| *** tristan has quit IRC | 04:56 | |
| *** tristan has joined #buildstream | 05:11 | |
| *** ChanServ sets mode: +o tristan | 05:12 | |
| *** bochecha has joined #buildstream | 07:29 | |
| gitlab-br-bot | buildstream: issue #96 ("Integration tests should not be in a separate repo") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/96 | 07:47 |
|---|---|---|
| *** jonathanmaw has joined #buildstream | 07:58 | |
| *** bochecha_ has joined #buildstream | 08:01 | |
| *** bochecha has quit IRC | 08:03 | |
| *** bochecha_ is now known as bochecha | 08:03 | |
| *** tlater has joined #buildstream | 08:17 | |
| *** ssam2 has joined #buildstream | 08:51 | |
| *** jonathanmaw has quit IRC | 08:54 | |
| *** jonathanmaw has joined #buildstream | 08:54 | |
| tristan | tlater, I think we misunderstood eachother on a fixme comment thing... we should still be able to install buildstream on non-linux unix | 08:56 |
| tlater | tristan: So setup.py should detect the host OS as well? | 08:57 |
| tristan | tlater, so setup.py requirements for ostree & bwrap should just be conditional to `if sys.platform.startswith('linux'):` as well | 08:57 |
| tristan | yeah, I mean it's just a one liner we're using anyway right now | 08:57 |
| tlater | Right, that should be fine | 08:57 |
| tristan | I was thinking also that we should probably be re-running the pytest suite with the force backend hack too... because we're not running that batch through the unix platform | 08:58 |
| tristan | but it's a small vector of tests that need dual-platform coverage | 08:59 |
| tristan | so, lets not worry about that right now I think | 08:59 |
| tristan | I want to land this asap, it looks quite ready to me | 08:59 |
| tristan | yeah looks like it would be a lot of work too | 09:00 |
| tristan | we'd have to have separate cache key test, yuk | 09:00 |
| tristan | Actually we probably need to make sure that the cache-key test doesnt run on non-linux | 09:01 |
| paulsherwood | tristan: before i forget... | 09:10 |
| paulsherwood | a downstream project demonstrated that ybd cache-key fails to take $PREFIX into account | 09:11 |
| tristan | paulsherwood, makes sense, because ybd exports it as an env var; we dont have that issue (env vars, if used are considered in the cache key; otherwise we use expansion of %{prefix} variables and only expanded strings are considered) | 09:32 |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 1 commit (last: setup.py: Make setup.py work on non-linux) https://gitlab.com/BuildStream/buildstream/commit/3d5796acfd4748b7b9d09ef410a5ebaf65146ea3 | 09:49 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 09:49 |
| tlater | tristan: A lot of those tests will fail on other platforms anyway. | 09:49 |
| tristan | tlater, which other ones can you think of ? | 09:49 |
| tlater | All ostree stuff, for example | 09:49 |
| tristan | Right, there is a strategy for that | 09:50 |
| tristan | though | 09:50 |
| tristan | tlater, look at tests/testlib/repo/ostree.py for example | 09:50 |
| tristan | tlater, when using an ostree scaffolding for testing the ostree source, we immediately skip the test | 09:51 |
| tristan | tlater, also tests/testlib/artifactshare.py has the same logic | 09:51 |
| tlater | Strange, I wonder why that didn't work for me when I ran the tests - maybe that was before the shift to this. | 09:52 |
| tristan | tlater, have access to any legitimately non-linux machine that you can test this on ? | 09:54 |
| tristan | I suppose it can be tested by forcing the backend and uninstalling bwrap/ostree | 09:55 |
| tlater | Right, that should probably be the easiest way | 09:56 |
| * tristan thinks that not running the pytest tests in CI for multiple platforms is not bad; but having the tests actually failing on non-linux would be embarrassingly lame :-S | 09:57 | |
| tlater | In that case just running them in the CI isn't an issue though, right? | 09:57 |
| tlater | If we already go through the effort of making them run... | 09:57 |
| tristan | Well, if we do run them in CI; it's certainly better | 09:57 |
| tlater | As long as nobody changes the artifact cache code again before merge this... | 09:58 |
| tlater | ;) | 09:58 |
| tristan | The push/pull stuff remains unimplemented for tarball based caches correct ? | 09:59 |
| tlater | Yes | 10:00 |
| tristan | in that case, I *think* all it means is disabling some tests which wont work; and when implementing tar cache sharing, we'll want it to support the local uris (like the ostree cache does); then we can run the same push/pull tests against a different artifact cache scaffolding | 10:00 |
| tristan | It should only mean this; the other thing is sandboxes, and we dont have any pytest tests which exercise sandboxes | 10:02 |
| tlater | I'll run the tests on the fallback platform, if something awful pops up I'll know :) | 10:02 |
| tristan | ok thanks ! | 10:03 |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 3 commits (last: Fix tests for other platforms) https://gitlab.com/BuildStream/buildstream/commit/a652263fe932823d6e5625a9b9176afad7d661f8 | 10:29 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 10:29 |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 15 commits (last: Fork and embed fusepy) https://gitlab.com/BuildStream/buildstream/commit/bf0f2863e8cea73e41d5d9efca3126e4075fc3ae | 10:30 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 10:31 |
| *** tiagogomes has joined #buildstream | 10:41 | |
| *** tlater has quit IRC | 10:56 | |
| *** tlater has joined #buildstream | 10:57 | |
| tlater | #81 is proving to be pretty tough :( | 11:31 |
| persia | Are you able to reproduce, at least? | 11:49 |
| tristan | persia, I think we are talking about MR 81 | 11:53 |
| tristan | not issue 81 | 11:53 |
| * tristan could be wrong | 11:54 | |
| tlater | No, issue 81 for that one. Needed to wait for MR 81 to finish testing. | 12:13 |
| tlater | I can definitely reproduce 81, but the permission issues are tricky to sort out neatly. | 12:14 |
| tristan | tlater, ok so with MR 81 - what are we doing, I noticed the pipeline doesnt have an added pytest run either | 12:16 |
| * tristan eager to merge hehe | 12:16 | |
| tlater | tristan: The tests seem to run fine, I'll add them to setup.py (here's hoping that doesn't break the carefully crafted coverage report) | 12:20 |
| * tlater is also eager to merge :) | 12:20 | |
| tlater | I meant the CI, not setup.py... Brain? | 12:21 |
| tristan | yep, to the .gitlab-ci.yml | 12:22 |
| tristan | yeah I suspect you've got coverage figured out now | 12:22 |
| tristan | its a nice feature the combined coverage :) | 12:23 |
| tlater | It seems that my DNS broke down, I'll push ASAP though. | 12:26 |
| tlater | Interesting that irc still works | 12:26 |
| tlater | (Perhaps it just pretends to and I'm talking to myself) | 12:30 |
| ssam2 | you are talking to yourself | 12:31 |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 2 commits (last: .gitlab-ci.yml: Add fallback tests to CI) https://gitlab.com/BuildStream/buildstream/commit/21b742186ffeab48e40bb2cfb6d3ae486da67492 | 12:35 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 12:35 |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 2 commits (last: .gitlab-ci.yml: Add fallback tests to CI) https://gitlab.com/BuildStream/buildstream/commit/c834564f6c3236f13ff357ddc8c728d94af42585 | 12:47 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 12:48 |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 2 commits (last: .gitlab-ci.yml: Add fallback tests to CI) https://gitlab.com/BuildStream/buildstream/commit/70398f22a7c1e8bb255fbb1286f96479acb0e608 | 12:55 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 12:55 |
| tlater | Hmm... Is there a way to fake the platform string in python? | 13:00 |
| tlater | I suppose maybe the setup.py should also listen to BST_FORCE_BACKEND | 13:00 |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 1 commit (last: setup.py: Make setup.py work on non-linux) https://gitlab.com/BuildStream/buildstream/commit/fc200481ab5f0b84cf930819fba61c4456c7ac15 | 13:02 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 13:02 |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 1 commit (last: setup.py: Make setup.py work on non-linux) https://gitlab.com/BuildStream/buildstream/commit/aedd000ad7dfd0deec824d8b6a1b960b9611467c | 13:03 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 13:03 |
| ssam2 | tlater, you can monkey patch anything | 13:32 |
| ssam2 | i think pytest has a 'monkeypatch' fixture that can do that for test cases | 13:32 |
| ssam2 | https://docs.pytest.org/en/latest/monkeypatch.html | 13:32 |
| tlater | ssam2: From outside python I mean, because setup.py doesn't want to run tests without dependencies in the first place :) | 13:32 |
| tlater | I guess importing setup in a strange way could work. | 13:33 |
| *** tristan has quit IRC | 13:33 | |
| gitlab-br-bot | push on buildstream@cross_platform (by Tristan Maat): 3 commits (last: Fix tests for other platforms) https://gitlab.com/BuildStream/buildstream/commit/206fc51a6cf57697f03e27508819eafaaee58317 | 13:46 |
| gitlab-br-bot | buildstream: merge request (cross_platform->master: Cross platform) #81 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/81 | 13:47 |
| bochecha | tlater: isn't that what PEP 518 is for? | 14:10 |
| gitlab-br-bot | push on buildstream@81-non-empty-read-only-directories-not-handled-during-bst-build-and-others (by Tristan Maat): 1 commit (last: Ensure that artifact file permissions are set in the right order) https://gitlab.com/BuildStream/buildstream/commit/487ac5e7ff91ced0220cfc48958efca8de826d4d | 14:11 |
| tlater | bochecha: I think that can only detect python dependencies. ostree/bubblewrap in this case | 14:15 |
| bochecha | tlater: yes, indeed | 14:15 |
| *** tristan has joined #buildstream | 14:27 | |
| gitlab-br-bot | push on buildstream@81-non-empty-read-only-directories-not-handled-during-bst-build-and-others (by Tristan Maat): 1 commit (last: Ensure that artifact file permissions are set in the right order) https://gitlab.com/BuildStream/buildstream/commit/79cb91f22c21d7bc64e18b8000d2e046e8d4c8d2 | 14:51 |
| bochecha | I have to say, overall I'm very happy about BuildStream. I started looking into it as an experiment, entirely unconvinced, thinking this would all be a huge waste of time... and now I'm looking at what seems to be a successful proof of concept, thinking that I'll need to convince my customer we should move to it | 14:54 |
| bochecha | good job everyone, I guess :) | 14:54 |
| tlater | \o/ | 14:55 |
| paulsherwood | w00t! :) | 14:57 |
| tristan | bochecha, very nice to hear that :D | 14:59 |
| bochecha | in FOSS we tend to only ever hear about complaints, never about positive experiences, so I'm trying to do my part to counter that tendency :) | 15:00 |
| paulsherwood | that's so true, bochecha :) | 15:01 |
| bochecha | although I'm not sure my feedback was entirely positive: BuildStream is actually giving me more work to do if we move to it! | 15:02 |
| bochecha | (but I guess that's not a terrible problem to have) | 15:02 |
| gitlab-br-bot | push on buildstream@81-non-empty-read-only-directories-not-handled-during-bst-build-and-others (by Tristan Maat): 1 commit (last: Ensure that artifact file permissions are set in the right order) https://gitlab.com/BuildStream/buildstream/commit/349a15cfa7c9e278bafeda5e4ac2710cf7eb902e | 15:07 |
| gitlab-br-bot | push on buildstream@81-non-empty-read-only-directories-not-handled-during-bst-build-and-others (by Tristan Maat): 1 commit (last: Ensure that artifact file permissions are set in the right order) https://gitlab.com/BuildStream/buildstream/commit/5b94d7f595c45f7ed6a9e11985a22c6afe1a9b7e | 15:14 |
| *** semanticdesign has joined #buildstream | 15:19 | |
| *** jonathanmaw has quit IRC | 16:18 | |
| *** tlater has quit IRC | 16:22 | |
| *** ssam2 has quit IRC | 18:10 | |
| *** xjuan has joined #buildstream | 18:35 | |
| *** tristan has quit IRC | 19:38 | |
| *** xjuan has quit IRC | 20:37 | |
| *** bochecha_ has joined #buildstream | 21:00 | |
| *** bochecha has quit IRC | 21:02 | |
| *** bochecha_ is now known as bochecha | 21:03 | |
| *** semanticdesign has quit IRC | 22:16 | |
| *** bochecha has quit IRC | 23:04 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!