not a kid anymore
[08:37:09] <Dark_Shikari> Sure he is.
[08:37:15] <peloverde> DonDiego: I said that to Berczi Gabor and I said that to Arpi, and I even had time to make two contributions in the process
[08:37:18] <MultimediaMike> Sure I am
[08:37:52] <kshishkov> MultimediaMike: where's the beef^Wcode?
[08:38:29] <MultimediaMike> Xan will totally be ready any day now
[08:38:47] <kshishkov> really?
[08:38:50] <MultimediaMike> Right after I finish the 669 loader
[08:39:05] <MultimediaMike> Which is an unrelated inside joke
[08:40:19] <kshishkov> isn't that some module format?
[08:40:23] <MultimediaMike> Maybe I should fix up ffvp8enc so that it beats x264
[08:40:25] <pross-au> what happened to the protracker guy?
[08:40:49] <DonDiego> peloverde: ok, sorry, i haven't read all the mails yet...
[08:40:54] <kshishkov> MultimediaMike: sorry, it's going to be xvp8 when BBB is finally ready to do it
[08:41:01] <spaam> pross-au: avseq?
[08:41:02] <MultimediaMike> Yeah 669 is a mod format
[08:41:10] <kshishkov> pross-au: it all got sweeped out by libavsynth, I fear
[08:41:32] <pross-au> who working on that kshishkov?
[08:41:49] <kshishkov> pross-au: BastyCDGS of course!
[08:42:17] <pross-au> still?
[08:42:41] <kshishkov> we still have some time till the heat death of th Universe
[08:43:44] <spaam> pross-au: do you wanna join his CDGS group with kirank and kshishkov ?
[08:44:08] <kshishkov> spaam: and you
[08:44:26] <spaam> i forgot that :)
[08:44:37] <pross-au> im already in it
[08:44:55] <spaam> good :D
[08:45:34] <kshishkov> pross-au: Bink first!
[08:47:01] <pJok> god morgon
[08:47:42] <kshishkov> goda morgnar, pJok
[08:48:01] <pJok> first mni on irc and then mike
[08:48:35] <pJok> hell froze over, pigs are flying... has world peace also come?
[08:48:50] <kshishkov> pJok: it's the other way round
[08:49:02] <spaam> maybe Döffinger will join irc soon also ;D
[08:49:04] <MultimediaMike> I was on Irc first; I just didn't stick around
[08:49:41] <pJok> mike, what i mean is, you've not been here for two years and you came back... mni has (to my knowledge) not been in here before
[08:50:18] * kshishkov remembers those times - single #ffmpeg channel, mostly everybody asking only how to recode his file into Fl*sh
[08:50:36] <Dark_Shikari> it's stil flash
[08:50:37] <Dark_Shikari> lol
[08:50:58] <kshishkov> but there's $channelname to, not only user channel
[08:51:02] <kshishkov> *too
[08:51:06] <MultimediaMike> "Join IRC or be stripped of your leadership " is a powerful motivator
[08:51:37] <kshishkov> MultimediaMike: sorry, you're not a leader anyway
[08:51:48] <Dark_Shikari> yes, ever since he was eaten by adobe
[08:51:59] <MultimediaMike> Not referring to me
[08:52:18] <MultimediaMike> I lead nothing
[08:52:36] <kshishkov> obviously, for you it should have something to do with game or QuickTime codecs
[08:52:37] <spaam> kshishkov: you can be the leader and force everyone to speak swedish and drink trocadero :)
[08:52:49] <pJok> mike, i thought you were leading a path to destruction (or at least Linux Flash support)
[08:53:03] <kshishkov> spaam: what? that would leave less trocadero to me! So I refuse
[08:53:18] <pJok> spaam, tried vira blåtira btw?
[08:53:56] <pJok> servus __av500__
[08:54:02] <pJok> or not
[08:54:21] <spaam> pJok: i dont like that one so much ;S
[08:54:52] <kshishkov> spaam: Guldus? Cuba cola? 21? Haiwa? Siddni?
[08:55:03] <pJok> spaam, i can't say i like it, but i don't dislike it either... its just odd that you can't seem to buy it anywhere in skåne... other than rosa pantern drugstore in ängelholm
[08:55:22] <kshishkov> spaam: and yes, Fruktsoda from Vasa has similar colour
[08:55:47] <pJok> i can usually only find Three Hearts soda from Halmstad everywhere
[08:55:55] <spaam> kshishkov: never heard about those .. only cuba cola
[08:56:14] <kshishkov> spaam: http://www.vasabryggeri.nu/
[08:56:20] <pJok> spaam, i think all those sodas are norrlans soda
[08:56:55] <spaam> ah yeah
[08:58:43] <KotH> salut MultimediaMike!
[08:58:43] <MultimediaMike> Is this really Arpi? http://www.facebook.com/people/Arpad-Gereoffy/100001021857658
[08:58:51] <KotH> MultimediaMike: how is it going in the new world?
[08:59:13] <KotH> MultimediaMike: yes
[09:00:01] <MultimediaMike> Koth- Attila?
[09:00:09] <KotH> juup, the one and only :)
[09:00:15] <DonDiego> MultimediaMike: http://linsze.hu/pic/arpi1.jpg
[09:00:24] <av500> MultimediaMike: hi Mike!
[09:00:28] <MultimediaMike> I missed you too
[09:00:48] <KotH> MultimediaMike: are you comming to LT this year?
[09:00:54] <KotH> MultimediaMike: i promise to come as well :)
[09:01:17] <MultimediaMike> Didn't plan too
[09:01:26] <KotH> :-/
[09:01:26] <MultimediaMike> Europe is so far
[09:01:44] <spaam> MultimediaMike: move ? :)
[09:01:45] <KotH> but we're at the center of the world! :)
[09:02:11] <MultimediaMike> Would love to come to fosdem
[09:02:22] <MultimediaMike> I like Brussels
[09:02:28] * KotH would like to as well... and beat the shit out of some people
[09:02:49] <KotH> mru: you should show MultimediaMike the picture of brussels we took :)
[09:02:56] <kshishkov> KotH: and do you know what country is at the centre of Europe?
[09:03:10] <KotH> kshishkov: switzerland ;-)
[09:03:17] <kshishkov> KotH: nope, Ukraine
[09:03:20] <MultimediaMike> So weird - is all this drama really bothering you people?
[09:03:42] <KotH> MultimediaMike: well... there has been a lot more going on than just on the mailinglist
[09:03:44] <kshishkov> KotH: there's even special monument standing at the geographical centre of Europe - made by Austrians
[09:03:46] <spaam> kshishkov: ukraine? i thought it was sweden.. ;/
[09:03:50] <KotH> MultimediaMike: and yes, it's bothering me
[09:04:00] <KotH> kshishkov: lol
[09:04:02] <kshishkov> spaam: that's centre of civilised world
[09:04:38] <kshishkov> KotH: that's completely true. And as you know half of Ukraine was under Austro-Hungarian rule.
[09:04:50] <kshishkov> it's mostly annoying
[09:05:09] <kshishkov> why not do something productive instead of flaming?
[09:05:34] <MultimediaMike> Being on one of the most reviled software teams on the planet, this drama looks ....small, I guess
[09:06:10] <superdump> that reminds me
[09:06:18] <Dark_Shikari> kshishkov: ffmpeg is all about people who do no real work
[09:06:25] <Dark_Shikari> michael (flames and reviews, no actual code)
[09:06:31] <bcoudu> ...
[09:06:33] <Dark_Shikari> me (bitches about things sucking, not much actual code)
[09:06:35] <av500> MultimediaMike: wrt reviled, when will flash be able to handle overlays?
[09:06:36] <kshishkov> Dark_Shikari: wasn't that #ffmpeg-devel ?
[09:06:37] <bcoudu> are we talking commit count ?
[09:06:38] <Dark_Shikari> mike (done nothing for 5 years)
[09:06:45] <Dark_Shikari> etc etc
[09:06:46] <Dark_Shikari> ;)
[09:06:51] * Dark_Shikari is semi-trolling
[09:07:10] <kshishkov> bcoudu: nope, just code and its functions
[09:07:11] <superdump> MultimediaMike: why don't you have feature and performance parity on Linux with Windows? and why hasn't the source code been released?
[09:07:56] <DonDiego> superdump: you never heard what the code looks like?
[09:08:01] <KotH> Dark_Shikari: and what would you say about me? lurking in the shadows and stabbing peoples backs?
[09:08:26] <superdump> DonDiego: i have :)
[09:08:47] <MultimediaMike> Overlays: glad you asked
[09:08:51] <av500> DonDiego: I can confirm
[09:08:52] <MultimediaMike> http://labs.adobe.com/technologies/flashplayer10/stagevideo.html
[09:09:59] <av500> MultimediaMike: and for that mobile OS from that search company?
[09:10:15] <MultimediaMike> When did I last do something on ffmpeg? Must have been those Theora optimizations a year ago
[09:10:57] <MultimediaMike> Unless you count ffvp8enc, which I wouldn't
[09:11:07] <kshishkov> MultimediaMike: http://wiki.multimedia.cx/index.php?title=Category:Undiscovered_Game_Formats
[09:11:16] <bcoudu> kshishkov, that should be about the same
[09:11:21] <Dark_Shikari> MultimediaMike: the theora code still sucks
[09:11:22] <Dark_Shikari> :>
[09:11:38] <kshishkov> Dark_Shikari: that's by codec design
[09:11:41] <MultimediaMike> Av500: what about them?
[09:11:54] <Dark_Shikari> kshishkov: no, the code sucks too
[09:11:57] <kshishkov> bcoudu: not exactly
[09:11:58] <Dark_Shikari> it does 4x 8x8 MC on 16x16 blocks
[09:12:02] <Dark_Shikari> even though 8x8 blocks are extremely rare
[09:12:32] <MultimediaMike> True, that was on my todo list
[09:12:57] <MultimediaMike> I thought dconrad would handle that
[09:12:58] <bcoudu> kshishkov, numbers ?
[09:13:10] <Dark_Shikari> MultimediaMike: apple ate him
[09:13:23] <MultimediaMike> He won't be handling anything related to ffmpeg
[09:13:44] <kshishkov> bcoudu: just the fact cosmetics and documentation-only commits exist should be enough
[09:14:34] <superdump> MultimediaMike: yeah but that's only for 32-bit so i'm using square
[09:15:25] <bcoudu> documentation counts IMHO
[09:16:15] <cartman> moin
[09:16:56] <kshishkov> grüß dich
[09:17:24] <cartman> nice discussion over ml
[09:21:00] <spaam> kshishkov: http://talk.newagtalk.com/forums/thread-view.asp?tid=165984&mid=1190655#M1190655  "Locomotive Engine Failure - Blown Piston"
[09:22:58] <kshishkov> spaam: well, it's rather ugly diesellok
[09:23:11] <spaam> yes. :)
[09:23:21] * kshishkov likes IORE
[09:28:15] <DonDiego> bye
[09:40:36] <j-b> ok, so you've reached DA point
[09:40:42] <j-b> what about conciliation now?
[09:41:16] <lu_zero> j-b: from the ffmpeg point of view everything is fine
[09:41:57] <j-b> lu_zero: you mean th code?
[09:41:58] <lu_zero> and trolls are just that
[09:41:58] <lu_zero> and I'm sick of discussing
[09:42:25] <bcoudu> ....
[09:42:51] <j-b> well, maybe you, lu_zero, should help people find a compromise?
[09:43:24] <MultimediaMike> J-b: is VLC Prepared to reconcile with Apple?
[09:43:30] <lu_zero> j-b: trying to find compromises is something I did in the past months...
[09:43:34] * kshishkov wonders: who knew that Niederösterreich had such an idiotic flag
[09:43:35] <j-b> MultimediaMike: oh, yes
[09:43:45] <j-b> MultimediaMike: I've mailed them again yesterday
[09:43:52] <lu_zero> so now I'm a bit discouraged on that part
[09:43:57] <j-b> MultimediaMike: I even mailed steve@ twice
[09:44:44] <MultimediaMike> Now you'll have to try timcook@ now
[09:45:01] <j-b> MultimediaMike: yeah, and that's a bit bad...
[09:45:21] <av500> j-b: asking for appl to change the appstore for gpl?
[09:45:35] <j-b> not for the gpl
[09:45:42] <j-b> just clarify one term
[09:45:47] <j-b> that they already apply in fact
[09:45:58] <j-b> but they don't want to write it down
[09:46:11] <jannau> it's of course not fine, we are losing developers but fear every mail no matter what's in it will accelerate the process
[09:47:34] <jannau> attila sugested that we (including michael) meet at fosdem
[09:47:49] <j-b> wow
[09:47:57] <j-b> are you making sense now?
[09:47:58] <j-b> no way
[09:48:21] <lu_zero> j-b: I know it won't work.
[09:48:28] <j-b> I disagree
[09:48:35] <Kovensky> talking about vlc / osx, how do I get rid of the floating player control =p </userquestion>
[09:48:47] <superdump> well, i'll be at fosdem
[09:48:52] <{V}> only flamebait arpi seems to have seen my attempt at finding a compromise :-/
[09:48:55] <lu_zero> brb
[09:48:56] <superdump> i guess michaelni won't be at fosdem
[09:49:20] <j-b> Kovensky: no fucking idea... there is a reason why I am rewriting this interface. Did you try the view something like cmd+m ?
[09:49:38] <superdump> bcoudu: will you be at fosdem?
[09:49:40] <Kovensky> oh, you're rewriting the cocoa interface? cool
[09:49:42] <kshishkov> superdump: what makes you think so?
[09:49:54] <j-b> Kovensky: to be a bit more up to date, yes
[09:50:25] <Kovensky> Cmd+m just minimizes ._.
[09:50:32] * Kovensky never used that shortcut before, just doesn't minimize windows
[09:50:38] <superdump> kshishkov: as far as i know, michaelni has not said he's going and perhaps it is a bit late to organise travel now, though perhaps not
[09:51:03] <bcoudu> superdump, nope
[09:51:14] <j-b> superdump: late? you are kidding...
[09:51:23] <kshishkov> superdump: has he ever went to some public event?
[09:51:40] <superdump> bcoudu: shame, it's been quite a while since we last met
[09:51:46] <j-b> he has to.
[09:55:32] <superdump> michaelni: is there any chance you could go to fosdem?
[10:03:06] <Tjoppen> libvorbis depends on libogg? that's.. interesting
[10:03:32] <j-b> I would have used another adjective
[10:49:41] <elenril> Dark_Shikari: srsly, don't feed the iive ;)
[10:52:26] <kshishkov> elenril: with all his faults he's at least FFmpeg dev
[10:54:52] <elenril> the last time he was active was the deathmatch in 2009 =p
[10:55:21] <elenril> who was it against btw?
[10:55:26] * elenril can't remember anymore
[10:55:27] <kshishkov> Diego
[10:55:31] <elenril> right
[10:55:48] <elenril> those were fun times
[10:55:55] <elenril> we should do it again
[10:57:50] <thresh> next week is ok for me to get enough pop-corn
[10:58:15] <lu_zero> sure
[10:58:23] <lu_zero> but probably it won't be that fun
[11:00:34] <j-b> thresh: can you bring some for FOSDEM? </sarcasm>
[11:01:04] <lu_zero> j-b: there is already plenty of beer
[11:01:20] <j-b> I'll live with beers
[11:02:26] * av500 prepares to shower j-b with french beer
[11:02:36] <j-b> I'd welcome that
[11:03:57] <cartman> cd ..
[11:04:05] <cartman> someone cd please :O
[11:04:32] <elenril> bash: ..: no such file or directory
[11:05:13] <{V}> elenril, good one :)
[11:05:37] <mru> morning
[11:06:03] <cartman> char* strerror(int inErrno)
[11:06:03] <cartman> {
[11:06:03] <cartman>   return "Unknown Error";
[11:06:04] <cartman> }
[11:06:07] <cartman> lol
[11:07:25] <lu_zero> cartman: O_o
[11:08:21] <jannau> cartman: at least you can't claim it's wrong
[11:08:33] <cartman> jannau: true
[11:39:59] <CIA-38> ffmpeg: Alex Converse <alex.converse at gmail.com> master * r5ce5dbc5f3 ffmpeg/libavcodec/ (dsputil.c dsputil.h):
[11:39:59] <CIA-38> ffmpeg: Make ff_float_to_int16*_c() static.
[11:39:59] <CIA-38> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[11:48:16] <wm4> what's the current state of the ffdrama? it seems everyone is fine with separated git repositories?
[11:48:26] <kshishkov> but one person
[11:49:00] <kshishkov> personally I'm going to keep ignoring it
[11:49:01] <mru> and hungarian trolls keep pouring out from under the bridges
[11:50:19] <mru> good choice
[11:50:41] <j-b> wm4: one git repository will die eventually
[11:50:57] <mru> right now one of them is clearly leading
[11:51:13] <kshishkov> mru: and another one is illegally taking over
[11:51:57] <j-b> define "legal"
[11:52:15] <wm4> there was also the issue who owns the ffmpeg trademark and domain...
[11:52:20] <mru> j-b: approved by MN
[11:52:24] <kshishkov> j-b: by "Quod Dici" law
[11:52:32] <mru> wm4: fabrice owns the ffmpeg name and domain
[11:52:33] <j-b> wm4: Fabrice
[11:52:38] <j-b> wm4: fortunately
[11:52:45] <j-b> since this is the biggest asset
[11:52:57] <wm4> yeah, I mean, did he put out a statement yet?
[11:53:06] <mru> no
[11:53:07] <j-b> he is too clever to put a statement now
[11:53:15] <wm4> heh
[11:53:30] <wm4> then I take it the whole situation is still in limbo
[11:53:52] <j-b> wm4: however, I know a good popcorn reseller </sarcasm>
[11:53:53] <mru> there are a bunch of trolls flaming away while the rest of us do useful work
[11:56:55] <jannau> you mean cosmetics, bikesheds and reverts which are still good enough to be ported over
[11:58:32] <mru> dammit, still slower than libvpx on one clip
[11:59:04] <mru> not that it really matters
[11:59:08] <mru> both are >130fps
[11:59:20] <kshishkov> oprofile to the help?
[11:59:32] <mru> oprofile isn't working too well on omap4
[11:59:34] <mru> yet
[11:59:39] <mru> I need to shout at ti more
[11:59:47] <lu_zero> perf is any better?
[11:59:52] <mru> no
[11:59:57] <kshishkov> mru: subcontract professional
[12:00:06] <mru> lu_zero: the interrupt configuration is missing
[12:00:22] <mru> so neither perf or oprofile can get any samples
[12:00:34] <lu_zero> I see
[12:00:55] <{V}> I don't think Fabrice is going to make any statement anytime soon. Has he even sent 1 email in the last year to any mailinglist (qemu, ffmpeg, tcc,..)
[12:01:09] <mru> not ffmpeg
[12:01:12] <mru> can't say for the others
[12:03:40] * lu_zero plays a bit with ruby-elf
[12:04:37] <jannau> gmane find mails on qemu lists from march 2010
[12:06:06] <cartman> rename it ffsmpeg
[12:06:50] <kshishkov> FFmedia?
[12:07:02] <kierank> FFohgod
[12:07:04] <cartman> For F*ck's Sake MPEG
[12:07:17] <mru> kierank: FFomg?
[12:07:24] <mru> wtFF
[12:07:25] <kierank> yes
[12:07:40] <kshishkov> but av500 thinks that the most important codec there is Bink, not MPEG-[124]
[12:08:20] <kierank> the most important codec is lego audio
[12:08:45] <kshishkov> indeed
[12:08:56] <kshishkov> and there's libavseq
[12:14:59] <mru> jannau: what were the magic headers patchwork reacts to?
[12:15:35] <kshishkov> mru: subject=open sesame
[12:17:25] <{V}> jannau, thanks. Not quite as long ago as I thought.
[12:21:28] <jannau> mru: X-Patchwork-State: Accepted
[12:21:55] <jannau> or Rejected or Changes Requested
[12:22:12] <jannau> the state is case sensitive
[12:22:47] * mru cooks up some emacs bindings for that
[12:24:09] <superdump> ffsmeg
[12:24:25] <kierank> ffspam
[12:24:42] <superdump> mru: you use emacs? interesting
[12:24:49] <superdump> for some reason i would have thought you would have used vim
[12:24:59] <mru> vim doesn't read email
[12:25:35] * mru was counting brackets before he could count his toes
[12:25:48] <thresh> I've heard it has that cool feature other software misses: editing
[12:25:53] <jannau> {V}: that were just 2 single mails with a pause of 1 1/2 years to the next
[12:28:19] <{V}> jannau, yes, I did the search on gmane myself after you mentioned march 2010. Even in one of the march 2010 messages he says he no longer follows qemu development
[12:29:20] <jannau> superdump: I use vim for composing mails and emacs for sources
[12:29:44] <superdump> i use vim solely for text editing
[12:30:12] <jannau> {V}: ah, I just looked at the dates
[12:30:23] <superdump> gvim actually (mostly because i couldn't get inter-process c+p working using vim in a terminal)
[12:31:23] <elenril> workfineforme in terminal
[12:32:02] <superdump> it used to
[12:32:12] <superdump> it's a user error somewhere, i just can't find it
[12:33:08] <elenril> you just need to use the X register
[12:33:41] <elenril> e.g. "+p to paste from X clipboard
[12:33:56] <mru> I like how emacs can open frames on multiple displays
[12:34:24] <mru> so can I move freely between desktop and laptop and keep on editing the same files / reading the same mail
[12:34:54] <thresh> what, they even reinvented screen(1)?
[12:35:14] <mru> ever heard of X?
[12:35:18] <mru> as in X11
[12:35:33] <thresh> who needs that to read mail or edit files?
[12:35:38] <elenril> why do you need X for editing text =p
[12:35:56] <thresh> counter-troll, aha
[12:36:06] <mru> I like it and that's all you need to know
[12:36:31] <mru> besides, a little bit of lisp is fun from time to time
[12:36:43] <wm4> I'm using kate, I guess that makes me totally uncool
[12:37:09] <KotH> wm4: you should date her instead ;)
[12:37:17] <mru> indeed
[12:37:22] <mru> women don't like being used
[12:37:42] <wm4> sorry I don't like it if women are implemented in C++
[12:37:45] <kshishkov> mru: because it's their responsibility?
[13:27:57] <Dark_Shikari> http://i.imgur.com/XqSzl.jpg
[13:30:26] <kierank_> ffmpeg leader criteria
[13:31:08] <kierank> lol what a surprise, it's military
[13:33:55] <mru> jannau: the patchwork header doesn't seem to work
[13:38:49] <Zor> Dark_Shikari: that is some amazing typesetting right there
[13:39:14] <jannau> mru: I'll look into it
[13:51:08] <Dark_Shikari> BBB: the intra predict section might be cleaner if you just create two copies of intra_predict
[13:51:11] <Dark_Shikari> with and without emuedge
[13:51:47] <Dark_Shikari> almost none of the code would be duplicated
[13:51:48] <Dark_Shikari> like 10%
[13:51:54] <Dark_Shikari> it'd be faster too
[13:54:57] <iive> sounds good to me.
[14:09:03] <Tjoppen> http://en.wikipedia.org/wiki/Inter_frame  pretty fancy
[14:32:34] <cartman> Where is Rich Felker guy, he must have smell the flames by now
[14:33:35] <mru> :)
[14:33:44] <lu_zero> probably he's wiser today
[14:34:31] <elenril> he lives in #mplayerdev permanently, ignoring us here
[14:35:42] <cartman> elenril: he is alive?!
[14:35:43] <cartman> damn
[14:35:51] <andoma> i remember him!
[14:36:03] <cartman> andoma: you bet you do
[14:36:05] <andoma> .. and endless mail threads
[14:36:06] <cartman> wise-ass Felker
[14:36:30] <cartman> even Bill Gates probably wrote more code compared to him
[14:36:31] <mru> the 3rd, don't forget that
[14:37:07] <cartman> fortunaly first 2 didn't make it :P
[14:41:28] <elenril> how about somebody warns/bans the trolls
[14:41:30] <elenril> from the ML
[14:41:44] <ohsix> then theres the naming problem
[14:41:49] <cartman> elenril: I believe in one thing
[14:41:53] <cartman> ignore them and eventually they die
[14:41:54] <cartman> :P
[14:42:21] <elenril> apparrently some people can't resist the temptation to respond
[14:42:32] <cartman> I cancelled twice in two days :P
[14:47:38] <BBB> Dark_Shikari: you mean intra_predict()? probably
[14:47:45] <BBB> Dark_Shikari: grew that way because I thought it'd share more
[14:47:52] <BBB> Dark_Shikari: first let me fix the aliasing issues
[14:47:57] <BBB> I have the branch back to master now
[14:48:04] <BBB> was working on porting patches from svn into branches
[14:51:08] <elenril> srsly, now it's just дурак - сам дурак level
[14:52:06] <kshishkov> elenril: so the first who quits it wins
[14:52:23] <BBB> Dark_Shikari: email sent
[14:52:36] <BBB> Dark_Shikari: will look into having two versions of intra_predict() as you suggested in a separate patch
[14:53:04] <iive> elenril: trolls live longer than humans ;)
[14:54:11] <elenril> i'm starting to really like the idea of moving the trolls to a separate ML
[14:56:02] <{V}> elenril, under the guise of "non-development discussions go here" ?
[14:56:43] <elenril> under the guise of "trolling not welcome on the dev ML" =p
[14:57:35] <cartman> iive: you should be ashamed of yourself for trolling :P
[14:58:26] <elenril> cartman: this is what happens to Cthulhu when the Great Old Ones return
[14:59:01] <cartman> Arpi & Gabu always trolled for their whole life
[14:59:57] <elenril> Ph'nglui mglw'nafh michaelni R'lyeh wgah'nagl fhtagn
[15:00:10] <CIA-38> ffmpeg: Luca Barbato <lu_zero at gentoo.org> master * ra8475bbdb6 ffmpeg/libavformat/ (8 files):
[15:00:10] <CIA-38> ffmpeg: os: replace select with poll
[15:00:10] <CIA-38> ffmpeg: Select has limitations on the fd values it could accept and silently
[15:00:10] <CIA-38> ffmpeg: breaks when it is reached.
[15:00:16] <CIA-38> ffmpeg: Luca Barbato <lu_zero at gentoo.org> master * rf81c7ac70a ffmpeg/libavformat/ (rtsp.c rtspdec.c):
[15:00:16] <CIA-38> ffmpeg: rtsp: make ff_sdp_parse return value forwarded
[15:00:16] <CIA-38> ffmpeg: the sdp demuxer did not forward it at all while the rtsp demuxer assumed
[15:00:16] <CIA-38> ffmpeg: a single kind of error
[15:06:19] <uau> kshishkov: btw there are better reasons to switch to mplayer2 than the morons who were associated with the older version
[15:06:21] <uau> like proper pause handling (can do things while the player stays paused), better matroska support including ordered chapters, improved VDPAU support, no use of internal ffmpeg symbols, exact non-keyframe-limited seeks, etc
[15:06:50] <kshishkov> uau: well, they created it
[15:07:10] <uau> yes they created it, but they had zero clue how to maintain software of the size they created
[15:07:49] <cartman> uau: there is mplayer2?!
[15:08:16] <uau> cartman: the git tree
[15:08:26] <cartman> damn I missed it
[15:08:32] <cartman> uau: maintained by you?
[15:08:34] <kshishkov> uau: you mean that glue between hacked libs?
[15:08:55] <uau> cartman: there's no release as mplayer2 yet (though should be soon)
[15:09:05] <cartman> uau: just give me the git :D
[15:09:09] <siretart> google already reports about 659.000 results for "mplayer2" :-)
[15:09:19] <cartman> I thought your branch was for i18n work only
[15:09:35] <elenril> siretart: lol
[15:09:47] <elenril> cartman: http://repo.or.cz/w/mplayer.git
[15:09:56] <cartman> elenril: cheers :)
[15:10:06] <uau> better use mplayer-build.git in most cases
[15:10:18] <elenril> yeah, was just going to suggest that
[15:10:19] <elenril> http://repo.or.cz/w/mplayer-build.git
[15:10:47] <uau> cartman: btw i think i already mentioned the git when talking about the translations before
[15:11:04] <cartman> uau: yeah I thought it was for i18n changes only
[15:11:40] <siretart> uau: is there a windows port of mplayer2?
[15:11:47] <cartman> this is what happens when you don't blog about these cool things :P
[15:11:58] <siretart> I see a lot of reports in google complaining about issues in MPLAYER2.EXE
[15:12:40] <cartman> siretart: yes thats the windows classic media player
[15:12:54] <uau> siretart: kovensky has done some windows builds - his last public build is from almost a year ago, i know he's worked on things since then, but i'm not sure about the current state
[15:13:10] <wm4> way too many <something>mplayer<something> out there...
[15:13:22] <Kovensky> 12:11.58 siretart: I see a lot of reports in google complaining about issues in MPLAYER2.EXE <-- that's microsoft's ancient windows media player 6.something
[15:13:28] <siretart> obviously my trolling against the name 'mplayer2' was too subtle...
[15:13:49] <elenril> siretart: we're using to more obvious trolls these days
[15:13:51] <uau> siretart: would you prefer just "Mplayer 2" then?
[15:14:09] <superdump> no
[15:14:15] <superdump> mplayer-uau is more unique :)
[15:14:22] <superdump> mplayer2 is a bad idea though
[15:14:27] <Kovensky> mprayer
[15:14:30] <microchip_> or uplayer ;)
[15:14:31] <superdump> too much noise in google results
[15:14:31] <wm4> uau: will you use a different binary filename too?
[15:14:39] <cartman> uplayer sounds nice
[15:14:46] <wm4> nouplayer
[15:14:49] <uau> superdump: i might not maintain it forever though...
[15:15:06] * siretart also likes uplayer. or nuplayer
[15:15:11] <superdump> uplayer already exists
[15:15:16] <microchip_> grrr
[15:15:19] <microchip_> fuck em
[15:15:30] <iive> nuplayer is really nice.
[15:15:35] <lu_zero> uplayer was a nice name
[15:15:40] <iive> it sounds like new player
[15:15:42] <lu_zero> eplayer maybe?
[15:15:44] <superdump> iPlayer
[15:15:46] <elenril> νplayer?
[15:15:56] <wm4> "mplayerx"?
[15:15:57] <uau> wm4: i don't intend to change filename immediately at least, to keep things compatible with GUIs
[15:15:57] * elenril stabs superdump 
[15:16:12] <siretart> iive: that's what I had in mind :-)
[15:16:25] <kierank2> New mplayer
[15:16:25] <microchip_> µplayer :D
[15:16:27] <kierank2> like New Labour
[15:16:31] <lu_zero> I'd keep it mplayer till it's a drop-in replacement though
[15:16:42] <iive> wm4: there is already mplayer-xp
[15:16:45] <wm4> uau: could provide a compatibility symlink as well
[15:16:47] <siretart> uau: which also means that "your" mplayer cannot be installed together with the existing mplayer package
[15:17:24] <wm4> mplayer-uau lol
[15:17:26] <uau> lu_zero: hmm? did you mean something else? that sentence doesn't make sense to me
[15:17:26] <lu_zero> siretart: if it's a drop-in replacement eselect/alternate would take care of it?
[15:17:35] <jai> iive: nick doesn't really maintain it iirc
[15:17:47] <iive> jai: still, it exists.
[15:17:51] <lu_zero> I'd keep the name mplayer for the executable
[15:18:30] <lu_zero> at least until you change it's interface (command line/ command pipe protocol)
[15:18:33] <siretart> hm, yes, playing with alternatives is an option. I just don't like alternatives in general, because it is a system wide setting and I'm kinda used to multi-user systems here
[15:18:36] <uau> lu_zero: that "till" didn't make sense IMO - use a different name if it's NOT?
[15:18:53] <uau> i guess you meant "while" then
[15:18:56] <superdump> anyway, this is discussion for #mplayerdev i guess
[15:18:59] <elenril> siretart: time to extend alternatives?
[15:19:06] <lu_zero> uau: I guess so
[15:19:07] <siretart> oh, indeed. let's move this discussion
[15:19:07] <elenril> to be usable by normal users
[15:19:11] <uau> superdump: there's already #mplayer2-dev
[15:19:17] * lu_zero runs
[15:19:19] <lu_zero> bbl
[15:19:21] <superdump> as you please
[15:20:21] <Kovensky> unrelated fact: "uau" means "wow" in portuguese (probably japanese too)
[15:20:24] <CIA-38> ffmpeg: Ronald S. Bultje <rsbultje at gmail.com> master * r9d4bdcb714 ffmpeg/libavcodec/vp8.c:
[15:20:24] <CIA-38> ffmpeg: Fix VP8 aliasing problems.
[15:20:24] <CIA-38> ffmpeg: Replace * (uint32_t *) buf accesses with AV_WN32A/AV_COPY32.
[15:25:12] <pJok> gotta love avisyth sometimes: Stream #0.1: Audio: pcm_f32le, 48000 Hz, 11873 channels, 1057058 kb/s
[15:25:45] <pJok> and i didn't know that ffmpeg supported 11873 channels
[15:26:18] <spaam> nice bitrate
[15:26:49] <pJok> indeed
[15:26:51] <microchip_> pJok: must have been designed for aliens with gazillion ears :p
[15:27:14] <pJok> it is supposed to be a silent audiotrack
[15:27:45] <mmu_man> then you'll get lots of silences :)
[15:41:46] <kierank> anybody know anything about igmp
[15:42:01] <kshishkov> google does
[15:42:51] <kierank> of course
[15:42:55] <cartman> it pings
[15:43:40] <cartman> j-b: vlc on Galaxy Tab looks good
[15:45:55] <cartman> uau: on OSX with mplayer-build.git: --cpu=host not supported with compiler gcc
[15:46:27] <uau> cartman: ancient OS X gcc i assume
[15:46:34] <cartman> uau: gcc 4.2.1
[15:46:37] <Compn> cartman : got latest xcode ?
[15:46:49] <cartman> sure --cpu=core2 works fine
[15:47:21] <cartman> I bet it doesn't extract host correctly
[15:47:40] <uau> cartman: mplayer-build uses ffmpeg's build system
[15:48:06] <jannau> cartman: I think -march=native was introduced in 4.3
[15:48:06] <uau> and ffmpeg does not have hacks to determine cpu with compilers that lack support
[15:48:27] <Kovensky> cartman: I workaround it by commenting out the --cpu=host line on the ffmpeg-config script
[15:48:28] <uau> i haven't added custom code to work around that
[15:48:51] <uau> cartman: add --cpu=core2 or whatever in ffmpeg_options
[15:48:54] <jannau> and that's that ffmpeg uses to extract the host cpu
[15:49:18] <Compn> does fate have an osx box ?>
[15:49:31] <Compn> or just mac-intel / ppc-linux
[15:49:38] <cartman> okies :)
[15:49:52] <uau> Compn: ffmpeg will compile if you leave out cpu-specific optimization completely
[15:50:22] <uau> so fate would likely work
[15:50:27] <Kovensky> I do have plenty of CPU time for running fate on this computer but I can't guarantee uptime or long-term stability =p
[15:50:44] <Kovensky> (hackintosh)
[15:51:05] <Kovensky> (CPU is a mostly unused i5 750, except when gaming (zomg))
[15:52:27] <Compn> uau : oh, these arent the default thing breaking ?
[15:52:38] <Compn> anyways, a roundup issue might be worth it
[15:52:54] <Kovensky> not really
[15:53:06] <Kovensky> there are only three "sane" behaviors
[15:53:11] <Kovensky> one is to upgrade your compiler
[15:53:23] <Kovensky> the other is to whine if you give the flag that will break on the compiler until you remove
[15:53:26] <uau> Compn: no - it's the ffmpeg "--cpu=host" switch erroring out if the compiler lacks support
[15:53:30] <Kovensky> the other is to whine about the flag but compile it anyway
[15:53:39] <Kovensky> the latter is clearly bordering on insane
[15:53:50] * cartman has gcc 4.6 trunk too
[15:53:57] <cartman> maybe I should just use clang ;)
[15:54:02] <Kovensky> it could also just not whine and ignore the --cpu=host flag, but I don't think that's standard practice =p
[15:54:18] * Kovensky wonders how reliable is clang++ nowadays
[15:54:28] <Kovensky> it can compile ffms2 too, but I didn't do any extensive testing
[15:54:32] <Compn> uau : so shouldnt there be a compiler test for that switch ?
[15:54:40] <Kovensky> (1.7 couldn't, newer revisions can)
[15:54:45] <Compn> checking to see if gcc is dumb .... yes
[15:55:19] <Kovensky> Compn: if there isn't a compiler test for that switch, then it'll just put -march=native on the CFLAGS, fail every subsequent test, and if it does manage to finish configuring, make will instafail
[15:55:52] <uau> Compn: you mean a separate test in the build repo scripts? it doesn't do such testing now, and doing two levels of tests could get ugly
[15:56:16] <Kovensky> uau: nah, he was thinking it was a ffmpeg bug
[15:56:27] <Compn> yes, in ffmpeg i mean
[15:56:32] <Compn> not your script
[15:56:52] <uau> Compn: i think it tests now and gives the error above if the compiler fails
[15:56:55] <Compn> ah
[15:56:58] <Compn> fair enough
[15:57:52] <Kovensky> so, does anyone want me to run fate on this could-be-unreliable-since-I-need-to-reboot-once-in-a-while-to-ressuscitate-one-of-my-old-HDDs-so-I-can-rescue-data-from-it-but-I-try-to-keep-uptime hackintosh? or do you have enough osx boxes?
[15:58:09] <cartman> configure.ac:33: error: possibly undefined macro: AC_DEFINE
[15:58:12] <cartman> fsck
[15:58:22] <Kovensky> cartman: msys?
[15:58:29] <cartman> Kovensky: OSX :P
[15:58:38] <Kovensky> lol
[15:58:41] <Compn> Kovensky : a cross-compile mingw would be better, since ramiro has left
[15:58:43] <Kovensky> I still didn't hit autotools issues here ._.
[15:58:50] <uau> cartman: missing some autotools stuff required by libass probably
[15:58:53] <Kovensky> Compn: eventually :)
[15:58:58] <cartman> uau: seems so *sniff*
[15:59:05] <Kovensky> Compn: http://github.com/Kovensky/AutoMinGW <-- gonna go back to working on it this weekend
[15:59:12] <jannau> Kovensky: x86 darwin gcc is missing
[15:59:13] <Kovensky> ramiro ragequit?
[15:59:24] <Kovensky> jannau: I'm on x86-64, but that can be fixed with an -m32
[15:59:33] <Compn> Kovensky : read the autobuilds site
[15:59:43] <jannau> Kovensky: I meant x86*
[15:59:51] <uau> cartman: btw most of the compiling discussion would probably be more appropriate for #mplayer2 / #mplayer2-dev
[16:00:11] <cartman> uau: I like keeping this channel off topic :D
[16:00:21] <jannau> but x86_32 might be also a good idea
[16:00:48] <cartman> uau: libass depends on pkg-config bleh
[16:01:12] <uau> you'll need pkg-config anyway
[16:01:50] <uau> at least to pass required flags from ffmpeg to static link the player (doing that sanely without pkg-config just wouldn't work...)
[16:02:18] * cartman cp's pkg.m4
[16:02:23] <Kovensky> jannau: how large are the vectors
[16:03:12] <jannau> ~400M currently
[16:05:06] <Kovensky> jannau: are there any setup instructions?
[16:08:07] <cartman> and it compiles :)
[16:08:29] <jannau> Kovensky: http://wiki.multimedia.cx/index.php?title=FATE
[16:14:42] * Kovensky leaves the rsync job running and goes to the last day of his internship
[16:38:28] <mru> lu_zero: ping
[16:50:08] <mru> the poll patch broke dos and os/2 builds
[16:52:08] <CIA-38> ffmpeg: Stefano Sabatini <stefano.sabatini-lala at poste.it> master * r3e5bc7ff6a ffmpeg/libavfilter/vf_setpts.c:
[16:52:08] <CIA-38> ffmpeg: In the start_frame() debug log, print the reference pos value rather than the evaluated value converted to int.
[16:52:08] <CIA-38> ffmpeg: That's required because -1 is evaluated as NAN, which converted back
[16:52:08] <CIA-38> ffmpeg: to int looks like a random number, this is especially annoying when
[16:52:09] <CIA-38> ffmpeg: debugging sources with undefined pos (as the video4linux2 device).
[16:52:09] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[16:52:10] <CIA-38> ffmpeg: Stefano Sabatini <stefano.sabatini-lala at poste.it> master * r3c802cabba ffmpeg/libavcodec/rawdec.c:
[16:52:10] <CIA-38> ffmpeg: In the rawvideo decoder, set pkt_pts in the output frame.
[16:52:11] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[16:52:11] <CIA-38> ffmpeg: Stefano Sabatini <stefano.sabatini-lala at poste.it> master * r85466e1e5f ffmpeg/doc/ (ffmpeg.texi ffplay.texi ffprobe.texi muxers.texi):
[16:52:12] <mru> ah, looks like a missing update in os_support.h
[16:52:12] <CIA-38> ffmpeg: Add muxers.texi file.
[16:52:12] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[16:52:21] <CIA-38> ffmpeg: Georgi Chorbadzhiyski <gf at unixsol.org> master * ra7827a17c6 ffmpeg/libavformat/mpegtsenc.c:
[16:52:21] <CIA-38> ffmpeg: In mpegts "reserved_future_use" field must be set to 1 in SDT table
[16:52:21] <CIA-38> ffmpeg: According to EN 300 468 section 3.1 (Definitions):
[16:52:21] <CIA-38> ffmpeg:  Unless otherwise specified within the present document all
[16:52:21] <CIA-38> ffmpeg:  "reserved_future_use" bits is set to "1".
[16:52:21] <CIA-38> ffmpeg: This was not the case for SDT generation so this patch fixes it.
[16:52:22] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[17:01:25] <CIA-38> ffmpeg: Alex Converse <alex.converse at gmail.com> master * re5c82df80e ffmpeg/libavcodec/aacdec.c:
[17:01:25] <CIA-38> ffmpeg: aacdec: Convert some loop copies into memcpy()s.
[17:01:25] <CIA-38> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[17:03:13] <CIA-38> ffmpeg: Mans Rullgard <mans at mansr.com> master * r79dca23dc2 ffmpeg/tests/ref/lavf/ts:
[17:03:13] <CIA-38> ffmpeg: Update mpegts test reference
[17:03:13] <CIA-38> ffmpeg: The output was changed by a7827a17c6b3388322350456d445c94b3a82cd25.
[17:03:13] <CIA-38> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[17:04:26] <CIA-38> ffmpeg: Stefano Sabatini <stefano.sabatini-lala at poste.it> master * r4fc9ff0ad6 ffmpeg/libavformat/img2.c:
[17:04:26] <CIA-38> ffmpeg: Make the image2 demuxer log more verbose
[17:04:26] <CIA-38> ffmpeg: Add an error message in case the user requests to write more than one file
[17:04:26] <CIA-38> ffmpeg: and the path does not contain a "%d" or "%0Nd" pattern.
[17:04:26] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[17:21:34] <jannau> Dark_Shikari: ping
[17:24:03] <CIA-38> ffmpeg: Mans Rullgard <mans at mansr.com> master * r362d8f7d9e ffmpeg/libavformat/ (os_support.c os_support.h): (log message trimmed)
[17:24:03] <CIA-38> ffmpeg: os_support: make poll() fallbacks conditional on CONFIG_NETWORK
[17:24:03] <CIA-38> ffmpeg: poll() is only used by networking code, so the fallback should
[17:24:03] <CIA-38> ffmpeg: only be built if networking is enabled. Also remove CONFIG_FFSERVER
[17:24:03] <CIA-38> ffmpeg: condition from the declarations.
[17:24:04] <CIA-38> ffmpeg: This should fix building on systems without poll(), broken
[17:24:05] <CIA-38> ffmpeg: by a8475bbdb64e638bd8161df9647876fd23f8a29a.
[17:25:23] <BBB> saste: ping
[17:27:18] <saste> BBB: pong
[17:27:32] <BBB> \o/
[17:53:13] <elenril> o_0 37 new mails in two hours?
[17:53:27] * elenril guesses it's all hugs and puppies
[17:53:49] <jannau> I think it's mostly review and patch mails
[17:54:07] <elenril> on ffmpeg-devel? inconceivable! ;)
[17:54:17] <elenril> michaelni: ping
[17:54:26] <mru> our new strategy: drown the trolls in patches
[17:54:43] <elenril> i like it
[17:54:47] <jannau> and get on gmane.org's top ten list
[17:55:06] <elenril> too bad i don't see any easy metadata work :)
[17:55:18] <mru> then do some metawork
[17:55:55] * elenril has to prepare for QFT exam anyway
[17:56:28] <mru> quantum flux trolling?
[17:56:50] * jannau doesn't trust the top ten list btw. http://dir.gmane.org/gmane.comp.hardware.beagleboard.general is on
[17:57:27] <mru> there is some correlation to reality
[17:57:30] <mru> but it's weak
[17:57:38] <elenril> quantum field theory
[17:57:43] <elenril> the evilest of all evils
[17:58:03] <mru> iKnow
[17:58:13] <elenril> (and a giant hack too)
[17:58:17] <mru> that's it's evil, I don't know it
[17:58:42] <mru> physics is nought but a hack
[17:59:13] <mmu_man> yeah who's flooding my mailbox ?
[18:39:46] <BBB> kshishkov: what was the review of the rm-multirate patch again? I've ported it to HEAD and it still works
[18:40:00] <BBB> kshishkov: If you review I'll try to find someone to implement the missing features so we can apply it
[18:53:21] <Compn> BBB : does seeking work with it ?
[18:53:34] <Compn> is this mulitrate demuxer or multirate over rtsp ?
[18:53:43] * Compn not paying attention
[19:01:14] <skal>  *yawn*
[19:01:19] <skal> good mourning
[19:01:36] <mru> hi skal
[19:02:01] <skal> hey mru
[19:03:41] <kierank> pffft centos has next to no packages :/
[19:04:00] <kshishkov> thresh: ^^^ see how civilised people feel about time before noon
[19:04:05] <mru> kierank: isn't that point?
[19:04:08] <kshishkov> BBB: I'll try to review it
[19:04:11] <lu_zero> hi skal
[19:04:22] <mru> it's a wannabe enterprise distro
[19:04:26] <kierank> mru: dunno
[19:04:32] <mru> and enterprise apps take over the entire machine anyway
[19:04:41] <mru> no point installing anything separately
[19:05:43] <lu_zero> is anybody on darwin/macosx?
[19:05:50] <lu_zero> j0sh: ?
[19:05:57] <kierank> I think we should have a daily ffmpeg drama newsletter
[19:06:14] <kierank> executive summary of the goings on
[19:07:47] <elenril> kierank: http://i.imgur.com/eSmOS.jpg
[19:07:54] <elenril> here's your executive summary
[19:07:59] <kshishkov> BBB: where is it?
[19:11:29] <mmu_man> elenril eh, and it's trollday today :)
[19:12:01] * {V} hoped it was friday
[19:12:37] <thresh> kshishkov: you mean they already killed all the morons and now mourn them?
[19:13:08] <kshishkov> thresh: could be (but not in Russia ever)
[19:13:37] <mru> thresh: that's mouroning
[19:15:10] <kshishkov> mru: you have shotgun, please fix this world
[19:15:23] * mru curses gcc
[19:15:34] <kshishkov> it's cursed already
[19:15:48] <kshishkov> and probably not only because of Stallman's bite
[19:17:00] <iive> try to ncurse it
[19:17:37] <lu_zero> mru: what's wrong now?
[19:18:08] <{V}> iive, adding dependencies to gcc does not seem like a good idea to me :p
[19:18:29] <mru> lu_zero: I wrote some inline asm for the vp56_rac functions
[19:18:36] <mru> now if I remove said asm again, it crashes
[19:18:59] <mru> building same code with armcc is fine
[19:19:30] <mru> and works on x86
[19:21:04] <lu_zero> O_o?
[19:21:22] <lu_zero> strange
[19:31:12] <BBB> kshishkov: I'll send it tonight
[19:31:13] <BBB> brb
[19:31:27] <kshishkov> ok
[19:35:37] <kierank> wow mike on irc
[19:35:54] <kierank> two michael's i've never seen on irc before
[19:39:36] <MultimediaMike> So you weren't here last night
[19:39:57] <Dark_Shikari> jannau: pong
[19:40:09] <mru> found the problem
[19:40:42] <MultimediaMike> You know, back in the day, most of the multimedia hackers were named Mike
[19:41:20] <kshishkov> back in what day?
[19:41:41] <MultimediaMike> First half of 2000s
[19:41:58] <Dark_Shikari> BBB: were you going to clean up the rest of that stuff?
[19:42:09] * kshishkov remember only Mike, Michael and Michel
[19:42:39] <lu_zero> hi MultimediaMike =)
[19:42:50] * mru wishes there were a michelle
[19:43:14] <MultimediaMike> On xine,  75% of the project leaders were named Mike
[19:43:28] <in3xes> If there are so many undisclosed formats how you guys managed to put them in ffmpeg? Reverse Engineering?
[19:43:41] * mru looks at MultimediaMike 
[19:43:52] <MultimediaMike> Mike,Miguel,  and Michael
[19:43:59] <MultimediaMike> Different Michael
[19:44:00] <{V}> lu_zero, vmrsss said the git.videolan version builds with --disable-filter=mp and as far as I know that almost the only difference between the two repos
[19:44:01] <elenril> mru: http://tvtropes.org/pmwiki/pmwiki.php/Main/ThereAreNoGirlsOnTheInternet
[19:44:10] <elenril> here, have a trope instead
[19:44:38] <kshishkov> MultimediaMike: for me Xine is just DVD player project
[19:44:47] <mru> elenril: I know a few girls on the internet
[19:44:53] <mru> ones I've met in real life
[19:44:59] <mru> so I know they're not something else
[19:45:03] <MultimediaMike> Yep, reverse engineer as necessary
[19:45:15] * mru prefers sideways engineering
[19:45:58] <MultimediaMike> A colleague always used to remind me that forward engineering was superior
[19:46:41] <kshishkov> but equally hard to find
[19:47:18] <iive> MultimediaMike: he may suffer from nih syndrome.
[19:48:08] <MultimediaMike> No, he just thought my RE exploits were silly
[19:48:28] <elenril> re is fun
[19:50:10] <MultimediaMike> And so educational
[19:50:21] <mru> reverse education
[19:51:15] <MultimediaMike> I contend that the fastest was to learn an ASM is to RE something
[19:51:21] <kshishkov> mru: judging from education people usually get it's an improvement
[19:51:51] <mru> MultimediaMike: is there any other way?
[19:51:56] <MultimediaMike> Trying to learn ARM ASM that way
[19:52:09] * mru learned z80 by reading hex dumps
[19:52:26] <mru> then editing the code in place to do what I wanted
[19:52:47] * kshishkov never had 8- or 16-bit machine
[19:52:50] <MultimediaMike> @mru read books that don't make a lot of sense out of context
[19:55:14] <MultimediaMike> I painstakingly learned x86 16-bit from books
[19:55:22] <MultimediaMike> Then forgot
[19:55:42] <mru> the only books on such things I've ever read are arch ref manuals
[19:55:42] * kshishkov has not managed to learn x86 asm from any of his two books
[19:55:49] <MultimediaMike> Then remembered quickly when starting. RE
[19:56:03] <ruggles> i had to learn SPARC asm in college. all i remember is that i hate m4.
[19:56:06] <kshishkov> mru: well, there are few geniuses (like you) out there
[19:57:06] <MultimediaMike> @ruggles. You had to program in m4?
[19:57:24] <ruggles> yeah. i guess the assembler had no preprocessor.
[19:57:33] <ruggles> that or the prof liked m4
[19:57:46] <mru> sparc as such isn't all too bad
[19:57:52] <mru> but they screwed up the 64-bit move a bit
[19:57:53] * kshishkov fears that anything written in m4 would turn into autotools script
[19:57:59] <MultimediaMike> I'M. relearning SH-4
[19:58:05] <mru> kshishkov: or a sendmail config file
[19:58:24] <kshishkov> mru: and which is worse?
[19:58:25] <mru> sh4 has a nice and small instruction set
[19:58:49] <mru> kshishkov: unknown, nobody has read both and lived
[19:58:50] <kshishkov> what SH-4 device can I buy for development?
[19:58:50] <MultimediaMike> I will one day make ffmpeg and fate run on my Dreamcast
[19:58:52] <lu_zero> kshishkov: autotools m4 isn't the worst m4 you would find around
[19:59:12] <kshishkov> lu_zero: why would I ever search for such thing?
[20:00:07] <lu_zero> kshishkov: not sure, sadism?
[20:00:09] <mru> MultimediaMike: encountered the ftrv instruction yet?
[20:00:11] <MultimediaMike> Kostya Dreamcast is cheap and easy to procure
[20:00:32] <kshishkov> MultimediaMike: I hope it's not bulky
[20:00:58] <MultimediaMike> But it's hard to program without the coder cable
[20:01:28] <MultimediaMike> DC is a very small console
[20:02:34] <kshishkov> hmm, a bit too slow for 720p H.264
[20:02:38] <mru> lu_zero: btw, os/2 is still failing
[20:02:41] <MultimediaMike> @mru. Yeah I used the ftrv before
[20:02:51] <lu_zero> ;_;
[20:03:20] <mru> lu_zero: seems like rtspenc.c isn't including os_support.h
[20:03:38] <MultimediaMike> DC SH-4 had some other custom instructions too
[20:04:01] <kshishkov> reminds me of MIPS with each chip with its own SIMD
[20:04:43] <MultimediaMike> Kostya indeed, the DC even struggled on basic MPEG 4 video
[20:05:05] <MultimediaMike> Console came out in 1999
[20:09:06] <MultimediaMike> KOSTYA:You should probably ask that dev of ffmpeg-devel which sh4 board he is using
[20:09:09] <mru> ruggles: I'll look at your fmtconvert patch in a while
[20:09:37] <ruggles> ok, thanks. i know it's huge... mostly it's just moving code around though.
[20:10:00] * j0sh is wiki'ing the dreamcast
[20:10:29] <j0sh> apparently the OS came on disk with each game... that's pretty cool
[20:10:57] <MultimediaMike> Omg you don't know the DC? :)
[20:11:18] <mru> disks, bah!  cartridges ftw
[20:11:21] <j0sh> well i know it, just not the guts
[20:11:29] <MultimediaMike> Minor footnote in gaming history by now, I guess
[20:11:49] <j0sh> def looks like it'd be fun to hack on tho
[20:12:26] <MultimediaMike> http://multimedia.cx/eggs/dreamcast-anniversary-programming/
[20:12:50] <MultimediaMike> A little post I did on that very topic
[20:15:13] <j0sh> ...must resist the temptation to ebay
[20:15:45] <j0sh> MultimediaMike: i got a crystalhd card because of the blogpost you made about one
[20:15:56] <j0sh> i have yet to use it, lol
[20:16:14] <MultimediaMike> What? Why would you do that?
[20:16:18] <MultimediaMike> :)
[20:16:23] * mru urges MultimediaMike to blog about something really expensive
[20:16:31] <MultimediaMike> I couldn't make it work
[20:17:09] <MultimediaMike> I don't know anyone who has made it work
[20:17:24] <MultimediaMike> I guess that's not true
[20:17:25] <j0sh> i have a pile of hardware in my "to hack with" pile that never goes down
[20:17:41] <MultimediaMike> Yeah, same
[20:18:17] <MultimediaMike> I still have a brand new vflash console,  for 2 years
[20:18:32] <mru> there are right now 9 arm boards on my desk
[20:18:39] <MultimediaMike> Nice
[20:18:56] <mru> that's the ones on my desk
[20:19:26] <MultimediaMike> My only programmable ARM device is again. The SC
[20:19:31] <MultimediaMike> DC
[20:19:50] <mru> get a beagleboard or something
[20:20:42] <MultimediaMike> I guess this nexus 1 I'm. Typing on right now might also count
[20:20:42] * j0sh has a beagleboard too
[20:21:03] <j0sh> i wanted to use it as a way to pick up neon
[20:21:29] <mru> I have 2k lines of vp8 neon asm coming soon
[20:21:34] <mru> you can review it :)
[20:21:54] <j0sh> nice, is it based off that vp8 neon code dump that came a few months ago
[20:22:56] <mru> yeah
[20:23:06] <mru> not many lines are unchanged though
[20:23:22] <j0sh> cool
[20:23:59] <jannau> Dark_Shikari: reping, Have you seen my reply to the h264 profiles? High 4:4:4 Intra and Predictive use the same profile_idc
[20:25:25] <Kovensky> heh, I wanna try my hand at neon too but no hardware
[20:25:42] <Kovensky> hopefully I can get my hands on some during my EE course
[20:26:06] <Dark_Shikari> jannau: and?
[20:26:22] <Kovensky> does qemu count? :)
[20:26:53] <_av500_> no
[20:27:34] <Dark_Shikari> mru: askdjflaksjdfkasjdlfkajsdlfjasldkfjas;ldkfja;lskdj;flaksdjf;laksjdf;laksjd;flakjsdkfjasdkfjsdlfkj
[20:27:43] <Dark_Shikari> intra predict in vp8 currently: ~2800 cycles
[20:27:55] <Dark_Shikari> templated, based on emu_edge: ~3600 cycles (WTF?)
[20:28:01] <Dark_Shikari> templated, but noinlined: ~2400 cycles
[20:28:03] <Dark_Shikari> GCC
[20:28:04] <Dark_Shikari> GCC
[20:28:06] <Dark_Shikari> DIE
[20:28:42] <_av500_> gcc rolls over and dies
[20:29:54] <jannau> Dark_Shikari: they look identical to the profile code and the patch is imho ok
[20:30:01] <Dark_Shikari> er...
[20:30:08] <Dark_Shikari> profile is not only determined by profile_idc
[20:30:14] <Dark_Shikari> profile_idc is a syntax element
[20:30:18] <Dark_Shikari> It is one of the things used to calculate profile
[20:30:31] <Dark_Shikari> It is the job of ffmpeg to Do The Right Thing and calculate the correct profile and export it.
[20:30:54] <jannau> it's currently the only thing that sets profile in avcodeccontext
[20:31:13] <Dark_Shikari> Then avccodeccontext should be fixed.
[20:31:35] <Dark_Shikari> er, the code should be
[20:31:39] <Dark_Shikari> if we're giong to export the profile we should do it correctly
[20:32:28] <jannau> ok
[20:33:00] <BBB> Dark_Shikari: you mean split intra_predict()?
[20:33:07] <BBB> Dark_Shikari: eventually, yes, but working on something else first
[20:33:41] <Dark_Shikari> ugh I can't bench anything
[20:33:44] <Dark_Shikari> everything is giving crazy results
[20:33:45] <BBB> but if you do it already :-p
[20:34:41] <_av500_> Dark_Shikari: btw, the h264 encoding that merbzt was bugging you about
[20:34:50] <_av500_> it was the decoder
[20:35:01] <_av500_> and its fixed now
[20:35:22] <Dark_Shikari> lol
[20:35:59] <_av500_> Dark_Shikari: today they delivered a bugfis for their mp3 decoder...
[20:36:09] <Dark_Shikari> lol
[20:36:13] <_av500_> yep
[20:36:15] <merbanan> _av500_: who ?
[20:36:22] <_av500_> merbanan: them
[20:36:30] <merbanan> ok
[20:42:04] <CIA-38> ffmpeg: Anton Khirnov <anton at khirnov.net> master * r042950542d ffmpeg/libavformat/asfdec.c:
[20:42:04] <CIA-38> ffmpeg: asfdec: ensure that the whole tag is read.
[20:42:04] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[20:42:15] <CIA-38> ffmpeg: Dave Yeo <daveryeo at telus.net> master * ra0788cc627 ffmpeg/libavformat/rtspenc.c:
[20:42:15] <CIA-38> ffmpeg: rtspenc: include os_support.h for system without HAVE_POLL_H
[20:42:15] <CIA-38> ffmpeg: fix compile on OS/2
[20:42:15] <CIA-38> ffmpeg: Signed-off-by: Janne Grunau <janne-ffmpeg at jannau.net>
[20:56:40] <Kovensky> lol, rsync apparently refuses timeouting
[20:56:52] <Kovensky> it still hadn't finished getting the vectors, I had to kill and restart it
[20:57:02] <Kovensky> it's now apparently going normally
[20:57:08] * Kovensky should have sshed sooner
[21:04:48] <soroushi> Hi, what is the behavior of swscale if the input video is interlaced and I ask for half width: 720 -> 288, will it just discard one field and return the other?
[21:05:11] <soroushi> err 576 -> 288
[21:05:55] <mru> did you mean height?
[21:06:07] <soroushi> yes, height
[21:06:43] <iive> soroushi: no
[21:06:58] <iive> you can use tfields and crop to do that
[21:07:54] <iive> ops, fil or il filters and crop
[21:08:07] <Vitor1001> BBB: your PPC patch breaks "make fate-vp8" in a previously working build
[21:08:10] <soroushi> I am using it in VLC and can set just mode parameter
[21:08:41] <iive> soroushi: if scale is instructed it would separate both fields, scale fields separately and then produce the resulted frame.
[21:09:24] <iive> this is the correct behaviour, because in real interlace frames it would cause color from one field bleeding into the other.
[21:09:28] <BBB> Vitor1001: so I heard, disregard it (I haven't applied it, have I?)
[21:09:40] <BBB> Vitor1001: I'd like to figure out what is broken right now though
[21:09:45] <BBB> if anyone can debug it that'd help a lot
[21:09:49] <Vitor1001> No, last time I saw it you were looking for someone to test
[21:10:04] <Vitor1001> Ask mru for a ppc account in his box
[21:10:08] <BBB> I sent an update patch that has brackets :)
[21:10:14] <BBB> mru says it's not broken on his box
[21:10:20] <BBB> so then I can't test that my patch fixes it...
[21:10:30] <Vitor1001> At least you can test if it doesn't break anything
[21:10:45] <soroushi> iive: I can't play with other than mode parameter with VLC, thanks anyway
[21:13:39] <iive> soroushi: hum, sorry, wrong channel, swscale won't do that automatically. you have to do it. bilinear filter doesn't discart pixels, nearest neighbour may do it.
[21:15:18] <Kovensky> so bilinear will effectively crappily deinterlace the source?
[21:15:56] <mru> fields should be scaled separately
[21:16:08] <mru> and care should be taken of chroma positions
[21:16:31] <soroushi> indeed, I use separatefields().selecteven().AssumeFrameBased() in avisynth and it give clear image and tried 10 mode of swscale and nearest neighbour gave the best image
[21:16:40] * Kovensky hopes h265 will finally kill interlacing
[21:17:13] <mru> didn't someone on this channel invent "russian interlacing" a few days ago?
[21:17:47] <Kovensky> thresh?
[21:17:54] <mru> or kshishkov
[21:17:54] <kierank3> it interlaces you somehow?
[21:18:17] <iive> soroushi: skipping half the fields is not good way to scale down height. It would create visible staircase artifact on diagonal edges.
[21:18:34] <iive> i mean, skipping half the lines.
[21:20:01] <soroushi> iive: my video do not many of them and I can live with it
[21:20:44] <soroushi> iive: I also have moving subtitle that discard field gives best quality
[21:21:14] <soroushi> and in swscale the best was nearest neighbour for that subtitle
[21:22:07] <iive> soroushi: dropping interlaced fields is OK.
[21:22:26] <iive> well, not really...
[21:22:37] <mru> better than dropping lines in general
[21:22:50] <mru> it shouldn't make aliasing any worse
[21:22:57] <jannau> Dark_Shikari: patches sent with proper annotation for high 4:4:4 predictive, constrained baseline and Intra profiles
[21:23:44] <mru> can we please stop writing "foo= bar"?
[21:23:50] <mru> and write "foo = bar" instead
[21:23:52] <soroushi> I will force upscaling later when video display and discard field and its nearest equivalent method in swscale gives better quality, especially for that subtitle
[21:24:08] <kierank3> mru: oh god yes finally
[21:24:23] <soroushi> iive: nearest equivalent(nearest neighbour) *
[21:25:09] <mru> kierank3: what happened to 1 and 2?
[21:25:14] <jannau> mru: I would gladly, just adjusted to neighbouring code
[21:25:31] <mru> jannau: there's sane code rigth beside it
[21:25:34] <mru> just do it
[21:25:49] <mru> if you like consistency, fix the ugly code in a separate commit
[21:26:26] <kierank> kierank1: :D
[21:26:50] <kierank1> hmmm i wonder why my other pc lost connection
[21:26:57] <mru> someone forgot to register his nick...
[21:27:18] <spaam> kierank is registered
[21:27:26] <kierank> kierank also :)
[21:29:10] <jannau> fixed locally. I won't change surrounding code now, michael will just complain over cosmetics
[21:29:27] <mru> let him
[21:29:38] <lu_zero> jannau: clean code is better code.
[21:29:50] <mru> this code is mostly maintained by Dark_Shikari and pengvado
[21:29:57] <mru> I'm sure they both prefer it clean
[21:30:51] <mmu_man_> so which git should I clone now ?
[21:30:58] <mmu_man_> </troll>
[21:31:02] <soroushi> speaking of upscaling, in some containers like flv and mp4 there is meta data that can be set to set different dimensions than real dimensions of video, onMetaData() in flv and tkhd in mp4 but players based on libavformat and libavcodec(ffplay, mplayer, vlc) do not obey meta dimensions
[21:31:12] <mru> mmu_man_: whichever works better for you
[21:31:15] <kierank> elenril ^
[21:31:49] <soroushi> flash obey dimensions set in onMetaData in flv but not in tkhd in mp4
[21:32:03] <mmu_man_> mru: <troll>like, the one I can push BeOS patches to ?</troll>
[21:32:08] <kierank> lol
[21:32:09] <soroushi> and quick time seems to obey dimensions set in tkhd
[21:32:13] <spaam> mru: i cant kill my comrade in CDGS.. the member count will drop... ;S
[21:32:18] <mru> mmu_man_: send patches and I'll review
[21:32:23] <mmu_man_> eh
[21:32:30] <lu_zero> mmu_man_: send patches and provide an haiku fate
[21:32:44] <mmu_man_> well I shall rewrite the a/v native input someday
[21:33:05] <lu_zero> it would have to be in C++ ?
[21:33:11] <mmu_man_> as for BeOS stuff I have a somehow working build, for that I can just set up one on github
[21:33:22] <mmu_man_> lu_zero: yes, the MediaKit API is C++
[21:33:41] <mru> mmu_man_: http://www.flickr.com/photos/av500/4767125825/
[21:33:43] <lu_zero> no plain C bindings?
[21:33:47] <mmu_man_> lu_zero: I don't believe the beosaudio.cpp ever caused any other platform trouble, did it?
[21:34:05] <mmu_man_> lu_zero: no, Haiku is a forward looking OS :P
[21:34:05] <lu_zero> mmu_man_: I have a patch for a C++ avdevice
[21:34:24] <lu_zero> mmu_man_: so I could use python ctypes bindings ^^;
[21:34:40] <mmu_man_> well we already use lavc and lavf in our MediaKit plugins btw
[21:34:50] <lu_zero> mmu_man_: jokes aside
[21:35:18] <mmu_man_> mru: Eternal curse on you all :P
[21:35:29] <mmu_man_> OMG, you did that on a mac
[21:35:33] <lu_zero> I the decklink patch added somehow the CXX bits I needed
[21:35:59] <mmu_man_> ROTFL a Mac with an FSFE fellowship sticker
[21:36:08] <lu_zero> and shouldn't be a problem
[21:36:40] <mru> if anything needs libstc++ or libgcc_eh there will be problems
[21:36:41] <mmu_man_> OMG, if all those phishing emails claiming my ISP ows me 89EUR were right I'd be rich
[21:37:11] <mmu_man_> why so ?
[21:37:26] <lu_zero> mru: if the base runtime is that one for everything, it shouldn't
[21:37:37] <mmu_man_> well the BeAPI doesn't use templates, std::* or exceptions, so...
[21:37:43] <lu_zero> mmu_man_: on linux linking to libstc++ is begging for troubles
[21:37:45] <mru> lu_zero: if something needs those, you need to link with g++
[21:38:01] <lu_zero> mru: I know =|
[21:38:37] <mru> and switching link command depending on some avdevice being on or off gets ugly quickly
[21:38:46] <mru> especially if building shared libs
[21:38:56] <mmu_man_> lu_zero: I suppose it's just because Linux freaks are afraid of it and thus do wrong things
[21:39:06] <mmu_man_> you always fear what you don't know
[21:39:32] <mru> yes, I don't know why, and it scares me
[21:40:00] <lu_zero> mmu_man_: more because libstdc++ by itself is ... icky
[21:40:16] <mmu_man_> I don't recall having much problem on BeOS with it
[21:40:45] <lu_zero> mmu_man_: sorry... as fellow beos user I recall gcc2 vs gcc3
[21:40:54] <mmu_man_> more probably because it was a separate lib in the OS and not left with gcc at the time
[21:41:03] <lu_zero> and haiku had it's share of fun with gcc3 vs gcc4
[21:41:09] <mmu_man_> lu_zero: that's !=
[21:41:13] <mmu_man_> it's about ABI changes
[21:41:35] <mmu_man_> mostly name mangling
[21:41:43] <lu_zero> mmu_man_: not just that
[21:42:03] <mmu_man_> vtable fmt and such
[21:42:12] <Compn> oh egypt pulled the bgp
[21:42:20] <mmu_man_> yeah
[21:42:27] <mru> Compn: border gateway plug?
[21:42:34] <lu_zero> as gentoo developer I had my share of pain because gcc-3 had some subtle changes not even versioned...
[21:42:35] <Compn> protocol
[21:42:38] <Compn> :)
[21:42:51] <mmu_man_> FDN (French Data Network) has set up free dialup access in France for them
[21:43:01] <Compn> the phones still work? :P
[21:43:04] <mmu_man_> http://www.fdn.fr/
[21:43:09] <Compn> and they called everyone ? :P
[21:43:21] <mmu_man_> I heard they were cut as well but I'm not sure
[21:43:31] <lu_zero> Compn: land phones so far should be fine
[21:43:32] <mmu_man_> at least they werent earlier when they set this up
[21:43:37] <lu_zero> cellphones got cut
[21:43:54] * lu_zero wonders about mesh-networking
[21:44:11] <Compn> does your isp have mesh networking lu_zero ?
[21:44:16] <mmu_man_> http://reflets.info/egypte-chronique-dun-massacre-pas-ordinaire-jan25-live/
[21:44:22] <Compn> e.g. just local isp customers to each other local-net
[21:44:26] <mmu_man_> some have set up a HAM radio link
[21:45:23] <mru> then it won't be long before we have SPAM radio
[21:45:33] <mmu_man_> :)))
[21:46:25] * Compn eats one can of spam per year
[21:46:28] <lu_zero> Compn: I have friends providing mesh solutions
[21:46:30] <Compn> you guys eat the spam ?
[21:46:38] <lu_zero> didn't
[21:46:41] <lu_zero> you did?
[21:47:16] <Compn> spiced pork and meat
[21:48:54] * elenril discovers some mpc7 in his collection
[21:49:05] <elenril> in wonder if we support metadata for it
[21:49:38] <mmu_man_> incidentally, I've cited shame.html several times to members of the french HADOPI, asking them why they didn't protect my copyright as they are supposed to >:-)
[21:50:34] <mmu_man_> they waste their time sending mails to kids doing P2P instead, that wouldn't have bought the disks anyway, and that aren't doing any harm to the moral rights of authors, actually helping them get advertisement...
[21:50:44] <Compn> did you get reply ?
[21:50:48] <mmu_man_> nope
[21:50:50] <mmu_man_> wonder why :)
[21:51:04] <Compn> also probably none of those in shame are in the hadopi jurisdiction :P
[21:51:07] <elenril> but but but PIRACY
[21:51:40] <elenril> won't anybody think of poor copyright lawyers
[21:51:41] <mmu_man_> Piracy is an incorrect word
[21:51:58] <elenril> arrr
[21:52:00] <mmu_man_> it's a semantic shift made to induce the idea that sharing is wrong
[21:52:03] <mmu_man_> anyway
[21:52:21] <mmu_man_> Copying is an act of love. Love is not subject to law. http://copyheart.org/
[21:52:25] <mmu_man_> :p
[21:52:35] <mmu_man_> ♡
[21:53:18] <mmu_man_> http://questioncopyright.org/minute_memes/copying_is_not_theft
[21:54:23] <mmu_man_> Compn: well they don't do P2P, but Orange certainly is in the french jurisdiction, so they could be tried if they really wanted to
[21:54:26] <elenril> i think you're preaching to the choir here ;)
[21:54:26] <mmu_man_> btu they don't care
[21:54:40] <mmu_man_> elenril: yeah it's just friday I guess :)
[21:54:58] <mmu_man_> still, go watch Copying is not theft, it's funny :)
[21:58:13] <elenril> michaelni: i case you feel like answering, i wanted to know about the playlist patch from last gsoc
[21:58:40] <elenril> i'm wondering what what happened to it
[21:59:30] <saste> elenril: it's still in soc
[21:59:48] <saste> elenril: nobody dared to touch it after gsoc
[22:00:05] <saste> elenril: i mean in the soc SVN repo
[22:00:34] <elenril> iirc the student went through several review rounds
[22:00:49] <elenril> then he sent one last version and no review came
[22:01:07] <saste> elenril: yes I remember the code was not bad
[22:01:18] <elenril> so i'm wondering if it's salvageable
[22:01:20] <saste> elenril: but then how much time passed since the last patch?
[22:02:29] <elenril> what difference does it make, we still don't have playlists ;)
[22:04:26] <lu_zero> elenril: care to rebase and put it in a branch/send it out?
[22:07:23] <mmu_man_> mru: btw, git: is a perfectly valid URI scheme, even though there is no RFC for it, and at least under Haiku it's fully supported by CheckItOut, so having <a href=""> would allow checking out in a click
[22:07:28] <mmu_man_> http://ffmpeg.org/download.html
[22:07:33] <mmu_man_> same for svn:
[22:08:12] <mmu_man_> though I think we could probably come up with something like maven's scm: canonical scheme as an RFC since http: can be used as well for git
[22:08:21] <mmu_man_> hmm for svn at least
[22:08:57] <jannau> http: shouldn't be used for git
[22:09:11] <mmu_man_> right, was thinking about svn
[22:09:15] <mru> only in a pinch
[22:09:22] <michaelni> elenril, IIRC i simply run out of time / motivation especially when it came to thinking about the architecture of it all, if someone else would do a round review on it i wouldnt be sad
[22:09:25] <mmu_man_> still, there is svn+ssh: for ex as well
[22:09:56] <mmu_man_> but svn can use http(s) too, which is usually wired back to the browser but well
[22:10:34] <elenril> michaelni: so it wasn't unusable horrible?
[22:10:38] <mmu_man_> oh, git: isn't listed on http://fr.wikipedia.org/wiki/Sch%C3%A9ma_d%27URI
[22:11:18] <mru> oh those scheming frenchmen...
[22:13:44] <microchip_> must be all those frogs
[22:14:07] <michaelni> elenril, i dont remember exactly but i think i didnt like the original architecture but after i complained it ended with a worse mix of 2 architectures IIRC  but its long ago so i might misremember
[22:15:51] * mmu_man_ doesn't eat frogs
[22:16:15] <mmu_man_> just like englishmen likely not always eat haggis
[22:16:21] <mmu_man_> scothish I mean
[22:16:45] <mru> scottish
[22:17:18] <kierank2> mmu_man_: you eat snails I assume
[22:17:38] <{V}> mmu_man, did you mention Orange specifically to those hadopi members you contacted?
[22:17:57] <mru> kierank2: snips and snails and puppydog tails
[22:18:48] <{V}> (not as just part of a list)
[22:18:59] * lu_zero eats frogs and snails from time to time
[22:20:45] <mmu_man_> kierank2: neither
[22:20:54] <mmu_man_> {V}: at least once yes
[22:20:57] <mmu_man_> but they just don't care
[22:21:09] <mmu_man_> I shall send a formal letter someday
[22:21:31] <mmu_man_> do we know which files they use ? need to make sure they use at least one I have © on
[22:21:47] <mmu_man_> let's see if the git HEAD still builds on OSX
[22:21:58] <mmu_man_> I have a pending change for the doc
[22:22:15] <mru> mmu_man_: fate says it builds
[22:22:19] <mmu_man_> mru: do we have FATE for OSX also ?
[22:22:20] <mmu_man_> ah ok
[22:22:27] <jannau> and even runs
[22:22:28] <mmu_man_> well do we build doc by default ?
[22:23:09] <jannau> I don't think so, at least not if the necessary tools are missing
[22:23:12] <mmu_man_> mru: oh IIRC it was due to texi2html installed from MacPort being too recent and rejecting an option
[22:23:26] <mmu_man_> -number specifically
[22:23:44] <mru> we fixed that
[22:24:55] <mmu_man_> there is still -number in the Makefile
[22:25:09] <mmu_man_> let's see
[22:26:44] <DonDiego> mmu_man_: i'm interested to hear if this is broken
[22:27:00] <DonDiego> unfortunately texi2html sucks :-/
[22:27:22] <mmu_man_> I suppose it does :p
[22:27:28] <mru> hmm, I thought we'd fixed that
[22:28:26] <mmu_man_> I recall sending a patch or something and it didn't get in
[22:28:33] <mmu_man_> or maybe here
[22:29:02] <DonDiego> mmu_man_: so what about the build?
[22:29:29] <jannau> there is no (High*) profile sample in fate-suite/h264-conformance?
[22:29:43] <mru> don't know
[22:29:50] <mru> grab one from ITU and add it
[22:30:11] <mru> http://www.itu.int/net/ITU-T/sigdb/spevideo/VideoForm-s.aspx?val=102002641
[22:30:36] <mmu_man_> HTML	doc/developer.html
[22:30:36] <mmu_man_> Option number is ambiguous (number-footnotes, number-sections)
[22:30:37] <mmu_man_> Try `texi2html --help' for more information
[22:30:51] <mru> jannau: and make sure it's correctly decoded of course
[22:30:57] <mmu_man_> texi2html --version
[22:30:57] <mmu_man_> 5.0
[22:31:30] <DonDiego> i have 1.78 here on my debian stable box :)
[22:31:36] <jannau> mru: we have to High and a High 4:4:4 Predictive sample in h264/
[22:31:48] <DonDiego> but the best ones are far older, on mphq we use 1.32 or so...
[22:31:50] <mru> 1.78 is the latest in gentoo
[22:31:56] <mru> must be a different version scale
[22:32:02] <mru> perhaps one that goes to 11
[22:32:42] <mmu_man_> anyway
[22:32:49] <mmu_man_> annoying
[22:32:51] <jannau> http://download.savannah.gnu.org/releases/texi2html/
[22:33:24] <mmu_man_> july 2010 :p
[22:33:40] <mru> oh, they did a version leap
[22:33:52] <mru> because versions <2 are so 1990s
[22:33:57] <mmu_man_> yeah
[22:33:57] <jannau>  NEWS: change the release number to 5.0, to match
[22:33:58] <jannau>         the next texinfo release.
[22:34:17] <DonDiego> ok, that makes sense
[22:34:18] <mmu_man_> then when they get to 9.99 they have to switch back :p
[22:34:26] <jannau> and it's still in cvs
[22:34:41] <DonDiego> not a project i would want to work on...
[22:34:46] <mru> let's make it a rule that every ffmpeg release shall have a number greater than any other released software at the time
[22:35:41] <mmu_man_> let's use +∞-N
[22:36:01] <jannau> I guess the plan might be to integrate texi2html into texinfo
[22:36:01] <CIA-38> ffmpeg: Peter Ross <pross at xvid.org> master * r4d54df8e07 ffmpeg/libavformat/ (mpegts.h mpegtsenc.c): (log message trimmed)
[22:36:01] <CIA-38> ffmpeg: mpegtsenc: support CODEC_ID_AAC_LATM
[22:36:01] <CIA-38> ffmpeg: $subject. Have used this for loopback testing with mpegts.c.
[22:36:01] <CIA-38> ffmpeg: -- Peter
[22:36:01] <CIA-38> ffmpeg: (A907 E02F A6E5 0CD2 34CD 20D2 6760 79C5 AC40 DD6B)
[22:36:02] <CIA-38> ffmpeg: [2. text/x-diff; 0001-mpegtsenc-support-CODEC_ID_AAC_LATM.patch]
[22:36:03] <CIA-38> ffmpeg: From 0f7f9db4b7da1793996af6dda84298507703759a Mon Sep 17 00:00:00 2001
[22:36:18] * mru curses
[22:36:24] <elenril> lu_zero: see branch playlists git://git.khirnov.net/git/ffmpeg
[22:36:29] <elenril> and it even compiles!
[22:36:39] <{V}> mru, gcc again?
[22:36:55] <jannau> no, people not using git send-email
[22:36:58] <mru> no, I fucked up the commit message
[22:37:15] <{V}> ohh
[22:37:16] * elenril should install gitorious or something
[22:37:24] <mru> I sent the whole message instead of just the attachment to git
[22:37:34] <pross-au> eh?
[22:37:42] <mru> I need to make some (local) commit hooks to catch such things
[22:37:42] <merbanan> elenril: was the asf stuff I sent you of any use ?
[22:38:08] <elenril> merbanan: well...i'm not really working on asf
[22:38:19] <elenril> the asf patches were just collateral damage
[22:38:20] <mru> pross-au: please use git send-email in future
[22:38:24] <merbanan> but it is metadata
[22:38:28] <merbanan> you love that stuff
[22:38:31] <elenril> i was actually working on mp3 muxer at the time
[22:38:32] <pross-au> k
[22:38:33] <mmu_man_> git send-mail sux
[22:38:53] <mmu_man_> it doesn't use the normal MUA, so mails aren't in sent/ folder..
[22:39:02] <mmu_man_> and git format-patch actually formats an email anyway
[22:39:03] <mmu_man_> sux
[22:39:09] <pross-au> mmu_man_: there's your calling then
[22:39:26] <mru> mmu_man_: you can have it send a copy to yourself
[22:39:34] <mmu_man_> bleh
[22:40:20] <wbs> mru: re git send-email, what way do you prefer for answering review comments inline as a normal reply to the review mail while sending a new updated patch? just comment on the mail and send the updated patch as a separate mail with git send-email?
[22:40:45] <mmu_man_> bleh, git send-mail doesn't even exist on OSX it seems
[22:40:52] <mmu_man_> I suppose it depends on linux crap
[22:41:18] <jannau> wbs: git send-email --annotate --in-reply-to=msg-id
[22:41:26] <elenril> mmu_man_: your trolling is too obvious, you've been reading the ML too much
[22:41:49] <wbs> jannau: yeah, but if I'd like to answer each of the comments in the previous mail inline as in normal mail discussions, too?
[22:41:50] <mmu_man_> elenril: I've been saying this way before the ml talked about switching to git
[22:42:00] <mru> wbs: if you attach a properly formatted patch to a review reply, that's rather obvious
[22:42:13] <wbs> mru: ok, that's what I've done so far
[22:42:14] <mmu_man_> there are several ugly incoherencies in git
[22:42:19] <mru> pross-au's patch here looked almost like just a send-email patch
[22:42:28] <mmu_man_> not talking about the option jungle :)
[22:42:29] <mmu_man_> anyway
[22:42:37] <jannau> wbs: and then copying your answer into the annotate editor and using ---8<---
[22:42:48] <mmu_man_> shall I send a patch for s/-number// ?
[22:42:50] <wbs> jannau: ah :-)
[22:42:56] <Yuvi> git send-email works on my mac...
[22:42:58] <mmu_man_> ELAZY
[22:43:15] <ohsix> -ELAZY
[22:43:16] <elenril> iirc send-email is a perl script
[22:43:26] <mmu_man_> ohsix: certainly not
[22:43:32] * pross-au is working on git-mutt-patch.sh ..
[22:43:34] <elenril> or something similarly horrible
[22:43:35] <mmu_man_> negating error codes is ugly
[22:43:40] <mmu_man_> and dangerous
[22:43:47] <mmu_man_> some OSes have them <0
[22:44:06] <mmu_man_> ah, was searching send-mail
[22:44:42] <mmu_man_> but I shouldn't have to set up a mail account in an SCM
[22:50:53] <mmu_man_> couldn't they just have send-email use the native clients...
[22:52:09] <jannau> it works out of the box with a proper sendmail compatible mta
[22:52:27] <jannau> no need to configure amail account
[22:52:46] <mmu_man_> except I don't think sendmail on OSX actually uses the Mail.app settings
[22:53:13] <mmu_man_> I suppose I could write a sendmail wrapper for Haiku
[22:53:18] <mmu_man_> that's ugly
[22:53:21] <Yuvi> mmu_man: have git send-email use an smtp sever
[22:53:30] <Yuvi> instead of sendmail
[22:54:10] <DonDiego> Yuvi: hey, long time no see :)
[22:54:42] <Yuvi> DonDiego: hi, I've been busy :)
[22:58:01] <Kovensky> 19:43.34 elenril: or something similarly horrible <-- me and TheFluff (mru too?) would like to have a chat with you
[22:58:04] * Kovensky runs
[22:58:17] <Kovensky> ohi Yuvi
[22:58:38] <TheFluff> >sendmail
[22:58:39] <TheFluff> sage
[22:59:34] <wbs> great, only 97 mails on the ML while being away for a few hours ;P
[22:59:39] <mru> going out, back later
[22:59:47] * elenril goes to sleep
[23:12:28] <mmu_man_> git send-email seems hung up
[23:12:53] <mmu_man_> Result: 250 2.0.0 Ok: queued as DC12F82261
[23:12:56] <mmu_man_> hmm
[23:13:07] <mmu_man_> ah :)
[23:13:36] <mmu_man_> hmm it forgot the Signed-off-by
[23:13:43] <astrange> is there a way to copy aac from ts->mov or mkv that works? ffmpeg streamcopy and mkvtoolnix don't
[23:13:43] <mmu_man_> man says it's on by default ??
[23:15:57] <mmu_man_> couldn't git clone set the sendemail.to on its own btw ?
[23:19:08] <mmu_man_> at least it seemed to have worked
[23:20:04] <Sean_McG> elenril: it's not authored as 1 file.
[23:20:41] <mmu_man_> hmm did it go through the ml ?
[23:21:22] <mmu_man_> ah yes
[23:23:40] <bcoudurier> astrange, what's wrong ?
[23:23:54] <bcoudurier> you have to absf the stream
[23:27:48] <astrange> ah
[23:27:57] <astrange> now it at least works in mplayer. only with faad though
[23:27:57] <astrange> [aac @ 0x10185fa00] Audio object type 4 is not supported.
[23:28:45] <astrange> the original ts doesn't show that error :/
[23:33:58] <superdump> 4 is LTP
[23:34:21] <superdump> a guy is working on that as an early small task for qualification for SoC
[23:34:38] <superdump> (rebasing and cleaning up LTP)
[23:34:59] <superdump> but i guess if it worked in the ts...
[23:39:55] <astrange> i really doubt tv tokyo is broadcasting LTP, it's a copy failure
[23:41:16] <DonDiego> gnite
[23:41:23] <Sean_McG> take care dude
[23:42:40] <kierank> astrange: japanese tv does a lot of weird non-standard things with aac audio
[23:49:44] <mmu_man_> mru: care to commit ?
[23:59:45] <jannau> mmu_man_: that would probably break the docs on the website which still uses texi2html 1.56k