| *** dftxbs3e_ has quit IRC | 01:05 | |
| *** dftxbs3e has joined #buildstream | 01:09 | |
| *** tristan has quit IRC | 01:41 | |
| *** tristan has joined #buildstream | 02:02 | |
| *** ChanServ sets mode: +o tristan | 02:02 | |
| tristan | nanonyme, since recently in master we support expanding variables all across the board, and removed functions like Element.node_subst_variables() which used to be needed for element plugins to explicitly substitute variables in loaded yaml | 02:04 |
|---|---|---|
| tristan | so in recent master, using variables in sources should "just work" | 02:04 |
| *** tristan has quit IRC | 06:44 | |
| *** tristan has joined #buildstream | 06:44 | |
| *** ChanServ sets mode: +o tristan | 06:44 | |
| *** santi has joined #buildstream | 09:43 | |
| *** Mark has quit IRC | 12:33 | |
| *** TobiasFella has quit IRC | 12:33 | |
| *** tristan has quit IRC | 12:50 | |
| *** tristan has joined #buildstream | 12:59 | |
| *** ChanServ sets mode: +o tristan | 12:59 | |
| tristan | juergbi, recently running the tests ends up spawning a lot of lingering buildbox-casd processes, is this normal/expected ? do they timeout and go away on their own or is it a bug ? | 13:27 |
| juergbi | lingering casd is certainly not expected in normal operation | 13:28 |
| juergbi | I've seen it at least with failing test cases as well | 13:29 |
| tristan | I just ran master and verified that it's not caused by my branch | 13:29 |
| tristan | but I seem to recall noticing this since the github migration (I probably wasn't paying attention before) | 13:29 |
| juergbi | for local test runs this should be irrelevant, though | 13:30 |
| tristan | Right now I've got only successful tests, on master, and at least 2 dozen casd processes | 13:30 |
| tristan | Oh right, I noticed this last week when running the RE tests in local docker-compose | 13:31 |
| tristan | I'm not sure I understand, it is expected with tests but not normal operation ? Is this because we have a 1-casd-per-artifact-cache scenario ? | 13:32 |
| tristan | Like, a casd will linger around in case of imminent connections, but one casd will be spawned per local CAS ? | 13:33 |
| juergbi | I wouldn't expect it with tests either but if they fail, odd things can happen | 13:36 |
| juergbi | and maybe some internal tests are missing cleanup but they shouldn't | 13:36 |
| juergbi | casd is spawned on each bst invocation | 13:36 |
| juergbi | but it should also be terminated before bst exits | 13:36 |
| tristan | I see | 13:48 |
| tristan | So it is a bug | 13:48 |
| nanonyme | tristan, cool. This is something I've recently been a bit annoyed at. Eg with Python updates there's bazillion places that need changing on every update | 15:27 |
| nanonyme | It can be reduced through variables but eg track doesn't support expanding currently | 15:27 |
| *** santi has quit IRC | 18:48 | |
| nanonyme | tristan, is there any plan to package buildbox and such for distros? | 19:17 |
| *** jward has quit IRC | 20:14 | |
| *** jward has joined #buildstream | 20:15 | |
| *** jward has quit IRC | 20:16 | |
| *** jward has joined #buildstream | 20:16 | |
| *** jward has quit IRC | 20:17 | |
| *** jward has joined #buildstream | 20:17 | |
| *** jward has quit IRC | 20:18 | |
| *** jward has joined #buildstream | 20:18 | |
| *** jward has quit IRC | 20:19 | |
| *** jward has joined #buildstream | 20:19 | |
| *** jward has quit IRC | 20:20 | |
| *** jward has joined #buildstream | 20:20 | |
| *** jward has quit IRC | 20:22 | |
| *** jward has joined #buildstream | 20:22 | |
| *** hergertme has quit IRC | 21:53 | |
| *** hergertme has joined #buildstream | 21:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!