| *** tristan has joined #buildstream | 01:57 | |
| *** narispo has quit IRC | 06:33 | |
| *** narispo has joined #buildstream | 06:33 | |
| *** narispo has quit IRC | 07:00 | |
| *** narispo has joined #buildstream | 07:02 | |
| *** slaf has quit IRC | 07:09 | |
| *** slaf has joined #buildstream | 07:11 | |
| *** slaf has joined #buildstream | 07:11 | |
| *** narispo has quit IRC | 07:11 | |
| *** slaf has joined #buildstream | 07:12 | |
| *** slaf has joined #buildstream | 07:12 | |
| *** slaf has joined #buildstream | 07:12 | |
| *** slaf has joined #buildstream | 07:12 | |
| *** slaf has joined #buildstream | 07:13 | |
| *** slaf has joined #buildstream | 07:13 | |
| *** narispo has joined #buildstream | 07:13 | |
| *** traveltissues has joined #buildstream | 08:18 | |
| *** mohan43u has quit IRC | 08:43 | |
| *** mohan43u has joined #buildstream | 08:46 | |
| *** paulsherwood has quit IRC | 09:01 | |
| *** paulsherwood has joined #buildstream | 09:02 | |
| *** bochecha has joined #buildstream | 09:04 | |
| *** phildawson_ has quit IRC | 09:27 | |
| *** phildawson_ has joined #buildstream | 09:27 | |
| benschubert | valentind: Sorry, was out end of last week. Thanks for the rational. We could also "just" use the public key for the source id :) | 09:39 |
|---|---|---|
| benschubert | juergbi: do we have extensive tests on cache cleanup with casd? (v0.0.4) On a sample of 200 builds, I roughly had 20% failing because of a cache too full | 09:39 |
| *** santi has joined #buildstream | 09:40 | |
| juergbi | benschubert: we do have tests, both on casd and buildstream side. they attempt to capture corner cases but we might still be missing some | 09:43 |
| juergbi | however, "cache too full" error is supposed to mean that the quota is too small for the session | 09:43 |
| juergbi | are you sure that is not the case? | 09:44 |
| benschubert | juergbi: I have 2To of disk space for buildbox-casd, the entire buildstream project built + sources is under 200Go | 09:44 |
| benschubert | The error is always on 'bst source fetch' also | 09:44 |
| juergbi | hm, that's very odd | 09:44 |
| juergbi | abderrahim[m] has reported #1264 where we fail cleanup too much | 09:45 |
| juergbi | but I'm not aware of an issue where we incorrectly report cache too full | 09:45 |
| benschubert | juergbi: ah, I mean the 2To are entirely full, but it's all available to casd to clean :) | 09:45 |
| juergbi | benschubert: can you check whether you have something in cas/tmp? | 09:46 |
| benschubert | yep, 1sec | 09:46 |
| juergbi | there is a buildbox-fuse bug about not cleaning up temp files. don't know whether you're using fuse | 09:46 |
| juergbi | fixed here https://gitlab.com/BuildGrid/buildbox/buildbox-fuse/-/merge_requests/27 | 09:46 |
| benschubert | mmh I don't have anything there | 09:47 |
| benschubert | but my entire cas/ dir is 66G | 09:47 |
| benschubert | Wut | 09:47 |
| juergbi | tons of stuff in the non-cas sources? | 09:48 |
| juergbi | that's (intentionally) not covered by cleanup | 09:48 |
| benschubert | checking | 09:48 |
| juergbi | btw: buildstream status bar should provide hints in such cases as it shows real time quota info from casd | 09:48 |
| benschubert | It's an automated process, I don't ge tthe status bar :) | 09:49 |
| juergbi | ah ok | 09:49 |
| benschubert | ok I guess I found the culprit: my sources | 09:49 |
| benschubert | I have over 1To of data in buildstream/sources/ | 09:49 |
| juergbi | so maybe lots of git repos with large history | 09:50 |
| benschubert | yep | 09:50 |
| benschubert | That's gonna be a very bad experience to users :/ | 09:50 |
| *** tme5 has joined #buildstream | 09:51 | |
| juergbi | remote asset API could help here | 09:51 |
| benschubert | but on the other hand I'm not sure we want to expose a cleanup mechanism to sources | 09:51 |
| benschubert | ah fair | 09:51 |
| juergbi | we could consider a configuration option where we don't keep the non-cas sources directory around at all | 09:52 |
| juergbi | i.e., delete the corresponding sources directory immediately after getting a source into CAS | 09:53 |
| valentind | Any body know about this issue with ruamel.yaml? https://gitlab.com/freedesktop-sdk/freedesktop-sdk/-/jobs/439213907 | 09:54 |
| benschubert | valentind: first time I see it :/ | 09:55 |
| benschubert | could you open an issue? | 09:55 |
| juergbi | doesn't look familiar to me either | 09:55 |
| juergbi | valentind: that's not master, is it? | 09:56 |
| valentind | Yes, but I want first to figure out why it is. If it is a specific element that we modified. Or if someone updated ruamel.yaml (I do not think we did). | 09:56 |
| valentind | It is 1.4 | 09:56 |
| juergbi | a lot changed in our yaml code since then, iirc | 09:56 |
| *** jonathanmaw has joined #buildstream | 09:57 | |
| benschubert | juergbi: it was indeed my sources :( | 09:58 |
| juergbi | benschubert: btw: are you or Chandan still planning to review !1816? otherwise I'll probably merge it soon as it's been up for quite a while and Sander has already taken a look | 09:58 |
| gitlab-br-bot | MR !1816: Reimplement support for incremental workspace builds https://gitlab.com/BuildStream/buildstream/-/merge_requests/1816 | 09:58 |
| benschubert | oh sorry, was a bit off buildstream those days. Let me have a quick look | 09:59 |
| juergbi | ta | 09:59 |
| benschubert | juergbi: any performance regression expected for the non workspace case? | 10:00 |
| *** tpollard has joined #buildstream | 10:02 | |
| juergbi | benschubert: no, there should be almost no impact on non-workspace builds | 10:02 |
| juergbi | one general change is the addition of a sources field to the artifact proto | 10:02 |
| juergbi | however, it's populated only if we're also caching buildtree | 10:02 |
| juergbi | and it should be very cheap | 10:02 |
| valentind | Found the issue. There was a comment in a (>): list | 10:04 |
| benschubert | valentind: that is probably also failing on master then, mind double checking/opening an issue? | 10:05 |
| valentind | Let me make a simple example. | 10:05 |
| gitlab-br-bot | BenjaminSchubert approved MR !1816 (juerg/incremental-workspace-build->master: Reimplement support for incremental workspace builds) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1816 | 10:06 |
| benschubert | juergbi: it's quite nice to follow, thanks! | 10:06 |
| *** phildawson-ct has joined #buildstream | 10:12 | |
| gitlab-br-bot | valentindavid opened issue #1265 (Multiline comment in a "(>)" crashes "bst track") on buildstream https://gitlab.com/BuildStream/buildstream/issues/1265 | 10:13 |
| valentind | benschubert, ^ | 10:13 |
| benschubert | valentind: thanks! | 10:18 |
| *** toscalix has joined #buildstream | 10:28 | |
| gitlab-br-bot | juergbi merged MR !1816 (juerg/incremental-workspace-build->master: Reimplement support for incremental workspace builds) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1816 | 10:37 |
| gitlab-br-bot | juergbi closed issue #985 (Workspaces via CAS) on buildstream https://gitlab.com/BuildStream/buildstream/issues/985 | 10:47 |
| benschubert | juergbi: also, any reason buildstream/build would not be empty at the end of a run? | 10:50 |
| juergbi | benschubert: except for segfaults, SIGKILL or similar, I don't think so | 10:52 |
| juergbi | we used to keep failed build roots around but that was dropped a long time ago | 10:52 |
| benschubert | ok, easy way opf recovering 30Go then x') | 10:57 |
| *** douglaswinship has joined #buildstream | 11:26 | |
| douglaswinship | Does anyone have experience using the bzr source plugin? I've never used Bazaar before. I'm trying to port a project to build with buildstream, and one of the sources is listed as a bazaar source. I can't seem to get the tracking right. | 11:29 |
| douglaswinship | the documentation I have lists a source url (http://bazaar.launchpad.net/~dbusmenu-team/libdbusmenu-qt/trunk/) and a revision (0.9.3+16.04.20160218-0ubuntu1). | 11:29 |
| douglaswinship | If I try to build with just the URL, the plugin crashes and complains that the key "track" was expected and not found. If I try to "track: 0.9.3+16.04.20160218-0ubuntu1", then it tells me that that's not a branch of the repository. | 11:30 |
| douglaswinship | "0.9.3" doesn't work either. | 11:30 |
| benschubert | douglaswinship: what does your element look like? | 11:31 |
| coldtom | douglaswinship: the branches are listed here https://code.launchpad.net/libdbusmenu-qt, i believe `trunk` is a branch (but that's master, so you probably want a release branch) | 11:32 |
| tpollard | you want 271 as the ref | 11:34 |
| tpollard | and track as trunk | 11:35 |
| WSalmon | any of the bst_external maintainers interested in https://gitlab.com/BuildStream/bst-external/-/merge_requests/117 ? once i start to move crash over to bst2 then i will try and get this functionality in to other plugins but as those plugins are currently in master of 1.4 changing them there seems like a bad idea | 11:35 |
| persia | WSalmon: Wouldn't ported versions live in https://gitlab.com/BuildStream/bst-plugins-experimental ? | 11:37 |
| WSalmon | yes that repo is for bst1.9/2 plugins so thats were i will work once crash is using bst2 but bst-experiment is were the bst1.4 plugins live atm | 11:39 |
| WSalmon | well exactly were the core plugins will end up is not 100% so were ever the remote plugin ends up is probably were this functionality will eventually live but bst-external seems like the right place for this plugin | 11:40 |
| benschubert | WSalmon: I wonder whether the trick with '{project_dir}' would do instead for this? (adding a variable that resolves to the project's directory) | 11:44 |
| coldtom | benschubert: how does one add such a variable at present ooi? or is this a feature that would need to be implemented in buildstream? | 11:45 |
| benschubert | coldtom: something we were talking about implementing | 11:45 |
| benschubert | but the ML thread stalled | 11:45 |
| coldtom | ah, aiui that ML thread would have been enough to solve WSalmon's issue | 11:46 |
| benschubert | yup :) | 11:46 |
| douglaswinship | benschubert: coldtom: got it resolved. I needed to move "trunk" off the end of the URL, and have _that_ as the value under 'track'. | 11:49 |
| douglaswinship | I also had https instead of http, which is apparently enough to break it. | 11:49 |
| douglaswinship | tpollard: 271 was close. When I run bst track, i get 271 as the ref, rather than as the tracking instruction. | 11:50 |
| tpollard | yep, that's the ref/rev of that release | 11:51 |
| WSalmon | yes, benschubert as for the best solution for master then the jury is very much still out. but for bst1.4 then a plugin seems like the way to go | 11:51 |
| douglaswinship | Of course, the element still doesn't build. But at least it's not a source issue :) On to the next bug... :D | 11:51 |
| *** xjuan has joined #buildstream | 13:16 | |
| *** phildawson has joined #buildstream | 14:00 | |
| *** phildawson-ct has quit IRC | 14:00 | |
| *** narispo has quit IRC | 14:25 | |
| *** narispo has joined #buildstream | 14:26 | |
| *** phildawson has quit IRC | 14:57 | |
| *** traveltissues has quit IRC | 15:02 | |
| *** phildawson has joined #buildstream | 15:03 | |
| gitlab-br-bot | abderrahimk opened issue #1266 (Can't load a project with an older workspace to a junction) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1266 | 15:21 |
| *** lachlan has joined #buildstream | 15:56 | |
| gitlab-br-bot | abderrahimk opened MR !1820 (abderrahim/simple-host-files->master: _project.py: fix the case where a simple string is passed to host-files) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1820 | 16:05 |
| *** lachlan has quit IRC | 16:05 | |
| gitlab-br-bot | BenjaminSchubert approved MR !1820 (abderrahim/simple-host-files->master: _project.py: fix the case where a simple string is passed to host-files) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1820 | 16:07 |
| *** lachlan has joined #buildstream | 16:47 | |
| *** lachlan has quit IRC | 16:57 | |
| *** lachlan has joined #buildstream | 16:59 | |
| *** lachlan has quit IRC | 17:03 | |
| *** lachlan has joined #buildstream | 17:08 | |
| *** lachlan has quit IRC | 17:13 | |
| *** lachlan has joined #buildstream | 17:15 | |
| *** lachlan has quit IRC | 17:32 | |
| *** lachlan has joined #buildstream | 17:38 | |
| *** phildawson has quit IRC | 17:42 | |
| *** lachlan has quit IRC | 17:43 | |
| *** tme5 has quit IRC | 17:58 | |
| *** lachlan has joined #buildstream | 18:02 | |
| *** rafaelff1[m] has quit IRC | 18:04 | |
| *** reuben640[m] has quit IRC | 18:05 | |
| *** m_22[m] has quit IRC | 18:05 | |
| *** segfault1[m] has quit IRC | 18:05 | |
| *** abderrahim[m]1 has quit IRC | 18:05 | |
| *** tchaik[m] has quit IRC | 18:05 | |
| *** Abderrahim[m] has quit IRC | 18:05 | |
| *** theawless[m] has quit IRC | 18:05 | |
| *** asingh_[m] has quit IRC | 18:05 | |
| *** Trevio[m] has quit IRC | 18:05 | |
| *** SamThursfield[m] has quit IRC | 18:05 | |
| *** awacheux[m] has quit IRC | 18:05 | |
| *** doras[m] has quit IRC | 18:05 | |
| *** cgm[m] has quit IRC | 18:05 | |
| *** connorshea[m] has quit IRC | 18:05 | |
| *** DineshBhattarai[m] has quit IRC | 18:05 | |
| *** krichter[m] has quit IRC | 18:05 | |
| *** mattiasb has quit IRC | 18:05 | |
| *** dbuch[m] has quit IRC | 18:05 | |
| *** pro[m] has quit IRC | 18:05 | |
| *** Demos[m] has quit IRC | 18:05 | |
| *** albfan[m] has quit IRC | 18:05 | |
| *** tlater[m] has quit IRC | 18:05 | |
| *** kailueke[m] has quit IRC | 18:05 | |
| *** jjardon[m] has quit IRC | 18:06 | |
| *** walterve[m][m] has quit IRC | 18:06 | |
| *** skullone[m] has quit IRC | 18:06 | |
| *** dylan-m has quit IRC | 18:06 | |
| *** lachlan has quit IRC | 18:08 | |
| *** brlogger has joined #buildstream | 18:12 | |
| *** rafaelff1[m] has joined #buildstream | 18:20 | |
| *** segfault1[m] has joined #buildstream | 18:21 | |
| *** lachlan has joined #buildstream | 18:22 | |
| *** segfault1[m] has quit IRC | 18:27 | |
| *** rafaelff1[m] has quit IRC | 18:27 | |
| *** abderrahim[m] has joined #buildstream | 18:28 | |
| *** jonathanmaw has quit IRC | 18:29 | |
| *** lachlan has quit IRC | 18:39 | |
| *** Abderrahim[m]1 has joined #buildstream | 18:57 | |
| *** asingh_[m] has joined #buildstream | 18:57 | |
| *** kailueke[m] has joined #buildstream | 18:57 | |
| *** tlater[m] has joined #buildstream | 18:57 | |
| *** albfan[m] has joined #buildstream | 18:57 | |
| *** rafaelff1[m] has joined #buildstream | 18:57 | |
| *** reuben640[m] has joined #buildstream | 18:57 | |
| *** doras[m] has joined #buildstream | 18:57 | |
| *** jjardon[m] has joined #buildstream | 18:57 | |
| *** SamThursfield[m] has joined #buildstream | 18:57 | |
| *** dbuch[m] has joined #buildstream | 18:57 | |
| *** tchaik[m] has joined #buildstream | 18:57 | |
| *** Demos[m] has joined #buildstream | 18:57 | |
| *** m_22[m] has joined #buildstream | 18:57 | |
| *** awacheux[m] has joined #buildstream | 18:57 | |
| *** cgm[m] has joined #buildstream | 18:57 | |
| *** mattiasb has joined #buildstream | 18:57 | |
| *** DineshBhattarai[m] has joined #buildstream | 18:57 | |
| *** theawless[m] has joined #buildstream | 18:57 | |
| *** segfault1[m] has joined #buildstream | 18:57 | |
| *** Trevio[m] has joined #buildstream | 18:57 | |
| *** walterve[m][m] has joined #buildstream | 18:57 | |
| *** dylan-m has joined #buildstream | 18:57 | |
| *** connorshea[m] has joined #buildstream | 18:57 | |
| *** skullone[m] has joined #buildstream | 18:57 | |
| *** pro[m] has joined #buildstream | 18:57 | |
| *** krichter[m] has joined #buildstream | 18:57 | |
| *** lachlan has joined #buildstream | 18:58 | |
| *** lachlan has quit IRC | 19:02 | |
| *** toscalix has quit IRC | 19:03 | |
| *** lachlan has joined #buildstream | 19:20 | |
| *** lachlan has quit IRC | 19:21 | |
| *** narispo has quit IRC | 20:13 | |
| *** narispo has joined #buildstream | 20:27 | |
| *** narispo has quit IRC | 20:33 | |
| *** narispo has joined #buildstream | 20:34 | |
| *** narispo has quit IRC | 20:44 | |
| *** narispo has joined #buildstream | 20:44 | |
| *** xjuan has quit IRC | 21:43 | |
| *** xjuan has joined #buildstream | 21:44 | |
| *** abderrahim[m] has quit IRC | 21:47 | |
| *** abderrahim[m] has joined #buildstream | 21:47 | |
| *** xjuan has quit IRC | 22:04 | |
| *** xjuan has joined #buildstream | 23:18 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!