| *** narispo has quit IRC | 00:44 | |
| *** ikerperez has joined #buildstream | 06:56 | |
| *** traveltissues has joined #buildstream | 07:40 | |
| gitlab-br-bot | traveltissues opened MR !1648 (traveltissues/undofilesareannoying->master: .gitignore: ignore undo files) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1648 | 08:04 |
|---|---|---|
| gitlab-br-bot | tpollard approved MR !1648 (traveltissues/undofilesareannoying->master: .gitignore: ignore undo files) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1648 | 08:06 |
| juergbi | traveltissues: shouldn't you rather add this to your global git ignore? | 08:14 |
| juergbi | not sure what's considered best practice in this regard | 08:14 |
| benschubert | Yeah, I don't think we want to end up with all editors files ignored there or it will become a mess :/ | 08:14 |
| benschubert | especially since we don't have any other such files currently :) | 08:21 |
| traveltissues | well tbf you could make the same argument for all the currently ignored files | 08:22 |
| traveltissues | but i understand your point | 08:23 |
| benschubert | traveltissues: the currently ignored files are files generated by the project itself, not files generated by people | 08:24 |
| *** santix has joined #buildstream | 08:41 | |
| *** tiagogomes has joined #buildstream | 08:41 | |
| *** adds68 has joined #buildstream | 08:53 | |
| *** jonathanmaw has joined #buildstream | 09:05 | |
| benschubert | https://gitlab.com/BuildStream/bst-plugins-experimental/merge_requests/48 to update bst-plugins-experimental with the latest breaking changes on the yaml side | 09:06 |
| benschubert | if someone has time to review | 09:06 |
| gitlab-br-bot | traveltissues opened issue #1172 (mypy failing on master from tempfile api change) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1172 | 09:28 |
| *** qinusty has joined #buildstream | 09:29 | |
| *** lachlan has joined #buildstream | 09:33 | |
| *** akvilebirgelyte has quit IRC | 09:56 | |
| *** akvilebirgelyte has joined #buildstream | 09:56 | |
| gitlab-br-bot | traveltissues closed MR !1648 (traveltissues/undofilesareannoying->master: .gitignore: ignore undo files) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1648 | 10:02 |
| gitlab-br-bot | cs-shadow opened issue #1173 (Tests fail with Python 3.8) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1173 | 10:05 |
| *** lachlan has quit IRC | 10:09 | |
| *** lachlan has joined #buildstream | 10:11 | |
| *** lachlan has quit IRC | 10:15 | |
| *** lachlan has joined #buildstream | 10:20 | |
| tlater[m] | Anyone else seeing this with mypy? https://gitlab.com/BuildStream/buildstream/-/jobs/323604176 | 10:23 |
| tlater[m] | That line hasn't been touched, and the variable is definitely supposed to be an Optional[string]. | 10:24 |
| * tlater[m] can't reproduce it locally either - did something in the test image change? Surely it couldn't have, I haven't rebased on top of master in a while... | 10:26 | |
| benschubert | mypy version maybe? | 10:27 |
| tpollard | traveltissues raised an issue about mypy I think | 10:27 |
| traveltissues | #1172 | 10:27 |
| benschubert | if you tox -r -e mypy? | 10:27 |
| tlater[m] | Heh, I've been running it manually because I didn't realize we had an env for it | 10:28 |
| * tlater[m] tries that command | 10:28 | |
| tlater[m] | No, doesn't show anything locally | 10:28 |
| tlater[m] | Yeah, https://gitlab.com/BuildStream/buildstream/issues/1172 is showing the exact same thing | 10:29 |
| tlater[m] | Ta traveltissues | 10:29 |
| * tlater[m] wonders how the test issue changed overnight, though | 10:29 | |
| tlater[m] | s/issue/image | 10:29 |
| *** akvilebirgelyte has quit IRC | 10:47 | |
| *** akvilebirgelyte has joined #buildstream | 10:48 | |
| cs-shadow | tlater[m]: we don't pin mypy version | 11:03 |
| cs-shadow | so tox pulls in the latest at runtime | 11:03 |
| tlater[m] | cs-shadow: From the issue it seems that the tox version changed | 11:04 |
| cs-shadow | ah! then I'm not sure how that would've happened | 11:04 |
| gitlab-br-bot | traveltissues opened MR !1649 (traveltissues/1172->master: tox.ini: pin mypy 0.730) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1649 | 11:20 |
| juergbi | benschubert: I'd appreciate feedback on my comment in #1126 | 11:56 |
| gitlab-br-bot | Issue #1126: Cache resolution is slow https://gitlab.com/BuildStream/buildstream/issues/1126 | 11:56 |
| *** narispo has joined #buildstream | 12:05 | |
| benschubert | juergbi: I'll have a look | 12:20 |
| *** santix has quit IRC | 12:20 | |
| gitlab-br-bot | cs-shadow approved MR !1649 (traveltissues/1172->master: tox.ini: pin mypy 0.730) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1649 | 12:20 |
| gitlab-br-bot | marge-bot123 closed issue #1172 (mypy failing on master from tempfile.mkstemp) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1172 | 12:25 |
| gitlab-br-bot | marge-bot123 merged MR !1649 (traveltissues/1172->master: tox.ini: pin mypy 0.730) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1649 | 12:25 |
| *** santix has joined #buildstream | 12:38 | |
| *** narispo has quit IRC | 13:07 | |
| *** narispo has joined #buildstream | 13:09 | |
| *** santix has quit IRC | 13:12 | |
| *** santix has joined #buildstream | 13:24 | |
| gitlab-br-bot | aevri opened (was WIP) MR !1650 (aevri/enable_spawn_ci->master: Enable a subset of `tests-spawn-multiprocessing-start-method`) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1650 | 13:26 |
| *** santix has quit IRC | 13:44 | |
| *** mohan43u has quit IRC | 13:48 | |
| *** mohan43u has joined #buildstream | 13:49 | |
| *** santix has joined #buildstream | 13:52 | |
| *** narispo has quit IRC | 14:12 | |
| *** narispo has joined #buildstream | 14:12 | |
| *** santix is now known as santi | 14:15 | |
| tlater[m] | Does tracking not change source keys? | 15:01 |
| * tlater[m] is tracing through what _update_state still does | 15:04 | |
| tlater[m] | And it seems to update elements when they finish tracking - I'd expect source keys to change, and that change to be reflected in update_cache_keys() | 15:04 |
| tlater[m] | But update_cache_keys() only seems to update the part of the key that's dictated by dependencies | 15:04 |
| * tlater[m] is a little confused about this | 15:04 | |
| juergbi | tlater[m]: for elements that are pending tracking, the source key is calculated only after tracking is done | 15:05 |
| juergbi | i.e., it's not changing from one key to another | 15:05 |
| tlater[m] | juergbi: I assume that applies to the *whole* cache key then? | 15:06 |
| juergbi | yes, cache key can only be calculated when the source key is available | 15:06 |
| tlater[m] | Awesome, ta juergbi | 15:06 |
| * tlater[m] wonders if we could calculate cache keys lazily | 15:07 | |
| *** santi has quit IRC | 15:07 | |
| tlater[m] | And then circumvent this whole thing... | 15:07 |
| juergbi | tlater[m]: as the cache key dependencies can be pretty deep, I don't think lazy calculation makes sense | 15:09 |
| juergbi | also, availability of a cache key may trigger further actions such as building | 15:09 |
| juergbi | our plan is to rather have a fine-grained push-based model | 15:10 |
| * tlater[m] is roughly aware of this, but struggles seeing where that is going | 15:12 | |
| juergbi | whenever something happens that may affect state, we update everything that can / needs to be updated | 15:12 |
| juergbi | in contrast to calling update_state() from all these points, we want to only trigger state updates that are relevant to the event | 15:13 |
| tlater[m] | Yup, that makes sense, but I don't really see much more opportunity for that | 15:14 |
| tlater[m] | Although I admittedly haven't finished reading through the code | 15:14 |
| *** santi has joined #buildstream | 15:15 | |
| juergbi | tlater[m]: a possible next step could be to directly call __update_source_state() instead of _update_state() from 'event handlers' that may have an influence on source state | 15:16 |
| gitlab-br-bot | traveltissues opened (was WIP) MR !1653 (traveltissues/notes->master: workspaces via sourcecache MR2.5) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1653 | 15:16 |
| juergbi | tlater[m]: and then call _update_state() from __update_source_state() if anything has changed | 15:16 |
| juergbi | and later replace further _update_state() calls with e.g. __update_cache_keys() or whatever else more specific thing that should be updated | 15:18 |
| juergbi | (and change __update_cache_keys() to trigger further updates as needed) | 15:18 |
| tlater[m] | I've actually started playing with this | 15:19 |
| tlater[m] | Hrm, I suppose I've just not been through all handlers yet and I'll find things that don't need `__update_source_state()` when I dig far enough | 15:19 |
| gitlab-br-bot | traveltissues opened (was WIP) MR !1651 (traveltissues/1161->master: remove some private api from workspace plugin) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1651 | 15:20 |
| tlater[m] | `_schedule_assemble()` seems to be an example of that | 15:20 |
| juergbi | yes, it's not always straight forward to see what depends on what | 15:21 |
| juergbi | tlater[m]: you probably do this anyway but due to this tricky dependencies, please try to keep each commit as small as possible | 15:21 |
| tlater[m] | juergbi: Yeah, I've not even started committing anything yet | 15:22 |
| tlater[m] | Just pulling a wire here and there and seeing what breaks | 15:22 |
| juergbi | sure, makes sense | 15:23 |
| tpollard | that's a good metaphor for working on update_state | 15:23 |
| tlater[m] | Hehe | 15:24 |
| * tlater[m] approaches most of BuildStream like this, tbf | 15:24 | |
| traveltissues | i'd appreciate someone taking a look at !1653 | 15:30 |
| traveltissues | benschubert, will you please have a look at !1651 when you have a chance | 15:31 |
| benschubert | sure | 15:32 |
| traveltissues | ty | 15:33 |
| *** santi has quit IRC | 15:46 | |
| *** santi has joined #buildstream | 15:53 | |
| *** bochecha has quit IRC | 15:56 | |
| aevri | If anyone else is excited about getting the spawn tests back on track, !1650 is ready for review. There's a small queue of small changes forming behind that one :) | 16:08 |
| gitlab-br-bot | MR !1650: Enable a subset of `tests-spawn-multiprocessing-start-method` https://gitlab.com/BuildStream/buildstream/merge_requests/1650 | 16:08 |
| * tlater[m] is brimming with excitement | 16:09 | |
| aevri | hehe | 16:10 |
| tlater[m] | ... partially because I think I'm done fighting with _update_state for the day ;p | 16:10 |
| aevri | While those are cooking, I'll probably get back to that casdprocessmanager change you wanted | 16:10 |
| aevri | aha, I hear it doesn't fight fair :) | 16:11 |
| tlater[m] | It has about 5 similarly-named friends who jump in once you think you've grokked it. | 16:12 |
| * tlater[m] is thankful others have handled its old form though. | 16:12 | |
| tpollard | I think python needs ___$method | 16:13 |
| tlater[m] | Hah! | 16:14 |
| aevri | I think I've been out of the loop a little while - does thumbs up on a review mean 'approved', or is it a general 'yay for this sort of thing'? | 16:31 |
| aevri | I've another review with some thumbs up but no explicit approval, I'm not sure if that means it's good to go :) | 16:31 |
| *** lachlan has quit IRC | 16:32 | |
| * cs-shadow personally uses it for 'yay for this sort of thing' | 16:36 | |
| juergbi | I don't think we've defined this in the project, so I suspect it's used in different ways | 16:39 |
| aevri | Alrighty, I'll press for explicit approval in those cases then, thanks! | 16:41 |
| traveltissues | marge is broken? | 16:48 |
| traveltissues | https://gitlab.com/BuildStream/buildstream/merge_requests/1653#note_232031703 | 16:48 |
| *** tiagogomes has quit IRC | 16:52 | |
| traveltissues | the runner seems to be falling over | 16:56 |
| traveltissues | https://gitlab.com/BuildStream/buildstream/-/jobs/323736664 | 16:56 |
| traveltissues | `fatal: unable to access 'https://gitlab.com/BuildStream/bst-plugins-experimental.git/': ` | 16:57 |
| gitlab-br-bot | BenjaminSchubert opened MR !1656 (bschubert/bump-bst-experimental-version->master: .gitlab-ci.yml: Update bst-plugins-experimental version) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1656 | 17:00 |
| *** jonathanmaw has quit IRC | 17:05 | |
| *** santi has quit IRC | 17:13 | |
| *** traveltissues has quit IRC | 17:13 | |
| gitlab-br-bot | cs-shadow approved MR !1656 (bschubert/bump-bst-experimental-version->master: .gitlab-ci.yml: Update bst-plugins-experimental version) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1656 | 17:34 |
| *** slaf has quit IRC | 17:35 | |
| *** slaf has joined #buildstream | 17:42 | |
| *** slaf has joined #buildstream | 17:42 | |
| *** slaf has joined #buildstream | 17:42 | |
| *** slaf has joined #buildstream | 17:43 | |
| *** slaf has joined #buildstream | 17:43 | |
| *** slaf has joined #buildstream | 17:43 | |
| *** slaf has joined #buildstream | 17:44 | |
| *** slaf has joined #buildstream | 17:44 | |
| *** slaf has joined #buildstream | 17:44 | |
| *** slaf has joined #buildstream | 17:45 | |
| benschubert | Does someone has access to marge? It is confused | 17:46 |
| gitlab-br-bot | jjardon opened issue #1174 (overnigth test are failing: AttributeError: 'buildstream.node.SequenceNode' object has no attribute '_strip_node_info') on buildstream https://gitlab.com/BuildStream/buildstream/issues/1174 | 18:50 |
| *** narispo has quit IRC | 18:53 | |
| *** narispo has joined #buildstream | 18:54 | |
| *** narispo has quit IRC | 19:14 | |
| *** narispo has joined #buildstream | 19:14 | |
| *** narispo has quit IRC | 19:17 | |
| *** narispo has joined #buildstream | 19:17 | |
| *** narispo has quit IRC | 19:38 | |
| *** narispo has joined #buildstream | 19:38 | |
| *** narispo has quit IRC | 22:04 | |
| *** narispo has joined #buildstream | 22:05 | |
| *** narispo has quit IRC | 22:54 | |
| *** narispo has joined #buildstream | 22:54 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!