[Ffmpeg-devel-irc] ffmpeg-devel.log.20160519

[burek]

[00:02:25 CEST] <jamrial> rcombs: is anything blocking your autobsf patchset?
[00:03:55 CEST] <cone-981> ffmpeg 03Michael Niedermayer 07master:7155bd1c94c2: avcodec/m101: Ask for samples with too small extradata_size
[00:03:56 CEST] <cone-981> ffmpeg 03Michael Niedermayer 07master:d08f2c172fd2: avformat/avidec: Detect index with too short entries
[00:03:59 CEST] <rcombs> jamrial: not that I can think of
[00:54:44 CEST] <Compn> Illya : yeah, i dont think many devs use thunderbird for mailing list stuff. its a nice client - a bit bloated... there may be scripts to make it work better with git
[00:57:05 CEST] <Illya> Compn: what do other devs use? I only really use email because of the ffmpeg-devel mailing list
[00:57:33 CEST] <DHE> you can mail directly from git. at least it'll be formatted right
[00:58:02 CEST] <DHE> there is a small amount of magic involved if, say, replying to an existing message
[00:59:12 CEST] <JEEB> sending: git send-email , grabbing: get the raw message and git am
[01:04:27 CEST] <Illya> yeah, I'm using git for sending
[01:10:48 CEST] <Compn> Illya : i use sylpheed-claws
[01:11:10 CEST] <Compn> i guess its just 'claws mail' now
[01:11:36 CEST] <Compn> oh i guess they are forks, nevermind
[01:11:43 CEST] <Compn> either slypheed ro claws :D
[01:12:20 CEST] <Compn> so slypheed is the one i prefer (looking at screenshots on wikipedia)
[01:12:28 CEST] <Compn> claws looks more like thunderbird
[01:13:54 CEST] <llogan> claws works fine. you can click on message and then choose "File -> Export selected to mbox file" then git am that resulting file
[01:14:05 CEST] <wm4> claws is the worst mail client I've ever used, although I'm still using it
[01:14:32 CEST] <llogan> i used to use mutt then i shaved my neckbeard
[01:14:46 CEST] <wm4> it tends to corrupt its index or redownload the whole mail index spontaneously for no reason
[01:15:03 CEST] <wm4> and it blocks the UI for network operations
[01:15:13 CEST] <llogan> that happens to me. a few times a year.
[01:15:27 CEST] <wm4> sometimes the index gets so corrupted, you see different mails than what you click on the list
[01:15:33 CEST] <llogan> that too
[01:15:34 CEST] <wm4> other than that it's not too bad I guess
[01:15:55 CEST] <wm4> (but they should really fix those bugs...)
[01:16:00 CEST] <llogan> i used to like it for the filtering and processign but then i got to move that to server side
[01:16:11 CEST] <wm4> also, so support for the imap push protocol or what it was
[01:16:21 CEST] <wm4> so you have to let it scan for mails all X minutes
[01:16:38 CEST] <wm4> and while it scans you can't do anything else
[01:17:12 CEST] <wm4> what sever side processing is there?
[01:17:13 CEST] <llogan> i still like it better than thunderbird
[01:17:25 CEST] <wm4> me too, apparently
[01:17:55 CEST] <llogan> i meant i moved to fastmail and started using their filters instead
[01:19:18 CEST] <llogan> their webmail client isn't bad, but no in-reply-to and some other annoyances. also can't view diff patches
[01:19:24 CEST] <cone-981> ffmpeg 03Michael Niedermayer 07master:4a81bddb4d9b: ffmpeg_filter: Fix AVBufferSrcParameters leak
[01:19:25 CEST] <cone-981> ffmpeg 03Michael Niedermayer 07master:641dccc2aa5e: avcodec/h264: Check init_get_bits8() for failure
[01:38:30 CEST] <Illya> llogan: I'm with fast mail, didn't realise they had server-side filters. That sounds cool, I'll look into that. Thunderbird works, more or less, for me. Wish Mail.app had a proper thread display then it'd be ideal 
[01:39:34 CEST] <Illya> And gtk UIs are amazingly smooth on OSX either
[01:40:10 CEST] <Compn> so you dont use email at all otherwise ? 
[01:40:18 CEST] <Compn> communicate via phone or facebook or skype etc ?
[01:40:37 CEST] <Compn> just curious as i'm old and still use irc :)
[01:42:25 CEST] <llogan> I think people use myspace these days. hey, you forgot to email me the other day. 72937,630 at compuserve.net
[01:43:14 CEST] <Compn> i actually had compuserve
[01:43:40 CEST] <llogan> same. and prodigy.
[01:43:50 CEST] <Compn> yup
[01:45:04 CEST] <Compn> first got internet in 1992. from a local newspaper!
[01:47:57 CEST] <Compn> Illya : i'm asking because i want to invest in new communication website stocks. twtr, fb, instagram , vine, etc
[01:49:58 CEST] <llogan> make sure to name the account "Money Burn Barrel"
[01:57:23 CEST] <Compn> fb opened at 29, dropped to 22, now at $100+
[01:57:55 CEST] <Compn> twtr not as good :P
[01:58:16 CEST] <iive> is that the code of twitter?
[01:58:31 CEST] <iive> are they publicly traded now?
[01:58:38 CEST] <Compn> yes
[01:59:24 CEST] <Compn> and yews
[01:59:26 CEST] <Compn> er yes*
[02:00:27 CEST] <Compn> but i think twitter opened at 17 climbed to 40
[02:04:46 CEST] <Compn> llogan : hows that legal weed going on in ak ? :P
[02:24:38 CEST] <llogan> Compn: good i guess, but i'm  a drunkard, not a pothead. a few businesses are opening soon. some sort of pot fest is happening soon.
[02:55:11 CEST] <Compn> llogan : you havent tried the brownies yet? :D
[03:30:20 CEST] <cone-981> ffmpeg 03Michael Niedermayer 07master:df01a29c1b15: avcodec/exr: Fix potential integer overflow
[05:08:05 CEST] <cone-981> ffmpeg 03James Almer 07master:e34e7a575eae: avformat/dtshddec: move data_start outside of DTSHDDemuxContext
[09:23:50 CEST] <Illya> Compn: I don't use email for anything except for signing up for the various accounts, and ffmpeg-devel. I pretty much exclusively communicate via IRC and facebook. Keep in mind, though, that I don't work, so this may have an impact on my usage of communication methods.
[09:27:41 CEST] <Shiz> Illya: protip from my side: git send-email with msmtp
[09:27:47 CEST] <Shiz> sends patches in perfect format
[09:28:06 CEST] <Illya> I'm using IMAP I think
[09:28:24 CEST] <Illya> oh, SMTP
[09:28:37 CEST] <Shiz> http://txt.shiz.me/OWU2N2I3ND
[09:28:39 CEST] <Shiz> this is what i use
[09:28:43 CEST] <Shiz> msmtp is a mailer program
[09:30:01 CEST] <Shiz> something like
[09:31:06 CEST] <Shiz> $ git format-patch <revisions> -o out/
[09:31:24 CEST] <Shiz> $ git send-email --to=somemailinglist at lists.mailinglists.com out/*.patch
[09:31:28 CEST] <Shiz> does it all right
[09:33:04 CEST] <Illya> -o sounds like it'd be useful, I have the email set in the .gitconfig. Anyways, I should find a git tutorial or something, I'm sure I'm missing loads of features
[09:33:46 CEST] <nevcairiel> i use msmtp as well, mostly because i'm on windows and it doesnt come with a default sendmail otherwise =p
[09:34:17 CEST] <nevcairiel> you dont need to format-patch before doing send-email though, it can do that for you if you specify a revision instead
[09:37:35 CEST] <Shiz> right
[09:37:39 CEST] <Shiz> i just like to double-check the .patch files
[09:37:41 CEST] <Shiz> :p
[09:38:24 CEST] <Shiz> -v is also a useful flag for specifying v2/v3/vN... patches
[09:38:48 CEST] <wm4> I usually send directly, but with --annotate
[09:38:54 CEST] <wm4> which also lets you revisit the patches
[09:39:07 CEST] <nevcairiel> thats what i do
[09:39:16 CEST] <Shiz> ah, thanks for the heads-up
[09:50:45 CEST] <cbsrobot> thun
[10:10:41 CEST] <michaelni> git.videolan.org is dead again?
[10:11:26 CEST] <j-b_> yes.
[10:11:42 CEST] <j-b_> The DoS is stronger than this machine. I guess we will kill it.
[10:12:03 CEST] <JEEB> :<
[10:12:16 CEST] <JEEB> jesus christ the bored kids
[10:21:13 CEST] <durandal_1707> attacking vlc, why I wonder, big frustrations maybe
[10:24:15 CEST] <Shiz> maybe upset vlc's rendering block'd their hentai mkvs
[13:34:49 CEST] <durandal_1707> can we move temporaly to github?
[13:35:48 CEST] <thardin> only if you're a time lord
[13:36:29 CEST] <durandal_1707> IIRC already happened?
[13:37:12 CEST] <thardin> git is distributed, so shouldn't be a biggie. bug tracker still works I take it
[13:38:39 CEST] <wm4> uh that would everyone require to have github access?
[13:39:23 CEST] <durandal_1707> I guess no need for devs and commits in next days :) happy holidays
[13:39:46 CEST] <nevcairiel> everyone can just keep developing thanks to git, hence the distributed part
[13:39:55 CEST] <nevcairiel> aand it gets pushed when its back online
[13:40:33 CEST] <durandal_1707> but, but I can't see what x did to y
[13:40:46 CEST] <JEEB> wat?
[13:41:17 CEST] <thardin> couldn't you just fiddle with the dns a bit?
[13:41:49 CEST] <thardin> point the existing domain to 127.0.0.1 until the attack is over, add git2.videolan.org in the meantime
[14:03:59 CEST] <durandal_1707> JEEB: what other dev did to some files
[14:04:31 CEST] <durandal_1707> This also means I can't fix bug because git is down
[14:06:20 CEST] <nevcairiel> you can fix bugs all the same
[14:07:55 CEST] <durandal_1707> but can't distribute fixes to others
[14:08:24 CEST] <nevcairiel> they'll survive for a couple hours
[14:09:12 CEST] <durandal_1707> hours will become days in no time
[15:23:13 CEST] <durandal_1707> wm4: for once a second I thought you posted diff :p
[15:26:31 CEST] <ubitux> seems ppl are pretty tense again
[15:28:19 CEST] <wm4> because they try, but to no avail
[15:28:23 CEST] <wm4> it's just the same shit as ever
[15:29:29 CEST] <mateo`> if all this tension if due to people not using git format-patch ..., just revoke ssh keys until they comply with this policy
[16:07:12 CEST] <durandal_1707> mateo`: and I'm not the one with those powers
[16:24:23 CEST] <mateo`> durandal_1707: who can do it ?
[16:27:00 CEST] <durandal_1707> dunno, probably michaelni 
[16:29:39 CEST] <mateo`> I mean (IMHO) it's just a matter of sending a (last) warning to the people who are concerned, and if it happens again, the key should be revoked, no feelings involved, it's just a technical rule.
[16:30:14 CEST] <mateo`> It just like writing code, you do shit, your compiler slaps you in the face.
[16:33:03 CEST] <Illya> mateo`: I wish. It'd be so much easier. I seem to write conforming code in terms of the compiler, but then it just doesn't work.
[16:34:24 CEST] <mateo`> :D
[16:46:57 CEST] <durandal_1707> I fixed bug Carl thinks is regression
[16:47:39 CEST] <wm4> that's not exclusive
[16:47:53 CEST] <wm4> there are many "fixes" for broken files that break legitimate files
[16:49:01 CEST] <durandal_1707> the amv sample from fate plays ok
[16:50:12 CEST] <nevcairiel> git.videolan.org works again btw, the days of peril are over =p
[16:51:25 CEST] <nevcairiel> the commit carl quoted to be the regression did one odd thing - it  removed the clipping of the value
[16:54:18 CEST] <michaelni> mateo`,  we could attempt to detect misformated patches and reject them in mailman with a message but iam no sure how to implement that exactly, its important that we dont loose "git am" good patches, iam no mail expert
[16:55:24 CEST] <michaelni> but i certainly can apply some patch to mailman code or add something into postfix config if someone has tested changes for that and people agree that we should do that
[16:58:15 CEST] <nevcairiel> thats too agressive, and we dont want to alienate new contributors that may not know any better yet, all we ask in this regard is that seasoned contributors bother to follow the rules, if that takes ignoring their patches until they are submitted properly (and of course no push access), then thats better than some fragile mailman hackery
[17:00:34 CEST] <mateo`> I agree
[17:00:53 CEST] <BtbN> also sounds like it could potentially have false-positives, like blocking replies with patches behind >
[17:01:08 CEST] <BtbN> I don't think that could be implemented in a perfect way.
[17:01:32 CEST] <ubitux> 1. how/who decides for the sanctions 2. who has the power to apply them?
[17:03:04 CEST] <michaelni> i can close a git account if theres some policy/agreement about it but i think ignoring misformated patches is enough here
[17:03:26 CEST] <nevcairiel> not if said person then just pushes them
[17:03:26 CEST] <michaelni> also iam happy to help people to fix their mail setup so send-email works
[17:04:05 CEST] <michaelni> nevcairiel, "if" but the patch should first be posted with commit message
[17:04:12 CEST] <michaelni> before it can be pushed
[17:04:28 CEST] <nevcairiel> but thats whats not happening
[17:04:45 CEST] <nevcairiel> yes, thats what should happen
[17:04:53 CEST] <nevcairiel> but its not, which is why we are having this entire discussion :D
[17:07:17 CEST] <BBB> michaelni: I think thats nicer, yes, lets help people setup their email so send-email works
[17:07:34 CEST] <BBB> I think all this punishment isnt helping, especially for new contributors who just dont know how we roll exactly
[17:07:50 CEST] <cone-084> ffmpeg 03Paul B Mahol 07master:a5c5ae5ed855: avcodec/iff: add support for ANBR files
[17:07:51 CEST] <cone-084> ffmpeg 03Paul B Mahol 07master:7494b49780b2: avformat/iff: parse DPAN chunk
[17:08:03 CEST] <mateo`> BBB: I think it's not about new contributors
[17:08:09 CEST] <BBB> I know, its about carl
[17:08:15 CEST] <wm4> <BBB> michaelni: I think thats nicer, yes, lets help people setup their email so send-email works <- isn't that what we've always done
[17:08:21 CEST] <BBB> yes
[17:08:24 CEST] <BBB> and carl keeps ignoring it
[17:08:25 CEST] <jamrial> most if not all new contributors so far needed to be told once to fix the way they send patches
[17:08:45 CEST] <BBB> I dont think we need to punish carl any more than just keep telling him to fix his mail client and otherwise not review his patches
[17:08:52 CEST] <wm4> so, punish carl
[17:08:53 CEST] <RiCON> i sent a patch using gmail once, and then configured git-email
[17:09:00 CEST] <BBB> the issue I have with carl is that he attacked derek
[17:09:21 CEST] <BBB> that is utterly unrelated to his mail client setup, is not remedied by him fixing his mail client, and also has a much higher cost to out project
[17:09:29 CEST] <BBB> (i.e. we lost derek)
[17:09:33 CEST] <ubitux> i remember several ppl trying to help carl figure out how to use pkg-config
[17:09:36 CEST] <ubitux> good luck with git send-email
[17:09:43 CEST] <ubitux> i don't think that's a technical issue
[17:09:47 CEST] <BBB> I dont care about carl using send-email or pkg-config
[17:09:51 CEST] <nevcairiel> we would be happy with format-patch
[17:10:03 CEST] <nevcairiel> but yes what BBB said, those attacks were entirely uncalled for
[17:10:05 CEST] <jamrial> BBB: systematically ignoring a rule because he thinks it shouldn't apply to him then subtly insulting or attacking those that politely point it out will not change just by telling him to stop for the hundredth time
[17:10:17 CEST] <BBB> the attack is the issue
[17:10:21 CEST] <BBB> ignoring the rule is minor
[17:10:23 CEST] <BBB> (imo)
[17:10:34 CEST] <BBB> so we should focus on the attack, not help him figure out send-email/pkg-config
[17:10:40 CEST] <BBB> he can do that himself if he cares
[17:10:52 CEST] <BBB> I mean, if you can code ffmpeg, you can certainly use pkg-config/git send-email
[17:11:07 CEST] <jamrial> he knows very well how to use git
[17:11:29 CEST] <ubitux> yes, it's obviously a will issue
[17:11:33 CEST] <jamrial> he uses git diff to make the stuff he sends
[17:11:42 CEST] <jamrial> he can obviously instead use git format-patch
[17:11:47 CEST] <jamrial> he just doesn't want to
[17:11:57 CEST] <BBB> anyway, my suggestion is to punish carl for attacking derek and stuff
[17:12:11 CEST] <BBB> but its just a suggestion
[17:13:15 CEST] <ubitux> and how do we decide?
[17:13:37 CEST] <iive> BBB: The stuff between Derek and Carl is not from yesterday.
[17:14:07 CEST] <BBB> ubitux: we vote :)
[17:14:09 CEST] <iive> I've seen too many time Derek entering the fray when Carl is involved.
[17:14:26 CEST] <iive> just to fuel the flames.
[17:14:34 CEST] <ubitux> BBB: on the ml with the [DECISION] thing?
[17:14:53 CEST] <ubitux> who starts it?
[17:14:56 CEST] <BBB> whatever method you prefer; we can vote on the ML, we can vote on the IRC meeting
[17:15:05 CEST] <BBB> we can do anonymous ballot voting
[17:15:47 CEST] <ubitux> as the lazy bum, i'm only willing to vote, nor participate in the elaboration process and infrastructure for voting
[17:15:49 CEST] <BBB> someone needs to re-generate the list of eligible voters
[17:16:08 CEST] <BBB> I think we still have that list from last time
[17:16:13 CEST] <ubitux> so as i said in the mail, the people who really wants that need to stand up
[17:16:15 CEST] <BBB> and then send email to these people and ask them to vote yes or no
[17:16:20 CEST] <BBB> ok
[17:16:33 CEST] <BBB> I believe paul asked for it
[17:16:38 CEST] <jamrial> BBB: attacking derek and being a passive aggressive asshole with paul are just the most recent examples
[17:17:00 CEST] <BBB> jamrial: you can create the list and send email also if you want
[17:17:10 CEST] <BBB> you can include any other examples you feel are appropriate
[17:17:12 CEST] <atomnuker> as of last time: http://ffmpeg.org/pipermail/ffmpeg-devel/2015-November/183803.html
[17:17:35 CEST] <BBB> very good
[17:17:39 CEST] <atomnuker> no one else was added since
[17:17:43 CEST] <BBB> so theres a list of voters, I think we know what we want to vote on
[17:17:50 CEST] <wm4> heh I'll probably fall of the number of commits list
[17:17:52 CEST] <BBB> so I think whoever wants to do it can now do it
[17:18:10 CEST] <ubitux> wm4: afaik you're part of the votee comittee
[17:18:33 CEST] <BBB> did llogan get two votes?
[17:18:43 CEST] <BBB> he was already on the list, but he was newly added later also
[17:18:46 CEST] <atomnuker> we decided to keep the list rolling so anyone with 50 commits done over a year is in
[17:18:46 CEST] <BBB> how did that happen?
[17:19:31 CEST] <BBB> anyway nevermind
[17:19:41 CEST] <jamrial> BBB: the list below are the new additions. the list above is the full list
[17:19:49 CEST] <BBB> ah good
[17:19:53 CEST] <jamrial> at least that's how i read it
[17:20:16 CEST] <michaelni> btw, somene might want to look at what google returns for carl daemon404 irc site:lists.ffmpeg.org
[17:21:00 CEST] <michaelni> <Daemon404> carl you fucktard [18:30] ...<Daemon404> carl is an asshole ...
[17:21:22 CEST] <michaelni> not taking sides at all, just wanted to point at it
[17:21:49 CEST] <wm4> suuure you're not taking sides
[17:22:27 CEST] <BBB> were not voting on punishing derek, were voting on punishing carl, right?
[17:22:35 CEST] <ubitux> michaelni: cehoyos has a much more passive/aggressive behaviour, derek is much more abrasive personality wise
[17:23:19 CEST] <ubitux> also, cehoyos is rarely on irc 
[17:23:25 CEST] <wm4> (only abrasive once he has to deal with fucktards)
[17:26:53 CEST] <michaelni> wm4, yes iam sure iam not taking sides, i dont like the hostilities from either side 
[17:27:52 CEST] <omerjerk> Hey
[17:28:01 CEST] <omerjerk> I get this warning while decoding - Using AVStream.codec to pass codec parameters to muxers is deprecated, use AVStream.codecpar instead.
[17:28:12 CEST] <omerjerk> What's the reason for this ?
[17:28:29 CEST] <wm4> API changes
[17:28:33 CEST] <omerjerk> I don't even remember when it started to appear. :/
[17:28:56 CEST] <RiCON> since it was deprecated, at least :)
[17:29:18 CEST] <RiCON> that's mostly a warning for developers
[17:29:21 CEST] <omerjerk> Any examples what to update in my code ?
[17:29:40 CEST] <nevcairiel> replace codec with codecpar
[17:30:47 CEST] <ubitux> omerjerk: it's a recent change
[17:30:57 CEST] <ubitux> not in any release afaik
[17:31:02 CEST] <omerjerk> okay.
[17:31:07 CEST] <omerjerk> https://github.com/omerjerk/FFmpeg/blob/float/libavcodec/alsdec.c 
[17:31:12 CEST] <omerjerk> What to update here ?
[17:31:23 CEST] <omerjerk> I can't find any variable named "codec"
[17:32:34 CEST] <pfelt> OT:  greatest of mornings everyone!
[17:32:48 CEST] <BBB> codecs dont need updating
[17:33:07 CEST] <BBB> some demuxers may need updating, and some application code may need updating to use codecpar instead of codec for lavf<>lavc interaction
[17:33:21 CEST] <wm4> omerjerk: it doesn't affect libavcodec
[17:33:35 CEST] <wm4> only libavformat, or anyone using libavformat
[17:33:39 CEST] <omerjerk> oh okay.
[17:34:08 CEST] <ubitux> "Working on porting libebur128 to FFmpeg now." pfff... :(
[18:15:35 CEST] <cone-084> ffmpeg 03Michael Niedermayer 07master:7ecfe4dc3634: avcodec/diracdec: Fix potential integer overflow
[18:43:30 CEST] <cone-084> ffmpeg 03Stefano Sabatini 07master:1ab8e03a457a: lavf/mpegtsenc: set metadata stream type and write descriptor for ID3 packets
[18:43:31 CEST] <cone-084> ffmpeg 03Stefano Sabatini 07master:4b38df82c25c: lavf/mpegts: add ID3 entry to the REGD_types array
[21:08:53 CEST] <durandal1170> can i get explicit review of amv patch?
[21:21:08 CEST] <cone-084> ffmpeg 03Michael Niedermayer 07master:699201ac0ef6: avcodec/options-test: Check avcodec_open2() return value
[21:49:33 CEST] <jamrial_> durandal1170: the ima amv patch should be ok
[23:48:33 CEST] <cone-084> ffmpeg 03Paul B Mahol 07master:14992421df88: avcodec/adpcm: pick correct step_index for IMA AMV
[00:00:00 CEST] --- Fri May 20 2016