[Ffmpeg-devel-irc] ffmpeg-devel.log.20170225

[burek]

[03:54:11 CET] <cone-413> ffmpeg 03Michael Niedermayer 07master:310d2af319d9: avcodec/pngdec: Fix runtime error: left shift of 152 by 24 places cannot be represented in type 'int'
[03:54:11 CET] <cone-413> ffmpeg 03Michael Niedermayer 07master:6bd79ba59f46: avcodec/amrwbdec: Fix 2 runtime errors: left shift of negative value -1
[04:02:26 CET] <cone-413> ffmpeg 03Steven Liu 07master:f73ef3748e83: avformat/hlsenc: fix hls_flags temp_file bug
[04:28:40 CET] <cone-413> ffmpeg 03Hendrik Leppkes 07master:3aef2fceff32: avformat/hlsenc: don't use %s for strftime on msvc
[11:51:03 CET] <nevcairiel> michaelni: 513a34943 broke fate on a variety of boxes, it would appear there is some other issue hidden there, ie. http://fate.ffmpeg.org/report.cgi?time=20170225072151&slot=x86_32-mingw-w64-dll-windows-native
[11:52:07 CET] <nevcairiel> in fact, basically every box that ran in the last couple hours is yellow now
[11:54:45 CET] <michaelni> nevcairiel, odd, i did test this one on several things
[11:55:17 CET] <nevcairiel> a quick check seems to show they all produce the same crc
[12:06:27 CET] <michaelni> valgrind shows nothing
[12:06:46 CET] <nevcairiel> and you still get the same crc?
[12:07:01 CET] <nevcairiel> thats odd
[12:10:40 CET] <michaelni> i can reproduce it with --enable-memory-poisoning
[12:12:08 CET] <nevcairiel> weird that valgrind didnt show it then
[12:12:36 CET] <michaelni> yes
[13:01:34 CET] Action: ubitux just realized his fate instances were stalled one month ago
[13:01:41 CET] <ubitux> they're all turning yellow now 
[13:10:32 CET] <cone-046> ffmpeg 03Michael Niedermayer 07master:d34bf886e963: avcodec/vp56: Implement very basic error concealment
[15:06:42 CET] <kierank> wm4: so now I am confused about the rules, michaelni does not maintain vp5 or vp6 yet has pushed without review as far as I can tell
[15:09:22 CET] <wm4> kierank: this seems to be a rule violation
[15:09:29 CET] <wm4> michaelni: what do we do to enforce the rules?
[15:15:42 CET] <wm4> kierank: yeah, it wasn't even on the ML
[15:15:46 CET] <wm4> unless I'm missing some posts
[15:16:25 CET] <wm4> the maintainer for vp5 and vp6 is "Aurelien Jacobs"
[15:16:41 CET] <wm4> I assume that makes him maintainer of vp56.c
[15:17:28 CET] <wm4> which I interpret as either "the rules don't mean shit" or "the MAINTAINERS file doesn't mean shit"
[15:17:36 CET] <wm4> that's not really a good situation
[15:17:38 CET] <Compn> aurel hasnt been around in a while
[15:17:47 CET] <Compn> where a while = years
[15:17:51 CET] <wm4> Compn: that affects most people list in MAINTAINERS
[15:18:03 CET] <Compn> the maintainers list is more for historical purposes
[15:18:09 CET] <Compn> in my opinion
[15:18:26 CET] <wm4> well according to the "rules", it affects whether or not you can push stuff without reviews
[15:18:44 CET] <wm4> and michaelni is "maintainer" of significant parts of the common code/headers
[15:18:55 CET] <Compn> right, but whats the rule when the maintainer is gone ?
[15:20:09 CET] <wm4> how do you even know whether the maintainer is "gone"
[15:20:17 CET] <wm4> many of them are still reachable by mail
[15:20:32 CET] <Compn> also, this is the kind of stuff that makes it not fun to work on the project for me. e.g. complaining about maintainers file, rules, review, patches, etc yet again. just letting you know my interest in the subject.
[15:21:07 CET] <Compn> review it after its committed, theres not much difference honestly
[15:21:20 CET] <Compn> before commit or after commit, the code is still code
[15:22:05 CET] <wm4> yeah, I'd prefer pushing my code without reviews too
[15:22:10 CET] <wm4> but I'm trying to follow the rules
[15:22:15 CET] <wm4> should that make me an idiot
[15:22:29 CET] <Compn> just ask michaelni again to post patches and move on
[15:22:43 CET] <wm4> it's _completely_ unclear how to behave is the maintainer is obviously not active anymore, or not very active
[15:22:53 CET] <Compn> so propose a rule
[15:25:21 CET] <wm4> Compn: or in other words, if michaelni ignores the rules, why should I respect them
[15:29:24 CET] <Compn> Do NOT commit to code actively maintained by others without permission
[15:29:34 CET] <Compn> since vp56 isnt actively maintained, i dont see a rule violation
[15:29:37 CET] <Compn> http://ffmpeg.org/developer.html#Development-Policy
[15:30:35 CET] <Compn> wm4 : maybe you've been reading some other ruleset ?
[15:31:00 CET] <Compn> in some other project
[15:31:05 CET] <Compn> maybe a fork
[15:33:49 CET] <wm4> so I have to guess whether it's "actively" maintained? also that would make it unmaintained code, not code maintained by michaelni - can you push changes to "unmaintained" code without reviews then?
[15:33:57 CET] <wm4> and without even sending it to the ML
[15:33:57 CET] <Compn> just look at git blame for vp56.c
[15:33:58 CET] <Compn> 2012
[15:34:05 CET] <wm4> what does that matter
[15:34:10 CET] <iive> wm4: is there some problem with the committed code?
[15:34:21 CET] <wm4> iive: who knows?
[15:34:36 CET] <Compn> wm4 : you are trying to apply a rule that is not ffmpeg policy to ffmpeg developers, it is infuriating
[15:34:44 CET] <iive> wm4: you are making it a problem, that's why I am asking you
[15:34:54 CET] <iive> is there any problem with this code
[15:34:58 CET] <Compn> wm4 : propose a new rule , or stop applying rules that dont exist to ffmpeg developers
[15:35:49 CET] <Compn> and yes, since there is no rule about pushing changes to unmaintained code... i say go for it
[15:36:07 CET] <Compn> if anyone complains, point to maintainers and say 'that file wasnt under anyones maintainership'
[15:36:10 CET] <JEEB> it's a valid point that there is no defined rule for "something is maintained", and thus you have to ballpark if the last change by that person was far back enough in the past
[15:36:12 CET] <wm4> <Compn> wm4 : you are trying to apply a rule that is not ffmpeg policy to ffmpeg developers, it is infuriating <- uh what
[15:36:21 CET] <wm4> I'm trying to fucking figure out what the rules are
[15:36:25 CET] <JEEB> ^ this
[15:36:27 CET] <wm4> because nobody fucking knows
[15:36:33 CET] <Compn> [09:09] <wm4> kierank: this seems to be a rule violation
[15:36:40 CET] <wm4> so far it looks like michaelni is doing whatever he pleases
[15:36:44 CET] <Compn> ^^^ no, you made a claim it broke a rule
[15:36:50 CET] <Compn> but there is no rule
[15:36:55 CET] <Compn> you are arguing in bad faith, my friend
[15:37:03 CET] <Compn> and it is shitty.
[15:37:06 CET] <wm4> Compn: well how the fucking fuck does it work then
[15:37:10 CET] <wm4> 1. the patch wasn't on the ML
[15:37:16 CET] <Compn> who says it has to be on the ML ?
[15:37:17 CET] <wm4> 2. it as pushed without a review
[15:37:24 CET] <Compn> who said patches have to be reviewed ?
[15:37:31 CET] <wm4> 3. the maintainer is not michaelni, but someone else
[15:37:34 CET] <JEEB> please, let him finish
[15:37:46 CET] <wm4> 4. the rules say code maintained by someone else has to be reviewed
[15:38:00 CET] <Compn> "actively maintained"
[15:38:12 CET] <wm4> 5. there's a timeout where you can push patches anyway if the maintainer doesn't react, but it's significantly over 0
[15:39:15 CET] <Compn> aurel is gone, afaict, so file is unmaintained , or at least, not actively maintained for years
[15:39:16 CET] <wm4> Do NOT commit to code actively maintained by others without permission.
[15:39:16 CET] <wm4> Send a patch to ffmpeg-devel. If no one answers within a reasonable
[15:39:16 CET] <wm4> time-frame (12h for build failures and security fixes, 3 days small changes,
[15:39:16 CET] <wm4> 1 week for big patches) then commit your patch if you think it is OK.
[15:39:33 CET] <wm4> is he gone? I could reach him by mail a few months ago
[15:39:49 CET] <Compn> i meant from ffmpeg, not gone from the internet or planet earth
[15:39:55 CET] <Compn> heh
[15:40:03 CET] <wm4> so why is he maintainer if he left?
[15:40:29 CET] <Compn> he was maintainer at the time when his name was added to maintainers file :)
[15:40:47 CET] <wm4> and why does this not warrant a review on the ML? we even post patches to the ML for code we maintain
[15:40:55 CET] <wm4> that's bullshit
[15:41:05 CET] <Compn> then dont post patches for code you maintain
[15:41:11 CET] <JEEB> jesus christ
[15:41:11 CET] <Compn> if you prefer not to
[15:41:32 CET] <Compn> its not a policy rule
[15:41:36 CET] <JEEB> the thing is, it would have been nice to get a patch and notice that the maintainer should be removed
[15:41:46 CET] <JEEB> that way it would have been clear that there is no maintainer
[15:41:52 CET] <iive> wm4: there is one rule. always act in good faith.
[15:42:07 CET] <JEEB> although to be honest, I'd still prefer patches on ML pretty much always
[15:42:15 CET] <Compn> then post patches if you prefer.
[15:42:16 CET] <JEEB> the timeouts in some cases seem OK enough
[15:42:48 CET] <iive> wm4: you are just trying to stir up shit and create controvercy.
[15:43:06 CET] <iive> wm4: if you have problem with michael, ask him directly and in private
[15:43:11 CET] <JEEB> I'm pretty sure he is just trying to point out that parts of the rules are currently "ballpark it"
[15:43:12 CET] <iive> do not make it public issue
[15:43:45 CET] <JEEB> and while the MAINTAINERS file seems to have some sort of weight to it, there is no effort in such cases to clean it up
[15:44:15 CET] <Compn> JEEB : no, he said a rule was violated when no rule was violated. if he wants to propose a rule, he can post to ffmpeg-devel with a new rule proposal
[15:44:31 CET] <wm4> all I get from this is "the rules don't mean shit"
[15:44:31 CET] <Compn> stop trying to control the facts, JEEB
[15:44:34 CET] <JEEB> Compn: it is not clear though, since the maintainer is still listed there
[15:44:41 CET] <Compn> "actively maintained"
[15:44:48 CET] <Compn> what part of "actively" do you not understand ?
[15:44:52 CET] <wm4> iive: no, rules are there because they are SUPPOSED to resolve conflict
[15:45:01 CET] <wm4> iive: if they can't do that they're not good rules
[15:45:04 CET] <iive> wm4: and what is the conflict in this case?
[15:45:13 CET] <jkqxz> iive: Compn:  You are trying to sweep aside these concerns by noting that it is mostly irrelevant /in this specific case/.
[15:45:17 CET] <iive> IS THERE A PROBLEM WITH THE CODE\
[15:45:19 CET] <wm4> or alternatively they don't mean anything (can ignore maintainers...)
[15:45:31 CET] <Compn> jkqxz : which concerns? that wm4 wants to propose a new rule or clarification of a rule?
[15:45:31 CET] <nevcairiel> jkqxz: he is not expressing general concerns, he is riding one specific case
[15:45:32 CET] <jkqxz> The specific case is not important, the general rule is.
[15:45:37 CET] <wm4> iive: it's not about code, it's about project development rules
[15:45:46 CET] <Compn> jkqxz : i specifically am not sweeping anything aside, and am encouraging wm4 to propose a new rule
[15:45:53 CET] <Compn> jkqxz : it insults me that you would say this.
[15:46:07 CET] <iive> wm4: you just said that they exist to resolve conflict
[15:46:11 CET] <JEEB> Compn: so if there's no bugs in the code for X months and I am not actively poking other components is that "actively maintaining"
[15:46:15 CET] <iive> but you are creating conflict, by using them.
[15:46:16 CET] <JEEB> (bugs that I know of)
[15:46:44 CET] <Compn> JEEB : i am not dealing with hypothetical questions at the moment in time , sorry
[15:46:47 CET] <JEEB> ok
[15:46:49 CET] <JEEB> no problem
[15:46:49 CET] <wm4> iive: there has been conflicts all along because of unclear/ignored rules
[15:47:37 CET] <Compn> wm4 : so propose a clarification of a rule
[15:48:04 CET] <iive> wm4: Do you want to destroy FFmpeg? Because this is the exactly same shit that broke FFmpeg apart
[15:49:10 CET] <iive> do you want michaelni to be kicked out of FFmpeg?
[15:49:12 CET] <wm4> iive: maybe the rules are not very good or not followed very well which _caused_ it to break apart already at least once?
[15:49:19 CET] <nevcairiel> If you want clearer rules, then propose those, without starting the discussion through attacks on  developers actions
[15:49:27 CET] <Compn> nevcairiel speaks the truth
[15:49:37 CET] <kierank> the rules are clear
[15:49:37 CET] <wm4> nevcairiel: which attack
[15:49:51 CET] <kierank> it's just "maintainers" is rooted in fantasy
[15:49:53 CET] <wm4> I only pointed out that michaelni disrespected Aurlien as maintainer
[15:50:01 CET] <JEEB> ok, something along the road of "if a component is considered not maintained a patch should be sent out to remove <whatever signifies that person as being a maintainer>", if we have noted something as being the sign of being a maintainer. in some cases people have pointed at the MAINTAINERS file, which is well-known to be not maintaned very well, and in some other cases other things have been pointed 
[15:50:07 CET] <JEEB> at.
[15:50:25 CET] <iive> wm4: if Aurlien has objected on that, I would have supported him.
[15:50:32 CET] <nevcairiel> wm4: and anyone with half a brain can see that he is not active anymore, so since I know you have one of those, you choose to ignore that, and thus acted in bad faith
[15:50:39 CET] <iive> what do YOU have to do with it?
[15:51:14 CET] <iive> rules are not laws
[15:51:26 CET] <JEEB> anyways, effectively the ballparking in this case was correct, but it's unfortunate that the patch didn't go through the usual workflow
[15:51:38 CET] <JEEB> and that's it
[15:51:50 CET] <wm4> nevcairiel: then why the fuck is he in the maintainers file?
[15:51:56 CET] <wm4> how can I determine whether someone is active?
[15:52:20 CET] <iive> wm4: rules are there to protect you
[15:52:33 CET] <Compn> git log file.c ... >5 years no commits = unmaintained
[15:52:45 CET] <Compn> git log file.c ... >5 years no commits from maintainer = unmaintained
[15:52:49 CET] <nevcairiel> because people don't tell you when they become inactive, they just go away, its just human nature
[15:52:55 CET] <JEEB> yes
[15:53:06 CET] <JEEB> which is why such files can be rather out-of-date
[15:53:39 CET] <Compn> sometimes people do come back as well :)
[15:53:48 CET] <Compn> rarely
[15:53:50 CET] <JEEB> of course
[15:54:13 CET] <iive> wm4: do you understand what "acting in good faith" means?
[15:59:56 CET] <wm4> sent a patch
[16:03:58 CET] <wm4> "Send a patch to ffmpeg-devel. If no one answers within a reasonable
[16:03:58 CET] <wm4> time-frame (12h for build failures and security fixes, 3 days small changes,
[16:03:58 CET] <wm4> 1 week for big patches) then commit your patch if you think it is OK.
[16:03:58 CET] <wm4> "
[16:04:02 CET] <wm4> is that the entire rule?
[16:04:14 CET] <wm4> e.g. 3 days of silence -> I can just push?
[16:05:54 CET] <iive> yes
[16:06:38 CET] <iive> it is good to ping the maintainer of the file, so he could review it sooner.
[16:08:10 CET] <wm4> so if you post a patch to the ML, and someone requests changes, but is not the maintainer, can I ignore those?
[16:09:04 CET] <nevcairiel> the entire r ules are designed around a concept of good faith, if you ask such things maybe we should design new rules that assume malice in every contributor
[16:09:41 CET] <wm4> again, rules are there to resolve conflicts... or else you could just make them "Act in good faith."
[16:09:57 CET] <wm4> I've seen review comments being ignored myself
[16:11:40 CET] <iive> well, it depends on the request and the consensus.
[16:11:53 CET] <wm4> this "good faith" thing is unclear as hell, because it expects you to fill the gaps in the rules yourself, which obviously doesn't work (kierank and I were assuming michaelni violated that good faith when he just pushed that vp56.c patch without even putting it on the ML)
[16:12:23 CET] <kierank> this is why the "Be excellent to each other" is nonsense
[16:12:30 CET] <kierank> open to interpretation
[16:12:37 CET] <nevcairiel> actually thats the wrong way around wm4, you assuming that is the bad faith =p
[16:12:41 CET] <durandal_1707> who cares about vp56?
[16:12:59 CET] <iive> nevcairiel: +1
[16:13:25 CET] <Compn> durandal_1707 : people who never admit that they were mistaken
[16:13:29 CET] <wm4> nevcairiel: uh no
[16:13:42 CET] <wm4> nevcairiel: I was assuming rules that I would have considered to be part of "good faith"
[16:13:51 CET] <wm4> nevcairiel: and _then_ I assumed michaelni acted in bad faith
[16:14:12 CET] <nevcairiel> you seem to not understand the concept of good faith
[16:14:12 CET] <durandal_1707> i pushed same thing to dnxhd. even without review months ago
[16:14:16 CET] <wm4> or does "good faith" mean we love us all and no rule gets ever violated
[16:14:19 CET] <iive> nevcairiel: +2
[16:15:10 CET] <wm4> nevcairiel: I couldn't see any good faith in a rule being (apparently) violated
[16:15:20 CET] <wm4> only after this discussion it's becoming clear he didn't actually violate any rules
[16:15:28 CET] <nevcairiel> thats because your first assumption is a rule violation
[16:15:36 CET] <wm4> <wm4> nevcairiel: I was assuming rules that I would have considered to be part of "good faith"
[16:15:36 CET] <wm4> <wm4> nevcairiel: and _then_ I assumed michaelni acted in bad faith
[16:15:36 CET] <nevcairiel> your entire mindset is based on bad faith, apparently
[16:15:43 CET] <kierank> what rules were not broken?
[16:15:50 CET] <nevcairiel> none were broken
[16:15:51 CET] <wm4> nevcairiel: it's bad faith from you assuming that
[16:15:57 CET] <nevcairiel> you just assume rules that dont actually exist
[16:16:08 CET] <kierank> "Send a patch to ffmpeg-devel."
[16:16:21 CET] <Compn> if the file is actively maintained by someone else...
[16:16:23 CET] <wm4> nevcairiel: jesus christ
[16:16:37 CET] <iive> Compn: you are not helping ;)
[16:16:46 CET] <Compn> iive : i could say the same to you :)
[16:17:41 CET] <durandal_1707> nobody reviews patches because of evil blood
[16:18:05 CET] <Compn> durandal_1707 : you committed things without review? have you no shame!?
[16:18:10 CET] <Compn> ehe
[16:19:05 CET] <durandal_1707> but patch is 100% correct
[16:19:51 CET] <Compn> i cant believe anyone uses our software , people committing things without review. so unprofessional :)
[16:20:23 CET] <Compn> ok enough trolling
[16:20:27 CET] <Compn> from me, bye
[16:20:56 CET] <wm4> what's I'm more surprised about is that iive disagrees with the patch
[16:21:52 CET] <iive> wm4: michael did wrong. He gets carried sometimes, it's not trival or security change.
[16:22:18 CET] <iive> wm4: however going against him in public is bad faith from you.
[16:22:21 CET] <wm4> well, non-trivial changes to unmaintained files don't require reviews, where do you see this?
[16:22:30 CET] <iive> aka, you brand him criminal and try to punish him.
[16:22:47 CET] <iive> and create huge fuss.
[16:23:29 CET] <iive> wm4: that's why I said to Compn that he is not helping, I do not agree with Compn viewpoint.
[16:23:34 CET] <iive> let it say it this way.
[16:24:21 CET] <iive> I do not think that Aurelien would object on this commit.
[16:24:47 CET] <wm4> whether he would or would not doesn't really matter
[16:24:56 CET] <iive> yes, it does.
[16:25:06 CET] <iive> the rules are there to protect you.
[16:25:19 CET] <wm4> iive: I guess I apologize for doing that and misinterpreting the rules, but I'm not taking back my claim that the rules are shitty and useless
[16:25:20 CET] <durandal_1707> iive: how do you know?
[16:25:24 CET] <iive> if you follow the rules, you are protected from critizm.
[16:25:41 CET] <wm4> that's what I said earlier, rules are there to resolve conflicts
[16:25:51 CET] <wm4> as long as there's no conflict, you don't need rules either
[16:25:58 CET] <iive> no, conflicts are not RESOLVED by rules.
[16:26:28 CET] <iive> and i asked you, what is the conflict here?
[16:26:29 CET] <wm4> ok, prevented
[16:26:50 CET] <wm4> as I said before, it has long been unclear what exactly are the rules
[16:26:56 CET] <iive> if the change is somehow broken, the michael is in hot water
[16:27:12 CET] <iive> because the rules would have protect him.
[16:27:51 CET] <iive> if he had followed him, he could say.. I sent patch, nobody found a problem in it. I didn't see it either. I tried my best.
[16:28:28 CET] <iive> him/them (rules)
[16:29:31 CET] <iive> some of us , the old developers, do trust michael to commit correct code, so we are not that outraged when he commits something directly
[16:30:06 CET] <wm4> I don't really trust him that
[16:30:23 CET] <wm4> I saw too many weird things which go back to his commits
[16:30:25 CET] <iive> bingo!
[16:30:30 CET] <wm4> that doesn't mean I think michaelni is a bad developer
[16:30:53 CET] <wm4> just that it might be good to have additional eyes on his patches if the patches deal with messy cases
[16:31:17 CET] <wm4> where "messy cases" means code or scenarios which are extra tricky
[16:34:44 CET] <wm4> I also acknowledge that Libav sometimes commits outright broken code (or sometimes code which doesn't even compile), but I still don't think having a minimum of other eyes staring at your changes is a bad thing
[16:35:04 CET] <wm4> (context: Libav has a policy to always review changes, even trivial ones)
[16:35:08 CET] <Compn> forced review does slow down development
[16:35:21 CET] <nevcairiel> libav reviews are sometimes quite ridicoulous when some persons just put a "maybe ok" under it and thats all it needs =p
[16:35:21 CET] <iive> they don't follow it that strictly, either.
[16:35:40 CET] <wm4> nevcairiel: yeah, I don't think that should be needed
[16:36:10 CET] <iive> yeh, or "requested changes done locally and pushed".
[16:37:10 CET] <wm4> something closer to the linux kernel development model might probably be good for a large project like ffmpeg
[16:37:25 CET] <wm4> _no_ code gets into the main tree without getting reviewed, but it still works pretty well
[16:38:13 CET] <BtbN> Something like gerrit. But good luck getting people to that.
[16:40:50 CET] <wm4> I wonder if BBB has an opinion about this
[16:43:37 CET] <jkqxz> I think it would be reasonable to say that all patches, however trivial, should be posted to the ML (or an ML, if ffmpeg-security is maybe more suitable for them, I don't know about that).
[16:44:16 CET] <BBB> ?
[16:44:17 CET] <jkqxz> Then the rules around pushing without review always apply directly.
[16:44:34 CET] <BBB> whats the problem?
[16:45:51 CET] <Compn> BBB : some devs want everything posted as patches to ml, and also possibly all patches reviewed before commit (basically libav rules)
[16:45:59 CET] <wm4> BBB: it's apparently allowed to push patches to "unmaintained" code without posting them to the ML, which... surprises me
[16:46:19 CET] <wm4> and I accused michaelni of a rule violation at first
[16:48:27 CET] <durandal_1707> i have patch which needs review
[16:53:17 CET] <wm4> durandal_1707: if nobody reviews it then you can really post it
[16:53:48 CET] <Compn> i think maybe durandal_1707 is asking wm4 to review said patch
[16:54:00 CET] <Compn> durandal_1707 : what is the mail patch subj or ml url ?
[16:55:32 CET] <durandal_1707> it is about aeval filter
[16:55:51 CET] <BBB> my understanding is that all patches are posted to the ML
[16:56:00 CET] <BBB> I distinctly remember having a discussion about that with michaelni
[16:56:10 CET] <BBB> I can look that up but I really dont feel like diggin in this old crap
[16:56:48 CET] <BBB> I also recognize that most old code is fundamentally unmaintained so usually these patches go in without actual review (even though a patch was posted) after the patch lingers on ML for 1-2 days
[16:57:25 CET] <BBB> this isnt any different from modern codecs btw
[16:57:31 CET] <BBB> if I post a vp9 patch, its hard to get review
[16:57:36 CET] <BBB> its generally hard to get review :)
[16:57:53 CET] <BBB> review is unsexy and tricky, but does take considerable time
[16:58:22 CET] <BBB> the libav model is not any better than ours, they may have mandatory review but most reviews are sure why not or ok I guess or maybe ok
[16:59:14 CET] <BBB> you get a second set of eyes but these eyes really didnt want to be there and all they checked for is that the style is diegoesque and theres no system(format c:) in there, theres no actual checking whether the patch is doing the right thing (remember cinepak patches a few weeks ago?)
[17:00:25 CET] <BtbN> There just aren't a lot of people proficient in those areas.
[17:00:39 CET] <BtbN> And those who are are usually the author of the patch, or they are busy.
[17:00:59 CET] <BBB> right
[17:01:19 CET] <BBB> so life is imperfect& Im starting to accept that.
[17:01:37 CET] <wm4> I'd propose every patch gets sent to the ML (except trivial typo or build fixes)
[17:02:00 CET] <wm4> but then you should be allowed to push a patch after 24h
[17:02:10 CET] <wm4> unless it's a big change and the maintainer didn't ack it
[17:02:17 CET] <BBB> I dont like the trivial typo/build fixes exception& if its that simple, ask someone on IRC for a quick ack and commit it in 5 seconds
[17:02:27 CET] <BBB> I agree the push-after-24h is a good idea
[17:02:53 CET] <BBB> but dont pay too much attention to me, I can live with whatever, Im a little less attached to all of this stuff nowadays
[17:03:16 CET] <BtbN> until someone sneaks in a complete rewrite of ffmpeg in rust, and nobody notices for 24h.
[17:03:41 CET] <BBB> he said unless-big-change
[17:03:59 CET] <BBB> which isnt very legal and can thus possibly be abused, but lets just go with it
[17:04:07 CET] <Compn> BBB : no one wants to go back to look at 2011 mails...
[17:04:10 CET] <BBB> if peolpe have commit rights, we assum we can trust them somewaht, right?
[17:04:42 CET] <wm4> except commit rights are given out like candy past the expiration date
[17:05:35 CET] <Compn> more committers speeds up development
[17:09:47 CET] <TerNer> Hi guys, I have succeed to mux .mp4 video to memory from raw .h264 frames
[17:10:24 CET] <TerNer> but I still have no idea how can I streaming it while muxing
[17:11:47 CET] <TerNer> afaik, moov should have 'duration of video' information if I have streaming muxed video
[17:11:50 CET] <BtbN> you should not use raw h264, where does that come from?
[17:12:02 CET] <BtbN> also, you can't stream mp4. As the header gets written at the end.
[17:12:14 CET] <JEEB> enable the option to fragment
[17:12:20 CET] <JEEB> and you will get movie fragments
[17:12:35 CET] <TerNer> BtbN: It comes from raw RGB32 BITMAP data
[17:12:48 CET] <JEEB> anyways, this is usage talk so #ffmpeg is more correct
[17:12:51 CET] <TerNer> JEEB: Is fragmented mp4 can streaming through http?
[17:12:51 CET] <BtbN> but that's not h264
[17:32:27 CET] <durandal_1707> ok, will push without patch review
[17:52:18 CET] <durandal_1707> nobody wants to review 2700 lines patch that i gonna commit right now?
[17:54:54 CET] <kierank> wm4: where is this rule about "unmaintained" code
[17:55:05 CET] <kierank> As far as I can tell the rules are made up here 
[17:58:12 CET] <iive> durandal_1707: you mean the patch you posted 30 minutes ago? Oh yes, pushing it now would really be an act of good faith, Not.
[17:58:46 CET] <JEEB> durandal_1707: so is it your authorship or kostya's?
[17:58:57 CET] <JEEB> I didn't notice an author field but that might be Gmail
[18:00:01 CET] <durandal_1707> iive: i posted patch that is about aevalsrc
[18:00:07 CET] <durandal_1707> long ago
[18:02:02 CET] <iive> avfilter/aeval: trim last frame's number of samples to match requested duration (24 feb) ?
[18:02:15 CET] <iive> that's yesterday, isn't it?
[18:02:41 CET] <BBB> kierank: I sometimes feel that way also
[18:08:35 CET] <durandal_1707> iive: yes
[18:56:32 CET] <cone-654> ffmpeg 03Carl Eugen Hoyos 07master:353f509ee348: lswr/rematrix: Remove an aggressive loop optimization.
[19:54:31 CET] <cone-654> ffmpeg 03Paul B Mahol 07master:dbc7f02a7272: avfilter/aeval: trim last frame's number of samples to match requested duration
[20:49:34 CET] <adeel> tired!
[21:12:16 CET] <durandal_1707> how writing sec patches can be fun at all?
[21:35:57 CET] <durandal_1707> what next video codecs should be added to lavc?
[21:36:10 CET] <durandal_1707> i added fmvc and scpr
[21:36:19 CET] <durandal_1707> whats next?
[21:41:46 CET] <jamrial> durandal_1707: https://trac.ffmpeg.org/ticket/4976
[21:43:07 CET] <durandal_1707> i hate game codecs, they are full of hacks
[21:43:36 CET] <durandal_1707> also i need bunch of samples
[21:44:03 CET] <durandal_1707> there is myriad bink versions
[21:44:28 CET] <durandal_1707> also kostya have patch for some of them
[22:17:40 CET] <cone-654> ffmpeg 03Michael Niedermayer 07master:5eb04570f660: avcodec/wavpack: Check post_shift
[22:17:41 CET] <cone-654> ffmpeg 03Michael Niedermayer 07master:58f3469cc6df: avcodec/wavpack: Fix 280:22: runtime error: left shift of negative value -1
[22:17:42 CET] <cone-654> ffmpeg 03Michael Niedermayer 07master:5804201cbac2: avutil/frame: Reimplement av_frame_new_side_data() without size=0 special case
[22:17:43 CET] <cone-654> ffmpeg 03Michael Niedermayer 07master:7e4f32f4e4b9: avutil/frame: Disallow zero sized frame side data
[22:24:57 CET] <RiCON> durandal_1707: the samples in screenpressor's site don't work yet
[22:25:45 CET] <durandal_1707> RiCON: those are old samples for old version
[22:26:05 CET] <RiCON> oic
[22:43:59 CET] <wm4> kierank: well, currently I'm in a state of extreme alcohol intoxication, but apparently that rule was always there
[22:44:18 CET] <wm4> seriously fuck alcohol, why is that stuff even legal
[22:45:22 CET] <wm4> I don't understand, why do legislators fail so hard
[22:45:45 CET] <wm4> that stuff is obviously 100% harmful
[22:47:19 CET] <atomnuker> hard alcohol or beer?
[22:53:39 CET] <wm4> I think it was a bad combintion of champagne and beet
[22:53:43 CET] <wm4> *beert
[22:53:46 CET] <wm4> *beer
[22:55:34 CET] <JEEB> ah, I know that feeling. except you get shots of salmiakki and vodka together with beer and brandy and whatever
[22:55:47 CET] <JEEB> mixing and matching is probably the worst
[22:58:50 CET] <atomnuker> "salmiakki (in Finland), is a variety of liquorice flavoured with ammonium chloride"
[22:59:47 CET] <atomnuker> if the alcohol doesn't finish you this will
[23:01:03 CET] <JEEB> atomnuker: we have nowadays everything from chocolate to ice cream with that
[00:00:00 CET] --- Sun Feb 26 2017