[Ffmpeg-devel-irc] ffmpeg-devel.log.20190819

[burek]

[00:18:51 CEST] <michaelni> kierank, if what i said was rude, is something only a neutral person could judge neither I nor you are neutral. What i can assure you is that i tried to write nice and polite replies. If i failed then iam sorry. No rudeness was intended if something does appear rude. 
[00:19:52 CEST] <kierank> I don't need to hear another riddle
[00:22:05 CEST] <Compnn> michaelni apologized if he came off as rude
[00:22:06 CEST] <kierank> As usual no compromise from you
[00:22:09 CEST] <Compnn> and you act like a shit kierank 
[00:22:33 CEST] <kierank> Oh you have a new nick, I need to add that one to ignore
[00:22:35 CEST] <Compnn> why do you even care what michael says
[00:22:37 CEST] <Compnn> lol
[00:22:43 CEST] <Compnn> michaelni, just add durandal_1707 to security list 
[00:23:03 CEST] <Compnn> dont matter if its just nonsense fuzz
[00:23:07 CEST] <Compnn> durandal_1707 wants to see it
[00:23:20 CEST] <thardin> how many are on that list?
[00:23:35 CEST] <Compnn> micheal posted a list to ml
[00:23:40 CEST] <kierank> iive and compn
[00:23:47 CEST] <Compnn> im not on security list
[00:24:00 CEST] <Compnn> nor do i want to be
[00:24:03 CEST] <kierank> For FFmpeg, and Diego from libav
[00:24:19 CEST] <Compnn> kierank thinks you mean irc ignore list
[00:24:38 CEST] <kierank> michaelni: so give Paul access to security like he wants
[00:24:49 CEST] <thardin> kierank: I mean the security list
[00:24:57 CEST] <kierank> Nobody every authorised you, you only inherit it from being leader
[00:25:05 CEST] <durandal_1707> i do not want it any more
[00:26:01 CEST] <Compnn> the number of adults decreases in this project
[00:26:11 CEST] <thardin> mmhm
[00:26:26 CEST] <Compnn> although i dont think acting like a child will grant immortality.
[00:27:50 CEST] <durandal_1707> from where you picked that one?
[00:28:36 CEST] <Compnn> durandal_1707, when does your vacation end ?
[00:29:33 CEST] <durandal_1707> never, i will never ever work with ffmpeg any more
[00:30:02 CEST] <durandal_1707>  i just need to complete gsoc stuff and bye bye
[00:32:15 CEST] <durandal_1707> and maybe sierpinski vsrc filter and some other small stuff
[00:33:10 CEST] <durandal_1707> like msrle encoder, the fun stuff as usual
[00:36:11 CEST] <kierank> durandal_1707: that is very sad
[00:36:52 CEST] <kierank> Many people leaving because of toxic atmosphete
[00:37:04 CEST] <durandal_1707> not at all, my life will not be miserable looking at certain patches
[00:37:20 CEST] <kierank> We should fork maybe
[00:37:23 CEST] <kierank> And you work on fork
[00:37:57 CEST] <kierank> Just pull patches and forget about stupid fuzz cofr
[00:37:59 CEST] <kierank> Code
[00:38:25 CEST] <kierank> Most timeout issues are theoretical
[00:39:30 CEST] <thardin> I find myself wondering who the customer is that wants all these fixes, that puts undue burden on other devs
[00:40:02 CEST] <thardin> or which free software project. what is the specific use case
[00:40:25 CEST] <kierank> Google of course
[00:40:25 CEST] <J_Darnley> Google.  Didn't someone mention about the load on poor Google's servers?
[00:40:30 CEST] <J_Darnley> snap
[00:40:49 CEST] <thardin> well then google can pay us
[00:41:06 CEST] <Compnn> good luck with fork , i wish ye well
[00:41:58 CEST] <Compnn> run into the same problem, its not the projects (ffmpeg, libav, ffmbc) its that the developers ego cant fit in the same hemisphere
[00:42:27 CEST] <thardin> surely they must know how to set up cpu limits on jobs
[00:42:40 CEST] <thardin> I do it all the time on hpc stuff
[00:42:53 CEST] <Compnn> google is hiring if you want to apply for a job
[00:43:01 CEST] <thardin> no thanks
[00:43:30 CEST] <kierank> Lemme stop you there thardin
[00:43:45 CEST] <kierank> let me explain what a Denial of Services is:
[00:43:57 CEST] <kierank> <explanation of DoS>
[00:44:01 CEST] <kierank> as nauseum
[00:44:14 CEST] <thardin> I see what you did there
[00:44:36 CEST] <kierank> That is basically the response anyone gets when they question the timeout patches
[00:45:56 CEST] <thardin> so I've noticed
[00:47:46 CEST] <thardin> I've also noticed most of them are linear slowdowns
[00:47:57 CEST] <kierank> Denial of patch review
[00:48:02 CEST] <kierank> DoPR
[00:48:15 CEST] <thardin> denial of more interesting work
[00:48:21 CEST] <kierank> Most of them the code has been written for that test case
[00:48:31 CEST] <kierank> And further test cases will need more hacks
[00:48:37 CEST] <kierank> And more random numbers tweaked
[00:49:36 CEST] <kierank> At least when I fuzzed I found real timeouts
[00:49:43 CEST] <kierank> Actual thread deadlocks
[00:49:48 CEST] <iive> is misspelling "Ad nauseam" part of the joke?
[00:50:16 CEST] <kierank> The Google fuzzed is single threaded though
[00:50:19 CEST] <kierank> So it won't find those
[00:53:38 CEST] <thardin> sleep time
[00:55:09 CEST] <J_Darnley> I liked I how I missed the few lines between "let me explain DoS" and "that's basically it"
[00:56:38 CEST] <Compnn> <kierank> Lemme stop you there thardin
[00:56:40 CEST] <Compnn> <kierank> let me explain what a Denial of Services is:
[00:56:40 CEST] <Compnn> <kierank> <explanation of DoS>
[00:56:40 CEST] <Compnn> * J_Darnley has quit (Read error: Connection reset by peer)
[00:56:40 CEST] <Compnn> <kierank> as nauseum
[00:56:40 CEST] <Compnn> <thardin> I see what you did there
[00:56:40 CEST] <Compnn> * J_Darnley (~J_Darnley at d51A44418.access.telenet.be) has joined
[00:57:20 CEST] <J_Darnley> perfect timing
[01:55:49 CEST] <mkver> Can avcodec_decode_subtitle2 actually modify its input packet? It makes a fuss about not modifying it, yet the packet parameter is not const.
[04:17:21 CEST] <Compn> was i too harsh on kierank ?
[04:31:14 CEST] <Compn> jdarnley, was i too hard on kierank ?
[05:45:19 CEST] <jamrial> kierank: neither compn or iive are on the security alias. i don't know where you got that from
[05:45:32 CEST] <jamrial> the list is in the maintainers' file
[08:25:30 CEST] <j-b> wow, again a nice and peaceful IRC week-end here...
[09:48:20 CEST] <Compn> j-b, how goes it ? 
[09:48:49 CEST] <Compn> same old adware using your project, and apple/microsoft and google ignoring imposters in app stores?
[09:48:55 CEST] Action: Compn been out of the loop
[09:54:56 CEST] <durandal_1707> kierank: when i can expect v360 patches pushed?
[09:55:11 CEST] <kierank> 10 mins
[09:58:05 CEST] <kierank> durandal_1707: patchv3?
[09:58:49 CEST] <kierank> dammit it's one of those inline patches
[09:58:57 CEST] <kierank> is there a way to get those as a file
[09:59:06 CEST] <kierank> some of us live in 2019 and don't use cli email clients
[09:59:27 CEST] <durandal_1707> patchwork?
[10:00:12 CEST] <durandal_1707> download as mbox
[10:03:39 CEST] <j-b> Compn: here ;)
[10:04:15 CEST] <durandal_1707> Compn is potplayer user
[10:04:18 CEST] <Compn> mmm? good to be somewhere.
[10:08:15 CEST] <kierank> done
[10:08:16 CEST] <cone-505> ffmpeg 03Eugene Lyapustin 07master:b26094e217d4: avfilter: add v360 filter
[10:08:16 CEST] <cone-505> ffmpeg 03Eugene Lyapustin 07master:d044252aac77: avfilter/vf_v360: add padding option for cubemap
[10:08:16 CEST] <cone-505> ffmpeg 03Eugene Lyapustin 07master:596ecc295fc4: avfilter/vf_v360: add dual fisheye format
[10:10:15 CEST] <durandal_1707> Hooray!!!
[10:17:29 CEST] <thardin> woohoo
[10:21:55 CEST] <Compn> heh, carl makes a decoder for a codec that we had a request 4 years ago for . nice job! https://trac.ffmpeg.org/ticket/8081
[10:24:23 CEST] <durandal_1707> Compn: will you add it to mplayer?
[10:25:21 CEST] <Compn> i dont have my usual computer with me durandal_1707, but yes i will sync up mplayer to ffmpeg when i get a chance
[10:27:36 CEST] <kierank> durandal_1707: is v360 important?
[10:31:01 CEST] <durandal_1707> kierank: to some people yes
[10:32:15 CEST] <durandal_1707> some of them use facebook filter for ffmpeg - they needed to build own ffmpeg
[10:32:50 CEST] <durandal_1707> hopefully they will now stop using facebook filter
[10:32:55 CEST] <JEEB> cool
[11:32:03 CEST] <jdarnley> Compn: I don't think so but I wan't following the conversation closely.
[14:30:39 CEST] <BtbN> Anyone interested in working on AMD hwaccel stuff? I still got a spare RX590 I can send around Europe. And to the US I guess, but shipping is quite expensive for that.
[14:58:16 CEST] <nevcairiel> i got a 560 for work a couple months ago, never got around to testing shit yet
[14:59:23 CEST] <nevcairiel> been busy with some annoying mobile topic
[15:40:48 CEST] <BtbN> I'm having so many issues with the Navi card I was sent that I'm very tempted to go back to Nvidia :/
[15:46:38 CEST] <jamrial> BtbN: 5700x? if you're trying the video encode parts, they replaced UVD/VCE with something new, so software supports may not be great just yet
[15:46:53 CEST] <BtbN> RX5700 (non-XT)
[15:47:06 CEST] <BtbN> Just general system health is terrible
[15:47:18 CEST] <BtbN> There is a documented known issue that "Enhanced Sync" causes system crashes
[15:47:24 CEST] <BtbN> yet it automatically re-enables each driver update
[15:47:35 CEST] <BtbN> For each Game/Application individually even.
[15:47:59 CEST] <BtbN> Plus, playing a video on Screen 2 and playing a Game on screen 1 is a certain path to weird artifacts or even bluescreens within 30 minutes
[15:48:15 CEST] <JEEB> \o/
[15:48:27 CEST] <J_Darnley> Wow.  Bluescreens?  In Current Year?
[15:49:40 CEST] <JEEB> J_Darnley: best paper train tickets? In Current Year? https://pbs.twimg.com/media/EB6LV8jW4AAnVtW.jpg:orig
[15:50:14 CEST] <J_Darnley> punched holes indicating more data than "used"?  wow
[15:50:26 CEST] <JEEB> yes xD
[15:50:47 CEST] <JEEB> 500+100+40 = 640 (two times 320 for the area I was travelling)
[15:50:52 CEST] <JEEB> then at the top the date
[15:51:00 CEST] <JEEB> after it there's the bitmask
[15:51:08 CEST] <JEEB> which is punched @ "one way"
[15:51:26 CEST] <JEEB> then hand-written comment with "large, 2" (two adults)
[15:51:49 CEST] <J_Darnley> never change, japan
[15:52:05 CEST] <BtbN> Aren't those Chinese glyphs?
[15:52:19 CEST] <J_Darnley> katakana seems to be present on a couple of the stations
[15:52:52 CEST] <J_Darnley> Otherwise, yes, they are chinese characters (or not but that is politics)
[15:54:13 CEST] <J_Darnley> oh, hiragana under the text field too
[15:55:11 CEST] <JEEB> BtbN: hanzi, kanji, however you clal them in which language :)
[16:38:04 CEST] <cone-505> ffmpeg 03Guo, Yejun 07master:29aeeb3e3e7b: libavfilter/dnn/dnn_backend_tf: fix typo that variable uninitialized.
[16:38:05 CEST] <cone-505> ffmpeg 03Guo, Yejun 07master:67889d4715d4: libavfilter/dnn/dnn_backend_tf: add tf.pad support for tensorflow backend with native model.
[16:38:06 CEST] <cone-505> ffmpeg 03Guo, Yejun 07master:d0fa1a58da71: FATE/dnn: let fate/dnn tests depend on ffmpeg static libraries
[17:42:45 CEST] <kierank> jamrial: those were the people I have ignored on irc
[17:43:27 CEST] <jamrial> ah, my bad
[17:43:35 CEST] <jamrial> guess i mixed the subjects of the discussion
[19:06:25 CEST] <durandal_1707> so no refcounting for fuzz and no cfr for codecs - remain in status quo - bravo' bravo' well played
[19:09:22 CEST] <kierank> durandal_1707: send a patch to the fuzzer
[19:23:32 CEST] <jamrial> durandal_1707: refcounting seems to be enforced internally with both the new and old decode API
[20:04:02 CEST] <thardin> huh, ibc spam
[20:07:45 CEST] <kierank> thardin: you mean you don't get dozens every day
[20:17:43 CEST] <thardin> nah only one per quarter or so
[20:19:37 CEST] <kierank> I get loads
[21:00:34 CEST] <durandal_1707> jamrial: not true
[22:25:49 CEST] <thardin> ffabrice latest hack: https://bellard.org/quickjs/
[00:00:00 CEST] --- Tue Aug 20 2019