IRC logs for #buildstream for Wednesday, 2019-09-18

« Tuesday, 2019-09-17 Index Thursday, 2019-09-19 »
*** wberrier has left #buildstream00:43
*** narispo has quit IRC02:01
*** narispo has joined #buildstream02:02
*** tristan has quit IRC05:42
*** tristan has joined #buildstream06:11
*** cs-shadow has quit IRC06:53
*** tristan has quit IRC06:55
*** slaf has quit IRC07:01
*** slaf has joined #buildstream07:04
*** slaf has joined #buildstream07:04
*** slaf has joined #buildstream07:04
*** slaf has joined #buildstream07:04
*** slaf has joined #buildstream07:05
*** slaf has joined #buildstream07:05
*** slaf has joined #buildstream07:05
*** slaf has joined #buildstream07:05
*** tristan has joined #buildstream07:30
*** ahmed89 has quit IRC07:58
benschubertjuergbi: Ok I have more info: the artifact is >60Go (Might want to tweak my flags to compile LLVM...), and the process that dies is buildbox-casd. Are we loading everythign in memory or something?08:25
juergbibenschubert: ok, thanks for the info. I hope not but there might be a bug08:44
juergbiupload batches are limited to 1 MB and we shouldn't prepare all batches in memory08:45
juergbimaybe I should try to look into this before the planned buildbox update for buildstream08:55
*** jonathanmaw has joined #buildstream09:01
*** bochecha has joined #buildstream09:01
benschubertjuergbi: it's quite an edge case, so iw ould not spend too much time to it :)09:13
*** traveltissues has joined #buildstream09:14
juergbibenschubert: while >60 GB could be considered an edge case for a single artifact, loading the whole artifact into memory should still be fixed09:14
benschubertagreed09:15
benschubert our caches will be glad about it :) but I would not delay the buildbox update too much for that, I think we gain more by having better exit from buildbox etc09:15
juergbisure, but there is anyway still a buildbox-casd MR pending review. will not (significantly) delay the buildbox update because of this09:20
benschubertgood for me then :)09:21
gitlab-br-bottlater closed issue #372 (Allow queues to run auxilliary jobs after an element's job finishes) on buildstream https://gitlab.com/BuildStream/buildstream/issues/37209:56
*** narispo has quit IRC10:15
*** narispo has joined #buildstream10:30
gitlab-br-bottraveltissues opened (was WIP) MR !1563 (traveltissues/985->master: workspaces via sourcecache) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/156310:34
*** ChanServ sets mode: +o tristan10:35
tristanjuergbi, benschubert ... (anyone in the room) ... Any objections that I push a 1.91.0 tag in master immediately (as we did for 1.3.0 initially, without any associated snapshot) ?10:38
tristanRationale: When you checkout BuildStream from master and run it, it currently reports that it is something like version 1.3.010:38
benschubertseems sensible, we can anyways have the snapshot published afterwards10:41
tristanright, with 1.3.x we rolled a 1.3.1 as the first one10:42
juergbitristan: no objections for tagging10:47
*** narispo has quit IRC11:25
*** bochecha has quit IRC11:25
*** tristan has quit IRC11:25
*** phildawson_ has quit IRC11:25
*** flatmush has quit IRC11:25
*** valentind has quit IRC11:25
*** juergbi has quit IRC11:25
*** paulsherwood has quit IRC11:25
*** bethw has quit IRC11:25
*** laurence has quit IRC11:25
*** connorshea[m] has quit IRC11:25
*** kailueke[m] has quit IRC11:25
*** reuben640[m] has quit IRC11:25
*** theawless[m] has quit IRC11:25
*** Demos[m] has quit IRC11:25
*** Trevinho[m] has quit IRC11:25
*** waltervargas[m] has quit IRC11:25
*** mrmcq2u[m] has quit IRC11:25
*** dineshdb[m] has quit IRC11:25
*** skullone[m] has quit IRC11:25
*** swick has quit IRC11:25
*** gimpnet-irc[m] has quit IRC11:25
*** asingh_[m] has quit IRC11:25
*** dbuch has quit IRC11:25
*** verdre[m] has quit IRC11:25
*** rafaelff[m] has quit IRC11:25
*** segfault3[m] has quit IRC11:25
*** slaf has quit IRC11:25
*** tiagogomes has quit IRC11:25
*** benbrown has quit IRC11:25
*** kapip has quit IRC11:25
*** aevri has quit IRC11:25
*** krichter[m] has quit IRC11:25
*** albfan[m] has quit IRC11:25
*** tlater[m] has quit IRC11:25
*** cgmcintyre[m] has quit IRC11:25
*** nielsdg has quit IRC11:25
*** doras has quit IRC11:25
*** mattiasb has quit IRC11:25
*** awacheux[m] has quit IRC11:25
*** jjardon[m] has quit IRC11:25
*** m_22[m] has quit IRC11:25
*** abderrahim[m] has quit IRC11:25
*** pro[m] has quit IRC11:25
*** ssssam[m] has quit IRC11:25
*** tchaik[m] has quit IRC11:25
*** jjardon has quit IRC11:25
*** benschubert has quit IRC11:25
*** lchlan has quit IRC11:25
*** persia has quit IRC11:25
*** hergertme has quit IRC11:25
*** ironfoot has quit IRC11:25
*** raoul_ has quit IRC11:25
*** traveltissues has quit IRC11:25
*** jonathanmaw has quit IRC11:25
*** tpollard has quit IRC11:25
*** samkirkham has quit IRC11:25
*** Kinnison has quit IRC11:25
*** SotK has quit IRC11:25
*** Iker has quit IRC11:25
*** qinusty has quit IRC11:25
*** adds68 has quit IRC11:25
*** jward has quit IRC11:25
*** WSalmon has quit IRC11:25
*** lantw44 has quit IRC11:25
*** mohan43u has quit IRC11:25
*** jswagner has quit IRC11:25
*** gitlab-br-bot has quit IRC11:25
*** Trevinho has quit IRC11:25
*** traveltissues has joined #buildstream11:25
*** jonathanmaw has joined #buildstream11:25
*** tpollard has joined #buildstream11:25
*** samkirkham has joined #buildstream11:25
*** Kinnison has joined #buildstream11:25
*** SotK has joined #buildstream11:25
*** Iker has joined #buildstream11:25
*** qinusty has joined #buildstream11:25
*** adds68 has joined #buildstream11:25
*** jward has joined #buildstream11:25
*** WSalmon has joined #buildstream11:25
*** lantw44 has joined #buildstream11:25
*** gitlab-br-bot has joined #buildstream11:25
*** Trevinho has joined #buildstream11:25
*** narispo has joined #buildstream11:25
*** laurence has joined #buildstream11:25
*** slaf has joined #buildstream11:25
*** tiagogomes has joined #buildstream11:25
*** benbrown has joined #buildstream11:25
*** kapip has joined #buildstream11:25
*** aevri has joined #buildstream11:25
*** krichter[m] has joined #buildstream11:25
*** albfan[m] has joined #buildstream11:25
*** tlater[m] has joined #buildstream11:25
*** cgmcintyre[m] has joined #buildstream11:25
*** nielsdg has joined #buildstream11:25
*** doras has joined #buildstream11:25
*** mattiasb has joined #buildstream11:25
*** awacheux[m] has joined #buildstream11:25
*** jjardon[m] has joined #buildstream11:25
*** m_22[m] has joined #buildstream11:25
*** abderrahim[m] has joined #buildstream11:25
*** pro[m] has joined #buildstream11:25
*** ssssam[m] has joined #buildstream11:25
*** tchaik[m] has joined #buildstream11:25
*** jjardon has joined #buildstream11:25
*** benschubert has joined #buildstream11:25
*** lchlan has joined #buildstream11:25
*** persia has joined #buildstream11:25
*** hergertme has joined #buildstream11:25
*** ironfoot has joined #buildstream11:25
*** raoul_ has joined #buildstream11:25
*** irc.acc.umu.se sets mode: +oo jjardon ironfoot11:25
*** paulsherwood has joined #buildstream11:25
*** flatmush has joined #buildstream11:25
*** juergbi has joined #buildstream11:25
*** swick has joined #buildstream11:25
*** bethw has joined #buildstream11:25
*** tristan has joined #buildstream11:25
*** phildawson_ has joined #buildstream11:25
*** bochecha has joined #buildstream11:25
*** mohan43u has joined #buildstream11:26
*** valentind has joined #buildstream11:26
*** jswagner has joined #buildstream11:27
*** ChanServ sets mode: +o tristan11:27
tristantag pushed11:52
*** segfault3[m] has joined #buildstream11:53
*** reuben640[m] has joined #buildstream11:59
*** theawless[m] has joined #buildstream12:02
*** kailueke[m] has joined #buildstream12:08
*** tristan has quit IRC12:09
*** connorshea[m] has joined #buildstream12:39
*** phoenix has joined #buildstream12:54
*** jonathanmaw has quit IRC13:04
*** phoenix has quit IRC13:13
*** waltervargas[m] has joined #buildstream13:16
*** mrmcq2u[m] has joined #buildstream13:24
*** Demos[m] has joined #buildstream13:29
*** jonathanmaw has joined #buildstream13:43
*** skullone[m] has joined #buildstream13:57
*** Trevinho[m] has joined #buildstream14:02
*** tpollard has quit IRC14:30
*** tpollard has joined #buildstream14:30
*** traveltissues has quit IRC14:32
*** traveltissues has joined #buildstream14:32
tlater[m]Hrm, I wonder if it would make sense to load plugins on-demand in the loader (i.e., whenever we encounter an element 'kind'). We could error out much more quickly in some cases if we did.14:43
*** asingh_[m] has joined #buildstream14:43
KinnisonRather than waiting until we instantiate the elements?14:44
*** bochecha has quit IRC14:47
*** gimpnet-irc[m] has joined #buildstream14:49
*** dineshdb[m] has joined #buildstream14:49
benschuberttlater[m]: but that would make the other error cases much slower? :)14:51
tlater[m]Kinnison: Yeah14:51
benschubertDont' get me wrong, I'd love to have a single loading step, parallelisable if possible14:51
benschubertbut ... that's a lot of effort14:52
KinnisonWe could fire off test-loads of plugins in spawned subprocesses14:52
tlater[m]Which error cases would be slower?14:52
Kinnisonso they happen in parallel with the rest of the load14:52
benschuberttlater[m]: cicles, unresolvable dependencies14:52
tlater[m]That would be interesting14:53
benschubertmisformed yaml files14:53
tlater[m]Ehh, if you use like 100 plugins I expect that not to slow down things too much14:53
tlater[m]And I don't think there are 100 plugins14:53
tlater[m]I'm thinking that because of this: https://gitlab.com/BuildStream/buildstream/commit/5edac8407254953eb778975a82f8bcbeed4608b714:54
tlater[m]It just seems like such a waste to load dependencies for an element that will never actually have dependencies, just to error out a minute later.14:55
tlater[m]When we already have all the information we need to decide whether to abort or not.14:56
*** tpollard has quit IRC15:10
*** cs-shadow has joined #buildstream15:14
*** verdre[m] has joined #buildstream15:17
*** phoenix has joined #buildstream15:19
*** rafaelff[m] has joined #buildstream15:20
*** narispo has quit IRC15:31
*** narispo has joined #buildstream15:31
*** tristan has joined #buildstream15:32
benschuberttlater[m]: if it doesn't slow down the successful case, I'm fine with that15:38
benschubertbut I'd much rather optimize for success than failure15:38
tlater[m]benschubert: That's the nice thing, I don't think it would, since we do that work later anyway.15:38
benschubertfair enough, if it's still as fast (or faster) I'm happy with it15:39
*** dbuch has joined #buildstream15:41
*** bochecha has joined #buildstream15:42
*** tpollard has joined #buildstream16:33
*** tiagogomes has quit IRC16:47
*** phoenix has quit IRC17:25
gitlab-br-bottlater opened MR !1608 (tlater/progress-tests->master: Improve assertions around element loading progress reporting) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/160817:33
*** jonathanmaw has quit IRC17:36
*** traveltissues has quit IRC17:44
*** bochecha has quit IRC17:52
*** jward has quit IRC18:19
*** jward has joined #buildstream18:19
*** tpollard has quit IRC18:23
*** narispo has quit IRC22:16
*** narispo has joined #buildstream22:17
*** narispo has quit IRC22:37
*** narispo has joined #buildstream22:37
« Tuesday, 2019-09-17 Index Thursday, 2019-09-19 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!