[FFmpeg-devel-irc] IRC log for 2011-02-04

irc at mansr.com irc at mansr.com
Sat Feb 5 01:00:09 CET 2011


[00:13:54] <pross-au> Err
[00:14:02] <pross-au> ftp> cd /MPlayer/incoming
[00:14:02] <pross-au> 550 Failed to change directory.
[00:14:13] <mru> try cd /incoming
[00:14:30] <pross-au> thanks!
[00:14:45] <pross-au> Is this a permanent change? (bugreports.html needs updating)
[00:15:04] <mru> diego moved it up a level, it does make sense
[00:15:17] <mru> I guess he forgot to update all references to it
[00:21:32] <michaelni> hi mans, ben said i have to reconcile with you for any compromis
[00:21:41] <michaelni> iam not so much interrested to join you guys than iam concerned about the community spliting up and leaving both sides
[00:21:49] <michaelni> ping mru
[00:22:04] <mru> explain "compromise"
[00:23:37] <mru> so far the only hints you've let slip have been totally unacceptable
[00:24:07] <mru> and "ben said ..." doesn't exactly convince me you understand what this is about
[00:24:56] <michaelni> compromise = "roots replaced by neutral people, stef,me,carl,reimar,baptiste joining commiters, clarification of leadership of new team, clarification of file maintainers vs commiters, some vission doc that lists goals of ffmpeg)
[00:25:11] <mru> unacceptable
[00:25:12] <michaelni> that is the idea for discussion
[00:25:23] <michaelni> what part?
[00:25:26] <mru> all of it
[00:25:52] <michaelni> you dont want to clarify how the new maintainers are lead?
[00:26:00] <mru> first of all, what has "root" got to do with it?
[00:26:09] <michaelni> 1 person, democraty, consensus?
[00:26:32] <michaelni> root abused their power to allow this without public discussion and vote
[00:26:41] <mru> oh but there was
[00:26:49] <mru> in october there was a vote
[00:27:01] <mru> the outcome was that you got to stay under certain conditions
[00:27:07] <mru> you failed to live up to those conditions
[00:27:11] <mru> these are now the consequences
[00:27:12] <iive> mru: you can't count?
[00:27:16] <mru> deal with it
[00:27:36] <michaelni> i did not fail the conditions
[00:27:43] <mru> iive: I can count to 5 and then kickban you
[00:28:00] <mru> michaelni: I can see this is pointless to discuss further
[00:28:03] <michaelni> besides there was a majority that unconditionally wanted to keep me
[00:28:10] <mru> I will not reply again
[00:28:19] <michaelni> as you wish
[00:28:49] <iive> mru: what would you wish michael to do, in order to accept reconciliation?
[00:34:32] <Jumpyshoes> BBB: did holger ever reply to that thread? about LGPL ok?
[00:43:05] <Dark_Shikari> michaelni: how about you get approval from other people?
[00:43:07] <Dark_Shikari> this is a democracy.
[00:43:17] <Dark_Shikari> it doesn't matter what mru thinks if nobody agrees with him.
[00:43:21] <Dark_Shikari> I certainly don't agree with him.
[01:39:30] <BBB> Jumpyshoes: he said he would, but the email didn't go through yet
[01:39:36] <BBB> Jumpyshoes: keep asking him to reply to that thread saying ok
[01:40:48] <BBB> oh and uhm, asking all roots to resign is rather silly
[01:42:02] <BBB> lu_zero: did you comment on "[PATCH] swscale: fix build with --enable-runtime-cpudetect --disable-mmx/mmx2/amd3dnow"? or should we start merging your branch now?
[01:44:46] <Jumpyshoes> BBB: okay
[01:46:17] <{V}> BBB, re: roots to resign, agreed and "<michaelni> that is the idea for discussion"
[01:46:43] <{V}> in other words a proposal.
[01:46:53] <BBB> it might help the discussion to enter into it with slightly more realistic goals
[01:47:47] <lu_zero> BBB: I wanted to check since my tree isn't ready at all right now
[01:47:58] <BBB> {V}: if a woman and a man enter into a bar with the woman saying "I'm ugly, fat and I want a rich, handsome beautiful wall-street banker who makes $10M/yr" and the man says "I'm a loser and unemployed and I want a victoria secret's photomodel with blond long hair", then I don't think much will come out of the discussion
[01:48:01] <lu_zero> (and possibly broken by the last commit in a case)
[01:48:41] * Sean_McG snickers
[01:49:01] <{V}> BBB, a mediator is needed.
[01:49:21] <lu_zero> or to be less sexist, a compromise could be done only if both parties have a middle ground and are up to agree with some of the other party requests
[01:49:29] <lu_zero> {V}: sadly not
[01:49:38] <BBB> lu_zero: well but then the statement doesn't sound as funny
[01:49:56] <lu_zero> I stated already that I thank you for your effort
[01:49:57] <BBB> but anyway I apologize for the sexism, none was intended
[01:50:01] <{V}> aiming is not a bad thing, as long as you're willing to settle for less
[01:50:14] <{V}> s/aiming/aiming high/
[01:50:29] <lu_zero> {V}: I had been into mediation a bit
[01:51:53] <lu_zero> you don't enter a compromise offending the other party AND requesting the other party annihilation
[01:52:43] <lu_zero> or even worst, shifting from two different moods
[01:56:16] <CIA-38> ffmpeg: Peter Ross <pross at xvid.org> master * rf61dee2fe4 ffmpeg/libavformat/wtv.c:
[01:56:17] <CIA-38> ffmpeg: wtv: filesystem implementation
[01:56:17] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[01:57:55] <bryno> which repo is the *official* repo?
[01:58:15] <lu_zero> bryno: git.ffmpeg.org
[01:58:45] <bryno> what's the difference between that one and the one at git.mansr.com ?
[01:58:57] <BBB> git.mansr.com is mans' personal tree
[01:59:03] <BBB> just like we all have our own personal trees
[01:59:06] <lu_zero> git.mansr.com contains mans' experiments
[01:59:10] <BBB> (github.com/rbultje !)
[01:59:13] <lu_zero> like mine at github
[01:59:18] <lu_zero> or ronald's ^^
[01:59:50] <CIA-38> ffmpeg: Nicolas George <nicolas.george at normalesup.org> master * r51b317d2e9 ffmpeg/libavformat/tcp.c:
[01:59:50] <CIA-38> ffmpeg: TCP: factor the poll() call
[01:59:50] <CIA-38> ffmpeg: Signed-off-by: Nicolas George <nicolas.george at normalesup.org>
[01:59:50] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[01:59:51] <lu_zero> once they are ready they'll be revised and sent as patchset in the ml
[02:00:29] <bryno> the mansr one looks like it reflects the one at ffmpeg.org.  the one at videolan.org has commits that aren't even made anywhere else
[02:00:45] <lu_zero> bryno: and vice versa possibly
[02:00:54] <lu_zero> btw check the branches
[02:01:03] <lu_zero> since the master is usually in line for everybody
[02:01:27] <bryno> it's becoming confusing. branches/forks everywhere
[02:01:35] <lu_zero> bryno: why?
[02:01:53] <Sean_McG> welcome to open source, enjoy the ride.
[02:02:06] <BBB> bryno: that's the whole point of git
[02:02:09] <lu_zero> you want something supposedly stable -> pick what a distribution you trust chose
[02:02:33] <lu_zero> want something more dynamic, pick the git.ffmpeg.org tree
[02:02:40] <bryno> not necessarily stable, but official
[02:02:56] <lu_zero> want to help on a specific field?
[02:03:24] <lu_zero> clone one of the branches in the public personal repos ^^
[02:03:31] <lu_zero> isn't that complex
[02:03:55] <bryno> i'm coming from the perspective of someone who uses the libraries, like how the libav mailing list is designated
[02:04:04] <lu_zero> git.ffmpeg.org repo aims to collect the changes and the bugfix
[02:04:43] <lu_zero> so if you track it you possibly have less surprises
[02:04:57] <bryno> there's 5 repos listed on the site whereas before it was only 1. just adds to some confusion
[02:05:19] <lu_zero> bryno: there had been always that many (and some more)
[02:06:05] <lu_zero> siretart poured a LOT of effort to convince us that releases should be done
[02:06:06] <bryno> i can imagine, but the site only referred to the main branch that most people grabbed
[02:06:16] <lu_zero> so you might consider that as well
[02:11:04] <BBB> mru: is there a reason you did not push "[PATCH] Make avfilter_graph_free() free the graph"?
[02:12:32] <mru> I was waiting for clarification on the api/abi change policy there
[02:12:35] <mru> I guess it's good to go
[02:12:39] <mru> agree?
[02:13:42] <mru> hmm, look at fate
[02:17:54] <mru> you really should have build tested that
[02:40:20] <BBB> uh, I did
[02:40:35] <mru> no, you did not
[02:40:36] <BBB> oh damnit I didn't wait for the build to finish
[02:40:46] <BBB> wtv.c finished so I thought it was good to go
[02:40:48] <BBB> darnit
[02:40:51] <BBB> learned something new
[02:40:57] <BBB> ok, will find the parts I forgot to apply
[02:41:01] * Dark_Shikari pokes BBB again
[02:41:02] <mru> and you didn't see it spew warnings in all direction?
[02:41:17] <mru> I just replied to the relevant patch
[02:46:28] <BBB> moving functions into internal.h ok?
[02:46:30] <BBB> Dark_Shikari: one second
[02:46:35] <mru> I guess
[02:46:40] <BBB> Dark_Shikari: let me fix yet another git messup of mine
[02:46:47] <Dark_Shikari> sure
[02:46:50] <Dark_Shikari> no rush
[02:47:57] <BBB> so how do I amend the patch without changing its authorship?
[02:48:36] <BBB> or should I change ownership now, and say "based on patch by ..."?
[02:48:40] <mru> edit files; git add files; git commit --amend
[02:51:24] <BBB> fixed
[02:51:38] <CIA-38> ffmpeg: Peter Ross <pross at xvid.org> master * re6fb5a4f78 ffmpeg/libavformat/ (internal.h utils.c wtv.c):
[02:51:38] <CIA-38> ffmpeg: add ff_index_search_timestamp and ff_add_index_entry
[02:51:38] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[02:53:55] <BBB> ok, Dark_Shikari, sorry
[02:54:14] <BBB> how should I help?
[02:55:45] <Dark_Shikari> Just save the patch and maybe bench it
[02:55:51] <Dark_Shikari> throw it in your folder or something.
[02:56:02] <Dark_Shikari> I don't like having patches I wrote that weren't useful (but could be useful) sit on my disk.
[02:56:25] <BBB> oh shit I missed it
[02:56:28] <BBB> ok, will check
[02:56:38] <CIA-38> ffmpeg: Stefano Sabatini <stefano.sabatini-lala at poste.it> master * r4359288c56 ffmpeg/ (4 files in 2 dirs):
[02:56:39] <CIA-38> ffmpeg: Make avfilter_graph_free() free the graph.
[02:56:39] <CIA-38> ffmpeg: Make avfilter_graph_free() free not only the internal structures, but
[02:56:39] <CIA-38> ffmpeg: also the allocated graph, and set the graph pointer to NULL for
[02:56:39] <CIA-38> ffmpeg: increased safety.
[02:56:39] <CIA-38> ffmpeg: Simplify usage.
[02:56:40] <CIA-38> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[02:56:49] <CIA-38> ffmpeg: Stefano Sabatini <stefano.sabatini-lala at poste.it> master * re8e5dde779 ffmpeg/libavfilter/graphparser.c: (log message trimmed)
[02:56:49] <CIA-38> ffmpeg: Make avfilter_graph_parse() not free the input graph
[02:56:49] <CIA-38> ffmpeg: Make avfilter_graph_parse() only release the internal structures
[02:56:49] <CIA-38> ffmpeg: allocated during the parsing, and leave to free the graph itself to
[02:56:49] <CIA-38> ffmpeg: the calling code.
[02:56:49] <CIA-38> ffmpeg: This approach looks cleaner, as the graph is not allocated by the
[02:56:50] <CIA-38> ffmpeg: function.
[03:01:08] <BBB> wbs: https://roundup.ffmpeg.org/issue2583 for you
[03:16:03] <Dark_Shikari> ok BBB I finally did something that helped performance (omg)
[03:16:16] <BBB> Dark_Shikari: you imlpemented ffmpeg-mt for vp8?
[03:16:35] <Dark_Shikari> lol
[03:17:25] <BBB> you added support for real slices?
[03:17:46] <BBB> as in, ndependent of each other, no refs to each other
[03:17:48] <Dark_Shikari> check my email
[03:17:48] <BBB> like in h264
[03:17:49] <BBB> :-p
[03:18:19] <BBB> dont see ot yet
[03:18:30] <BBB> there it is
[03:19:16] <Dark_Shikari> it's like 8 cycles faster or something
[03:20:04] <BBB> I'm confused, 8 cycles is noteworthy but 30 cycles like your last patch is not?
[03:20:09] <BBB> or was that sarcastic? :-p
[03:20:11] <Dark_Shikari> Relative
[03:20:14] <Dark_Shikari> this is 8 cycles out of a 30 cycle function
[03:26:20] <BBB> mru: http://patches.ffmpeg.org/patch/688/ shoud be committed
[03:26:27] <BBB> Dark_Shikari: oh that's not bad
[03:28:15] <BBB> Dark_Shikari: that passes make fate-vp8?
[03:28:50] <BBB> I thought we had to cache the mvmode somehow for the next blocks
[03:29:05] <BBB> (I mean near vs. nearest)
[03:33:30] <BBB> mru: also "[FFmpeg-cvslog] Implement av_samples_alloc() and av_samples_fill_arrays()" should maybe be applied?
[03:37:39] <Dark_Shikari> yes it passes
[03:37:44] <Dark_Shikari> mvmode is not needed for future mbs
[03:37:47] <Dark_Shikari> also
[03:37:49] <Dark_Shikari> #define POW2CLIP(x,max) (((x) & ~max) ? (-(x))>>31 & max : (x));
[03:37:52] <Dark_Shikari> Do we have a macro that does this?
[03:39:53] <Dark_Shikari> this makes the filter_level clip faster.
[03:40:26] <BBB> static inline, I can see gcc calculating x 4x
[03:40:48] <Dark_Shikari> this is from x264_clip_pixel
[03:40:50] <BBB> av_clip_(u)ont16/32/64 are implemented as per above
[03:40:54] <BBB> but there's no generic version of it
[03:40:57] <Dark_Shikari> yes, but we need this to clip to 63
[03:41:03] <Dark_Shikari> I don't want to bikeshed this :>
[03:41:05] <BBB> hehe :)
[03:41:08] <Dark_Shikari> also filter_level is already calculated
[03:41:11] <BBB> just make it a local macro
[03:41:12] <Dark_Shikari> #define POW2CLIP(x,max) (((x) & ~max) ? (-(x))>>31 & max : (x)); filter_level = POW2CLIP(filter_level, 63);
[03:41:23] <BBB> it's fine with me
[03:41:28] <BBB> is it faster?
[03:41:38] <BBB> btw is gaikai going to use vp8? you work a lot on vp8 lately
[03:42:06] <Dark_Shikari> no
[03:42:24] <Dark_Shikari> yes its faster, ~1-2 clocks
[03:50:56] <BBB> ok, reviewed those two
[03:51:04] <BBB> wifey wants to sleep, so I'll bench your last patch tomorrow
[03:51:16] <Dark_Shikari> you really don't have to bench it, I already did
[03:51:18] <Dark_Shikari> oh
[03:51:21] <Dark_Shikari> you mean the sse4 one?
[03:51:34] <Jumpyshoes> i was going to ask what wifey was <_< >_>
[03:51:42] <BBB> Dark_Shikari: yes that one
[03:51:53] <Dark_Shikari> Jumpyshoes: <insert photo of dakimakura here>
[03:51:57] <BBB> Jumpyshoes: you're too young for that kind of stuff :-p
[03:52:00] <Dark_Shikari> BBB: ok, no problem, it's not important
[03:52:10] <Jumpyshoes> Dark_Shikari: nothing wrong with those :P
[03:52:12] <BBB> Dark_Shikari: vp8 performance is fun :)
[03:52:41] <Jumpyshoes> btw, is there a public xvp8 tree anywhere?
[03:52:46] <Dark_Shikari> yes
[03:52:52] <BBB> it's not uptodate
[03:52:57] <Jumpyshoes> oh
[03:52:59] * BBB lazy
[04:24:41] <CIA-38> ffmpeg: Jason Garrett-Glaser <jason at x264.com> master * rdd18c9a050 ffmpeg/libavcodec/ (vp8.c vp8data.h): VP8: simplify lf_delta mb mode logic
[04:24:51] <CIA-38> ffmpeg: Jason Garrett-Glaser <jason at x264.com> master * ra1b227bb53 ffmpeg/libavcodec/vp8.c: VP8: faster filter_level clip
[05:42:02] <Dark_Shikari> mru: why is the troll still not banned?
[05:46:24] <ubitux> in libavutil/tree.h, there is a "if(*next) av_freep(next)", according to the context, do you think it is a "typo" for "if(next) av_freep(next)", or we should just remove it with av_freep(next)?
[05:46:45] <ubitux> (it's in a comment, not a functionnal code this is why i ask)
[05:47:28] <ubitux> i would have removed the if if it was code
[05:48:24] * elenril yawns
[05:48:33] <elenril> awesome, more drama
[05:49:28] <peloverde_> Is gabu's last post enough to get him banned?
[05:50:19] <peloverde_> Any list admins around here?
[05:53:25] <Dark_Shikari> I've repeatedly asked mru why the troll isn't banned
[05:53:56] <wooster> what'd he post?
[05:55:08] <elenril> Dark_Shikari: can mru even do that?
[05:55:21] <elenril> i think doesn't admin the ML
[05:55:29] <ubitux> http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2011-February/105633.html
[05:55:32] <ubitux> here is the post.
[05:55:52] <Dark_Shikari> he's been constantly trolling the ml
[05:55:54] <Dark_Shikari> for like a week
[05:59:18] <peloverde_> who are list admins?
[05:59:39] <Dark_Shikari> no idea
[05:59:41] <Dark_Shikari> bcoudurier: you know?
[06:03:35] <uau> KotH has maintained some of the mail stuff at least
[06:39:47] <Dark_Shikari> mru: I'd like to troll you for a moment
[06:39:50] <Dark_Shikari> so I just realized, in vp8
[06:39:54] <Dark_Shikari> mbedge_lim = bedge_lim + 4
[06:39:58] <Dark_Shikari> we should not be passing this as a parameter
[06:40:06] <Dark_Shikari> because it's derived trivially from one of the other parameters.
[06:40:34] <Dark_Shikari> ... oh wait.  there's no single loopfilter DSP function that takes both.  I think
[06:40:40] <Dark_Shikari> phew.
[06:40:43] <Dark_Shikari> don't have to break your asm.
[07:00:22] <elenril> o_0 what's up with all the vp8 patches
[07:02:26] <Dark_Shikari> me being bored
[07:02:58] <lu_zero> yawn
[07:03:04] <lu_zero> good morning
[07:03:09] * lu_zero heads to the airport
[07:04:35] <elenril> Dark_Shikari: help merging -mt =p
[07:08:01] <elenril> btw it'd be really nice if somebody deprecated all the stuff that's to be removed from AVCodecContext
[07:36:27] <elenril> :/ srsly, what does this troll have to do to get banned
[07:42:12] <cartman> moin
[07:58:51] <spaam> God morgon
[07:58:59] <pross-au_> Evening
[07:59:15] <pJok> ohayou gozaimasu
[08:02:33] <siretart> Dark_Shikari: AFAIUI, benoit- is the list moderator, not mru
[08:02:38] <Dark_Shikari> ah
[08:02:40] <Dark_Shikari> benoit-: ^
[08:03:23] <siretart> btw, who cares about vc1 these days? https://roundup.ffmpeg.org/issue2584
[08:04:02] <Dark_Shikari> ask jumpyshoes, he fixes these things
[08:04:22] <siretart> ok, thanks
[08:04:32] <Dark_Shikari> I'll try to see if I can replicate btw
[08:04:45] <cartman> man, get a fscking real SSL certificate for god's sake
[08:05:06] <siretart> at least a cacert one would do for me
[08:05:27] <Dark_Shikari> siretart: doesn't replicate here.
[08:05:42] <Dark_Shikari> "Bits overconsumption" ---> looks like it might be an overread problem
[08:05:48] <siretart> okay, that indicates that some commit in master fixed it then?
[08:05:51] <Dark_Shikari> No
[08:05:59] <Dark_Shikari> It could just indicate that ffmpeg.c has enough padding to avoid it.
[08:06:32] <Dark_Shikari> (if that's the issue.  it might not be.)
[08:06:38] <Dark_Shikari> someone should valgrind it
[08:06:41] <siretart> well, it was observed in vlc, this would rule out a change in ffmpeg.c
[08:07:02] <spaam> cartman: will you fix OpenGL support for ffplay when roundup have a good cert? ;)
[08:07:16] <cartman> spaam: I am on my way fixing myself
[08:07:18] <Dark_Shikari> siretart: that isn't what I said
[08:07:27] <Dark_Shikari> I don't think anything was changed at all.
[08:07:44] <spaam> cartman: ohh
[08:07:51] <cartman> spaam: for any specific problem though, reimar wrote MPlayer's OpenGL vout drivers
[08:07:59] <siretart> Dark_Shikari: ah, so you're basically saying it needs more investigation? - ok
[08:08:01] <cartman> maybe assign to him :)
[08:08:04] <Dark_Shikari> yes
[08:08:32] <spaam> cartman: i know. but you code some opengl stuff now. so it will be easy for you ;)
[08:08:46] <cartman> spaam: baaaah
[08:09:19] <spaam> cartman: pJok can help you. :)
[08:09:26] <cartman> spaam: oh :D
[08:10:14] <pJok> wa?
[08:11:02] <cartman> pJok: spaam says you are an OpenGL wizard
[08:11:16] <pJok> i am sure his tabcomplete is borked ;)
[08:11:28] * pJok is about as wizardry as Clippy
[08:11:40] <cartman> good :P
[08:12:02] <pJok> i think he meant peloverde rather than me
[08:12:12] <pJok> im a techie, not a coder ;)
[08:13:17] <cartman> No worries I am not writing OpenGL either :P
[08:13:32] <peloverde> I don't think he means me, I haven't done any OpenGL since first year at university
[08:13:48] <av500> cartman: its not me either
[08:13:57] <cartman> av500: I guessed that much :)
[08:14:17] <cartman> http://dilbert.com/fast/2011-02-04
[08:14:17] <cartman> lol
[08:24:56] <wbs> BBB: I'll have a look
[08:29:39] <kshishkov> siretart: I can at least name you one person who doesn't care about VC-1
[08:31:02] <spaam> kshishkov: will we see interlaced support for vc-1 this year? ;D
[08:31:22] <cartman> spaam: interlaced bink-b first
[08:31:49] <spaam> cartman: Nice
[08:32:08] <kshishkov> spaam: he's right (especially since such bink files don't exist)
[08:32:24] <spaam> kshishkov: maybe we can create one?
[08:33:13] <kshishkov> hmm, good idea, make Dark_Shikari work on bink-b encoder
[08:33:57] <KotH> salut
[08:34:35] <kshishkov> god morgon
[08:36:50] <cartman> moin KotH
[08:42:50] <Tjoppen> hm. sample_aspect_ratio as a name results in an ambiguous acronym
[08:43:22] <kshishkov> ?
[08:43:33] <Tjoppen> SAR can often mean "stored aspect ratio"
[08:43:42] <Tjoppen> hence DAR = SAR * PAR
[08:43:56] <Tjoppen> stead of SAR * SAR :o
[08:43:59] <Tjoppen> *instead
[08:44:33] <kshishkov> just kill the ones who proposed "stored" thingy
[08:45:13] <superdump> indeed
[08:45:23] <superdump> i haven't heard of SAR being stored aspect ratio
[08:45:27] <superdump> always sample aspect ratio
[08:45:55] <Tjoppen> what a good name for the essence's aspect then? EAR I suppose
[08:46:34] <kshishkov> and I'm pretty sure you can tons of meanings for every TLA, look at Stefano's signature as example
[08:46:43] <kshishkov> Tjoppen: IORE
[08:47:25] <superdump> isn't the essence aspect ratio then easily calculable from the stored width/height?
[08:47:45] <peloverde> deep down I know kshishkov secretly wants to finish AAC
[08:48:51] <kshishkov> peloverde: actually I'd like my coworker to do it instead
[08:51:34] <peloverde> and I would like Abraham Lincoln to do it but sometimes we have to stick to what's reasonable
[08:52:33] <kshishkov> peloverde: not working at all? That always was reasonable to me
[09:07:10] <mru> Tjoppen: if not stored, then what?
[09:07:13] <mru> forgotten?
[09:07:50] <mru> peloverde: you know who he works with, right?
[09:07:58] <av500> "suggested" aspect ratio
[09:08:07] <mru> supposed
[09:08:42] <wbs> mru: who does he work with, menno bakker or someone similar?
[09:08:47] <av500> ivan
[09:09:01] <mru> wbs: ivan dimkovic
[09:09:15] <wbs> oh, even better :-)
[09:09:19] <kshishkov> indeed
[09:09:53] <av500> and since we probably share some ancestry way back, it makes me total cool too
[09:09:54] * cartman wonders if kshishkov works for the guys who burned Rome
[09:10:17] <av500> cartman: close
[09:10:48] <peloverde> I'm aware
[09:12:38] <kshishkov> av500: 26 tram stops from here so it's not close :P
[09:13:37] <kshishkov> cartman: I don't I'd be able to work there
[09:13:53] <cartman> can't parse that
[09:13:55] <av500> kshishkov: 26 tram stops to Rome?
[09:14:14] <mru> all trams stop in rome, no?
[09:15:04] <kshishkov> mru: not the ones in small village on west coast (aka Göteborg)
[09:19:15] <KotH> Dark_Shikari, uau: yes, mru and me are root, but we do not manage the ffmpeg mailinglists. hence it would be inaproriate for us to ban someone. especially if the ml admins wont do that
[09:20:08] <Dark_Shikari> so who does?
[09:20:14] <spaam> Compn ?
[09:20:19] <KotH> ffmpeg-devel-owner at mphq :)
[09:23:12] <cartman> re
[09:23:24] <cartman> Nero AAC encoder is from 2009 it seems , pff
[09:24:03] <mru> their avc decoder is nice
[09:25:12] * cartman pets his nvidia card for that
[09:25:54] <kshishkov> mru: that makes me wonder - you know the man who wrote it so is it that hard to write fast and nice H.264 decoder given enough time?
[09:26:26] * cartman just had a DejaVu
[09:26:28] <mru> well, he's a philosopher...
[09:26:43] <cartman> psychic powers?
[09:27:33] <kshishkov> mru: exactly, so can't proper engineer write it too?
[09:27:39] <mru> he smokes cigars, claims the code is revealed to him in the smoke
[09:27:44] <cartman> LOL
[09:27:56] <cartman> I worked with a guy like that
[09:27:58] <mru> only last bit is made up
[09:28:15] <cartman> http://cdn1.cnnturk.com/Handlers/file_.ashx?FileID=428734 looks similar btw.
[09:29:49] <j-b> 'morning
[09:29:54] <cartman> moin j-b
[09:30:05] <peloverde> It would be cool if our AVC decoder were fast
[09:30:18] <peloverde> Dark_Shikari: didn't you have some ideas on how to restructure it?
[09:30:20] <j-b> +10
[09:30:29] <cartman> peloverde: then lots of commercial companies would go bankrupt
[09:30:30] <av500> and what about the H264 decoder?
[09:31:21] <kshishkov> av500: maybe writing fast h264 decoder is not that hard
[09:31:33] <cartman> just need correct cigars
[09:31:46] <av500> kshishkov: I want a fast mpeg4-part10 decoder too
[09:31:49] <Dark_Shikari> peloverde: well I've already done some optimizations
[09:31:52] <Dark_Shikari> ffmpeg-mt is more important though
[09:32:13] <Dark_Shikari> One thing I want to do: port what I did in vp8, zero the dct blocks in idct
[09:32:18] <Dark_Shikari> problem: requires changes in ALL h264 asm
[09:32:20] <Dark_Shikari> ppc, x86, and arm
[09:32:25] <Dark_Shikari> and whatever else
[09:32:45] <mru> those are the only ones iirc
[09:32:49] <mru> let's do it if it helps
[09:32:51] <kshishkov> av500: make your Indians do it!
[09:33:02] <cartman> av500 works with Chinese
[09:33:11] <av500> cartman: wrong
[09:33:13] <mru> cartman: but do they work with you?
[09:33:22] <cartman> you mean "wong!"
[09:33:28] <cartman> thats how you do it
[09:33:42] <cartman> mru: if you pay enough they pretend to do so
[09:33:44] <av500> wlong again!
[09:34:01] <Dark_Shikari> mru: you know what I'm talking about, right?
[09:34:04] <Dark_Shikari> i.e. to avoid clear_blocks
[09:34:06] <Dark_Shikari> as in vp8
[09:34:12] <cartman> av500: so indians it is?
[09:34:19] <av500> cartman: for codecs, yes
[09:34:23] <cartman> av500: interesting
[09:34:42] <Dark_Shikari> speaking of indians
[09:34:44] <cartman> av500: edge cases that ffmpeg doesn't handle?
[09:34:47] <Dark_Shikari> I just had one randomly PM me about a gstreamer question
[09:34:50] <Dark_Shikari> unrelated to x264
[09:34:57] <Dark_Shikari> I think it's one of those guys who PMs everyone on the /who list until he gets a response
[09:35:00] <Dark_Shikari> instead of asking it in channel
[09:35:00] <av500> Dark_Shikari: just send him teh codez
[09:35:29] <cartman> he just needs a complete "sample" :p
[09:36:36] <cartman> wbs: my Android 2.1 crash bug is 7 days old and going :P
[09:36:48] <j-b> pfff, you guys are still fighting over admin/roots? In those days of cloud? pfff
[09:36:56] <wbs> j-b: lol
[09:37:01] <wbs> cartman: awwww, poor you ;P
[09:37:04] <kshishkov> j-b: not us
[09:37:05] <cartman> We all have a "Got root?" tshirt
[09:37:08] <cartman> we need it.
[09:37:08] <wbs> cartman: you haven't even debugged it and sent a patch yet! ;P
[09:37:20] <cartman> wbs: uh oh its inside the libstdc++ :(
[09:38:12] <cartman> wbs: its Google's fault anyway. Obviously their famous QA doesn't test < 2.2
[09:38:16] <j-b> wbs: well, seriously. Finding git hosting is fucking easy (github, gitorious, or even git.v.o). Hosting the website is a joke (it is 20 static pages). Hosting mailing list is of almost no maintainance and outsourceable quite easily.
[09:38:27] <j-b> wbs: so, what's the rest? fate? ok.
[09:38:59] <cartman> j-b: you don't get it.
[09:39:07] <cartman> its about the name "FFmpeg" in the end
[09:39:42] <j-b> cartman: so? administrating a domain doesn't require a server.
[09:39:43] <mru> j-b: and the 50GB of samples?
[09:39:59] <j-b> mru: that's all?
[09:40:03] <mru> also, I don't trust github
[09:40:08] <j-b> mru: jones.v.o has a 1TB disk
[09:40:21] <cartman> a disk could be donated
[09:40:26] <j-b> sure
[09:40:32] <mru> github often has downtime and they lost data at least once
[09:40:43] * twnqx eyes a stack of decommisioned 1TB drives to his right
[09:40:54] <j-b> mru: use gitorious. use us. use $whatever
[09:40:57] <mru> it's nice as as a distribution channel, no more
[09:41:28] <av500> mru: I can host your 50gb of samples on one A70H
[09:41:30] <j-b> I just mean that roots/admin discussion is a non-issue.
[09:41:42] <mru> that I agree with
[09:41:51] <j-b> it isn't 2001 anymore
[09:41:57] <mru> true
[09:42:23] <twnqx> just use more than 1 server and sync daily or something
[09:42:25] <kshishkov> mru: sourceforget.net!
[09:42:28] <cartman> still ownership of the domain will be important
[09:42:29] <twnqx> to reduce the dependency
[09:42:31] <mru> when we set this thing up, the only vaguely viable public hosting was sourceforgery
[09:43:00] <j-b> cartman: sure, and this is NOT a roots/admin issue§.
[09:43:04] <mru> and their service was so poor we had to abandon it
[09:43:06] <cartman> j-b: right
[09:43:32] <vipw> try savannah, it's even worse
[09:43:58] <j-b> if you don't expect the QoS that michaelni expected in the last mail a few months ago, it is quite easy
[09:43:58] <mru> that's why I said vaguely viable
[09:44:42] <mru> we've had much better availability stats than most of the public offerings over the last 5 years
[09:45:06] <peloverde> we also block half the internet
[09:45:15] <av500> and hungary
[09:45:19] <mru> mostly china
[09:45:29] <j-b> I can say the same
[09:45:35] <mru> but _they_ block half the internet too
[09:45:37] <j-b> "we've had much better availability stats than most of the public offerings"
[09:45:42] <peloverde> When i started my current job I had to ask KotH to unblock me
[09:45:45] <mru> j-b: I'm sure you have
[09:46:11] <DonDiego> note that i'm working on getting said half of the internet unblocked
[09:46:15] <j-b> mru: especially since we splitted development hosting|www hosting|other_stupid_websites
[09:46:18] <peloverde> Samples and the website should have different rules
[09:46:39] <DonDiego> peloverde: what rules do you mean?
[09:47:03] <mru> rules of engagement
[09:47:04] <peloverde> it should be ok to block half of china from samples, but not from the website
[09:47:10] <j-b> peloverde: why are people blocking anything?
[09:47:26] <mru> peloverde: DonDiego is working on exactly that
[09:47:42] <peloverde> j-b: some people try to mirror all of samples eating a tremendous amount of bandwidth
[09:47:48] <mru> j-b: because chinese leeches suck up all our bandwidth otherwise
[09:48:08] <j-b> peloverde: you don't have unlimited bandwidth?
[09:48:23] <mru> there's no hard limit
[09:48:29] <cartman> there is no unlimited bandwidth in reality
[09:48:31] <DonDiego> peloverde: half-way done, the website already resides on another IP, i still need to find a solution for rsync and ftp
[09:48:41] <peloverde> DonDiego: awesome
[09:48:56] <peloverde> will the website be ipv6 ready?
[09:49:03] <DonDiego> hehe
[09:49:08] <DonDiego> i never tried, good question
[09:49:18] <mru> we don't have a public ipv6 address
[09:49:29] <av500> DonDiego: make it v8 while you are at it...
[09:49:41] <mru> with golden packets
[09:49:58] <j-b> traffic on ipv6 on www.v.o was around 0.5% last month
[09:49:58] <av500> alternate route packets
[09:50:11] <wbs> j-b: that's surprisingly much :-)
[09:50:21] <j-b> wbs: indeed
[09:50:47] <peloverde> j-b: give it a year or two
[09:50:54] <j-b> wbs: but since India is counting for 8% of the traffic (1,5% last year) it might be the reason
[09:51:09] <spaam> j-b: www.v.o do not have aaaa records ;S
[09:51:18] <j-b> spaam: because I deactivated it
[09:51:21] <spaam> why?
[09:51:33] <j-b> because I want a better allocation
[09:51:43] <j-b> spaam: so I have new IPv6 now
[09:51:51] <spaam> ok :)
[09:51:51] <j-b> spaam: so I have new IPv6 adresses now
[09:52:23] <j-b> spaam: seeing the traffic was more than I expected, I asked our provider for more. He said: "I hope this is for a BSD server".
[09:53:19] <DonDiego> how is this related to bsd?
[09:53:57] <mru> perhaps bsd handles high ipv6 traffic  better
[09:54:06] <j-b> DonDiego: just basic trolling from the sysadmin that is a BSD fan
[09:54:11] <mru> linux certainly has rough edges
[09:54:18] <spaam> : )
[09:54:24] <av500> please no bsd trolling now
[09:54:33] <twnqx> mru: in what ways?
[09:54:38] <av500> kshishkov: trains, pleaseeeee!
[09:54:43] <mru> we caught linux doing tons of unaligned accesses in ipv6 code recently
[09:54:55] <twnqx> i mean... i'm using ipv6 for a few years on my linux boxes, and never used up >500mbit with it, but...
[09:55:13] <mru> it works for sure
[09:55:18] <av500> twnqx: your fringe uses cases do not reflect ffmpeg.org reality :)
[09:55:27] <mru> but I can very well imagine that it's not as well optimised as the ipv4 paths
[09:55:40] <twnqx> sure, single high-bandwidth download is different from webserver reality
[09:55:57] <av500> ff.org must handle all the known internet pulling git at once
[09:55:59] <twnqx> mru: so i assume you're not talking x86/x86_64?
[09:56:12] <mru> this was on arm
[09:56:36] <kshishkov> are there ARM SOCs with gigabit ethernet?
[09:56:45] <twnqx> av500: i am hosting a private gentoo portag emirror... i switched to using the disk as write-mostly with a ramdisk as raid 1 >_>
[09:56:50] <mru> using a kernel patched to warn about alignment traps instead of silently fixing them
[09:56:58] <mru> kshishkov: yes, the sheeva for example
[09:57:17] <kshishkov> mru: ah, the one that needs fan as well
[09:57:26] <spaam> twnqx: gentoo fan? ;D
[09:57:35] <mru> kshishkov: that's not the only one
[09:58:00] <twnqx> spaam: user. i have enough machines with it to run an own mirror for those
[09:58:04] <kshishkov> mru: what else?
[09:58:12] <cartman> setting up an IPv6 tunnel was easy
[09:58:18] <mru> don't remember the details
[09:58:19] * kshishkov has not seen ARM server CPUs yet
[09:58:24] <spaam> twnqx: ok :)
[09:58:30] <mru> gbit is not just for servers
[09:58:35] <av500> kshishkov: ask canonical
[09:58:49] <av500> they have a bunch of PXA270 based build servers :)
[09:59:11] <kshishkov> mru: I also mean ARM-based chips with need of fan cooling
[10:00:37] <pross-au> kshishkov: there are really cheap 1Us avail
[10:01:13] <kshishkov> pross-au: I don't use big iron
[10:01:25] <kshishkov> except at work
[10:06:54] <ubitux> I'm looking for a (simple) audio-only demuxer+decoder I could use as base for writting one, do you have one in mind?
[10:07:38] <kshishkov> any game format
[10:07:46] <mru> great, combine it with my video-only player and we'll have completeness
[10:07:47] <pross-au> sunau
[10:08:50] <kshishkov> mru: why haven't you added audio out for omapfbplay? is it too messy with OSS/ALSA/NAS/whatever?
[10:09:07] <ubitux> kshishkov: one random in mind?
[10:09:22] <mru> alsa and libavcodec don't fit together well
[10:10:06] <kshishkov> mru: only alsa and pulseaudio fit together
[10:10:16] <kshishkov> ubitux: westwoodaudio.c
[10:10:24] <ubitux> thanks
[10:10:52] <kshishkov> ubitux: or try vocdec.c
[10:10:55] <mru> alsa is stupid and won't let me just mmap the damn buffer
[10:11:10] <mru> it insists on some silly function being called before and after each buffer fill
[10:11:19] <kshishkov> mru: first three words were enough
[10:11:56] <ubitux> kshishkov: right, perfect, thanks :)
[10:11:57] <CIA-38> ffmpeg: Clément Bœsch <ubitux at gmail.com> master * r523d9407d5 ffmpeg/ (3 files in 3 dirs):
[10:11:57] <CIA-38> ffmpeg: Remove a few if (p) av_freep(&p) forms
[10:11:57] <CIA-38> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[10:12:08] <CIA-38> ffmpeg: Clément Bœsch <ubitux at gmail.com> master * r290849e2a4 ffmpeg/ (libavfilter/defaults.c libavformat/avidec.c):
[10:12:08] <CIA-38> ffmpeg: Remove forgotten if (p) av_free(p) forms
[10:12:08] <CIA-38> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[10:12:19] <kshishkov> mru: and I presume you've never looked at standard Window audio output API (before DShow), they had that approach for ages
[10:12:33] <mru> of course not
[10:12:53] <kshishkov> well, I felt in heaven after trying OSS output later
[10:13:12] <mru> I don't like functions with dozens of arguments and an entire essay for a name
[10:13:16] <mru> +Ex
[10:13:50] <av500> mru: so, use OSS emulation on ALSA :)
[10:13:51] <cartman> mru: no sound then? :)
[10:13:53] <kshishkov> even if it's extrememly complicated to invoke several ioctl()s and then use write() I'd prefer OSS
[10:14:01] <av500> kshishkov: +1
[10:14:10] <kshishkov> cartman: haven't you guessed it?
[10:14:11] <mru> cartman: some way or other I suppose I'll do it regardless
[10:14:20] <mru> but it probably means using an intermediate buffer
[10:14:21] <mru> yuck
[10:14:27] <cartman> good good :p
[10:15:21] <av500> mru: cmon, one buffer for audio
[10:15:52] <av500> mru: you are after 16 sample latency now?
[10:21:07] * kshishkov suspects mru would need to allocate many buffers to allow realtime playback of Monkey Audio on BeagleBoard
[10:22:37] <av500> one for each simian?
[10:25:12] <Kovensky> 07:13.51 av500: mru: so, use OSS emulation on ALSA :) <-- did alsa ever fix their OSS emulation or do they keep using its brokeness to claim that OSS is a POS
[10:25:21] <twnqx> kshishkov: which reminds me... .ape is supported by ffmpeg, right?
[10:25:27] <mru> Kovensky: don't know
[10:25:29] <mru> twnqx: yes
[10:25:52] <av500> Kovensky: no idea
[10:26:04] <av500> Kovensky: I doubt they fixed it
[10:27:15] <kshishkov> av500: it's Monkey's Audio not because it employs monkeys in decoder but merely because it used monkeys in design and development
[10:34:05] <kshishkov> BTW, anyone wants decoder writter specially for git.ffmpeg.org?
[10:34:43] <Kovensky> most people with sound issues on linux are either ALSA trolls, people that believe the ALSA trolls, or freetards
[10:35:20] <mru> alsa can actually be coerced into doing the right thing
[10:35:29] <kshishkov> with pulseaudio?
[10:35:31] <mru> if you replace the default configuration
[10:35:49] <mru> set the default device to hw:0 or similar
[10:36:10] <mru> that shuts off all it's nasty resampling, mixing, ipc backdoors, etc
[10:36:15] <mru> -'
[10:39:50] <Kovensky> mru: you still need dmix to be able to play more than one sound simultaneously though
[10:40:18] <mru> no you don't
[10:40:25] <cartman> Kovensky: or you get a Creative soundcard
[10:40:29] <cartman> and it'll do hw mixing
[10:40:35] <kshishkov> indeed
[10:40:36] <mru> or any card with hw mixing
[10:40:43] <mru> if you really care about it that much
[10:40:50] <mru> otherwise you make do with one sound at a time
[10:40:51] <Kovensky> won't help with my realtek hda :(
[10:40:53] <cartman> yeah those are cheap nowadays
[10:40:56] <Kovensky> and anyway
[10:41:02] <Kovensky> OSS4 can do mixing in software
[10:41:07] <av500> yes
[10:41:14] <av500> and it does mixing where it belong
[10:41:15] <cartman> HDA series are crap soundcards
[10:41:17] <cartman> buggy shit
[10:41:22] <mru> and dmix is the worst possible way of doing it
[10:41:40] <kshishkov> cartman: they design good network cards though
[10:41:45] <cartman> kshishkov: true
[10:41:49] <mru> who?
[10:41:53] <Kovensky> HDA is an intel spec
[10:41:59] <cartman> ACPI too
[10:42:00] <cartman> :P
[10:42:03] <Kovensky> (aka Azalia)
[10:42:08] <mru> the intel NICs are DEC designs originally
[10:42:11] <kshishkov> mru: Realtek 8139 IIRC - the most famous NIC ever
[10:42:20] <mru> infamous
[10:42:26] <cartman> modprobe 8139too
[10:42:27] <cartman> :p
[10:42:39] <Kovensky> there's an 8139too.ko?
[10:42:45] <mru> my motherboard has an onboard rtl8111
[10:42:50] <merbzt> I have a 8169 in my fitpc
[10:42:54] <cartman> Kovensky: yes
[10:42:55] * kshishkov remembers his surprise when LG NIC worked with DEC driver
[10:42:55] <mru> 3/4 boots it doesn't even show up on the pci bus
[10:42:56] <pross-au> kshishkov: pftt, everyone knows it was ne2k
[10:43:12] <cartman> pross-au: ne2k was pretty stable for me
[10:43:20] <Kovensky> quick, lspci equivalent for osx =p
[10:43:21] <cartman> my first custom ethernet card
[10:43:21] <merbzt> let's say it doesn't really work well
[10:43:33] <cartman> Kovensky: System information :P
[10:43:45] <Kovensky> cartman: can I run it from command line? :D
[10:43:46] <kshishkov> at least Realteks had hardware bandwidth limiter connected to CPU load
[11:32:49] <DonDiego> saste: have you thought about creating a libavfilter topic branch?
[11:33:42] <saste> DonDiego: yes maybe I'll do
[11:33:56] <mru> I think that would be a good idea
[11:34:10] <saste> DonDiego: but some patches should be committed right now
[11:34:26] <saste> the aspect ratio fix and the movie source
[11:34:32] <saste> lots of users asking for that
[11:35:15] <mru> nothing should be committed without proper review, no matter how many are asking for it
[11:35:24] <mru> and the aspect ratio patch had review comments
[11:35:55] <saste> mru: of course I mean *after* review
[11:45:32] <Tjoppen> I've got some avc-intra samples to upload. should I put them in the usual spot, namely the ftp?
[11:49:21] <kshishkov> yes
[11:49:38] <kshishkov> preferably with decoder patch sent to ML :)
[11:50:19] <Tjoppen> hehe :)
[11:53:13] <Tjoppen> for avc-intra 50 it looks like an intra prediction issue, while avc-intra 100 just looks completely wrong. partly because it's 4:2:2
[11:54:02] <Dark_Shikari> avc-intra 50 is 10-bit
[11:54:05] <Dark_Shikari> ffmpeg can't decode that properly
[11:55:36] <Tjoppen> ah, that too
[11:55:44] <Tjoppen> 50 ought to work though
[11:55:51] <Dark_Shikari> why?
[11:55:53] <Dark_Shikari> 50 is 10-bit
[11:57:20] <Tjoppen> ah, I completely missed that
[11:58:35] <Tjoppen> so the artifacts I see are probably things like the lowest two bits being truncated accumulating to visible errors?
[11:58:59] <Tjoppen> also, it's interesting that the decoder outputs something semi-sane
[11:59:16] <Dark_Shikari> not surprising at all
[11:59:19] <Dark_Shikari> its not truncation
[11:59:22] <Dark_Shikari> it's the opposite
[11:59:24] <Dark_Shikari> overflow/saturation
[11:59:30] <Dark_Shikari> 10 bits of range instead of 8
[12:01:52] <kshishkov> are you going to fix that?
[12:02:17] <Dark_Shikari> irock is working on it, and uploaded his tree to github the other day
[12:02:24] <Dark_Shikari> help is welcome.
[12:04:22] <Tjoppen> would you have a url to said repo?
[12:04:58] <Compn> spaam : nope, i asked a few ffmpeg-devel admins but none said they needed my help...
[12:05:09] <Compn> michaelni and bcoudurier are admins iirc
[12:07:13] <Dark_Shikari> https://github.com/irock/FFmpeg
[12:08:51] <Tjoppen> cool. thanks
[12:09:22] <Tjoppen> uhm, /MPlayer/incoming no longer exists - should samples go in /incoming instead?
[12:09:32] <kshishkov> of course
[12:09:40] <kshishkov> it was just moved up
[12:10:03] <Tjoppen> ok. the site should be updated
[12:10:36] <kshishkov> flame Diego
[12:11:17] <Tjoppen> or I'll just put a patch on the ml for it
[12:11:50] <Tjoppen> haha, awesome. avid outputs 10-bit 4:2:2 as normal 8-bit uyvy with the low-order bits stored separately
[12:12:50] <Compn> i asked DonDiego to make symlinks
[12:12:54] <Compn> but that didnt happen i guess ;\
[12:12:57] <DonDiego> Tjoppen: which url did i forget to update?
[12:13:18] <DonDiego> i grepped everything for MPlayer/incoming i think
[12:13:45] <Tjoppen> http://ffmpeg.org/bugreports.html
[12:13:57] <Tjoppen> "cd -> /MPlayer/incoming"
[12:15:47] <DonDiego> bcoudurier: why are you still asleep at the wheel and gabu is still not banned or moderated?
[12:15:58] <DonDiego> sorry to bring this up in public, but this cannot continue
[12:16:14] <Dark_Shikari> I already complained
[12:16:17] <Dark_Shikari> I complained repeatedly
[12:16:19] <Dark_Shikari> nothing has happened
[12:16:24] <Dark_Shikari> our ML admins are incompetent
[12:18:06] <Tjoppen> samples uploaded to /incoming/avcintra
[12:19:22] <DonDiego> Tjoppen: oops, i committed locally but forgot to push, fixed - thanks for noticing
[12:19:43] <Compn> DonDiego : what about symlinks from MPlayer/incoming to /incoming ?
[12:20:00] <Compn> yes / no ?
[12:20:17] <DonDiego> i see little point now that the docs are updated
[12:20:40] <Compn> did you update every issue on roundup and bugzilla ?
[12:20:46] <Compn> containing the old url
[12:20:50] <DonDiego> no, why should i?
[12:21:02] <DonDiego> incoming is not something you can link to
[12:21:37] <Compn> ftp://upload.ffmpeg.org/MPlayer/incoming ?
[12:21:39] <Compn> what ?
[12:22:00] <DonDiego> you cannot link to files in there
[12:22:21] <Compn> but you can link users to upload files there, and thats what carl has been doing for ages in the tracker
[12:22:44] <DonDiego> well, now you send them to the new url...
[12:23:02] <Compn> is there something technically wrong with doing a symlink ?
[12:23:11] <DonDiego> no
[12:23:20] <DonDiego> but i want to actively deprecate the old location
[12:23:34] <Compn> it is actively depreciated
[12:23:38] <Compn> since you removed the urls
[12:23:54] <Compn> but to keep our old users afloat, who maybe wont check the new url, it might be nice to keep an old symlink alive
[12:26:27] <DonDiego> i don't see any users being misled
[12:26:52] <Compn> are you looking at ftp logs for MPlayer/incoming ?
[12:27:14] <DonDiego> those old bugtracker urls were relevant when they got posted, but i don't think people follow them anymore
[12:27:20] <DonDiego> i can do that later
[12:28:12] <Compn> its just a good idea to have a symlink for a dir that we've used for incoming for 5+ years. cant say i didnt try...
[12:28:39] <DonDiego> i'll look at the logs later
[12:28:45] <DonDiego> ping me next week please
[12:28:54] <DonDiego> i need to pack bags for fosdem now and run out
[12:32:35] <Compn> bcoudurier : are you still accepting paypal for donations? i thought you stopped but now i cant remember. at least BBB seems to think you still are.
[12:38:44] * elenril does an EvilLaugh
[12:38:53] * elenril now officially knows quantum field theory
[12:39:06] <cartman> it'll all be obsolote later on :P
[12:40:01] <_av500_> its all meta knowledge...
[12:40:09] <elenril> time to destroy the world
[12:41:43] <cartman> michaelni: if this flames don't end I'll use your mails to train GMail spam filter
[12:41:46] <cartman> seriously
[12:42:00] <Dark_Shikari> good idea
[12:44:25] <kshishkov> hmm? simple procmail rule on "leader.*" would suffice
[12:52:45] <Dark_Shikari> how do I send just one patch for git send-email?
[12:52:49] <Dark_Shikari> HEAD~2 sends the last two commits
[12:52:54] <Dark_Shikari> I want to send the commit "HEAD~2"
[12:52:54] <elenril> git send-email -1
[12:52:59] <Dark_Shikari> not HEAD~1
[12:53:13] <mru> git send-email -1 $refspec
[12:53:27] <mru> e.g. git send-email -1 HEAD~2
[12:54:17] <elenril> won't that send HEAD~3?
[12:54:30] <mru> not iirc
[12:54:42] <Dark_Shikari> BBB: I confirmed that gcc doesn't do that
[12:55:27] <elenril> mru: tested, it sends HEAD~3
[12:55:43] <mru> that's absurd
[12:55:44] <BBB> Dark_Shikari: oh shit, that's bad, ok well ok more reason to apply then I guess
[12:55:51] <elenril> mru: why, it's what i'd expect
[12:56:02] <CIA-38> ffmpeg: Jason Garrett-Glaser <jason at x264.com> master * r79dec1541b ffmpeg/libavcodec/vp8.c:
[12:56:02] <CIA-38> ffmpeg: VP8: faster deblock strength calculation
[12:56:02] <CIA-38> ffmpeg: Convert hev_thresh logic to a LUT, simplify mbedge_lim calculation.
[12:56:03] <CIA-38> ffmpeg: Jason Garrett-Glaser <jason at x264.com> master * r8a2c99b486 ffmpeg/libavcodec/vp8.c: VP8: slightly faster loopfilter sharpness logic
[12:56:04] <elenril> err..wait, i'm wrong
[12:56:06] <BBB> Dark_Shikari: does it change bla*2 to bla+bla?
[12:56:06] <elenril> nvm
[12:56:06] <mru> elenril: why would you expect that?
[12:56:14] * elenril fails at basic math :)
[12:56:27] <elenril> (though i have a good excuse)
[12:56:32] <Dark_Shikari> BBB: yes, of course
[12:56:45] <Dark_Shikari> gcc is just dumb
[12:56:53] <mru> <insert sarcasm about quantum fields>
[12:56:57] <Kovensky> elenril: "well, it's correct, for low enough values of 3"?
[12:57:26] <elenril> mru: i'm allowed to say stupid things for the rest of the day after a hard exam ;)
[13:00:58] * pJok turns on his quantum mechanics generation unit
[13:26:07] <_av500_> geez this train is full of fosdem
[13:28:08] <vipw> don't worry, it washes off
[13:28:36] <jannau> _av500_: it was the obvious connection for the beer event
[13:30:58] <thresh> :(
[13:34:55] <pJok> _av500_, you wouldn't happen to know which init file mounts the sd card under android?
[13:35:36] <_av500_> pJok: vold
[13:36:00] <_av500_> the volume deaemon
[13:37:10] <pJok> ahh
[13:37:35] <pJok> im pondering on if i can get it to initialize earlier so i can actually use it for internal storage
[13:43:05] <jannau> av500: where are you in ice 14? Can I bring anything (coffee + cake) from the station?
[13:43:30] <pJok> _av500_, seems like you need to recompile the kernel for it... damnit
[13:47:53] <_av500_> jannau: car 36
[13:48:12] <mru> long train...
[13:48:37] <jannau> mru: they start at 20
[13:48:44] <_av500_> jannau: black tea please
[13:49:31] <jannau> ok
[13:49:40] * cartman notes that _av500_ is more Turkish than anything else
[13:51:27] <jannau> I'm already in car 36
[13:53:15] <_av500_> jannau: i hope they dont decide to merge the cars physically
[13:53:37] <_av500_> cartman: not sugar with a drop of tea in it
[13:53:53] <cartman> thats me
[13:53:54] <cartman> :p
[13:55:13] <jannau> _av500_: unlikely, different platforms
[13:55:42] <_av500_> jannau: watch DB :)
[13:57:41] <jannau> yes, the rheinbrücke just before the station is infamous for derailing trains
[14:08:37] <BBB> wbs: if that patch to issue 2583 works, submit it to ML please ;-)
[14:15:16] <Dark_Shikari> BBB: fun...
[14:15:23] <Dark_Shikari> latest ffmpeg and libvpx, gcc 4.3, win32
[14:15:26] <Dark_Shikari> libvpx: 34fps
[14:15:27] <Dark_Shikari> ffvp8: 48fps
[14:15:44] <Dark_Shikari> (゚∀゚)アハハ八八ノヽノヽノヽノ \ / \/ \
[14:16:05] <kierank> lol
[14:16:14] <Dark_Shikari> one wonders how libvpx could be so bad
[14:17:00] <mru> nih driven development
[14:17:13] <cartman> nice
[14:18:11] <kshishkov> I thought they copy ideas from us as well
[14:18:18] <Dark_Shikari> Total work put into vp8.c, not counting adding features that libvpx doesn't have (e.g. emu edge) in the past 6 months: a day or two
[14:18:28] <Dark_Shikari> total work put into libvpx decoding: who knows how many man months
[14:18:31] <Dark_Shikari> progress: libvpx still sucks
[14:18:39] <cartman> thats how the economy works :p
[14:22:40] <cartman> Dark_Shikari: what you are discarding that, FFmpeg already has basic & intermediate infrastructure for codec optimization
[14:23:01] <Dark_Shikari> not really relevant, vp8 doesn't use any of it
[14:23:02] <cartman> they possibly didn't want whole ffmpeg dependency for some reason
[14:23:14] <Dark_Shikari> except basic stuff like intreadwrite
[14:23:32] <cartman> no dsp function used?
[14:23:35] <cartman> none at all?
[14:23:36] <Dark_Shikari> sure -- vp8 ones
[14:23:50] <cartman> and those in turn are standalone? :)
[14:23:53] <Dark_Shikari> yes
[14:24:02] <cartman> oh well nicely done then ;-)
[14:24:03] <Dark_Shikari> the only dsp functions it borrows are intra pred
[14:24:05] <Dark_Shikari> which are dead trivial
[14:24:32] <Dark_Shikari> but still, the point is -- they have 6-10 years of an entire dev team on it
[14:24:49] <Dark_Shikari> we have 3 people, none of which were anywhere close to full-time for anywhere more than a week or three.
[14:24:50] <cartman> team of indians maybe
[14:25:02] <kshishkov> Chinese
[14:26:19] <pJok> Dark_Shikari, outsource to india!
[14:26:43] <kierank> pJok: they already learnt that lesson for mbaff
[14:26:54] <cartman> Google is a complex infrastructure
[14:27:05] <cartman> its not easy to find a reason for this libvpx thing for example
[14:27:38] <Kovensky> libvpx is not google's fault, at least not directly
[14:27:47] <Kovensky> it spent a lot more time in duc--on2's hands
[14:27:51] <pJok> cartman, at least android was set in motion for them to have another platform to make money off and put google search on
[14:28:08] <cartman> pJok: Android has its own share of problems too
[14:28:09] <Dark_Shikari> cartman: no, it's simple
[14:28:13] <Dark_Shikari> 1) incompetent coders
[14:28:27] <pJok> cartman, of course... java is one of them ;)
[14:28:29] <Dark_Shikari> 2) open source model (read: collaborative development) is orders of magnitude more efficient than closed
[14:28:40] <cartman> 1 cannot be easily solved
[14:28:47] <cartman> their recruitement process sucks
[14:29:06] <KotH> 1 can be easily solved: hire only competent staff
[14:29:18] <cartman> KotH: where competent is defined by*
[14:29:19] <cartman> ?
[14:29:22] <KotH> ok, for this the whole HR crew needs to be replaced by competent people
[14:29:22] <cartman> DNA?
[14:29:23] <cartman> :P
[14:30:04] <KotH> cartman: did you know what every hire at google has to be interviewed by an engineer working in the field where the hiree is been interviewed for?
[14:30:20] <cartman> KotH: of course, I did an interview with them
[14:30:22] <KotH> cartman: ie, at least the engineer gets an idea whether the guy is competent or not
[14:30:49] <_av500_> i would says their hiring works as they now have BBB
[14:30:52] <KotH> but: HR is so incompetent that most engineers refuse to waste their time with people totaly unfit for the job
[14:32:19] <cartman> well I can't care less as long as they maintain GMail
[14:34:34] <_av500_> maybe BBB can make it not mangle git patches
[14:35:56] * cartman watches dancing kame
[15:11:10] <jannau> this train is indeed full of fosdem, even the 1st class
[15:11:18] * jannau waves to _av500_ 
[15:11:34] <kierank> is there wifi on that train?
[15:11:36] <kierank> or 3g?
[15:13:09] <jannau> 3g but I guess we could easily reach full covered with mobile hotspots
[15:13:16] <jannau> coverage
[15:13:20] <kshishkov> av500 usually brings his own
[15:13:30] <cartman> leech his 3G
[15:14:50] * jannau sits next to another haiku dev
[15:17:37] <kshishkov> jannau: what? there are two of them?
[15:17:49] * _av500_ waves to jannau 
[15:17:52] <jannau> apparently
[15:17:58] <cartman> _av500_ is lagging
[15:18:22] <_av500_> cartma is slacking
[15:18:25] <_av500_> +n
[15:18:26] <cartman> see
[15:18:33] <cartman> _av500_: I am compiling man
[15:18:39] <cartman> make -j32 while I chat
[15:18:47] <jannau> hmm, just a single wlan on board, wpa2 encrypted, essid "lgf"
[15:18:56] <Flameeyes> cartman: have you not applied the 200loc patch?
[15:19:02] <_av500_> cartman: you dont need to rebuild android in order to install one app
[15:19:07] <cartman> Flameeyes: kernel one? :)
[15:19:11] <Flameeyes> yeah
[15:19:15] <cartman> _av500_: just in case!
[15:19:29] <cartman> Flameeyes: ah sure, thats how I can talk && make -j32 :D
[15:19:49] <cartman> _av500_: you never compiled Qt, it shows ;>
[15:19:54] <_av500_> Flameeyes: there is much more elegant solution in 2 lines of bash, no?
[15:20:00] <Flameeyes> _av500_: no...
[15:20:10] <_av500_> cartman: qte2.x all the time :)
[15:20:15] <Flameeyes> there is a "wannabe elegant" solution that shown that whoever came up with the cgroup idea was on crack
[15:20:23] <jannau> _av500_: we could troll lennart tomorrow
[15:20:25] <cartman> _av500_: I've a minion over @ FOSDEM, ;P
[15:20:38] <Flameeyes> because the main "original" user of cgroups (lxc) is incompatible with what Lennart, Kay and Greg decided would be the official way in the kernel
[15:21:03] <Flameeyes> to the point that the whole shebang has a huge FAIL sticker over it from my pov
[15:21:37] <cartman> >>I get this error: “glibc detected”
[15:21:38] <cartman> wow
[15:21:48] <cartman> fatal error
[15:21:58] <_av500_> cartman: run!
[15:22:00] * cartman loves StackOverflow
[15:22:39] <kierank> cartman: you ask them to send you the codez?
[15:22:49] <cartman> nah the guy is right
[15:22:51] <cartman>  *** glibc detected *** ./a.out: free(): invalid next size (fast): 0x09f931f0 ***
[15:22:58] <cartman> how fucked up is that output
[15:23:02] <Flameeyes> very fucked up
[15:23:10] <cartman> glibc detected what
[15:23:15] <cartman> stupid Drepper
[15:23:15] <Flameeyes> but not as fucked up as the "repo" tool
[15:23:30] <cartman> Flameeyes: git fuckage ;>
[15:23:39] <Flameeyes> cartman: nono it's repo the fuckage
[15:23:47] <cartman> heheh
[15:24:14] <Flameeyes> "repo branches" -> first entry is "automake111"; "repo checkout automake111" -> "no project has branch automake111"
[15:24:15] <Flameeyes> wtf
[15:24:34] <cartman> thats a nice version of automake :D
[15:24:53] <Flameeyes> [that's 1.11, not 1.1.1 :P]
[15:25:51] <cartman> v111 was better :>
[15:26:48] <Kovensky> <@Flameeyes> to the point that the whole shebang has a huge FAIL sticker over it from my pov <-- yes it was, and the whole idea was copied from a BFS feature but shoehorned in whatever infrastructure the kernel already had that seemed appropriate enough IIRC
[15:27:19] <cartman> Kovensky, Flameeyes whats the shebang story?
[15:27:27] <Flameeyes> Kovensky: the huge FAIL sticker is not over the 200loc patch but over cgroups
[15:27:46] <Flameeyes> cartman: http://blog.flameeyes.eu/2011/01/10/cgroups-woes have fun :D
[15:27:58] <cartman> ah always full of information :D
[15:28:56] * _av500_ thinks Flameeyes is a team of blog writers
[15:29:30] <Flameeyes> _av500_: no I just lack a life at over 100mt from a keyboard :P
[15:29:44] <_av500_> 100 sounds to ,uch
[15:29:48] <_av500_> much
[15:29:56] <_av500_> that would allow you to go to a pub or so
[15:29:58] <Kovensky> Flameeyes: talking about your blag, when I was reading some older entries on your blag from google reader I kept getting error messages with this logo: http://puu.sh/Frs
[15:30:28] <Flameeyes> Kovensky: it should be fixed now, that's basically a webapp total failure
[15:30:47] <cartman> Flameeyes: ah so cgroups is the new sysfs
[15:30:49] <Kovensky> I enjoyed the irony from the logo :)
[15:31:11] <Flameeyes> yeah "just works" and rails don't go hand in hand
[15:31:45] <Flameeyes> fwiw http://blog.flameeyes.eu/2011/02/04/a-deji-vu-is-usually-a-glitch-in-the-rails also gave me a few more itches about Rails in general
[15:32:08] <Flameeyes> sooner or later I'll look into converting my blog to something else...likely going to try the JSP stuff again, at the time I tried it it was quite high-performance
[15:32:30] <Kovensky> lol JSP
[15:32:43] <Flameeyes> what's the alternative? :) Wordpress?
[15:32:44] <Kovensky> ever looked at Mojolicious?
[15:32:52] <cartman> Flameeyes: plain texy
[15:32:55] <cartman> text*
[15:33:11] <Flameeyes> Kovensky: oh god.. perl? :P I can't read that stuff!! >_<
[15:33:40] <Flameeyes> seriously my choice would end up in a language I can sort-of understand and write.. and I can read Java better than Python
[15:33:42] <Kovensky> and you go and decide to use... Java? >_<
[15:34:00] <cartman> Python is Read/Write unlike Perl
[15:34:15] <Flameeyes> cartman: my python sucks, and lu_zero can testify that
[15:34:56] <cartman> Flameeyes: use whatever works approach :-)
[15:35:19] <Flameeyes> yeah that's about it.. typo worked for me for a few years now...
[15:36:03] <Kovensky> < cartman> Python is Read/Write unlike Perl <-- blasphemy!
[15:36:28] <Flameeyes> guuuh Roller is now Apache?
[15:36:56] <cartman> after Sun went the way of well... Oracle
[15:38:23] <Kovensky> talking about sun / oracle, I remember someone saying that zfs wasn't GPL-licensed because the engineers refused the GPL
[15:38:37] <Kovensky> but why do people think that GPL is the only free software license? o_O
[15:38:38] <cartman> Engineers can't refuse shit
[15:38:40] <cartman> management does
[15:52:00] <kshishkov> Daemon404: sorry, no work on multichannel WavPack for you
[15:52:09] <Daemon404> lol, not here about that
[15:52:17] * Daemon404 sae the commit though.
[15:52:20] <Daemon404> s/sae/saw/
[15:53:18] <Daemon404> actually here to poke mru about some arm-related docs.
[15:53:54] <thresh> he's probably getting drunk already
[15:54:16] <kshishkov> maybe it's too early for him
[15:54:24] <kshishkov> thresh: BTW, where are you?
[15:54:31] <thresh> kshishkov: at home :( flu.
[15:55:31] <kshishkov> thresh: yes, mru was right saying two Konstantins for FOSDEM is too much
[15:56:02] <thresh> kshishkov: ha ha, there would be another Kostya
[15:56:14] <thresh> a friend of mine, mozilla-russia guy
[15:56:43] <thresh> suppose the universe decided three Kostyas is too much :)
[15:57:21] <kshishkov> thresh: then my University wouldn't exist - there were three Kostyas in my group there (including me)
[16:02:27] <cartman> have a nice weekend!
[16:08:49] <BBB> Dark_Shikari: ouch that's quite bad (for them :-p)
[16:10:11] <BBB> Dark_Shikari: I wonder if can beat them further after -mt
[16:10:18] <BBB> astrange: ping ping ping please rebase your patch to master
[16:10:34] <Kovensky> lol
[16:21:45] <Compn> Error occurred while sending the message:
[16:21:45] <Compn> 421 Connection rate too high, try again later [R0203001]
[16:22:12] <Compn> my isp's email server has been down 5+ hours
[16:28:43] <CIA-38> ffmpeg: Martin Storsjö <martin at martin.st> master * r1338dc0823 ffmpeg/libavformat/ (movenchint.c rtpenc_chain.c rtsp.c sapenc.c): (log message trimmed)
[16:28:43] <CIA-38> ffmpeg: libavformat: Use avcodec_copy_context for chained muxers
[16:28:43] <CIA-38> ffmpeg: This avoids having the chained AVStream->codec point to the same
[16:28:43] <CIA-38> ffmpeg: AVCodecContext owned by the outer AVStream. The downside is that
[16:28:44] <CIA-38> ffmpeg: changes to the AVCodecContext made after calling av_write_header
[16:28:44] <CIA-38> ffmpeg: cannot be detected automatically within the chained muxer.
[16:28:45] <CIA-38> ffmpeg: This avoids having to manually unlink the chained AVStream->codec
[16:30:57] <CIA-38> ffmpeg: Martin Storsjö <martin at martin.st> master * rf124b087ee ffmpeg/libavformat/ (avformat.h utils.c version.h):
[16:30:57] <CIA-38> ffmpeg: libavformat: Add a function for freeing an AVFormatContext
[16:30:57] <CIA-38> ffmpeg: This function is useful for freeing data structures allocated by
[16:30:57] <CIA-38> ffmpeg: muxers, which currently have to be freed manually by the caller.
[16:30:57] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[16:37:46] <spaam> Compn: ok :)
[16:39:17] <CIA-38> ffmpeg: Martin Storsjö <martin at martin.st> master * rff19748977 ffmpeg/doc/APIchanges:
[16:39:17] <CIA-38> ffmpeg: Add an APIchanges entry for avformat_free_context
[16:39:17] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[16:40:31] <CIA-38> ffmpeg: Martin Storsjö <martin at martin.st> master * rb22dbb291d ffmpeg/ (5 files in 2 dirs):
[16:40:31] <CIA-38> ffmpeg: Use avformat_free_context for cleaning up muxers
[16:40:31] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[16:41:39] <CIA-38> ffmpeg: Martin Storsjö <martin at martin.st> master * r397ffde115 ffmpeg/libavformat/rtpenc_chain.c:
[16:41:39] <CIA-38> ffmpeg: rtpenc_chain: Don't copy the time_base back to the caller
[16:41:39] <CIA-38> ffmpeg: If required, the caller can do this itself. ff_write_chained rescales
[16:41:39] <CIA-38> ffmpeg: timestamps as necessary, and all current callers of rtpenc_chain
[16:41:39] <CIA-38> ffmpeg: use ff_write_chained, making this timebase copy unnecessary.
[16:41:40] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[16:43:16] <CIA-38> ffmpeg: Martin Storsjö <martin at martin.st> master * r5306bf41a6 ffmpeg/libavformat/ (Makefile movenchint.c):
[16:43:16] <CIA-38> ffmpeg: movenchint: Use rtpenc_chain for setting up the chained RTP muxer
[16:43:16] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[16:46:58] <Daemon404> Compn, your first mistake was using your ISP's servers in the first place.
[17:10:49] <Compn> yes, i know
[17:10:51] <Compn> blame the victim
[17:11:57] <ohsix> yea, especially when one of the things your isp purports to do for your money is get you some mail
[17:23:21] <peloverde> quiet morning
[17:23:38] <CIA-38> ffmpeg: Martin Storsjö <martin at martin.st> master * r185a155e57 ffmpeg/libavformat/applehttp.c:
[17:23:38] <CIA-38> ffmpeg: applehttp: Handle absolute paths relative to the current server
[17:23:38] <CIA-38> ffmpeg: This fixes roundup issue 2583.
[17:23:38] <CIA-38> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[17:26:52] * elenril should've gone too
[17:28:40] <BBB> wbs: hi peloverde
[17:28:42] <BBB> ops
[17:28:51] <elenril> lol
[17:28:53] <BBB> that was a weird combination of two sentences in different timeframes
[17:29:00] <BBB> let's go back
[17:29:17] <BBB> wbs: can you merge the two checks in url_make_absolute() or somehow share code between the codepaths in there?
[17:29:22] <BBB> wbs: it looks like spaghetti now
[17:29:31] * elenril likes spaghetti
[17:29:32] <BBB> peloverde: fosdem, probably
[17:30:12] <peloverde> most likely
[17:34:51] <BBB> Dark_Shikari: which testsample for your patch from yesterday?
[17:37:44] <BBB> Dark_Shikari: quick test on sintel shows that it's significantly slower (!!)
[17:37:53] <BBB> 44 vs. 51 cycles
[18:00:54] <uau> Flameeyes: what's the problem with cgroups? that blog post doesn't seem to say much beyond some previously written software not working nicely with current API
[18:02:02] <Flameeyes> uau: beside that the "previously written" is the one for which cgroups were implemented to begin with
[18:02:38] <uau> well that still doesn't by itself identify any significant problem IMO
[18:03:19] <elenril> Flameeyes: written any book yet? ;)
[18:03:32] <Flameeyes> and that the "current API" was decided single-handedly without checking whether the rest of the software worked at all, while the preivous interface would have done nicely for both the old and the new software alike
[18:03:55] <Flameeyes> elenril: no, I'm not a native english speaker, publishers don't want me
[18:04:38] <elenril> how evil of them
[18:05:28] <uau> Flameeyes: so the changes caused problems for some software, still doesn't tell anything about possible problems with the current version
[18:05:35] * kierank wonder why readline does a double free
[18:05:37] <Flameeyes> ...
[18:06:14] <elenril> kierank: more free ⇒ more better
[18:26:11] * elenril wonders if it's possible to make RA addresses lower priority than DHCPv6 addresses
[18:26:35] <kierank> typical...readline doesn't double free under valgrind
[18:30:00] * _av500_ waves from fosdem
[18:31:17] * Tjoppen waves from umeå
[18:32:29] * siretart waves to fosdem
[18:48:08] <wbs> BBB: the if checks can't really easily be merged, they do similarlooking stuff, but not really similar enough to be factorized. if you're comfortable with c string handling, it's not all that bad :-)
[18:48:18] <wbs> BBB: thanks for applying anyway :-)
[18:52:49] <wbs> and _finally_ we got the muxer cleanup function... so that libav users don't suddenly start leaking memory just because libavformat started doing something it didn't do before
[18:57:44] <BBB> that's indeed good
[18:57:45] <BBB> much-needed
[18:58:28] <wbs> the memory leaks that I found the earlier days would have been fixed by that. and they've been running in production at $dayjob for a few months ;P
[18:59:04] <wbs> I usually valgrind it all when I upgrade ffmpeg there, but apparently I didn't valgrind the configuration that uses those muxers
[19:01:44] <kshishkov> well, is anybody going to review Xan4 decoder patch?
[19:02:40] <BBB> kshishkov: I guess I can...
[19:02:46] <BBB> what was the thradname again?
[19:03:08] <kshishkov> [PATCH 0/2] Origin Wing Commander IV decoder or something like that
[19:20:49] <kierank> is there any way of getting lavf to timeout if it can't open an udp stream
[19:22:15] <wbs> kierank: hmm, that might be useful. if we'd get easily settable protocol specific avoptions, that might be a good candidate for that. (we've got protocol specific avoptions already, they're just not easily settable from the command line)
[19:25:35] <j-b> michaelni: are you coming to FOSDEM?
[19:39:35] * elenril kicks debian's isc-dhcp-client maintainer
[19:40:16] <elenril> what an awesome idea to release with dhcpv6 client completely broken :?
[19:41:29] <michaelni> j-b, with the whole current mess and accusations and lies against me, i really prefer to stay far away from FOSS
[19:41:30] <ohsix> isn't stateless autoconfiguration the beans
[19:41:54] <michaelni> I thought many of the people to be my friends ...
[19:42:36] <Compn> oh someone mentioned michaelni had a wikipedia page!
[19:42:40] * Compn goes to look at that
[19:43:37] * michaelni is not in the mood to read more libel so ill not look at wikipedia today
[19:43:43] <Compn> ooh
[19:43:45] <Compn> yeah better not
[19:43:50] * Compn looks at ffmpeg page
[19:43:53] <Compn> yeesh
[19:43:55] <ohsix> is it libel if you never see it
[19:44:03] <Compn> shcrodingers libel
[19:45:56] <michaelni> ohsix, its is if i "see" it indirectly too, like someone else sees it and tells me or changes his behavior depending on it
[19:46:12] <michaelni> ;)
[19:46:27] <ohsix> unknowable things aren't worth fretting over
[19:46:37] <Compn> argh, isp news server still down
[19:46:40] <Compn> time to use gmane.
[19:46:45] <Compn> er isp mail server
[19:47:28] <ohsix> gmane is awesome; i read stuff there, getting main is for jerks
[19:49:20] * michaelni fails to find a wikipedia page about himself
[19:50:23] <Compn> yes, maybe it was a joke ?
[19:50:25] <Compn> hmm
[19:52:47] <michaelni> Anyway, if anyone catches someone spreading lies or libel about me, please tell me. I no doubt am inpolite and occasioanlly an asshole but some of the recent things IMHO go too far, i was not behind everything that did not work out like everyone wanted
[19:54:39] <ohsix> how is anyone supposed to know if it's a lie
[19:54:59] <michaelni> ask for evidence & facts
[19:55:13] <michaelni> mailing list postings irc logs whatever
[19:55:24] <ohsix> how do they know theres even a basis for asking evidence to be presented
[19:55:54] <iive> vmrsss ??
[19:56:44] <ohsix> reputations and names are like trademarks; got to defend them yourself if you register ;]
[19:56:47] <j-b> michaelni: well, I just thought it to be a cool place to find a solution
[19:57:36] <michaelni> j-b, how far is fosdem from vienna?
[19:57:51] <michaelni> in terms of time/money
[19:58:02] <michaelni> and from when to when is fosdem?
[19:58:18] <j-b> tomorrow and sunday
[19:59:07] <michaelni> thats a little narrow for me in terms of planing and packing and finding a cheap way to get there
[19:59:35] <j-b> 99euro to get there
[19:59:39] <Compn> maybe could get some kind of foundation to pay for that...
[19:59:47] <kshishkov> j-b: from where?
[19:59:54] <j-b> vinna
[19:59:55] <j-b> vienna
[20:00:06] <j-b> kshishkov: tomorrow, 9:40
[20:00:15] <michaelni> 99 with what? train airplane?
[20:00:23] <j-b> airplane, ofc
[20:00:24] <michaelni> 9:40 iam sleepimg
[20:00:31] <kierank_> lol
[20:00:37] <kshishkov> j-b: 9:40 is ETA for me
[20:00:46] <j-b> 21:10 20EUR
[20:00:57] <thresh> 99 euro for a plane flight, that's ridiculously low
[20:00:58] <j-b> kshishkov: 9:40am
[20:01:04] <BBB> wbs: https://roundup.ffmpeg.org/issue2586
[20:01:22] <j-b> kshishkov: 21:10, 20EUR from Vinna
[20:01:45] <j-b> +e
[20:01:57] <kshishkov> j-b: Ryan Airways? 120 EUR if you want to get by actual plane and such?
[20:02:17] <j-b> kshishkov: http://quicktrip.brusselsairlines.com, of course
[20:02:29] <j-b> and 125EUR to come back from BRU to VIE
[20:03:01] <michaelni> j-b, they ll kill me i dont need a flight back, just cheap life insurance
[20:03:11] <j-b> michaelni: they won't.
[20:03:41] <kierank> kshishkov: nothing wrong with ryanair
[20:03:54] <j-b> they are clever enough not to do so
[20:04:16] <michaelni> j-b, yeah it will look like an accident
[20:04:30] <j-b> michaelni: come on....
[20:04:36] <kierank> michaelni: nobody is going to kill you
[20:04:39] <j-b> michaelni: FFmpeg needs you
[20:04:41] <michaelni> iam joking
[20:04:49] * thresh loves ryanair
[20:04:51] <j-b> for the exact same reason than FFmpeg needs mru
[20:04:54] <michaelni> i know that but iam trying to find an excuse
[20:05:00] <j-b> and FFmpeg needs Dark_Shikari
[20:05:21] <j-b> michaelni: seriously, come. even just on the sunday.
[20:05:32] <j-b> We'll find a way to discuss.
[20:05:43] <j-b> FFmpeg is too important for this mess not to end;
[20:06:03] <Sean_McG> FWIW, I agree.
[20:09:42] <ruggles> can ffmpeg still add streams after read_header() if a packet for a new stream is encountered?
[20:10:15] <wbs> ruggles: iirc yes, there's a demuxer flag that says whether it usually does that, too
[20:10:50] <Compn> what happened last time michaelni went to a convention? it saved ffmpeg for a few years ?
[20:10:53] <Compn> ehe
[20:11:01] * Compn forgets what convention it was
[20:11:15] <ruggles> wbs: interesting. how does ffmpeg (the app) handle that?
[20:11:41] <michaelni> Compn, i was at a convention?
[20:12:00] <wbs> ruggles: I guess it keeps track of which ones have a decoder opened yet or not, if you get a packet for a stream that you haven't opened the decoder for yet, just open it (and decide what to do with it)
[20:13:18] <wbs> BBB: I've got code in progress for handling that case
[20:14:16] <skal> +1
[20:14:18] <elenril> michaelni: yes, but we erased your memory of it
[20:14:55] <michaelni> elenril, do you have proof? ;)
[20:15:49] <j-b> proofs can be erased :D
[20:15:59] <j-b> photo can be photoshopped
[20:16:32] <michaelni> j-b, you first need a photo from me ;)
[20:17:19] <j-b> pff...
[20:17:29] <j-b> Anyway, you should come.
[20:17:38] <j-b> and I can guarantee your safety and a normal debate
[20:17:50] <michaelni> j-b, if it was in walking distance i would
[20:18:00] <j-b> it is a plane distance
[20:20:12] <iive> how much time would it take with train?
[20:20:58] <BBB> michaelni: I'm pretty sure foundation can cover your flight expenses if you ask
[20:21:27] <michaelni> BBB 100 euro is no problem for me
[20:21:45] <michaelni> but iam not a traveller
[20:21:54] <BBB> expenses so you guys can sit in a nice restaurant and talk shit out
[20:22:01] <BBB> under j-b's guidance, I'm sure
[20:22:42] <j-b> I can make people talk
[20:22:44] <j-b> :)
[20:22:53] <BBB> j-b: wine does a lot of good stuff to people
[20:22:54] <ubitux> btw, no one has news from Jacob?
[20:22:56] <BBB> so does wodka
[20:23:05] <j-b> BBB: good idea, vodka
[20:23:05] <iive> it is much harder to avoid answer when talking face to face. at least without looking really bad.
[20:23:15] <michaelni> BBB wine causes headache at best
[20:23:21] <j-b> iive: and it avoids the misunderstandings of the ml
[20:23:32] <BBB> j-b: please show michaelni the difference between good wine and baf wine
[20:23:35] <j-b> michaelni: you haven't tested some good wine then;
[20:23:41] <BBB> :)
[20:23:48] <j-b> BBB: I have a very nice bottle here
[20:24:01] <j-b> but if I sneak it out, my g/f is going to kill me
[20:24:08] <BBB> from the ffineyard?
[20:24:40] <j-b> Chateau Margaux
[20:25:23] <j-b> michaelni: please, come;
[20:28:16] <thresh> hmm, 300 euro for a bottle
[20:28:36] <thresh> and that seems to be the cheapest one of the whole line
[20:28:54] <j-b> more or less, yes
[20:29:11] <thresh> you french are *weird*
[20:29:30] <j-b> I know
[20:29:52] <j-b> but, I'll live with it
[20:30:41] <kshishkov> j-b: but some Russians even commit suicide with your wine
[20:31:04] <Flameeyes> j-b: you french still have the curse of the peritél over your heads! >_<
[20:31:17] <j-b> Flameeyes: :)
[20:31:26] <j-b> Flameeyes: where are you now, btw?
[20:31:36] <Flameeyes> j-b: in my office as usual :|
[20:31:45] <j-b> .it ? veneto?
[20:31:53] <Flameeyes> playing with ruby-elf and finding bugs in elflickers
[20:32:01] <Flameeyes> yup, Venice inland (Mestre)
[20:33:14] <thresh> kshishkov: d
[20:33:21] <thresh> err
[20:33:34] <j-b> Flameeyes: lucky you
[20:33:41] <Flameeyes> j-b: not sure of that :)
[20:33:48] <thresh> kshishkov: you mean fill the whole pool with a wine like that and then drown there?
[20:34:48] <kshishkov> thresh: nope, IIRC last year some guy bought a bottle of some young French wine for $50000 and drank it. It was effectively poison
[20:38:29] <siretart> michaelni: I'm flying from nuernberg, and it's about 200EUR from here. 120EUR from vienna seems increadibly cheap to me. If you can, I'd love to meet you in bruessels
[20:39:00] <j-b> michaelni: what can I say/do to make you come?
[20:39:09] <j-b> michaelni: what can I say/do to make you write a MVC decoder? :D
[20:39:16] <j-b> should I kill 2 kittens?
[20:39:33] <kshishkov> 3 goats
[20:39:44] <Flameeyes> siretart: hey since you're around.. how's going with OpenVZ? :) /me longing for IPv6
[20:39:46] <j-b> I can do that
[20:40:06] <Sean_McG> I heard the last IPv4 block was handed out a few days ago.
[20:40:24] <Sean_McG> it's the beginning of the end
[20:40:41] <wbs> good thing we've got quite good IPv6 support in ffmpeg nowadays
[20:40:46] <iive> j-b: try to bribe him with 2000€
[20:40:59] <j-b> iive: to come? or for MVC?
[20:41:03] <Flameeyes> Sean_McG: my ISP still rather provide me with 6 IPv4 than giving me an IPv6 subnet... it's not yet the beginning of the end
[20:41:03] <iive> mvc
[20:41:12] <Anaerin> Sean_McG, A friend of mine has 3 /16's for sale, if you need IPs.
[20:41:16] <Sean_McG> Flameeyes: the providers are the problem.
[20:41:24] <j-b> iive: then, it is fucking cheap
[20:41:36] <j-b> iive: even I, can pay it
[20:42:27] <Sean_McG> anyways, I'mma shut down my VMs, back 'em up and then take a nap
[20:42:29] <Tjoppen> new regular ordinary swedish meal time \o/
[20:42:35] <Sean_McG> you folks take it easy, enjoy FOSDEM
[20:42:58] <iive> j-b: well, it's michael's final word.
[20:43:08] <j-b> iive: sure...
[20:43:21] <j-b> but I believe than even I can pay 3 time that price
[20:43:24] <iive> btw, what is mvc?
[20:43:34] <siretart> Flameeyes: good news and bad news: good news is that we do now have ipv6, but openvz's ipv6 support sucks. hard, so I fear we won't be able to offer this soon :-(
[20:44:06] <Flameeyes> uhm, ouch on that openvz support.. but what is it that you won't offer? openvz or ipv6 entirely?
[20:44:29] <thresh> ipv6 works nicely in openvz
[20:45:17] <siretart> thresh: unfortunately not with venet interfaces
[20:45:54] <siretart> and configuring individual bridges for each customer is out of question
[20:46:43] <siretart> Flameeyes: we are still trying to figure out what we can technically do. ipv6 with kvm for example is no problem and something that you can have right now
[20:46:55] <siretart> but we really should take this to /query, I guess
[20:47:54] <Flameeyes> sorry was just curious about the situation with that given the amount of hubhub over ipv6 lately, nothing really important now :)
[20:48:05] <iive> j-b: stereoscopic 3D ?
[20:48:07] <thresh> siretart: they should work
[20:48:08] <kierank> lu_zero: is cancelling the thread lavf is running in the way to stop udp stream opening from blocking?
[20:48:14] <thresh> once i'm at work, will test
[20:48:15] <j-b> iive: yes
[20:48:33] <j-b> iive: not to mention a correct 2D-3D filter in GPL
[20:49:36] <siretart> thresh: I did experiment with that. it totally messes up the host routing and in my tests was rather unreliable. veth interfaces OTOH work just nice
[20:49:53] <j-b> wbs: how do I know if FFplay will be able to seek in a file?
[20:50:26] <thresh> siretart: I see
[20:50:35] <wbs> j-b: hmmm.. there's like 2-3 different methods for seeking in ffplay
[20:50:41] <kshishkov> j-b: avc intra please
[20:50:54] <kshishkov> j-b: and please help Peter with Bink-b
[20:50:59] <j-b> kshishkov: lol
[20:51:11] <j-b> wbs: nice.
[20:51:27] <kierank> j-b: well someone's already working on avc-intra 50 (I think that's the one with 10-bit precision)
[20:51:52] <kierank> i mean kshishkov
[20:52:02] <j-b> wbs: av_seek_frame
[20:52:28] <kshishkov> kierank: iKnow
[20:53:01] <j-b> I really should setup a bounties page for codecs on videolan.org
[20:53:12] <wbs> j-b: ffplay seems to use avformat_seek_file actually
[20:53:21] <kierank> j-b: i suggested something like kickstarter.com for ffmpeg
[20:54:05] <kshishkov> j-b: what, you're playing for each .dll ?
[20:54:39] <j-b> kshishkov: ?
[20:55:17] <kshishkov> well, codec bounty == bounty for codec == bounty for codec even in DShow .dll?\
[20:55:35] <j-b> no
[20:55:46] <kshishkov> only VfW dll?
[20:55:58] <j-b> feature bounty = bounty for codec = codec inside lavc for GPL
[20:57:06] <Sean_McG> wow...Amazon is amazing... I ordered Dr. Who seasons 3 & 4 on Wednesday afternoon, they already arrived today.
[20:57:43] <elenril> beware, they're watching you
[20:58:00] * kierank watches elenril
[20:58:02] <elenril> they knew you'd order it so they sent it sooner
[20:58:11] <Sean_McG> eheheh
[20:58:17] <kshishkov> maybe they were just very happy to get rid of it
[20:58:54] <pJok> kshishkov, http://pucko.se/Images/ProductImages_200/17287.jpg
[21:00:16] <Sean_McG> what is that, chocolate milk?
[21:00:55] <pJok> swedish chocolate milk
[21:01:06] <pJok> which isn't really true, since its made in denmark
[21:45:51] <wbs> Flameeyes: the lscube bugzilla db seems to be down, in case you didn't know
[21:46:13] <Flameeyes> wbs: thanks... not sure if I have a clue on what my access on that box should be but will check
[21:46:18] <Flameeyes> since I guess lu is in bruxelles right now
[21:46:29] <kshishkov> BBB: thanks for review
[22:10:22] <ruggles> pross-au: any insight on issue 2556?
[22:13:22] <pross-au> mmm
[22:14:14] <pross-au> for some samples, the stream_guid info is wrong, therefore the preference given to stream2_guid chunks
[22:14:43] <ruggles> that seems ok. but some streams seem to not have stream_guid info at all.
[22:14:55] <pross-au> oh wow
[22:15:14] <pross-au> downloading..
[22:16:29] <ruggles> if you change line 801 in wtv.c to accept negative stream index you'll see lots more streams. but most are empty (no data ever sent with that sid).
[22:17:08] <pross-au> Yes
[22:17:10] <ruggles> but in this case the ac3 stream is not empty. and in one of the other samples there is a non-empty stream but it's just silence.
[22:17:46] <pross-au> ah
[22:17:50] <pross-au> interesting
[22:20:05] <BBB> kshishkov: byand-large it looks good to me, all I could see is some potential security issues, but nothing big
[22:20:17] <BBB> kshishkov: so on next review I'll probably ok, wait for 1-2 days and commit
[22:20:23] <BBB> unless someone else wants changes
[22:20:39] <ruggles> pross-au: maybe the stream_guid streams are supposed to be primary, but something about the way that file was recorded marked the mp2 stream as primary instead?
[22:21:28] <Sean_McG> even though it's described as commentary for the visually impaired?
[22:22:00] <ruggles> dunno... maybe whoever recorded it set some setting in the capture software to choose that stream?
[22:22:37] <pross-au> ill test it with the windows7 SBE
[22:24:11] <BBB> fok.nl
[22:24:14] <BBB> oops
[22:24:17] <BBB> wrong window :-p
[22:25:58] <BBB> I should build something into this chat client so command-T switches to chrome and then adds a new tab
[22:47:41] <kshishkov> BBB: updated. And just in case - I can't commit it anywhere except my local git repo
[23:03:14] <astrange> BBB: will attempt to. which part failed?
[23:03:34] <BBB> I think particular parts of #1 and a later patch had already been applied
[23:03:41] <BBB> kshishkov: I'll commit for you
[23:07:01] <kshishkov> BBB: thanks
[23:13:42] <pross-au> Q: say my format has 4 audio streams, how do i mark one (say the 3rd one) as 'primary'
[23:14:28] <Sean_McG> isn't that container dependant, with some containers not supporting it?
[23:15:09] <ohsix> he probably knows that; just wants to know how to flag it
[23:15:14] <kshishkov> pross-au: add it first I think
[23:15:31] <pross-au> kshishkov: fuck, i had hoped not..
[23:16:08] <kierank> how do i use av_close_input_file as part of a thread cleanup?
[23:16:08] <kshishkov> just don't care about it
[23:16:54] <kshishkov> pross-au: I'm not sure if we have primary stream definition at all and user may not care as well
[23:17:04] <pross-au> Ahah. We have AVStream->disposition
[23:17:58] <pross-au> the problem is, there is a wtv file with a 'silence' audio stream first, and the the actual audio stream
[23:19:55] <pross-au> i had re'd an 'ignore this stream' flag in the format, but turns out not to be reliable
[23:20:11] <kierank> pross-au: is there no language code?
[23:20:41] <pross-au> both streams say 'Eng'
[23:20:45] <kierank> hmm
[23:20:48] <pross-au> indeed!
[23:21:02] <kshishkov> what? not reliable binary format from M$? Impossible!
[23:21:28] <kierank> there's two ways of flagging audio description in mpegts. i guess your file uses the second way. flag in the iso-639 descriptor
[23:21:28] <ohsix> there's always NE
[23:21:29] <pross-au> we must demand better kshishkov
[23:22:24] <pross-au> kierank: in wtv, the language code is embedded in an MPEG TS Descriptor (escapsulated in a chunk)
[23:22:55] <pross-au> this wtv stuff is holding up important IDCT stuff
[23:22:58] <kierank> in the 639 descriptor there's a flag
[23:23:03] <pross-au> oh
[23:23:04] <kierank> that's separate to the language
[23:23:26] <kierank> http://dl.dropbox.com/u/2701213/Specs/ISO_IEC_13818-1_2007_PDF_version_%28en%29.pdf
[23:23:28] <kierank> audio_type
[23:23:50] <kierank> some countries they use it. in britain we use the language code "NAR"
[23:23:53] <kierank> in*
[23:24:19] <pross-au> thanks for that
[23:25:26] <The_Tick> hey Yuvi
[23:27:26] <The_Tick> hi I'm the project manager for perian
[23:27:34] <The_Tick> we use ffmpeg and have for years
[23:27:49] <The_Tick> I'm seeing this "compromise" email on the mailing list
[23:28:17] <The_Tick> I'm wondering if we could help by finding hosting to help alleviate some tension possibly? :)
[23:28:56] <The_Tick> I have good contacts at a few hosting sites from having managed adium and growl
[23:29:11] <kshishkov> I fear it's not about that
[23:29:30] <The_Tick> oh no
[23:29:34] <The_Tick> what's it about?
[23:29:49] <The_Tick> maybe a almost impartial third party could help
[23:32:20] <kierank> not an easy situation to fix
[23:32:32] <kshishkov> well, judging from current emails one party demands are completely not acceptable for other party
[23:33:21] <The_Tick> oh that sort of thing
[23:34:12] <The_Tick> well what are the valid points on both sides?
[23:34:29] * The_Tick continues to read the emails, but there seems to be a lot of backstory
[23:35:27] <kshishkov> the story is simple - some developers grew tired of the way FFmpeg was developed and decided to restart it under new set of rules
[23:35:43] <The_Tick> not a fork but a reboot?
[23:36:06] <kshishkov> depends on whom you're listening to
[23:36:55] <kshishkov> "leader" believes it was illegitimate takeover and wants full control on infrastructure
[23:37:38] <_av500_> kshishkov: where are yozu?
[23:37:45] <kshishkov> mostly because git.videolan.org repository is not listed as main one on ffmpeg.org
[23:38:05] <kshishkov> _av500_: still at home. My train is in 3 hours
[23:38:09] <The_Tick> it's the git clone used by videolan for their custom patches right?
[23:38:18] <kshishkov> nope
[23:38:31] <kshishkov> it's git repo set up specially for FFmpeg transition to Git
[23:38:31] <_av500_> kshishkov: ah
[23:38:38] <The_Tick> ahh
[23:39:23] <kshishkov> and "leader" has admin rights here. And commits from there are posted to FFmpeg-cvslog still
[23:40:00] <The_Tick> what's the problem with not listing the repository used for testing?
[23:40:09] <kierank> there's also the issue that the "leader" is head of the foundation
[23:40:29] <kshishkov> kierank: not head
[23:40:36] <kierank> really?
[23:40:38] <kierank> what is he?
[23:41:14] <kshishkov> The_Tick: it's listed http://ffmpeg.org/download.html but not as the primary one
[23:41:40] <kshishkov> kierank: one of directors. Ben is the president IIRC
[23:41:57] <The_Tick> it's git, there is no main repo :)
[23:42:15] <kierank> kshishkov: i see
[23:43:17] <The_Tick> so that's it, it's not listed as the main repo, that's the argument?
[23:43:47] <kshishkov> partially
[23:43:56] <The_Tick> that seems easily resolved
[23:44:02] <The_Tick> what's the other problems?
[23:44:29] <_av500_> The_Tick: lots
[23:44:49] <kshishkov> something that looks to me like Michael wanting full control back again. Or almost full for starters
[23:44:56] <The_Tick> what I've found on the projects I've managed
[23:45:00] <The_Tick> is when we were all arguing
[23:45:18] <The_Tick> it was good to bring in someone who had no real benefit to either side, air it all, and brainstorm solutions that worked for everyone
[23:46:16] <The_Tick> ok so there's a control problem
[23:46:25] <The_Tick> any other problems?
[23:46:36] <kshishkov> well, to me it all sounded like "Return to me, I'll forgive everything but three people - I find that unacceptable - Me too - +1"
[23:47:11] <michaelni> kshishkov, you are misunderstanding something
[23:47:27] * michaelni didnt read the backlog ...
[23:47:32] <ohsix> third party gooood, people can ignore ill will but it'll still be there; and their perceptions are colored
[23:47:52] <kshishkov> michaelni: could be, that's why I point it's my view or opinion
[23:47:57] <kierank> The_Tick: I tried to be the third party for an evening
[23:48:06] <The_Tick> michaelni: essentially I'm trying to get to the bottom of the problems in a method that's worked well for me in the past so that I can feel comfortable with us continuing to use ffmpeg for perian
[23:48:59] <kierank> The_Tick: I don't think it'll be possible to solve these problems on irc
[23:49:08] <kierank> phone possibly but face-to-face much more likely
[23:49:17] <kshishkov> face-to-faces
[23:49:25] <The_Tick> kierank: I've solved harder problems on irc :)
[23:49:33] <ohsix> faec to faeces
[23:49:42] <The_Tick> whatever you guys have tried so far
[23:49:45] <The_Tick> has it worked?
[23:50:01] <kierank> The_Tick: social problems?
[23:50:02] <kshishkov> ohsix: yep, it's been that way so far
[23:50:05] <kierank> or technical problems?
[23:50:06] <The_Tick> kierank: yes
[23:50:08] <The_Tick> to both
[23:50:45] <The_Tick> michaelni: you said kshishkov misunderstood something
[23:50:48] <The_Tick> can you clarify what that is?
[23:51:23] <michaelni> "Return to me, I'll forgive everything but three people"
[23:51:26] <michaelni> thats nonsense
[23:51:51] <michaelni> I want noone to return to me
[23:52:10] <kshishkov> and nothing?
[23:52:42] <michaelni> I want to be able to work in a friendly environment without being draged down in some peoples philosophical views
[23:52:53] <pross-au> kierank: is the meaning of lang=NAR defined somewhere
[23:53:02] <The_Tick> michaelni: such as?
[23:53:15] <ohsix> mismatching philosophy, you want your own to supersede?
[23:53:40] <kshishkov> The_Tick: would you believe that people in other repo can say the same?
[23:53:47] <ohsix> i've seen a lot of words about troll driven development, it sounds like you don't like that; heh (but it's also all it runs on)
[23:53:52] <kierank> pross-au: it's audio description for the visually impaired (aka audio narrative). It's defined in the UK D-Book which costs about £10k to get
[23:53:55] <michaelni> philosophical views like "ohh no libmpcodecs, ohh you forget a whitespace ohh this must be capitalized ohh we want to remove the he264 encdoer
[23:54:12] <kierank> pross-au: other countries might have standardised it, i dunno
[23:54:22] <The_Tick> michaelni: I don't know that those are philisophical differences so much as how people like to view their code
[23:54:31] <iive> kshishkov: i think michael "demans" the current root team to don't be root anymore. I haven't seen anything about kicking them out of the project.
[23:54:34] <The_Tick> philisophical differences being things like
[23:54:40] <The_Tick> "we want to go bsd and not gpl anymore"
[23:54:44] <kshishkov> kierank: standards often cost as much in all countries
[23:54:58] <iive> The_Tick: how about "voting doesn't work" "meritrocracy is the way to go"?
[23:55:03] <michaelni> The_Tick, thats not a problem they cant even if they wanted
[23:55:18] <kierank> kshishkov: well some of the D-books like norway and italy are available for free
[23:55:21] <The_Tick> ok, hold on
[23:55:33] <ohsix> if you pick merit you have to follow through and rely on people not having motives
[23:55:33] <The_Tick> michaelni: let's flesh out the rest of your point here
[23:55:37] <kshishkov> iive: well, at least we have original source here to clarify
[23:55:40] <michaelni> The_Tick, whitespace nitpicking at the level that we had was philosophy not just normal
[23:55:41] <The_Tick> everyone else let him do that
[23:56:01] <The_Tick> michaelni: so you want people to be less pedantic?
[23:56:01] <Sean_McG> stupid polarizing crap like that
[23:56:02] <michaelni> also rejecting 3 times as many filters is not a technical decission
[23:56:04] <pross-au> kierank: thanks, thats plenty
[23:56:33] <michaelni> The_Tick, i want them to do the work if THEY want the formating to be so pedantic
[23:56:48] <The_Tick> michaelni: do you guys have a coding standards document?
[23:57:13] <michaelni> somewhere but reallity was much stricter in the end
[23:57:31] <The_Tick> so maybe to resolve that point, you guys could update and agree upon the changes
[23:57:35] <The_Tick> and stick to them
[23:57:41] <ohsix> cant' the typo/spelling/formatting crap be easily dismissed with a commit hook?
[23:57:48] <kshishkov> it is
[23:57:57] <kshishkov> except for typos
[23:58:26] <kshishkov> especially since certain guy can't even give an answer
[23:58:36] <kshishkov> only awnsers
[23:58:41] <ohsix> on lists i've read where people actually get things done theres usually 2-3 passes and they do include white space and other comments from people; theres no discussion over it, they roll another patch
[23:59:16] <kierank> kshishkov: the whole innofficial and unofficial thing was ridiculous


More information about the FFmpeg-devel-irc mailing list