[Ffmpeg-devel-irc] ffmpeg-devel.log.20141006

[burek]

[00:02] <rcombs> cygwin's gdb does no better, and I'm about 2 inches from compiling ffmpeg with cygwin
[00:02] <rcombs> actually, fuck it, what's the worst that could happen, I'll give it a go
[00:03] <Compn> rcombs : are you sure you are running the ffmpeg_g binary in gdb ?
[00:03] <Compn> dumb question but it happens :)
[00:03] <rcombs> Compn: quite certain
[00:04] <rcombs> and that it was configured with `--enable-debug --disable-stripping`
[00:04] <Compn> because gdb works on win32
[00:04] <Compn> usually 
[00:04] <Compn> what gdb did you install ?
[00:04] <Compn> gdb --version
[00:05] <Compn> you said two versions ? :\
[00:05] <rcombs> I've tried cygwin's 7.8, minnow's 7.6, and the 7.6 JEEB linked
[00:05] <rcombs> all give the exact same behavior here
[00:05] <pross> rcombs: what did you build ffmpeg with?
[00:05] <rcombs> mingw
[00:05] <Compn> what compiler you mean ?
[00:05] <rcombs> gcc
[00:06] <pross> that does not compute then
[00:08] <Compn> wasnt there an autobuild of ffmpeg_g somewhere ?
[00:09] <Compn> no i guess not
[00:09] <Compn> rcombs : cross-compile ffmpeg on linux and test htat binary :p
[00:10] Action: rcombs dies
[00:13] <cehoyos> rcombs: Did you test current git head? Could it be related to today's (yesterday's?) win32 2G file support?
[00:14] <rcombs> cehoyos: no, I'm on a fork
[00:18] <cone-408> ffmpeg.git 03Michael Niedermayer 07master:02f7665a3ce7: avformat/mov: do not overwrite extradata in mov_read_glbl()
[00:49] <J_Darnley> Did metadata ever get added to libavfilter filter graphs?
[00:50] <J_Darnley> I mean metadata from the input files/streams.
[00:55] <nevcairiel> you can pass side data in frames
[01:30] <cone-408> ffmpeg.git 03Michael Niedermayer 07master:f9fefa499f0a: swresample/swresample: fix sample drop loop end condition
[01:31] <cone-408> ffmpeg.git 03Michael Niedermayer 07master:6b347f519d53: swresample/swresample: replace always true if() by av_assert0()
[01:59] <cehoyos> Compn (et al): Did you already post this shitload yesterday?
[01:59] <cone-408> ffmpeg.git 03Michael Niedermayer 07release/2.2:a28c276b8dcc: swresample/swresample: fix sample drop loop end condition
[01:59] <cone-408> ffmpeg.git 03James Almer 07release/2.2:f253fa95529a: x86/synth_filter: add missing HAVE_YASM guard
[01:59] <cehoyos> Because if yes, I am terribly sorry that I didn't read it: I would have commented in that case!
[02:00] <cehoyos> Because this amount of dirty lies (from diry thieves - and yes, "dirty" *is* an offense!) makes is clear why all this discussion is a complete loss ot time.
[02:01] <cehoyos> Where did anybody ever apologize?
[02:02] <cehoyos> What rules were ever broken?
[02:03] <Compn> i posted it in the meeting yep
[02:03] <Compn> and diego thinks you would punch him if you met in real life :D
[02:03] <cehoyos> And could somebody please explain who asked (where) for money? Afair, avconv threw all the money out of the window just to make everybody life more difficult (except the ones who made profit).
[02:03] <cehoyos> Sorry for not reading it: I am sure everybody would have been delighted to read my comments yesterday!
[02:04] <Compn> yes , i was wondering where your comments were, thats why i've been making jokes at you this week :)
[02:04] <Compn> the comment about michael wants to be paid ? i said it. i think it was a mail michael wrote about bug bounties
[02:04] <wm4> cehoyos: you're one of the last ones who still insist on adding fuel to the flames
[02:04] <cehoyos> So he claims that he apologized because he fears I would punch him if he would not spread such lies? What kind of logic is this?
[02:05] <cehoyos> Really? Did you read the excerpt?
[02:05] <Compn> no, he claims to have apologized privately to everyone about the fork/split
[02:05] Action: wm4 votes for putting cehoyos and diego into a room and throwing away the key
[02:05] <Compn> lol
[02:05] <cehoyos> So he did apologize to you?
[02:05] <Compn> yes at last year vdd
[02:05] <Compn> and online
[02:05] <Compn> which is why i stopped caring about the fork and let it be
[02:06] <cehoyos> That's good, but in your excerpt it says that he apologized in public, where was that?
[02:06] <Compn> moved on with ffmpeg and not looked back
[02:06] <cehoyos> Please share the link!
[02:06] <Compn> that i dont know
[02:07] <cehoyos> And please note that I am surprised about "moving on": Yesterday (only?) past issues were discussed, or do I misremember?
[02:07] <cehoyos> What is it that you don't know?
[02:07] <Compn> i dont know where diego apologized in public. i didnt see it
[02:08] <cehoyos> Well, you say so in your excerpt: It is of course possible that it wasn't him who lied, but afaict somebody doesn't say the truth or am I wrong?
[02:08] <Compn> its not my excerpt , i didnt write it
[02:08] <Compn> i was there in the meeting though
[02:08] <cehoyos> So the writer lied? May I ask who it was?
[02:09] <Compn> diego did say whats in the transcript iirc , so he apologized again at the meeting
[02:09] <Compn> i'm not sure what you are asking here
[02:09] <kierank> do these apologies really matter
[02:10] <kierank> about things from 4 years ago
[02:10] <cehoyos> In the excerpt it says that he had already apologized in public: Afaik, this is (one more) blatant lie.
[02:10] <Compn> kierank : it matters to carl i guess
[02:10] <Compn> i've moved on
[02:10] <cehoyos> But perhaps I am wrong and you can explain when / where he apologized in public?
[02:10] <Compn> you'd have to ask diego where he did that carl
[02:10] <cehoyos> No, this is of course not about things that happened three years ago: It is about lies that were told last week.
[02:11] <Compn> cehoyos : but probably you should give up caring about diego
[02:11] <Compn> i think outside people do not care about fights between you guys
[02:11] <Compn> at least thats what was said at the meeting
[02:12] <Compn> by various downstream project developers
[02:13] <cehoyos> Well, I believe that Diego is a dirty swine (yes, this is another offense) who is a copyright violator, who is part of group of thieves who tried to takeover the FFmpeg project and who is spreading lies: I therefore definitely don't care about him. But unfortunately, you posted these lies on the FFmpeg bug tracker without any kind of additional information. Now I am curious how such things happen.
[02:13] <cehoyos> If nobody cares about these fights, why are you posting such bullshit without any explanations?
[02:14] <Compn> the same free speech that i uphold for you, i uphold for libav developers. i've stopped counting who is right and wrong 
[02:14] <wm4> (lol open source thieves)
[02:14] <Compn> (shhh lol)
[02:14] <Compn> i posted the transcript because it has various ideas and concerns by the larger community about us and libav
[02:15] <cehoyos> People who read this must believe that avconv patches are reviewed (we all know there are nearly no reviews there), that Michael requests money (I don't remember that) and that Diego ever apologized for anything. None of this is true to the best of my knowledge.
[02:15] <BBB> I need some way to save highlights...
[02:15] <BBB> my chat client is getting really buggy lately, it just loses everything
[02:15] <Compn> cehoyos : one sec i'll find michael money email
[02:15] <BBB> does anyone happen to have a way to find the last highlight for me?
[02:15] <Compn> BBB : try not to troll so hard ya know? :P
[02:15] <cehoyos> I know which mail you are taling about.
[02:15] <Compn> and i was the one who brought it up
[02:15] <Compn> i said that in the meeting. me
[02:15] <cehoyos> I am not a native speaker but I don't remember him asking for money.
[02:15] <Compn> brought up michaels email
[02:16] <Compn> erm
[02:16] <cehoyos> Maybe I need a translator.
[02:16] <kierank> BBB: 
[02:16] <kierank> 11:19 AM <"ubitux> pa: i think BBB did the original matroska demuxer
[02:16] <Compn> maybe i misread
[02:16] <BBB> kierank: ty!
[02:16] <Compn> cehoyos : https://www.mail-archive.com/ffmpeg-devel@ffmpeg.org/msg02304.html
[02:17] <cehoyos> This doesn't say (not even remotely) "Michael was requesting to be paid for his work on ffmpeg."
[02:18] <Compn> At least as far as iam concerned, it certainly would
[02:18] <Compn> motivate me to get _some_ payment for what i do 7 days a week
[02:18] <Compn> and surely the 0$ for FFmpeg (not counting the occasional consulting
[02:18] <Compn> stuff) is not going to work out forever for me
[02:18] <Compn> though thats drifting a bit off topic ...
[02:18] <Compn> cehoyos : what does this mean ^^
[02:19] <Compn> we could always ask michaelni , i could be wrong.
[02:19] <Compn> i didnt mean to insult anyone of course
[02:19] <cehoyos> In the context of the email, it doesn't (not even remotely as said) mean "I request to be paid"
[02:20] <Compn> could you explain to me what it means then ?
[02:21] <iive> who won't be happy to be paid for doing his hobby?
[02:21] <Compn> i'd be happy to be paid for doing this ffmpeg stuff :)
[02:21] <cehoyos> I don't think "insult" is the right word (it is not illegitimate to ask for money). But mentioning in a discussion that starts with the (wrongful) claim that Michael didn't abide to the rules that he asks for money puts FFmpeg in nearly as bad a light as my little outbreaks here.
[02:21] <iive> honestly, he says "for what i do 7 days a week".
[02:22] <Compn> wheres the claim that michael didnt abide to some rules ?
[02:22] <Compn> i think you misread the transcript
[02:22] <iive> url?
[02:22] <Compn> iive : http://paste.ubuntu.com/8488924/
[02:23] <cehoyos> (The unusual word "abide" is only once in the text.)
[02:23] <Compn> "abide" isnt in the transcript 
[02:23] <cehoyos> The whole deal started with denoting Michael because he was not abiding the rules. <-- We all know this is a lie, it was already explained often enough.
[02:23] <Compn> oh abiding
[02:23] <cehoyos> But it is not a lie spread three years ago, but (apparently) last week.
[02:23] <Compn> let libav live in the past
[02:24] <Compn> and we (ffmpeg) can live in the future
[02:24] <cehoyos> And of course a code of conduct makes not much sense if it is only meant to be valid for FFmpeg but not for avconv.
[02:24] <Compn> lol
[02:24] <cehoyos> Ok: Why are you posting such links, I don't understand that if I read your comments here.
[02:25] <Compn> [20:14] <Compn> i posted the transcript because it has various ideas and concerns by the larger community about us and libav
[02:25] <cehoyos> (And sorry, llogan already explained that yesterday using much better wording.)
[02:26] <Compn> about lucas' comments about 2011, yeah. i dont care anymore. let them say whatever they want. people will judge libav and ffmpeg and decide who is right or wrong
[02:26] <Compn> you calling people theives and liars all the time isnt helping, at least in my opinion.
[02:26] <cehoyos> But you posted the usual lies without any explanations on a page that is also read by people who have only limited (or no) knowledge about what happened: They will not only believe that avconv reviews patches but also that the development pace is high (is "high" correct as an interpretation?)
[02:27] <Compn> but ... like i say, i'm for free speech. so you go right on doing what you want.
[02:27] <iive> "Luca Barbato: What's libav? Well, we started with mainly a technical disagreement on how to lead a project. The whole deal started with denoting Michael because he was not abiding the rules."
[02:27] <Compn> again, let them find out themselves. i dont care
[02:27] <iive> Compn: what have you done?
[02:27] <wm4> if the Libav development pace wasn't high, the continuous merging wouldn't put stress on ffmpeg
[02:27] <cehoyos> I am also for free speech but I really don't understand why you are posting such articles as if you would agree with everything it says?
[02:28] <Compn> cehoyos : thats actually funny what you are saying
[02:28] <cehoyos> ?
[02:28] <Compn> because, diego says that by ffmpeg not disagreeing with you calling them swine, it means we agree with you
[02:28] <Compn> tacit agreement
[02:28] <Compn> cehoyos : Diego: one point for nicolas, it's very improtant that you don't just have it, but you enforc it.
[02:28] <Compn> If you just ignore this kind of behavior then it's a tacit form of agreement.
[02:28] <cehoyos> I don't think the FFmpeg developers agree with me (I am actually quite sure at least some don't).
[02:29] <Compn> thats why i think its funny you say it
[02:29] <Compn> because you and diego agree on the "tacit agreement" thing
[02:29] <Compn> (anyone want to say anything here?)
[02:29] <cehoyos> I am not sure how Diegos interpretations (that I couldn't care less about) are related to you posting links to articles (I thought!) were written by you.
[02:30] <Compn> its a transcript written by someone who i'm not sure i should post their name. as you may go and bother them on email...
[02:30] <cehoyos> No, you misunderstand: I thought you *wrote* the article.
[02:30] <Compn> ok
[02:30] <cehoyos> And I believe that this is how somebody else would interpret your link on trac.
[02:31] <cehoyos> As said: I expected them to claim stuff like that and I not surprised (and this is why I am not sure what a code of conduct should be good for).
[02:31] <Compn> ok
[02:31] <Compn> yes , thats why most of us agreed in the meeting that a code of conduct would be ignored and useless
[02:31] <Compn> in #ffmpeg-meeting yesterday
[02:32] <cehoyos> I am just surprised why we post such claims on our bug tracker without at least saying that while some things are simply not true, others are at least disputed.
[02:32] <Compn> changing the subject
[02:32] <cehoyos> But much better would have been to just ignore it - as you suggested, but unfortunately you didn't "abide" (is that a correct usage of the word?)
[02:33] <Compn> yes its correct. and you are correct, by reposting this, i am technically bringing up the past because the past was brought up by them.
[02:33] <cehoyos> You of course don't have to tell anybody's name, we accept anonymous contributions (the difference is that we don't tell everybody we wouldn't accept them on the very same day).
[02:34] <Compn> cehoyos : have you emailed new libav contributors and tried to convince them that libav team was full of thieves and liars ?
[02:34] <Compn> because that claim was made...
[02:34] <Compn> but they didnt mention who emailed them
[02:34] <Compn> so i'm guessing here 
[02:35] <iive> "I've seen ffmpeg r0each out and insult people making libav patches, and I've seen libav people get upset about ffmpeg stuff"
[02:36] <cehoyos> No, I email many people and try to inform them about the true background of avconv, not specifically contributors.
[02:36] <Compn> maybe i used wrong word 'contributors' 
[02:36] <cehoyos> But as you know I did email Katherina (and iirc, I wished her all the best at least twice).
[02:36] <Compn> ok because i talked to her , i wasnt 100% sure
[02:36] <Compn> i'm not sure if that freaked her out or not
[02:37] <Compn> your emails might have freaked her out, i mean
[02:37] <cehoyos> But I am sure I didn't write anything to her that I don't regularly write on public mailing list so there are no secrets afair.
[02:37] <Compn> i'm just letting you know that your emails maybe freaking people out instead of your intended effect of teaching them the truth
[02:37] <wm4> I can confirm that mini complained to me when I sent a patch to libav once instead to ffmpeg (even though I'm only an occasional contributor) - such things might be taken quite sourly by Libav, even if mini doesn't necessarily have ill intentions with this
[02:38] <cehoyos> I finally found something important that I agree with and that very well reflects my feelings: "People spreading misinformation to outside people."
[02:39] <cehoyos> wm4: Was that the patch that caused a regression? Maybe because they forgot to review it?
[02:39] <wm4> they don't forget to review, and that patch was thoroughly tested
[02:39] <Compn> we've made insults against libav in this channel before, probably at Daemon404 , who has submitted patches to both
[02:40] <Compn> like i said, i stopped dwelling in the past
[02:40] <Compn> trying to anyhow
[02:40] <wm4> Daemon404 just insults everyone as he likes
[02:40] <Compn> now i need to stop insulting libav :D
[02:41] <pross> Compn: its hard not to!
[02:42] <iive> i have a rare talent/curse. I can insult people without using insults, just by telling the truth.
[02:42] <iive> can I still do that?
[02:42] <pross> iive: another way, some people are asking us to forget the truth
[02:43] <Compn> like i said, that meeting was very facepalm
[02:44] <pross> how were the nibbles?
[02:46] <wm4> Compn: how was the meeting
[02:46] <wm4> did they throw looks that can kill
[02:46] <wm4> or whatever that idiom is in english
[02:46] <iive> he wasn't there, he sent a drone instead.
[02:46] <cehoyos> iive: It appears I have the oposite talent: I just tell the truth and everybody thinks I wanted to insult somebody;-)
[02:47] <Compn> iive : i was there
[02:48] <iive> you came to europe \o/
[02:49] <Compn> wm4 : no, it was same as vdd13 meeting. everyone just talking normally.
[02:49] <Compn> wm4 : but libav really dislikes carl and his swine comments :D
[02:49] <wm4> I dislike these comments too
[02:49] <wm4> I don't know anybody who likes them
[02:49] <Compn> i should say everyone dislikes the insults
[02:49] <Compn> all developers and 3rd party people that were there
[02:50] <Compn> said it explicitly
[02:50] Action: wm4 stares at cehoyos 
[02:50] <Compn> lol
[02:50] <Compn> wm4 : i think 3 people asked for your real name though
[02:51] <Compn> 'who is wm4' 'no one knows'
[02:51] <iive> cehoyos: you are doing it wrong. e.g. swine is clearly an insult. now, pigs are actually quite smart animals. You are clearly insulting them. They have never done anything bad to the open source! So stop using that word.
[02:51] <wm4> not like my real name would add any information (but you can find it publicly on the internet if you look hard enough)
[02:51] <Compn> wm4 : i think i spent about 5 minutes looking for your name, but gave up because i realized i only remember irc nicknames anyway
[02:52] <wm4> lol
[02:52] <wm4> cehoyos: if you think you can't work with Diego because something money something, state that, but we don't need insults here
[02:53] <Compn> iive : i was in paris for vdd13 too. ;P
[02:53] <wm4> or we're going to have to say thing to them like "only cehoyos uses insults"
[02:53] <wm4> *things
[02:53] <Compn> wm4 : i actually asked carl if he would work with diego via email a few weeks ago
[02:54] <iive> Compn: I had no idea.
[02:54] <cehoyos> Of course "swine" was an insult, I used it three years ago and would have forgotten about it if I had not been told about it yesterday.
[02:55] <Compn> iive : europe is fun. i was thinking of visiting further east, maybe visit amsterdam or even further, but i didnt have enough time.
[02:55] <cehoyos> "something money something"? I don't understand (do you speak German?)
[02:55] <wm4> it doesn't mean anything
[02:55] <wm4> it's a placeholder basically
[02:55] <cehoyos> Then I don't understand...
[02:55] <iive> probably the bribe attempt.
[02:55] <Compn> cehoyos : wm4 means what problem do you have with diego
[02:56] <Compn> wm4 : and i think the answer is , the same problem he had in 2011 with diego 
[02:56] <wm4> I know what problem he has, I just don't care
[02:56] <Compn> ah
[02:56] <wm4> (anymore)
[02:56] <Compn> ok then
[02:56] <cehoyos> That he regularly (including the link Compn shared) insults me by claiming things about avconv and FFmpeg that are simply not true
[02:56] <cehoyos> (Like avconv doing "real" reviews and FFmpeg doing "less" reviews.)
[02:57] <cehoyos> iive: No, definitely not.
[02:57] <Compn> we just have to "turn the other cheek" and ignore these insults
[02:57] <wm4> duh, ubitux (who clearly is a FFmpeg developer) complained that Libav requires _too much_ reviewing
[02:57] <Compn> at least, thats what google says
[02:57] <cehoyos> No.
[02:57] <cehoyos> avconv requires cosmetic fixes, patches that were not written by Michael typically don't get any other review.
[02:57] <Compn> because responding to these insults makes us look "bad"
[02:58] <cehoyos> Please remember vc1: Crashed with *all* samples we had.
[02:58] <Compn> "and hurts ffmpeg's reputation"
[02:58] <wm4> cehoyos: did it crash avconv too?
[02:58] <wm4> or was it a botched merge
[02:58] <cehoyos> Please remember j2k: We are still suffering from this patch...
[02:58] <Compn> no, just ignore it and move on! :)
[02:59] <wm4> and of course blindly/semi-automatically merging changes from an essentially foreign codebase is going to hurt sometimes
[02:59] <cehoyos> Ther were approximately three "botched merges" in 3.5 years: I really, really don't understand why you are repeating this claim again and again.
[02:59] <wm4> although I think mini is doing an awesome job for keeping up
[02:59] <Compn> maybe j-b could explain his position on the fighting
[02:59] <cehoyos> Particularly now when development speed in avconv is nearly 0, and the merges are much easier now.
[02:59] <iive> wm4: is mini short of michaelni ?
[02:59] <wm4> iive: yes
[03:00] <cehoyos> (Of course I was only meaning the commits to the avconv codebase not the merges.)
[03:00] <iive> oh...
[03:00] <cehoyos> iive: It is actually an intended insult afair.
[03:00] <iive> Ow
[03:00] <wm4> mini? no it's not an insult
[03:00] <Compn> i also find it strange people shorten michael 's name in this way
[03:01] <cehoyos> No?
[03:01] <wm4> no
[03:01] <cehoyos> No?
[03:01] <wm4> no
[03:01] <Compn> i dont see you saying luba or DoDi 
[03:01] <Compn> or AnKi
[03:01] <wm4> Compn: luca
[03:01] <cehoyos> Yes, it was meant by the avconv developers as an insult, it may not be interpreted as such by Michael, I don't know.
[03:01] <wm4> hell even cehoyos' nick has his first two names abbreviated with 1 letter each
[03:02] <cehoyos> Yes, it is my normal user name that I chose
[03:02] <Compn> wm4 : Luca Barbato is to "LuBa" as michael niedermayer is to "mini"
[03:02] <cehoyos> Michael chose michaelni afaik
[03:02] <iive> or they could have used initial, mn
[03:02] <Compn> and yet i've never seen someone say 'luba'
[03:02] <wm4> "michaelni" is also acceptable, but his full name is obviously too hard to write
[03:03] <wm4> so of course people are going to use abbreviations
[03:03] <cehoyos> iive: I believe you are right, this was also used in the past
[03:03] <Compn> like i said, i ignore all this crap
[03:03] <cehoyos> wm4: I honestly wonder: Do you don't know it or do you don't understand it?
[03:03] <wm4> cehoyos: ?
[03:04] <cehoyos> The abbreviation was chosen as a possible insult (as said, it may not be interpreted as such)
[03:04] <cehoyos> compn: JB is responsible for the j2k mess...
[03:04] <wm4> seems like you're trying hard to interpret insults into everything
[03:05] <cehoyos> Really?
[03:05] <Compn> cehoyos : like i said, move on already. no one cares :P
[03:05] <cehoyos> Please try to find the first usage of this shortage of Michael's name.
[03:05] <cehoyos> Apparently people do care: They spent a whole meeting to wine about possible insults.
[03:06] <Compn> lool
[03:06] <iive> yeh.
[03:06] <Compn> yeah well 
[03:06] <wm4> michaelni: is calling you "MiNi" an insult? [ ] yes [ ] no
[03:06] <cehoyos> Not I am quite surprised (and feeling insulted): Why did nobody protest about the insults from the avconv developers?
[03:06] <Compn> cehoyos : stefano, thilo , reimar and me were there 
[03:06] <wm4> who's thilo again?
[03:06] <cehoyos> wm4: You misunderstand - Michael may not care about it - it was *meant* as an insult.
[03:07] <Compn> wm4 : i dont think he is on irc , at least..... i dont remember
[03:07] <Compn> __gb__ was there too
[03:07] <cehoyos> Thilo was an author of the MPEG4-ALS decoder.
[03:07] <Compn> __gb__ : its all your fault! 
[03:07] <Compn> ehe
[03:07] <Compn> cant remember if Daemon404 was there this year or if i'm confusing him with someone else
[03:07] <cehoyos> Compn: I know but I don't understand you: Is this an answer to something / to what?
[03:08] <cehoyos> He was there
[03:08] <Compn> cehoyos : you were asking why we didnt object to avconv statements
[03:08] <Compn> i was telling you who was there to possibly object. but i cant speak for others
[03:09] <BBB> oh there was a meeting, now I get it
[03:09] <BBB> is there a log/summary?
[03:10] <wm4> BBB: summary: "everyone is mad"
[03:10] <BBB> everyone in ffmpeg?
[03:10] <Compn> BBB : http://paste.ubuntu.com/8488924/
[03:10] <BBB> \o/
[03:10] <BBB> ty
[03:10] <Compn> there was also an #ffmpeg-meeting yesterday, but doesnt include what we're talking about
[03:12] <Compn> merbanan was there, i think?
[03:12] <Compn> cant remember if he came late or left early
[03:12] <Compn> j-b was in the building
[03:13] <Compn> for some reason i thought it was a technical meeting 
[03:14] <Compn> not 'all downstream projects and google tell us to stop fighting' meeting :P
[03:16] <Compn> because if i knew that was the meeting i wouldnt have gone hehe
[03:16] <BBB> who is Google? Thierry?
[03:16] <Compn> i think
[03:16] <Compn> tall guy
[03:16] <Compn> at youtube
[03:20] <BBB> light beard, french?
[03:20] <BBB> I bet it was thierry
[03:20] <BBB> ok
[03:20] <BBB> well that was a mostly useless document
[03:21] <BBB> I agree with his statement; you guys need to all forget about the past - all of you
[03:21] <BBB> if you have any will in you to ever make things right, reset the past to zero and start over, in a social context
[03:21] <BBB> but I dont see that happening
[03:21] <Compn> is what i'm trying to explain to carl at the moment
[03:22] <wm4> BBB++
[03:22] <wm4> (BBC?)
[03:22] <BBB> :D
[03:23] <wm4> but even of ffmpeg and libav devs suddenly fall into love with each other - the source diverted too much
[03:23] <wm4> and everyone dislikes each other's sources
[03:23] <Compn> yes it was brought up at the meeting too
[03:23] <BBB> well yeah thats what forks do
[03:23] <BBB> they diverge
[03:23] <cehoyos> Did they really diverge?
[03:23] <BBB> so the longer you keep bickering about the past, the more it will diverge
[03:24] <iive> well... ffmpeg already have most of libav merged.
[03:24] <cehoyos> Isn't Michael from time to time doing cosmetic commits and make the source files more similar?
[03:24] <Compn> BBB : if you are saying that to someone in particular, you should tell them directly ;p
[03:24] <wm4> there's also the git history
[03:24] <wm4> K
[03:24] <BBB> cehoyos: libav might still object to git history polution (or authorship pollution in annotate)
[03:24] <wm4> oops
[03:24] <cehoyos> (He is doing it to make his merges simpler, but still: Is it really true that the sources have diverged or has FFmepg simply more source code?)
[03:24] <BBB> wm4++
[03:24] <wm4> Libav folks rightfully dislike the ffmpeg git history
[03:25] <cehoyos> They do object, of course!
[03:25] <iive> wm4: we can always move back to svn! it would ensure non-diverging linear history!
[03:25] <Compn> edit git source to remove 'merged' from gitlog :P
[03:25] <wm4> and of course ffmpeg devs do not want to switch to the libav repo
[03:25] <cehoyos> But unfortunately, I feel exactly the same way: avconv source code is often missing authorship while authorship is more often correct in FFmpeg
[03:25] <cehoyos> So while I agree with the argument I come to a different concolusion...
[03:26] <BBB> and thats why well never truly merge
[03:26] <BBB> ok Im gonna do something useful again
[03:26] <wm4> BBB: D:
[03:26] <cehoyos> Sorry;-)
[03:26] <BBB> ?
[03:26] <wm4> but yeah.
[03:26] <iive> really, what is the point of using distributed vcs like git, if you can't merge stuff?
[03:26] <wm4> that's the other core issue
[03:27] <cehoyos> iive: It is even better: Iirc, in the original "mission statement" the advantages of git were an important point: I strongly suspect they expected at the time that Michael wouldn't like git in some ways...
[03:27] <iive> yeh...
[03:28] <cehoyos> Suddenly, using git features was not jsut bad, it even meant that FFmpeg sources cannot be used.
[03:28] <BBB> wm4: I dont have any illusion that after some magic irc discussion, some people here will <3 libav again
[03:28] <iive> but michael managed to exploit the strong sides of git to his utmost advantage.
[03:28] <cehoyos> But even that isn't a real surprise: After all, at the time FFmpeg development was too slow for them, now they consider it too fast...
[03:28] <BBB> wm4: but hey, I guess I send my patches here and they get in so Im ok with it
[03:28] <cehoyos> iive: Defintely;-)
[03:30] <BBB> I actually think I know how to get the projects together
[03:30] <BBB> just set up a bounty for michael, worth a shitton of money, like $100k or so, with the goal being to merge ffmpeg into libav and then rename libav back to ffmpeg and drop ffmpeg, or some variant thereof; I bet he might do it
[03:31] <Compn> why not just git pull libav/master in that case ?
[03:32] <wm4> BBB: heh
[03:32] <Compn> or do you mean submitting 300k of code to libav-devel ?
[03:32] <BBB> Compn: they miss a ton of features
[03:32] <BBB> yes
[03:32] <Compn> ohhhhh lol
[03:32] <Compn> sure, but hes still banned on that mailing list
[03:32] <Compn> :P
[03:32] Action: Compn runs
[03:32] <BBB> good point, that would need fixing
[03:32] <BBB> but overall this would work
[03:33] <BBB> anyone got $100k handy?
[03:33] <BBB> lets ask mr. Google
[03:33] <iive> BBB: i got scared for a moment when I read "set up bounty for michael"
[03:33] <Compn> aside from stalling development for a year BBB
[03:33] <Compn> complete stall of entire projects
[03:33] <BBB> Compn: nah, michael is a lot faster than that
[03:33] <Compn> libav review can keep up with him ?
[03:34] <Compn> i'm not making insult, i dont know.
[03:34] <iive> they can't.
[03:35] <BBB> I bet if you propose this and ask for commit access for him in the resulting repo, it would be an overall win for everyone
[03:35] <Compn> anyways,  i like the idea BBB
[03:35] <BBB> its more constructive than anything Ive seen being proposed here so far
[03:35] <BBB> you guys come up with something else thats not totally impossible :)
[03:36] <Compn> did you see the 3rd fork idea ?
[03:36] <Compn> that picks from both ffmpeg and libav ? run by 3rd party , probably sponsored or corporate
[03:37] <iive> BBB: libav people still hold ban on him in libav-devel maillist
[03:37] <iive> what makes you think they would give him write access?
[03:38] <iive> Compn: oracle, facebook or microsoft?
[03:39] <Compn> mpeg-la 
[03:39] <Compn> dolby
[03:39] <Compn> fraunhoffer
[03:39] <Compn> :P
[03:40] <Compn> but really i need to do something else
[03:40] <Compn> cehoyos : a whole room of people told us to stop fighting :)
[03:41] <cehoyos> Compn: ??
[03:41] <Compn> the meeting
[03:41] <Compn> it was about 40-50 people 
[03:42] <cehoyos> Okl there where three avconv developers, ten supportesrs and four people from FFmpeg: The others were spectators?
[03:42] <cehoyos> I really have troubles understanding all this...
[03:42] <cehoyos> And since you were highlighting me: What this an answer to something I said? Sorry, it is late and I don't understand...
[03:43] <Compn> it was full of vlc , handbrake, xencoder, and other developers in the multimedia open source . some people had opinions, some didnt speak up.
[03:44] <cehoyos> Ok, I understand this.
[03:44] <Compn> cehoyos : i highlighted you because you maybe one person who is still fighting ?
[03:45] <pross> are these unofficial minutes/transcript? or official ones approved by the attendees?
[03:45] <Compn> pross : the only transcript that was written i think
[03:45] <pross> who wrote it? i mean, is it biased?
[03:45] <cehoyos> But I don't understand what above means: While I still (levely) remember the emails "the split is such a good idea, development speed will improve (by people who had the power to stop it immediately!), I am not surprised people have realised it has its drawbacks: What can we do about it? - Nothing, unfortunatley (contrary to vlc and handbrake).
[03:46] <Compn> pross : it looks accurate but has some errors of course. its hard to type in real time
[03:46] <pross> ok
[03:46] <jamrial> wonder if anyone recorded it
[03:46] <Compn> i dont think it has any bias, no
[03:47] <iive> hard to type that fast on mobile
[03:47] <Compn> he was on a laptop
[03:47] <pross> 40-50 is a big turn out IMHO
[03:47] <jamrial> forty people means around forty cellphones and laptops. someone probably recorded the whole thing
[03:48] <jamrial> which could be used to check the transcript
[03:48] <Compn> pross :  i think the entire conf was 80 people ?
[03:48] <Compn> could count the names if it matters
[03:48] <pross> jamrial: better would be a scribe, that takes the key points that _everyone_ agrees to at the end of the meeting
[03:48] Action: pross note to self. volunteer at the next libav ffmpeg death match.
[03:49] <Compn> it was in google hq, i'm sure google has secret microphones recording at all times :P
[03:50] <pross> one option that nobody talks about is 'inertia'. eventually, one side will wither out and die.
[03:50] <Compn> well, some of us are secretly hoping for that :p
[03:51] <pross> Compn: what! you havent laid bets already?
[03:51] <Compn> i still think its funny everyone talks about libav fork
[03:51] <Compn> but no one talks about ffmbc fork
[03:52] <Compn> which is older, and has more features
[03:52] <pross> yep
[03:52] <Compn> and a developer we need to merge :P
[03:53] <Compn> a developer that can work with ffmpeg even
[03:53] <pross> i am working on fork #4
[03:53] <Compn> which, i'm not sure libav will work with ffmpeg devs...
[03:53] <cone-408> ffmpeg.git 03Michael Niedermayer 07release/2.2:b05d35504074: Update for 2.2.9
[03:54] <pross> they did once. what happened?
[03:54] <pross> s/they/some/g
[03:55] <Compn> i dont know
[03:56] <Compn> nor do i care
[03:56] <Compn> libreoffice / openoffice , lets just be seperate and ignore each other
[03:56] <Compn> (also libreoffice is better than openoffice, imo)
[03:57] <pross> second that motion
[03:59] <cehoyos> Just in case you don't know: libreoffice and openoffice share boots on exhibitions...
[03:59] <cehoyos> So I am not sure a comparison makes sense;-)
[03:59] <cehoyos> booths...
[04:00] <pross> now that is a step in the right direction. we can enconomize on sticker printer too :)
[04:02] <iive> imho, all these getting together efforts are having the opposite effect
[04:02] <iive> both project are better working on their own
[04:03] <pross> (sorry, was being sarcastic)
[04:03] <cehoyos> Good night!
[04:19] <cone-408> ffmpeg.git 03Michael Niedermayer 07fatal: ambiguous argument 'refs/tags/n2.2.9': unknown revision or path not in the working tree.
[04:19] <cone-408> Use '--' to separate paths from revisions
[04:19] <cone-408> refs/tags/n2.2.9:HEAD: swresample/swresample: replace always true if() by av_assert0()
[09:08] <ubitux> that was some lenghty discussion
[09:16] <knoch> hello, how to fix a pts and dts negative value in an AVPacket returned by av_read_frame() ?
[09:28] <j-b> 'morning
[10:12] <hima> ubitux: where can i find the definition of the function ff_text_init_buf()
[10:15] <ubitux> hima: try git grep ff_text_init_buf
[10:19] <hima> ok ubitux can you please explain the task the function vplayer_probe() does 
[10:20] <ubitux> just like any other probe function, it's used to detect the file type
[10:20] <ubitux> a probe function returns a score of confidence in wheither it matches the format or not
[10:20] <knoch> hello, how to fix a pts and dts negative value in an AVPacket returned by av_read_frame() ?
[10:21] <hima> ohh okkay got it! 
[10:36] <hima> ubitux: what does ff_text_r8 do?
[10:55] <ubitux> hima: look at the doxy comment in libavformat/subtitles.h
[10:56] <ubitux> this ff_text api is useful only if you want to support utf-16 file
[11:04] <hima> okay so i dont need to use it for kate right?
[11:06] <ubitux> probably not
[11:06] <ubitux> check the specs maybe
[11:11] <hima> ok ubitux. in the srtdec.c file in the function get_pts why do we need to run a loop?
[11:13] <ubitux> hima: it's not really optimal, and it's from legacy code a bit
[11:14] <ubitux> hima: in the case of srt you have a first line with a number
[11:14] <ubitux> then a line for the timestamp
[11:14] <ubitux> then the payload
[11:18] <hima> okay so when i=1 it is in the first line where there is a number is it?
[11:19] <ubitux> i=0 is the number, and i=1 is the timestamp line, i'd say
[11:19] <ubitux> add some debugging and see by yourself
[11:20] <hima> yeah my mistake right i=0 the number okay i will check 
[11:21] <ubitux> it really is clumsy, but it's because the surrounding code was more complex a while ago iirc
[11:21] <ubitux> maybe it's trying to handle badly formatted srt
[11:21] <ubitux> with "forgotten" number above the timestamp
[11:22] <ubitux> srt is probably the most common subtitles format in the wild
[11:22] <ubitux> and since it's humanly editable
[11:22] <ubitux> all kind of broken srt are to be found in the wild
[11:23] <hima> ohh okay 
[11:27] <Daemon404> [02:07] <@cehoyos> Compn: I know but I don't understand you: Is this an answer to something / to what? <-- i sat in the back of that horrible meeting
[11:27] <Daemon404> er..
[11:27] <Daemon404> fucking... wrong line
[11:27] <Daemon404> sorry cehoyos
[11:27] <cehoyos> I know, I understand
[11:27] <Daemon404> [02:07] < Compn> cant remember if Daemon404 was there this year or if i'm confusing him with someone else <-- this one
[11:27] <cehoyos> (I answered that you were there)
[11:28] <Daemon404> oh ok.
[11:28] Action: Daemon404 sat quietly in the back with thierry.
[11:28] <knoch> is there anyone able to answer me ?
[11:28] <cehoyos> If you want to provide a patch; Certainly
[11:28] <Daemon404> knoch, there is no fix
[11:28] <Daemon404> a negative pts and dts can be valid in some containers
[11:29] <knoch> such as mpegts ?
[11:29] <Daemon404> doesn't sound so correct in mpegts
[11:30] <cehoyos> Just make sure you don't see AV_NOPTS_VALUE which is not a negative number, but no pts
[11:30] <cehoyos> FFmpeg cannot read timestamps correctly for some H.264 streams
[11:30] <cehoyos> This is not easy to fix.
[11:31] <knoch> this is a H264 stream
[11:31] <cehoyos> Do you get AV_NOPTS_VALUE?
[11:32] <knoch> just checked : yes
[11:33] <knoch> so there is no pts
[11:34] <knoch> how ffmpeg handle this ? ffplay is able to play this file, perfectly
[11:34] <knoch> but my application can't
[11:37] <cehoyos> Are you sure ffplay plays the file perfectly? Or is it just guessing 25fps?
[11:37] <cehoyos> You can use -fflags +genpts but there is no garantee it will make a difference
[11:37] <ubitux> av_frame_get_best_effort_timestamp()?
[11:37] <orcus> When did -framerate appear? The man pages still refer to -r instead in examples
[11:38] <cehoyos> There is an output option -r and an input option -r
[11:39] <knoch> how can I know if ffmpeg is guessing the framerate ?
[11:39] <orcus> Unless I'm doing something wrong, output -r drops frames to achieve "framerate", rather than scaling how long each frame is shown for
[11:40] <orcus> Which is unintuitive, and makes the examples from man pages look poor
[11:47] <cone-839> ffmpeg.git 03Rémi Denis-Courmont 07master:319424d25c53: vdpau: add helper for VDPAU to libav error codes conversion
[11:47] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:567c0dcee674: Merge commit '319424d25c53f82b87187ce03ed984d438f1bee6'
[11:56] <cehoyos> orcus: Please read the man page again, an update was pushed a short time ago to clarify that -r is something else than -framerate
[12:03] <Daemon404> g 31
[12:03] <Daemon404> ... dammit
[12:03] <orcus> @cehoyos: Ah, I may be unclear: I see that that, but I feel that the example "For creating a video from many images" should be updated to use -framerate, as that is what I expect the example to do
[12:04] <orcus> @cehoyos: as in, the example ``worked'' (i.e. gave output that I thought looked nice) a few months (?) ago, and now doesn't - it judders
[12:05] <orcus> @cehoyos: whereas with -framerate, it works as it did earlier (unless my memory is faulty!)
[12:11] <cehoyos> Then please send a patch to the development mailing list updating the documentation accordingly (I suspect this is why you came here).
[12:15] <orcus> @cehoyos: ok, will do
[12:20] <cone-839> ffmpeg.git 03Rémi Denis-Courmont 07master:fcc1022611f7: vdpau: factor out common end-of-frame handling
[12:20] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:b64b719ad5c2: Merge commit 'fcc1022611f79c2f3aa2f392a5ce14c74be9c1d7'
[12:43] <cone-839> ffmpeg.git 03Rémi Denis-Courmont 07master:ce083282f0a8: vdpau: common support for managing the VdpDecoder in avcodec
[12:43] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:10b6d7462cbf: Merge commit 'ce083282f0a8b7d63c4047c30b7bac498f9806dd'
[12:55] <BBB> is there a meeting log from the #ffmpeg-meeting thing saturday somewhere, or is a summary coming to the ML?
[13:04] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:8df41976b7cd: avcodec/vdpau_internal: add comment to #endif
[13:30] <cehoyos> orcus: Sorry, I thought you found another instance of a sample command where the input option -r was used instead of the input option -framerate
[13:30] <cehoyos> I am not convinced that there is an output option -framerate, in any case, -r should be kept imo because it will also work (identically) with older versions of FFmpeg
[13:31] <cehoyos> And I don't know of a wish to replace the output option -r with -framerate
[14:16] <cone-839> ffmpeg.git 03Rémi Denis-Courmont 07master:89ac99ba5f2d: vdpau: pass codec-specific parameters from hwaccel
[14:16] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:0ba887bbf4d9: Merge commit '89ac99ba5f2dc9f69ad3bc294753930eb0b3e4a4'
[14:16] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:ec6a855b3a6b: avcodec/vdpau: warn if the user application has not setup avctx->hwaccel_context instead of potentially crashing
[14:46] <cone-839> ffmpeg.git 03Rémi Denis-Courmont 07master:502cde409ca5: vdpau: force reinitialization when output resolution changes
[14:46] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:067d11bf71e8: Merge commit '502cde409ca5ee97ef70c2cdede88b9101746ff6'
[15:08] <arwa> Hey...Can anyone help me with the command "make -j<whatever> ffmpeg" in writing_filters ? Its giving me errors
[15:10] <cone-839> ffmpeg.git 03Rémi Denis-Courmont 07master:e3e158e81f06: vdpau: add av_vdpau_bind_context()
[15:11] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:a61899a0f13c: Merge commit 'e3e158e81f0666b8fe66be9ce1cad63a535920e0'
[15:11] <ubitux> arwa: <whatever> is a number, right
[15:11] <arwa> okay...what number do i need to enter?
[15:12] <ubitux> depends on your cpu
[15:12] <arwa> how do i check it? 
[15:12] <ubitux> are you on linux?
[15:12] <arwa> yes ubuntu12.04
[15:12] <ubitux> what does nproc returns?
[15:12] <arwa> 2
[15:13] <ubitux> alright, then try "make -j3" (3 being nproc + 1, common heuristic)
[15:13] <ubitux> it's the number of parallel jobs
[15:14] <arwa> oohhh...!!! It refers to no of parallel jobs!! thanks :D
[15:15] <arwa> but its still giving me errors ---> Undefined reference to 'avdevice_configuration',,,,,,etc
[15:17] <cone-839> ffmpeg.git 03Rémi Denis-Courmont 07master:bef067f88c74: vdpau: check video surface and decoder capabilities
[15:17] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:573d3330107b: Merge commit 'bef067f88c74190cdf7e76d12f02e12e069974aa'
[15:30] <reynaldo> oh, missed saturday's meeting
[15:31] <reynaldo> do we have the meeting log somewhere I can take a look?
[15:32] <michaelni> pester saste, he wanted to post one IIRC
[15:33] <michaelni> these things really should be posted immedeatly after the meeting
[15:34] <reynaldo> michaelni: will do, did you attend?
[15:37] <michaelni> yes
[15:38] <michaelni> i think for future meetings someone else should take over posting the log, saste seems always too busy
[15:39] <reynaldo> saste: hello o/
[15:39] <saste> reynaldo, hey
[15:39] <reynaldo> saste: do you happen to have saturday's meeting log available somewhere ?
[15:39] <saste> reynaldo, yes, I'm going to publish it soon
[15:39] <reynaldo> I missed the anouncement (and the meeting) :/
[15:39] <saste> give me a few minutes
[15:39] <reynaldo> saste: cool, thanks
[15:39] <reynaldo> surething
[15:41] <ubitux> hey.
[15:42] <shreshtha_> hey!
[15:42] <ubitux> a third opw candidate came up for the subtitles task (hello shreshtha_)
[15:42] <ubitux> so... anyone has an idea for a qualification task? i'm currently out of ideas
[15:43] <reynaldo> shreshtha_: hey, welcome ;)
[15:44] <shreshtha_> hey! :)
[15:44] <reynaldo> ubitux: for the already listed projects or in general ?
[15:46] <reynaldo> I'm guessing qualis should be tied to the applicant's choosed project(s), which would put mentors as the primary source for them
[15:46] <ubitux> reynaldo: it's for the subtitles task
[15:46] <reynaldo> noted
[15:47] <saste> reynaldo, done
[15:48] <saste> reynaldo, we missed you, was you not able to attend or did you just miss the call?
[15:48] <reynaldo> missed the call sadly, not that active during weekends :/
[15:48] <ubitux> ...maybe i could propose a bug fix but that's kind of evil
[15:49] <reynaldo> saste: please CC me next time around, not a big deal anyway as you were around and know the OPW drill quite well
[15:49] <saste> reynaldo, sure will do
[15:49] <reynaldo> ubitux: maybe some triagging work on unconfirmed subtitle bugs for starters? then select 2-3 to work on ?
[15:49] <reynaldo> saste: thanks
[15:50] <saste> reynaldo, if you want we can schedule a meeting about OPW if you think it is needed
[15:50] <benoit-> saste: thanks for the log
[15:50] <reynaldo> saste: let me take a look at the log first
[15:50] <reynaldo> & thanks for the log :)
[15:51] <ubitux> reynaldo: i'd like to propose something a bit more interesting, and easier
[15:52] <reynaldo> ubitux: oh, sure. just the fairly basic (and obvious if I might say) idea that come to mind
[15:53] <reynaldo> :)
[15:54] <ubitux> maybe i could check for bugs in lrcdec and ask for a fix
[15:54] <ubitux> ah something interesting could be to fuzz some subtitles formats
[15:54] <ubitux> make them crash, and fix that
[15:55] <ubitux> shreshtha_: would that be fine with you?
[15:58] <ubitux> we'll say something like: find 3 crashes or bug in the current subtitles demuxers or decoders (using zzuf for example)
[15:58] <ubitux> i can help you finding ones if you don't
[15:58] <ubitux> they will guide you into fixing that
[15:59] <shreshtha_> sounds interesting! Presetly I am a bit deficient in the knowledge of subtitles, can u help me with that
[15:59] <ubitux> that's not exactly development but well
[15:59] <ubitux> yeah sure
[15:59] <ubitux> just a moment
[16:00] <ubitux> shreshtha_: so, akira4 and hima are the 2 others opw candidates for subtitles (they're not present currently but sometimes they join)
[16:01] <ubitux> both of them are working on writing a demuxer for one subtitles
[16:01] <ubitux> a few days ago i explained the subtitles chain to akira4 in this channel
[16:01] <ubitux> here is the backlog: http://b.pkh.me/opw-subtitles.log
[16:01] <ubitux> i recommend you read that first, and then ask me questions
[16:01] <shreshtha_> sure
[16:02] <ubitux> i'm sorry i don't have any other simple subtitles demuxer i could think of
[16:02] <shreshtha_> I'll go through this and get back to u :)
[16:02] <ubitux> currently akira4 is on stl format, and hima is on a simple version of kate
[16:03] <ubitux> if you find one yourself unsupported, you can probably do it instead of what i propose
[16:03] <ubitux> otherwise we'll probably go for the fuzzing/crashing/bugging thing
[16:04] <shreshtha_> okay..i will try doing this
[16:07] <benoit-> ubitux: "developers assigning themselves to tickets might help": I'm going to try to do that... sarting now :)
[16:08] <ubitux> benoit-: cool, thanks
[16:21] <arwa> I tried following the instructions in writing_filters .....Its now giving me error [[AVFilterGraph @ 0xaf5f260] No such filter: 'try'] when I try running the filter, where try is the name of the filter.
[16:32] <cone-839> ffmpeg.git 03Rémi Denis-Courmont 07master:8de1d67967a9: avconv_vdpau: update to new VDPAU interface
[16:32] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:403133ab5eb3: ffmpeg: add vdpau_old to allow continued testing of the older (but not oldest) API
[16:32] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:0e57c051181c: avcodec/vdpau: fix render2() check
[16:32] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:4d8356b67880: Merge commit '8de1d67967a9f9e22c66cb0c1e518ae4f30d07dd'
[17:24] <J_Darnley> arwa: did you add an entry for your filter in allfilters.c?
[17:49] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:67ddf21611b9: avcodec/vdpau: do not dereference hwctx before checking it for NULL
[18:00] <ubitux> michaelni: swr works on unaligned data for the output, right?
[18:07] <wm4> urgh the new vdpau shit got pushed
[18:08] <nevcairiel> you know how it goes, vlc-dude says, libav do
[18:13] <cehoyos> lol!
[18:13] <cehoyos> The old vdpau code (the only I want to support) still works, the "new" api crashes MPlayer atm
[18:14] <nevcairiel> thats more than likely mplayers fault
[18:14] <cehoyos> Correction: Fixed by Michael in his latest commit
[18:16] <kierank> nevcairiel: and libav does, ffmpeg do
[18:16] <cehoyos> The original VDPAU api wasn't dropped...
[18:17] <wm4> cehoyos: the new vdpau code (I mean hwaccel interface) is still slightly better than the old one, because now all hwaccels are uniform
[18:17] <wm4> but I can't say I like this hwaccel stuff much
[18:17] <wm4> separate decoders seem much simpler
[18:17] <nevcairiel> duplicates a lot of efforts tho
[18:18] <wm4> (instead of selecting hwaccel backend and pixel formats)
[18:18] <nevcairiel> and the way the old "separate" vdpau decoder was build is just a giant hackjob
[18:18] <wm4> in any case, maybe mplayer should switch now...
[18:19] <wm4> because now there are 3 vdpau APIs
[18:30] <akira4> ubitux, did you review the stl demuxer?
[19:04] <ubitux> akira4: i was waiting for you to do it :p
[19:05] <ubitux> akira4: give me... 15 minutes to get back on it
[19:05] <akira4> sure thing.
[19:07] <ubitux> akira4: did you do any change on it since last time?
[19:07] <akira4> no actually..I have a test coming up so couldn't do much
[19:07] <ubitux> sure ok
[19:09] <ubitux> akira4: first, the boring part, make sure you follow the coding style of the project (see http://ffmpeg.org/developer.html#Coding-Rules-1); you're violating it in a few places (no tabs, space placements, etc)
[19:09] <ubitux> look at the rest of the code and keep it consistent with that
[19:10] <akira4> okay. I'll fix that.
[19:10] <ubitux> 2. in the get_pts() function, >= 6 looks wrong
[19:11] <ubitux> 3. in that same sscanf, make sure len is initialized (it's not counted in the sscanf counter in most implementation, so you can't rely on that)
[19:12] <akira4> okay.
[19:13] <ubitux> 4. the rest of the code looks correct so far, but i think it's not tolerant enough; i saw some files where the ',' weren't separated by spaces
[19:14] <ubitux> here is an example: http://www.eso.org/~lchriste/trans/eyes/subtitles/soundtrack.stl
[19:14] <ubitux> (the first one when i type "subtitles ext:stl" on google)
[19:14] <ubitux> alright, so that's it for the code itself
[19:14] <ubitux> now a few more things
[19:15] <ubitux> 5. please add an entry in doc/general.texi (look for the other subtitles entries); you can use make doc/general.html to check the output
[19:16] <akira4> what is it for?
[19:16] <ubitux> documenting that it's supported
[19:16] <akira4> okay
[19:16] <ubitux> 6. mention it in the Changelog
[19:17] <ubitux> alright so for now that should be enough;
[19:17] <ubitux> after you've done this
[19:17] <ubitux> we'll see how to add a codec
[19:17] <ubitux> add a simple test in our test suite
[19:17] <ubitux> and submit the patch
[19:17] <ubitux> and you'll be done
[19:18] <akira4> I see. Thanks a lot :) I'll get to it.
[19:18] <ubitux> akira4: for 3., it has security concerns, try git show 138902dfb for a similar example
[19:19] <akira4> Got it. I'll ask if I have any doubts.
[19:22] <ubitux> akira4: ah i forgot, please also add the copyright boilerplate
[19:22] <ubitux> on top of the .c
[19:22] <akira4> sure. 
[19:26] <periahmadi> @saste Hey, Im looking into applying for the Gnome OPW and Im interested in working on one of ffmpegs projects, for which youre listed as a mentor. I was wondering how much of a knowledge of C I would need to complete the application process. Thanks!
[19:28] <ubitux> did you find a project you would be interested in in the list?
[19:29] <periahmadi> MPEG-4 Audio Lossless Coding (ALS) encoder if that questions was directed at me
[19:31] <J_Darnley> Wasn't someone already working on that?
[19:32] <ubitux> there is one potential applicant yes, but well
[19:32] <ubitux> we have 3 for subtitles
[19:32] <ubitux> :D
[19:32] <nevcairiel> curious that ALS gets so much interest, especially since encoders are not the easiest task of the bunch
[19:33] <periahmadi> Should I look into something with no potential applicant?
[19:33] <ubitux> i don't think it really matters
[19:33] <nevcairiel> if thats the task you want to do and feel confident that you can, no reason not to try to get it
[19:33] <periahmadi> I dont know why others are interested but Im interested in learning audio coding in genereal
[19:34] <J_Darnley> If the other is only "potential" then I guess it doesn't hurt.
[19:34] <J_Darnley> I wonder what the drop-out rate will be compared to gsoc.
[19:40] <periahmadi> Im just not sure how much starting knowledge any of these projects require. The OPW website qualifies this as a small contribution, which is pretty subjective.. Im on a time limit to commit as the deadline to apply is coming up in a few weeks and Im interested in ffmpeg but not sure Im qualified to make a contribution in the timeframe. Are there any ffmpeg projects that are more suited to a noob than the ALS? 
[19:41] <J_Darnley> Well a new encoder will be small relative to the rest of FFmpeg, but it will be Hard(TM).
[19:42] <J_Darnley> You will need to speak C quite well already.
[19:42] <wm4> don't we have qualification tasks vs. real task?
[19:42] <wm4> qualification tasks being something simpler and smaller
[19:43] <J_Darnley> You will have to deal with memory, pointers, and all the things people dislike about C.
[19:43] <J_Darnley> (I don't know why though)
[19:44] <periahmadi> haha
[19:45] <periahmadi> Ah I see, so the qualification task is whats required to apply and the project is the ultimate end to the OPW work?
[19:46] <wm4> I think so
[19:47] <wm4> e.g. the two who applied for the subtitles task are both implementing parsers for a trivial subtitle format
[19:47] <wm4> as qualification
[19:49] <wm4> actually the wiki page says: "Qualification Task: Add floating point support to MPEG-4 ALS decoder"
[19:49] <jamrial> floating point for a lossless decoder?
[19:52] <wm4> *shrug*
[19:52] <J_Darnley> As I discovered last time it was mentioned, ALS does use floats!
[19:54] <periahmadi> So if im a total novice in C, would the qualification tasks help me develep the rest of what I need for the final project or should I look into another project?
[19:56] <j-b> I'm afraid ffmpeg is hard code
[19:56] <periahmadi> ah ok thats what I thought
[19:57] <periahmadi> thanks everyone for your help
[19:57] <wm4> periahmadi: the OPW wiki page says: "Qualification Task: Add floating point support to MPEG-4 ALS decoder"
[19:57] <wm4> j-b: is it really so hard?
[19:57] <j-b> wm4: yes.
[19:57] <periahmadi> haha :)
[19:58] <wm4> j-b: measured on what?
[19:58] <wm4> I think C is a pretty simple language
[19:58] <j-b> on all the other open source projects
[19:58] <j-b> FFmpeg C level is hard.
[20:01] <J_Darnley> periahmadi: if you still want to try, I suggest you contact the mentor(s) and see if you can be lead through the decoder and where to begin the qualification task
[20:02] <periahmadi> Thanks, Ill shoot him/her an email. I have a feeling Im way too underqualified though
[20:06] <wm4> periahmadi: what's your level of C? do you know how to use printf/malloc/free and some of these things?
[20:08] <periahmadi> hmm&my level is like equivilant to a codecademy course. very basic
[20:08] <periahmadi> Ive also not practiced in years so I need to refamiliarize myself
[20:09] <wm4> yeah, then this could be hard
[20:10] <periahmadi> ok
[20:10] <wm4> I'm not really a fan of discouraging people before they even try, though
[20:10] <saste> periahmadi, sorry for the slow reply
[20:11] <arwa> Can anyone help me out with writing filters?
[20:11] <periahmadi> I appreciate it! No discouragement here. I was looking for a pretty clear idea of what Id be getting myself into
[20:11] <saste> periahmadi, C-coding skill is probably secondary, assuming you can learn fast and you're motivated enough
[20:11] <periahmadi> @saste its cool everyone heres been very helpful
[20:12] <saste> periahmadi, first problem is to find a task, then you can take it a sort of challenge or test, you'll be able to test yourself and learn something
[20:12] <saste> periahmadi, then keep in mind that we have a limited number of available slots, just 1 or maybe 2
[20:13] <saste> so it depends on you, if you think that the selection process could be useful to learn something
[20:13] <periahmadi> ah ok, 1 or 2 for each project? 
[20:13] <saste> periahmadi, no each project has a different number of slots, depending on their sponsors
[20:13] <periahmadi> ok 
[20:14] <saste> arwa, what's the problem?
[20:16] <arwa> I read the writing_filters doc....I tried following it....but its giving me errors when I try to run the filter ---> [AVFilterGraph @ 0xaf5f260] No such filter: 'foobar'
[20:17] <saste> arwa, did you update the Makefile and allfilters.c?
[20:17] <saste> also what if you run ffmpeg -filters?
[20:18] <saste> is the filter listed?
[20:18] <arwa> yes I did
[20:18] <saste> can you show your patch?
[20:18] <saste> git diff
[20:18] <saste> you can check in the log and see how filters commits are done?
[20:18] <saste> for example
[20:19] <saste> git show 422619646ea0e938188a49a06226831cc42e2a6a
[20:19] <arwa> okay...btw I ran ffmpeg -filters....The filter is not mentioned there
[20:19] <saste> and of course you need to recompile ffmpeg in order to see the filter
[20:20] <arwa> i did it by ./configure and make -j3 ffmpeg
[20:20] <saste> arwa, BTW what document are your following?
[20:20] <arwa> do i need to do something more?
[20:20] <saste> arwa, no, but are you sure you're running the modified version of ffmpeg?
[20:21] <saste> also, are you able to compile ffmpeg with no errors?
[20:21] <saste> usually when you write a new filter there are many compilation issues to fix
[20:21] <saste> did you already fix them? sorry for the silly question
[20:21] <arwa> At first, I was not able to compile it withour errors...but now its working....its not giving me any errors
[20:21] <arwa> without*
[20:21] <saste> ok, so it's compiling
[20:22] <arwa> yes
[20:22] <saste> now, are you sure you're running the correct ffmpeg version?
[20:22] <saste> you should specify the complete path of the compiled ffmpeg
[20:22] <arwa> how do i ensure that?
[20:22] <saste> which -a ffmpeg
[20:23] <saste> or you cd to the ffmpeg source dir, and run: ./ffmpeg -filters
[20:24] <arwa> okay
[20:24] <arwa> thaks
[20:24] <arwa> its working now :D
[20:24] <arwa> there were two versions...and i was using the wrong one !!
[20:24] <arwa> thanks alot!!
[20:25] <saste> arwa, np
[20:26] <arwa> Well...I was thinking of starting with a filter....Can you suggest me some cool ideas?
[20:26] <arwa> I was thinking of implementing a fish eye lens view for the video?
[20:27] <arwa> Is it okay? Or should I go with something standard?
[20:30] <saste> arwa, i'm leaving in a few minutes, can you ask on list?
[20:30] <arwa> which list?
[20:31] <saste> or leave your email address (here or as a private message)
[20:31] <saste> arwa: ffmpeg-devel
[20:31] <J_Darnley> What's "standard"?  I think fish-eye distortion is already covered.
[20:31] <ubitux> don't we have already fish eye filters?
[20:32] <ubitux> libavfilter/vf_lenscorrection.c and libavfilter/vf_vignette.c are in the spirit
[20:33] <arwa> Ohhh...!! Okay
[20:33] <arwa> So, any suggestions on filters?
[20:33] <J_Darnley> You can finish my work of porting eq and eq2 and merging those two with hue
[20:34] <J_Darnley> port another mp filter
[20:34] <arwa> Okay....Can you give me a detailed info?? Or some material to read on this?
[20:35] <J_Darnley> port something from avisynth
[20:35] <J_Darnley> There's the OPW page on the wiki
[20:35] <J_Darnley> That oen't list specific ideas.
[20:35] <J_Darnley> *doesn't
[20:36] <J_Darnley> port something from sox.
[20:36] <J_Darnley> I believe the choice is yours
[20:36] <saste> arwa: http://trac.ffmpeg.org/wiki/SponsoringPrograms/OPW/2014-12
[20:36] <wm4> seems some people still like mplayer's af_hrtf
[20:36] <saste> ask if you want more specific ideas, possibly on the ffmpeg-devel ML
[20:37] <J_Darnley> wm4: I wonder how that compres with sox's earwax
[20:37] <wm4> no idea
[20:37] <wm4> but I think it's the only non-trivial filter that's still in mplayer but not libavfilter
[20:39] <arwa> I am actually cool with image processing....I have never tried sound processing :/ 
[20:41] <ubitux> arwa: mmh
[20:41] <ubitux> got one for you
[20:41] <ubitux> what about xbr?
[20:41] Action: J_Darnley thinks video is easier: no floats, no complex, no phase
[20:41] <ubitux> arwa: https://trac.ffmpeg.org/ticket/3404
[20:41] <ubitux> see my first comment
[20:42] <ubitux> the hqx filter is implemented, but xbr isn't
[20:42] <ubitux> and your fortunate, xbr is actually documented
[20:42] <arwa> okay
[20:42] <arwa> xBR or xBRZ??
[20:42] <ubitux> never heard of xBRZ
[20:43] <gnafu> I could guess what it is.
[20:43] <arwa> just a enhanced version of xbr
[20:43] <arwa> could you provide me the doc for it?
[20:43] <ubitux> yeah, i was looking for it but it seems the forum link moved
[20:44] <ubitux> just a moment
[20:44] <arwa> cool :D
[20:44] <ubitux> http://web.archive.org/web/20140904180543/http://board.byuu.org/viewtopic.php?f=10&t=2248
[20:48] <ubitux> arwa: note that i have no real preferences between xbr over xbrz
[20:48] <arwa> i think i will go with the documented one :P
[20:49] <ubitux> sure thing
[20:49] <ubitux> it has extensions
[20:49] <ubitux> make sure it's bitexact with the reference though
[20:50] <ubitux> arwa: btw, are you an opw applicant?
[20:52] <arwa> yes
[20:53] <ubitux> are you looking for the libavfilter improvements?
[20:53] <arwa> yes
[20:53] <ubitux> cool, okay
[21:23] <cone-839> ffmpeg.git 03Manfred Georg 07master:577899a6458c: lavc: specify the behavior of av_lockmgr_register on failure.
[21:23] <cone-839> ffmpeg.git 03Michael Niedermayer 07master:d47dd84391f2: Merge commit '577899a6458ccad9026eb268f10dc0b39c224c8d'
[21:44] <kierank> Anyways there will be a multimedia track at fosdem. 
[22:23] <mediocregopher_> Hey everyone! I'm having trouble in ffmpeg 2.4.2 getting libvpx streaming from ffserver
[22:23] <mediocregopher_> I keep getting the error: [libvpx @ 0x20a21a0] CQ level 0 must be between minimum and maximum quantizer value (10-42)
[22:23] <mediocregopher_> my qmin is 10 and qmax is 42, so those parts make sense
[22:24] <mediocregopher_> but I can't figure out how to set the cq-level
[22:24] <mediocregopher_> it doesn't seem to be a valid AVOptionVideo setting
[22:25] <nevcairiel> my guess is it maps to -q
[22:25] <nevcairiel> try -q:v number
[22:27] <kepstin-laptop> mediocregopher_: actually, it looks like it's mapped to the "-crf" option
[22:28] <kepstin-laptop> why is your qmin so high, tho? that'll probably degrade your video quality a fair bit.
[22:28] <mediocregopher_> huh, thought I tried the crf option, I'll try it again
[22:28] <mediocregopher_> and I just copied this off of some blog, I'm kind of a noob <_<
[22:29] <nevcairiel> indeed, looks like crf is it
[22:29] <kepstin-laptop> mediocregopher_: probably want to start at https://trac.ffmpeg.org/wiki/Encode/VP8
[22:32] <kepstin-laptop> mediocregopher_: and also http://wiki.webmproject.org/ffmpeg if you want the full list of ffmpeg options that do things in libvpx.
[22:33] <mediocregopher_> awesome, thanks guys! I'll hack at it a bit more and see if I can't get it
[22:33] <kepstin-laptop> although that's the command line options, not api fields, but still it should be a start.
[23:13] <mediocregopher_> So I'm still having trouble getting it to work, here's the server configuration I'm using: http://gobin.io/NKHl and here's the ffmpeg command/output: http://gobin.io/zOyX
[00:00] --- Tue Oct  7 2014