| *** tristan has joined #buildstream | 05:17 | |
| *** ChanServ sets mode: +o tristan | 05:17 | |
| juergbi | nanonyme: the plan is to have stable buildbox-* branches in time for bst 2.0. buildbox 1.0 may not be released in time, however, in that case we can create a 0.x stable branch | 07:16 |
|---|---|---|
| juergbi | there is almost no interface breakage in buildbox-casd/run master as it is, however, there is no guarantee right now. and the buildbox-common library API/ABI is not stable | 07:17 |
| juergbi | (buildstream doesn't directly use the library API, though) | 07:17 |
| juergbi | I expect it to be acceptable for 2.x releases after 2.0 to require upgrading to a more recent buildbox stable branch (e.g. from a 0.x branch to a 1.0 branch). however, 2.0.x releases shouldn't require that | 07:20 |
| juergbi | (buildbox 1.0 might even be fully backwards compatible with 0.x with regards to the interfaces buildstream uses) | 07:21 |
| gitlab-br-bot | devcurmudgeon closed issue #474 (BST needs to support its historic cache-key algorithms...) on buildstream https://gitlab.com/BuildStream/buildstream/-/issues/474 | 08:01 |
| *** santi has joined #buildstream | 08:37 | |
| benschubert | tristan: do you have a rough idea on when you will tackle #2070? So that I know if I should focus and plugins tests for now instead :) (On the scheduler, there is still _one_ concurrency issue I'm trying to fix) | 09:25 |
| tristan | benschubert, I'm sorry, I've been side tracked wrestling with github all week | 09:40 |
| tristan | I think I can get a new branch with tests for #2070 this weekend | 09:40 |
| benschubert | ah no worries ! Just want to know how I prioritize my work :) I'll switch to source tests in the meantime | 09:41 |
| *** tristan has quit IRC | 09:54 | |
| *** cs-shadow has joined #buildstream | 10:00 | |
| gitlab-br-bot | BenjaminSchubert opened MR !2085 (bschubert/simplify-workspace-test->master: Simplify source workspace tests) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/2085 | 11:46 |
| benschubert | ^ should be a quick review if someone has time | 11:47 |
| *** tpollard has joined #buildstream | 15:22 | |
| *** cs-shadow has quit IRC | 15:41 | |
| *** tpollard has quit IRC | 16:00 | |
| *** santi has quit IRC | 17:13 | |
| *** santi has joined #buildstream | 17:13 | |
| *** santi has quit IRC | 17:16 | |
| *** cs-shadow has joined #buildstream | 17:17 | |
| *** tristan has joined #buildstream | 18:32 | |
| *** ChanServ sets mode: +o tristan | 18:32 | |
| *** tristan has quit IRC | 18:38 | |
| nanonyme | benschubert, is there something there that removes temp dirs after usage? | 19:44 |
| benschubert | nanonyme: pytest should do it's cleanup on the next run | 19:44 |
| nanonyme | Ack. I've taken the pattern in Another Project that I have a fixture that just does with tempfile.TemporaryDir() as tmpdir: yield tmpdir | 19:45 |
| nanonyme | Python then nicely removes the directory after test | 19:46 |
| nanonyme | Well, tempdir.TemporaryDirectoy(dir=".") but anyway | 19:46 |
| nanonyme | Oops, tempfile.andsoforth. I shouldn't try to write code this late | 19:47 |
| benschubert | nanonyme: pytest 'tmpdir' allows to keep them so that you can debug if there is a problem, I dislike removing them as you run | 19:47 |
| nanonyme | Ah, okay | 19:47 |
| nanonyme | That's also a quite good point | 19:47 |
| nanonyme | In the Other Project we had common issue that workspaces became so large (cumulatively) over time that it's best to just emove all non-SCM data after build if build didn't fail | 19:49 |
| nanonyme | Quickly scanned through the code but while seemed reasonable, not really familiar enough with codebase to properly review | 19:51 |
| benschubert | Thanks :) Sure, let's get someone who's familiar with it to review it (disclaimer: I'm not either) | 19:53 |
| *** cs-shadow has quit IRC | 21:27 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!