IRC logs for #buildstream for Monday, 2019-08-05

« Sunday, 2019-08-04 Index Tuesday, 2019-08-06 »
*** tristan__ has joined #buildstream03:53
gitlab-br-bottristanvb merged MR !1517 (tristan/bst-1/misc-cleanup->bst-1: Backport some misc fixes (bst-1)) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/151705:02
*** tristan__ has quit IRC05:08
*** tristan_ has joined #buildstream05:15
*** tristan_ has quit IRC05:16
*** tristan_ has joined #buildstream05:50
*** tristan_ has quit IRC05:53
*** narispo has quit IRC06:27
*** narispo has joined #buildstream06:28
*** kapip has joined #buildstream06:51
*** narispo has quit IRC07:33
*** narispo has joined #buildstream07:33
*** bochecha has joined #buildstream08:49
*** bochecha has quit IRC08:54
*** bochecha has joined #buildstream08:54
*** bochecha has quit IRC08:59
*** bochecha has joined #buildstream09:00
*** bochecha has joined #buildstream09:00
*** tpollard has joined #buildstream09:04
*** jonathanmaw has joined #buildstream09:09
*** bochecha has quit IRC09:10
*** bochecha has joined #buildstream09:10
*** cs-shadow has joined #buildstream09:16
*** bochecha has quit IRC09:20
*** bochecha has joined #buildstream09:21
*** lachlan has joined #buildstream09:30
*** lachlan has quit IRC09:46
*** lachlan has joined #buildstream09:46
laurencetristan, persia, valentind - I am actually not going to be available at all for the meeting tomorrow, I have a clash I can't get out of.10:29
laurenceSorry about that.10:29
*** kapip has quit IRC10:29
*** jjardon changes topic to "BuildStream 1.2.8 is out ! | https://gitlab.com/BuildStream/buildstream | Docs: https://docs.buildstream.build/ | IRC logs: https://irclogs.baserock.org/buildstream | Mailing List: https://mail.gnome.org/mailman/listinfo/buildstream-list | Roadmap: https://wiki.gnome.org/Projects/BuildStream/Roadmaps"10:38
*** lachlan has quit IRC10:41
*** lachlan has joined #buildstream10:45
*** toscalix has joined #buildstream10:53
*** toscalix has joined #buildstream10:54
*** lachlan has quit IRC11:00
*** mohan43u has quit IRC11:13
*** mohan43u has joined #buildstream11:17
*** lachlan has joined #buildstream11:19
*** lachlan has quit IRC11:29
*** lachlan has joined #buildstream11:33
gitlab-br-botcoldtom opened MR !1520 (coldtom/move-cmake->master: Move cmake plugin to bst-plugins-experimental) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/152011:40
*** kapip has joined #buildstream11:47
gitlab-br-botjjardon opened issue #1097 (bst 1.x overnigth tests are failing: SyntaxError: invalid syntax) on buildstream https://gitlab.com/BuildStream/buildstream/issues/109712:04
tlater[m]So I just ran into the question of what it means to "expire" artifacts when they are split between two caches12:05
tlater[m]Hooray for distributed systems issues12:05
tlater[m]In essence; we'll need to manage CAS blobs in one server and artifact metadata pointing to those blobs in another12:06
tlater[m]We won't have infinite storage, so we'll continue expiring artifacts whenever we run out of space in LRU fashion12:07
tlater[m]If we make each server independently expire its data, the two servers become de-synced, and things like the upcoming `bst show` will give us artifacts that don't exist anymore/ignore artifacts that still exist12:07
tlater[m]How do we solve this?12:07
tlater[m]And how do we communicate with CASes that are not bst-artifact-server?12:07
persiaIf there is a way to check existence (without download), `bst show` could verify the blobs existed before reporting on them.12:08
persiaBut if CAS has separate expiry, partial-artifact effects are likely to appear, where only some blobs from a given artifact are unexpired in CAS.12:09
tlater[m]persia: And I suppose blobs that still exist but belong to an "expired" artifact will simply be cleaned up after a while...12:10
tlater[m]Hrm, let me see if there is such a thing in the protocol12:10
tlater[m]persia: Perhaps, but I expect that in those cases it would make sense to treat the full artifact as expired12:11
tlater[m]Since any artifact that is "used" will use all of its blobs12:11
tlater[m]Until we change that of course12:11
persiaI thought I remembered discussion about partial artifacts.  Am I confused?12:12
*** phoenix has joined #buildstream12:12
persiaMy understanding was that it might be interesting to expire logs or certain rarely-used blobs, but not the core dependency blobs.12:12
tpollardthere was the hope to support partial artifacts in remotes yes12:14
tlater[m]persia: Hrm, that is interesting. But I suspect it'll take enough work on the client end that removing a check for "do we have all artifacts" will be the least of our worries.12:15
tlater[m]s/artifacts/blobs/ I should say12:15
* tlater[m] still needs to make up better words12:16
* persia hopes someone who actually knows something about "partial artifacts" can contribute something useful, as otherwise this discussion is unlikely to be particularly helpful12:23
tpollardI did about buildtrees, setting different expiry rates for a given set of blobs not so much12:26
tlater[m]persia: I do have vague recollections of this, and I roughly know bst's artifact format12:27
tlater[m]Given that I just don't think it's much of a concern whether we currently treat these artifacts as fully expired or not12:27
tlater[m]Since it'll take a fair bit of work to get them to split nicely anyway, and this will be a relatively small part of that12:28
tlater[m]After all, unless I start thinking about splitting artifacts for this patch, it's not going to be useful to think about treating them differently yet.12:28
tlater[m]... and I do think blob-wise expiry rates are a little out of scope :)12:28
persiatlater[m]: Right.  My thought was that it would be sensible to coordinate: either documenting the work-to-be-done in the ticket (that I currently cannot find) about partial artifacts, or using any thought already put into that as guidance as the best immediate compromise for the split artifact server stuff.12:29
persiaFor current and immediate stuff, it doesn't matter, but looking at the longer-term model, it makes sense to do things in a way that don't need to be redone *too* much :)12:29
tlater[m]iirc we hadn't put any work into it yet, besides the basic idea. But I'll concede that I wasn't as active earlier this year.12:30
persiaYou may be entirely right :)  I'm mostly just remembering things and asking questions, rather than arguing any particular course from firm conviction or certain knowledge.12:30
*** bochecha_ has joined #buildstream13:17
*** bochecha has quit IRC13:18
*** bochecha_ is now known as bochecha13:18
*** tristan has quit IRC13:32
*** lachlan has quit IRC14:14
*** lachlan has joined #buildstream14:35
gitlab-br-botjjardon opened MR !1521 (jjardon/test_bst_external->master: Fix bst-external to be python 3.5 compatible) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/152114:37
*** bochecha_ has joined #buildstream14:46
*** bochecha has quit IRC14:48
*** bochecha_ is now known as bochecha14:48
*** lachlan has quit IRC15:12
ironfootI've just tried to set "cachedir" in my ~/.config/buildstream.conf, but I get an error:15:12
ironfootError loading user configuration: ~/.config/buildstream.conf [line 5 column 10]: Unexpected key: cachedir15:12
ironfootwas just following: https://buildstream.gitlab.io/buildstream/using_config.html#default-configuration15:12
ironfootbst --version: 1.2.715:13
coldtomhttps://docs.buildstream.build/using_config.html#default-configuration is the docs for 1.2.x15:14
ironfootta! was looking for that15:14
*** tristan has joined #buildstream15:20
*** lachlan has joined #buildstream15:25
gitlab-br-botjjardon opened (was WIP) MR !1521 (jjardon/test_bst_external->bst-1.2: Use a python 3.5 compatible bst-external version) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/152115:26
gitlab-br-botjjardon merged MR !1521 (jjardon/test_bst_external->bst-1.2: Use a python 3.5 compatible bst-external version) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/152115:27
tlater[m]Has the suggestions feature disappeared from gitlab? :(15:27
*** tristan_ has joined #buildstream15:29
jjardonjonathanmaw: hey , are you still the bst-external maintainer? Would it be possible to have a new release?15:47
* jonathanmaw is not. jennis is maintainer, though15:48
jonathanmawhmm, looks like I forgot to take myself off the MAINTAINERS file15:48
jonathanmawjennis: would you mind removing that while you're doing the new release?15:49
jennisSure15:49
*** lachlan has quit IRC15:53
jennisLooks like jjardon has already done this?15:53
jonathanmawah, perhaps I haven't pulled from the repo in a whiel15:54
jennislooks like he's done it just now ^15:55
jennisjjardon?15:55
jjardonsorry, my muscle memory failed me and I've just pushed; jennis do you mind if I tag 0.16.0 and we are done?15:56
jennisHeh, no worries15:56
alexandrufazakasjennis I can still see jonathanmaw in the maintainers file right now https://gitlab.com/BuildStream/bst-external/blob/master/MAINTAINERS15:56
jennisYeah alexandrufazakas I'll still need to make that change15:56
alexandrufazakasAnd jjardon didn't touch that file in his last commit afaict?15:56
jennisI was talking about the release (:15:57
alexandrufazakasOops, my bad jennis :)15:57
*** kapip has quit IRC16:07
*** lachlan has joined #buildstream16:11
*** raoul has quit IRC16:25
*** toscalix has quit IRC16:34
*** lachlan has quit IRC16:40
*** lachlan has joined #buildstream16:53
gitlab-br-botjjardon opened issue #1098 (bst 1.x overnigth tests are failing: Message handling out of sync) on buildstream https://gitlab.com/BuildStream/buildstream/issues/109817:03
*** lachlan has quit IRC17:05
*** tristan has quit IRC17:07
*** tristan has joined #buildstream17:09
gitlab-br-botjjardon opened issue #1099 (code_quality job is broken) on buildstream https://gitlab.com/BuildStream/buildstream/issues/109917:10
*** tristan has quit IRC17:11
*** tristan has joined #buildstream17:12
*** tristan has quit IRC17:13
*** jonathanmaw has quit IRC17:17
gitlab-br-botjjardon opened MR !1522 (jjardon/code_quality->master: .gitlab-ci.yml: Use new way to define code_quality job) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/152217:17
*** lachlan has joined #buildstream17:19
*** lachlan has quit IRC17:23
*** tristan has joined #buildstream17:24
*** tristan has quit IRC17:27
*** tristan has joined #buildstream17:28
*** tristan has quit IRC17:29
*** tristan has joined #buildstream17:29
*** tristan has quit IRC17:30
*** tristan has joined #buildstream17:31
*** tristan has quit IRC17:32
*** tristan has joined #buildstream17:34
*** tristan has quit IRC17:35
*** lachlan has joined #buildstream17:37
*** jonathanmaw has joined #buildstream17:39
*** jonathanmaw has quit IRC17:44
*** lachlan has quit IRC17:47
gitlab-br-botmarge-bot123 closed issue #1099 (code_quality job is broken) on buildstream https://gitlab.com/BuildStream/buildstream/issues/109917:51
gitlab-br-botmarge-bot123 merged MR !1522 (jjardon/code_quality->master: .gitlab-ci.yml: Use new way to define code_quality job) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/152217:51
*** lachlan has joined #buildstream17:52
*** lachlan has quit IRC17:59
*** lachlan has joined #buildstream18:11
*** traveltissues has joined #buildstream18:46
*** traveltissues has quit IRC18:46
*** lachlan has quit IRC18:50
*** lachlan has joined #buildstream19:10
*** lachlan has quit IRC19:32
*** lachlan has joined #buildstream19:35
*** lachlan has quit IRC19:53
*** lachlan has joined #buildstream19:53
*** lachlan has quit IRC20:08
*** phoenix has quit IRC20:17
*** phoenix has joined #buildstream20:25
*** phoenix has quit IRC20:40
*** tristan has joined #buildstream22:25
*** tristan has quit IRC23:08
*** tristan has joined #buildstream23:08
*** bochecha has quit IRC23:31
*** bochecha has joined #buildstream23:31
*** tristan has quit IRC23:36
*** tristan has joined #buildstream23:38
*** bochecha has quit IRC23:41
*** bochecha has joined #buildstream23:42
*** bochecha has quit IRC23:57
*** bochecha has joined #buildstream23:57
*** bochecha has quit IRC23:57
*** bochecha has joined #buildstream23:58
« Sunday, 2019-08-04 Index Tuesday, 2019-08-06 »

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