| *** tristan has joined #buildstream | 00:35 | |
| *** nimish has joined #buildstream | 02:28 | |
| *** tristan has quit IRC | 03:52 | |
| *** alatiera has joined #buildstream | 05:41 | |
| *** alatiera has joined #buildstream | 05:44 | |
| *** ahmed89 has quit IRC | 06:26 | |
| *** ahmed89 has joined #buildstream | 06:34 | |
| *** finn has quit IRC | 06:58 | |
| *** nimish has quit IRC | 07:16 | |
| *** tpollard has quit IRC | 07:16 | |
| *** nimish has joined #buildstream | 07:45 | |
| *** ahmed89 has quit IRC | 07:50 | |
| *** alatiera_ has joined #buildstream | 08:15 | |
| *** alatiera has quit IRC | 08:17 | |
| *** finn has joined #buildstream | 08:19 | |
| *** tpollard has joined #buildstream | 08:19 | |
| *** nimish has quit IRC | 08:20 | |
| *** nimish has joined #buildstream | 08:20 | |
| *** alatiera has joined #buildstream | 08:21 | |
| *** alatiera has joined #buildstream | 08:24 | |
| *** nimish has quit IRC | 08:25 | |
| *** nimish has joined #buildstream | 08:25 | |
| *** nimish has quit IRC | 08:31 | |
| Kinnison | So long as the topological ordering is consistent, the exact ordering between unorderable elements should not matter and we should be able to cope with that in the test suite really :/ | 08:47 |
|---|---|---|
| Kinnison | To that end I'd be tempted to explicitly introduce randomness into the ordering of elements to be built during testing, simply to ensure no test expects a specific linear ordering unless it has specified it in the dependency graph | 08:48 |
| *** nimish has joined #buildstream | 08:53 | |
| *** toscalix has joined #buildstream | 08:54 | |
| *** nimish has quit IRC | 09:18 | |
| *** nimish has joined #buildstream | 09:18 | |
| *** raoul_ has joined #buildstream | 09:24 | |
| jennis | I've added a test for our artifact completions, which, if I have the artifact that this test actually tests (tests/completions/project/elements/import-bin.bst) in *my* local cache, my test passes, if not, it fails (as I noticed it failed in CI). | 09:24 |
| jennis | Seems, odd and wondered whether anyone has any ideas / could take a look? https://gitlab.com/BuildStream/buildstream/merge_requests/1054 | 09:24 |
| gitlab-br-bot | jennis opened MR !1054 (jennis/add_artifacts_completion->master: completions.py: Add a test for our artifact ref autocompletions) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1054 | 09:25 |
| jennis | The test is different from the other completion tests in the sense that we must build the artifact before we try to auto-complete its ref | 09:25 |
| *** phildawson_ has joined #buildstream | 09:26 | |
| gitlab-br-bot | jennis opened (was WIP) MR !1049 (jennis/deprecate_bst_checkout->master: Deprecate `bst checkout` for `bst artifact checkout`) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1049 | 09:55 |
| *** nimish has quit IRC | 09:56 | |
| *** jonathanmaw has joined #buildstream | 09:58 | |
| *** nimish has joined #buildstream | 10:15 | |
| *** lachlan has joined #buildstream | 10:25 | |
| *** nimish has quit IRC | 10:50 | |
| *** nimish has joined #buildstream | 10:51 | |
| *** raoul_ has quit IRC | 11:04 | |
| *** lachlan has quit IRC | 11:11 | |
| *** lachlan has joined #buildstream | 11:13 | |
| *** raoul_ has joined #buildstream | 11:18 | |
| *** phildawson_ has quit IRC | 11:33 | |
| *** phildawson_ has joined #buildstream | 11:33 | |
| *** lachlan has quit IRC | 11:43 | |
| *** lachlan has joined #buildstream | 11:51 | |
| *** phildawson_ has quit IRC | 11:52 | |
| *** phildawson_ has joined #buildstream | 11:52 | |
| gitlab-br-bot | tpollard opened (was WIP) MR !1050 (tpollard/829->master: Download buildtrees on demand for bst shell --use-buildtree) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1050 | 12:01 |
| *** nimish has quit IRC | 12:26 | |
| *** nimish has joined #buildstream | 12:26 | |
| *** lachlan has quit IRC | 12:31 | |
| *** raoul_ has quit IRC | 12:33 | |
| *** lachlan has joined #buildstream | 12:38 | |
| *** nimish has quit IRC | 12:41 | |
| *** nimish has joined #buildstream | 12:41 | |
| *** nimish has quit IRC | 12:46 | |
| *** nimish has joined #buildstream | 12:47 | |
| *** nimish has quit IRC | 12:52 | |
| *** nimish has joined #buildstream | 12:52 | |
| *** alatiera has quit IRC | 13:06 | |
| *** alatiera has joined #buildstream | 13:06 | |
| *** raoul_ has joined #buildstream | 13:34 | |
| gitlab-br-bot | juergbi merged MR !1034 (juerg/arch->master: Accept architecture aliases) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1034 | 13:34 |
| *** finn_ has joined #buildstream | 13:45 | |
| *** finn has quit IRC | 13:46 | |
| laurence | lachlan, should the above fox the current baserock test in master? ^^ | 13:53 |
| laurence | s/master/benchmarks/ | 13:53 |
| laurence | s/fox/fix/ | 13:54 |
| lachlan | laurence: Looks like it | 14:02 |
| laurence | cool | 14:03 |
| * lachlan prods a new build to check | 14:13 | |
| gitlab-br-bot | aevri opened (was WIP) MR !1055 (aevri/contrib-typos->master: contributing: fix 'oprtation' and some other typos) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1055 | 14:13 |
| *** nimish has quit IRC | 14:22 | |
| *** nimish has joined #buildstream | 14:22 | |
| *** ctolentino has joined #buildstream | 14:28 | |
| *** alatiera_ has joined #buildstream | 14:50 | |
| *** alatiera has quit IRC | 14:50 | |
| *** alatiera_ is now known as alatiera | 14:51 | |
| gitlab-br-bot | aevri merged MR !1055 (aevri/contrib-typos->master: contributing: fix 'oprtation' and some other typos) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1055 | 15:07 |
| *** nimish has quit IRC | 15:07 | |
| *** nimish has joined #buildstream | 15:08 | |
| *** cs-shadow has quit IRC | 15:12 | |
| *** alatiera has quit IRC | 15:13 | |
| *** alatiera has joined #buildstream | 15:14 | |
| *** lachlan has quit IRC | 15:45 | |
| *** lachlan has joined #buildstream | 15:47 | |
| *** alatiera has quit IRC | 15:49 | |
| *** alatiera has joined #buildstream | 15:50 | |
| *** tristan has joined #buildstream | 16:01 | |
| *** ChanServ sets mode: +o tristan | 16:03 | |
| *** cs-shadow has joined #buildstream | 16:14 | |
| tristan | juergbi, I'm curious about subproject element sorting in Loader._sort_dependencies() | 16:25 |
| tristan | juergbi, I know it's a while ago, but is this just to ensure constant sort order when elements in a local and subproject have the same name ? | 16:26 |
| *** alatiera has quit IRC | 16:27 | |
| tristan | I might just update the comment to reflect that (it appears to me like it is unintentionally beyond the element name check) | 16:27 |
| *** nimish has quit IRC | 16:28 | |
| *** nimish has joined #buildstream | 16:28 | |
| *** alatiera has joined #buildstream | 16:30 | |
| *** sambishop has joined #buildstream | 16:40 | |
| tristan | oh awesome, the arch compat branch dropped ! | 16:40 |
| *** nimish has quit IRC | 16:48 | |
| *** nimish has joined #buildstream | 16:48 | |
| gitlab-br-bot | jmacarthur opened MR !1056 (jmac/make_marked_directories->master: _sandboxremote.py: Create marked directories before running) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1056 | 16:58 |
| juergbi | tristan: yes, that's just in the unlikely case that everything else is equal, to have defined order | 17:00 |
| juergbi | or is this maybe impossible to reach? I don't remember otoh | 17:00 |
| laurence | if anyone sees a tlater around, pls let him know i have a few quick questions for him (have mailed him anyway) | 17:04 |
| *** nimish has quit IRC | 17:08 | |
| *** nimish has joined #buildstream | 17:09 | |
| tristan | juergbi, now that you mention it | 17:10 |
| tristan | juergbi, it appears that you could have removed that part with 399f09b74 ? | 17:10 |
| juergbi | that's the commit that added junctions | 17:11 |
| tristan | oh wait, it's the same commit | 17:11 |
| * tristan lost in the annotate log | 17:11 | |
| tristan | Ah no, but strange... | 17:12 |
| tristan | anyway, you are saying the only reason it's there is for uniqueness | 17:12 |
| tristan | And I suppose the constant order of deps affects the artifact cache key so changing that has a cost | 17:13 |
| tristan | it's reachable btw, just confusing that above the 'full_name' is used, and below this other thing is used | 17:13 |
| tristan | juergbi, since you're around; do you have any idea what can cause the build order to be random ? | 17:15 |
| tristan | just in case you have a hunch :) | 17:16 |
| tristan | or, if anyone is around and has a hunch of course :) | 17:16 |
| juergbi | tristan: you mean if you run the scheduler with max parallel fetchers/builders etc. set to 1? | 17:16 |
| juergbi | for any other configuration, there is no guarantee of order anyway | 17:17 |
| juergbi | or what am I missing? | 17:17 |
| tristan | yeah that's exactly what I'm doing | 17:17 |
| tristan | well there are some guarantees, but there is currently no exact guarantee no | 17:17 |
| *** lachlan has quit IRC | 17:18 | |
| tristan | I'm hoping to change that so that we can use the strict guarantee to conveniently assert other expectations | 17:18 |
| tristan | juergbi, context is that right now there is a problem, elements are not getting promoted to the build queue | 17:18 |
| tristan | fast enough | 17:18 |
| juergbi | hm, is one of the data structures the scheduler uses implemented with a random component? some hash tables might do that | 17:19 |
| tristan | we go through like all the elements in the pull queue and a lot of fetches before ever hitting a build (with freedesktop-sdk as a sample) | 17:19 |
| juergbi | ok, but that's a more fundamental logic bug somewhere | 17:19 |
| tristan | comments in https://gitlab.com/BuildStream/buildstream/issues/824 | 17:19 |
| juergbi | i.e., we don't need guaranteed build order for this | 17:19 |
| juergbi | but it might help track the issue down, I suppose | 17:20 |
| tristan | juergbi, Indeed I agree | 17:20 |
| juergbi | regression from the cache expiry related changes? | 17:20 |
| juergbi | I thought it was working properly originally | 17:20 |
| tristan | I thought it was at first | 17:20 |
| tristan | but now I'm tending to think it's about build of build dependencies and the dynamic enabling of them | 17:21 |
| tristan | how that effects the loops, etc, may be a cause of this | 17:21 |
| juergbi | ah, right, that was another possibly related change | 17:21 |
| *** alatiera_ has joined #buildstream | 17:23 | |
| *** alatiera has quit IRC | 17:24 | |
| *** alatiera_ is now known as alatiera | 17:25 | |
| *** nimish has quit IRC | 17:29 | |
| *** nimish has joined #buildstream | 17:29 | |
| *** finn_ has quit IRC | 17:31 | |
| *** lachlan has joined #buildstream | 17:33 | |
| *** lachlan has quit IRC | 17:43 | |
| *** nimish has quit IRC | 17:44 | |
| *** nimish has joined #buildstream | 17:44 | |
| *** lachlan has joined #buildstream | 17:49 | |
| *** nimish has quit IRC | 17:49 | |
| *** nimish has joined #buildstream | 17:50 | |
| *** raoul_ has quit IRC | 18:03 | |
| *** tristan has quit IRC | 18:04 | |
| *** nimish has quit IRC | 18:05 | |
| *** tristan has joined #buildstream | 18:06 | |
| *** ChanServ sets mode: +o tristan | 18:06 | |
| *** tpollard has quit IRC | 18:10 | |
| *** jonathanmaw has quit IRC | 18:30 | |
| *** lachlan has quit IRC | 18:50 | |
| gitlab-br-bot | tristanvb opened issue #854 (Coverage data seems to get lost, maybe) on buildstream https://gitlab.com/BuildStream/buildstream/issues/854 | 18:53 |
| *** xjuan has joined #buildstream | 19:16 | |
| *** alatiera has quit IRC | 19:37 | |
| *** alatiera has joined #buildstream | 20:02 | |
| *** tristan has quit IRC | 20:07 | |
| *** tristan has joined #buildstream | 20:08 | |
| gitlab-br-bot | tristanvb opened MR !1058 (tristan/element-processing-order->master: Make build plan element list stable) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1058 | 20:10 |
| *** tristan has quit IRC | 20:22 | |
| *** toscalix has quit IRC | 20:22 | |
| *** alatiera_ has joined #buildstream | 20:23 | |
| *** alatiera has quit IRC | 20:23 | |
| *** alatiera_ is now known as alatiera | 20:24 | |
| *** alatiera has quit IRC | 21:16 | |
| *** ctolentino has quit IRC | 21:28 | |
| *** benschubert has quit IRC | 22:24 | |
| *** xjuan has quit IRC | 23:45 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!