IRC logs for #buildstream for Wednesday, 2019-12-04

« Tuesday, 2019-12-03 Index Thursday, 2019-12-05 »
*** narispo has quit IRC00:30
*** narispo has joined #buildstream00:30
*** mohan43u has joined #buildstream05:23
*** traveltissues has joined #buildstream07:37
*** tpollard has joined #buildstream09:03
*** rdale has joined #buildstream09:15
*** santi has joined #buildstream09:47
*** tme5 has joined #buildstream09:52
*** phildawson has joined #buildstream10:09
*** jonathanmaw has joined #buildstream10:25
*** lachlan has joined #buildstream10:28
*** lachlan has quit IRC10:48
*** santi has quit IRC10:48
*** lachlan has joined #buildstream10:56
*** lachlan has quit IRC11:03
*** santi has joined #buildstream11:14
*** lachlan has joined #buildstream11:24
*** lachlan has quit IRC11:28
*** cs-shadow has joined #buildstream11:43
*** lachlan has joined #buildstream11:57
juergbitlater[m]: in case you're interested in the userchroot docker image structure: https://gitlab.com/BuildStream/buildstream-docker-images/merge_requests/148#note_25417786811:58
tme5is it right for a source plugin to start a timed activity while checking if it's cache is valid?12:05
juergbitme5: in get_consistency()?12:06
tme5the git source needs to clone the repo to check whether it is fully cached, due to submodules. At the moment this is done silently, so later if it went to fetch nothing would happen12:07
tme5juergbi, yes12:07
juergbi!1739 is tangentially related to this12:09
gitlab-br-botMR !1739: WIP: Optimize consistency and state handling https://gitlab.com/BuildStream/buildstream/merge_requests/173912:09
tme5was wondering if i could change that to do a non-silent fetch. But then the user might panic that something is happening when they do a bst show12:09
juergbiwith !1739 that check would be done in the new is_cached() method12:09
juergbiand that should be less frequently called than get_consistency()12:10
juergbiI think it might still sometimes be invoked in bst show, though12:10
tme5but it will still be called in bst invocation that aren't mean to do anything12:11
juergbido you already have an example output where this shows up in bst show?12:12
tme5perhaps what i'll do is try to make it so get_consistency of the git plugin doesn't actually need to download anything12:12
juergbiit should never download anything12:13
juergbiget_consistency() should always be local12:13
juergbibut it could still take a while12:13
tme5it isn't local for the git plugin12:13
juergbihm, that's odd12:14
juergbimaybe a bug? or I'm forgotten something12:14
tme5get_consistency calls _have_all_refs calls _refresh_submodules calls ensure12:14
tme5_GitMirror.ensure, which does a git clone12:15
juergbitme5: _have_all_refs() only calls _refresh_submodules() if has_ref() returns True12:17
tme5ah yes was just about to say12:17
juergbiand it only returns True if it's already local, or not?12:17
tme5yes you're correct12:17
juergbiand if self.mirror already exists, ensure() won't do anything12:17
tme5ok12:19
tme5thanks for having a look, that gives me confidence to use the fetch method on the mirror instead of ensure12:19
tme5(in this case of what i'm implementing :P)12:20
*** toscalix has joined #buildstream12:29
*** lachlan has quit IRC12:30
*** santi has quit IRC12:46
*** santi has joined #buildstream12:46
*** toscalix has quit IRC12:51
*** lachlan has joined #buildstream12:58
*** toscalix has joined #buildstream13:00
*** lachlan has quit IRC13:02
*** toscalix has quit IRC13:38
*** toscalix has joined #buildstream13:39
*** santi has quit IRC13:48
*** santi has joined #buildstream13:57
tme5does anyone know off the top of their heads why the user is warned when not specifying a submodule in git source config?14:48
coldtomprobably because it means you're either fetching something that's not explicitly specified, or the build may fail from a missing submodule14:49
tme5that's what I mean, fetching something not explicitly specified14:59
coldtomhttps://gitlab.com/BuildStream/buildstream/issues/783 might give some context15:14
tme5I see, that has, tyvm15:33
gitlab-br-botBenjaminSchubert opened MR !1746 (bschubert/ci-improvements->master: Multiple CI Improvements and fixes) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/174615:42
benschubertSeems like the wsl runners are outdated? WOuld someone mind updating them?15:47
*** phildawson_ has joined #buildstream16:03
*** phildawson has quit IRC16:04
*** lachlan has joined #buildstream16:08
gitlab-br-botBenjaminSchubert opened MR !1747 (bschubert/ci-improvements->master: Multiple CI improvements) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/174716:08
benschubertpfiou finally got it right ^ some CI improvements :)16:10
*** lachlan has quit IRC16:14
tlater[m]CI config is a PITA to test :D16:21
tlater[m]benschubert: None of the jobs seem to be running on your branch16:22
tlater[m]Only tests-wsl, but both of them need clicking16:23
tlater[m]Huh, but they do run on the commit16:24
juergbiit creates two pipelines. one is 'detached' for wsl/win3216:26
tlater[m]That one seems to be the one associated with the MR though16:27
juergbiyes16:27
juergbihttps://gitlab.com/gitlab-org/gitlab/issues/3475616:28
juergbi"Well, with such bug, rules: are barely usable if at all. We've dropped rules: from our pipelines altogether due to this behavior."16:28
tlater[m]Ergh, I figure that's what benschubert means by "awkward"16:30
benschubertOk, I guess we'd rather drop that commit for now then?16:31
tlater[m]Yeah, though it's a shame16:32
* tlater[m] figures marge would fail to marge that kind of thing anyway16:33
juergbihm, now it looks correct16:37
juergbidid it just take a while?16:37
juergbiwell, there is still the detached pipeline, no idea what that's for16:40
*** santi has quit IRC16:40
juergbimaybe we can use a wsl/win32 job templates instead to at least fix the duplication?16:40
* tpollard wonders if anyone benefits for the UI showing status messages by default 16:53
tpollards/for/from16:53
*** santi has joined #buildstream16:56
tpollardparticularly the verbose running command multiline ones16:59
coldtomi quite like the verbose running command multiline ones, it makes it easier to see at a glance where an element is up to17:01
tlater[m]Yeah, I agree with coldtom there17:01
tlater[m]If it didn't show those, I'd probably feel like it's too magic17:02
tlater[m]And doubt it's actually doing anything17:02
tpollardinteresting17:03
coldtomit breaks things up a little, adds some variety to the log17:05
*** lachlan has joined #buildstream17:05
tpollardI think the running configure/build/install etc is useful for the feeling like it's doing something, I'm not sure getting a long list of configure flags is needed though17:05
tlater[m]Probably not :)17:06
tpollardthis is specifically the terminal output btw, verbosity is not touched for the actual logs17:06
tpollardthanks for the input!17:09
benschubertjuergbi: tlater[m] ok, I guess let's drop that commit :) I'll revisit when there is a new upadte to this17:11
tlater[m]benschubert: You can keep a FIXME linking to the issue17:11
tlater[m]Or maybe just open an issue :D17:11
benschubertI'll go for the issue17:11
*** lachlan has quit IRC17:17
benschubertok, PR updated17:19
gitlab-br-botBenjaminSchubert opened issue #1227 (Consolidate CI jobs) on buildstream https://gitlab.com/BuildStream/buildstream/issues/122717:24
gitlab-br-bottlater approved MR !1747 (bschubert/ci-improvements->master: Multiple CI improvements) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/174717:25
*** rdale has quit IRC17:39
*** tme5 has quit IRC17:42
*** lachlan has joined #buildstream17:42
*** tpollard has quit IRC17:45
*** lachlan has quit IRC17:46
*** santi has quit IRC18:02
*** toscalix has quit IRC18:17
*** toscalix has joined #buildstream18:17
*** toscalix has quit IRC18:18
gitlab-br-botjuergbi merged MR !1747 (bschubert/ci-improvements->master: Multiple CI improvements) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/174718:21
*** toscalix has joined #buildstream18:24
*** traveltissues has quit IRC18:27
*** phildawson_ has quit IRC19:04
*** narispo has quit IRC19:08
*** narispo has joined #buildstream19:08
*** lachlan has joined #buildstream19:13
*** lachlan has quit IRC19:19
*** jonathanmaw has quit IRC19:25
*** phoenix has joined #buildstream19:42
*** phildawson_ has joined #buildstream19:51
*** phildawson_ has quit IRC19:54
*** phoenix has quit IRC20:04
*** toscalix has quit IRC20:05
*** toscalix has joined #buildstream20:05
*** lachlan has joined #buildstream20:53
*** lachlan has quit IRC21:00
*** toscalix has quit IRC21:14
*** cs-shadow has quit IRC21:21
*** narispo has quit IRC21:22
*** narispo has joined #buildstream21:22
*** narispo has quit IRC21:31
*** narispo has joined #buildstream21:32
*** narispo has quit IRC23:16
*** narispo has joined #buildstream23:16
*** narispo has joined #buildstream23:17
*** narispo has quit IRC23:56
*** narispo has joined #buildstream23:59
« Tuesday, 2019-12-03 Index Thursday, 2019-12-05 »

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