| *** narispo has quit IRC | 05:08 | |
| *** narispo has joined #buildstream | 05:08 | |
| *** tristan has quit IRC | 05:09 | |
| *** samkirkham has joined #buildstream | 07:14 | |
| *** tristan has joined #buildstream | 07:35 | |
| *** tristan has quit IRC | 07:42 | |
| *** ikerperez has joined #buildstream | 07:44 | |
| juergbi | it seems buildstream-docker-images CI is failing for a week or so. connection to docker daemon was refused / docker daemon not running. is anyone familiar with this issue? | 08:05 |
|---|---|---|
| *** adds68 has joined #buildstream | 08:09 | |
| *** bochecha has joined #buildstream | 08:11 | |
| benschubert | juergbi: ah thanks! I know the problem, let me make a PR for that | 08:16 |
| juergbi | ta | 08:16 |
| benschubert | juergbi: https://gitlab.com/BuildStream/buildstream-docker-images/merge_requests/120 that should do it | 08:20 |
| alexandrufazakas | jennis, tlater[m]: given the discussion yesterday on #1091, where are we going with it? Should the box be ticked in #773? | 08:33 |
| gitlab-br-bot | Issue #1091: Create a custom artifact-ref click parameter type https://gitlab.com/BuildStream/buildstream/issues/1091 | 08:33 |
| gitlab-br-bot | Issue #773: The command-line should provide access to artifacts https://gitlab.com/BuildStream/buildstream/issues/773 | 08:33 |
| *** jonathanmaw has joined #buildstream | 08:35 | |
| jennis | alexandrufazakas, I think the two tab-completion related boxes can be checked | 08:36 |
| jennis | s/checked/ticked | 08:36 |
| alexandrufazakas | sweet | 08:37 |
| alexandrufazakas | jennis: wait | 08:37 |
| alexandrufazakas | you mean the tab completoin one and the glob expression one? | 08:37 |
| *** jonathanmaw_ has joined #buildstream | 08:41 | |
| *** jonathanmaw has quit IRC | 08:43 | |
| jennis | Yes, them two | 08:46 |
| jennis | Glob expressions work | 08:46 |
| jennis | e.g, if I want to delete *all* artifacts from my autotools project, `bst artifact delete autotools/*` will do this | 08:47 |
| alexandrufazakas | Yeah, I saw that _classify_artifacts splits them into globs/non-globs and they're later expanded, I just watned to make sure :D | 08:47 |
| * alexandrufazakas ticks them off while he can still do this | 08:47 | |
| alexandrufazakas | I'll leave a comment on the issue explaining we already have these then, jennis? | 08:48 |
| jennis | Yep :) Perhaps a link to the tests too, we should have tests for both | 08:48 |
| jennis | If not, we'd need to write | 08:48 |
| jennis | I already linked the tab completion of artifacts tests on #1091 | 08:48 |
| alexandrufazakas | Yep | 08:49 |
| alexandrufazakas | I'll see if we have any for globbing | 08:49 |
| jennis | alexandrufazakas, just had a quick glance, we gloab an artifact log command, but not an artifact delete one | 08:51 |
| jennis | I think it'd be nice if we hada test for this | 08:51 |
| *** jonathanmaw_ has quit IRC | 08:51 | |
| *** jonathanmaw has joined #buildstream | 08:51 | |
| jennis | The test could be: build all elements in a project, ensure they're cached, then use glob expression to delete all artifacts in the project and verify they're no longer in the cache | 08:52 |
| alexandrufazakas | jennis: the test you mentioned is tests/froentend/artifact.py::test_artifact_log right? | 08:52 |
| jennis | yep | 08:52 |
| alexandrufazakas | Sure, I'll try to do that | 08:53 |
| alexandrufazakas | Thanks for the help | 08:53 |
| jennis | But it'd be nicer/clearer to have a test which removes artifacts via a glob (IMO) | 08:53 |
| alexandrufazakas | jennis: Yup. I'll try to write one myself :) | 08:54 |
| jennis | sweet, thanks! | 08:54 |
| jennis | I feel like the Project.junction attribute is poorly named, from looking at the code (_project.py), it seems like we're referring to the parent project, rather than the junction (which I'd consider to be a child project...) | 09:02 |
| jennis | Perhaps I'm misunderstanding this | 09:02 |
| *** phildawson has joined #buildstream | 09:04 | |
| benschubert | juergbi: issue on the docker images is fixed | 09:18 |
| juergbi | great, thanks | 09:18 |
| tpollard | do we have a fix for installs failing after the ruamel change? I've tried clean installing but I still get the NoneType from userdata | 09:37 |
| tpollard | Actually I can install, but then I can't run any bst command that loads userdata | 09:38 |
| benschubert | tpollard: uninstall ruamel.yaml.clib and reinstall it | 09:38 |
| benschubert | I have no idea why, but I've been bitten by this one too | 09:39 |
| tpollard | that did the trick benschubert, how odd! | 09:41 |
| tpollard | thanks :) | 09:41 |
| benschubert | Yep,I have no idea where that comes from :( | 09:42 |
| *** tpollard has quit IRC | 10:13 | |
| *** tpollard has joined #buildstream | 10:14 | |
| *** jonathanmaw_ has joined #buildstream | 10:18 | |
| alexandrufazakas | laurence: You need a personal access token, ticking the first 3 boxes so it can query buildstream group, repo and also user api. Then run the script as `./update_committers ${token}` | 10:19 |
| alexandrufazakas | ah | 10:19 |
| alexandrufazakas | come think of it, it might not be perfect this time | 10:19 |
| alexandrufazakas | since I use `git diff --quiet` | 10:19 |
| alexandrufazakas | and the file does not exist before this | 10:19 |
| laurence | this is related to the committer policy ? | 10:20 |
| *** jonathanmaw has quit IRC | 10:20 | |
| laurence | and specifically the committers file? | 10:20 |
| laurence | Right, so, reading the additions to CONTRIBUTING, the script exists here: https://gitlab.com/BuildStream/buildstream/blob/master/contrib/update_committers.py and this needs running for the first creation of the COMMITTERS file | 10:27 |
| laurence | That makes sense :) | 10:27 |
| alexandrufazakas | Yes. Although I *think* the commit and MR part won't work properly since COMMITTERS.rst doesn't exist before hand in the repository, my bad about that. | 10:28 |
| laurence | juergbi, as someone with original committer status, would you be so kind as to run the script for the first time? ironing out the issues that alexandrufazakas mentions | 10:31 |
| laurence | I think it would be best done by an original committer | 10:31 |
| laurence | (if that makes sense) | 10:31 |
| juergbi | alexandrufazakas: jinja2.exceptions.TemplateNotFound: COMITTERS.rst.j2 | 10:36 |
| alexandrufazakas | juergbi I'll have a look :/ I've used it multiple time myself before committing it | 10:37 |
| juergbi | alexandrufazakas: the file in git is called COMMITTERS.rst.j2 and the script looks for COMITTERS.rst.j2 (only one M) | 10:38 |
| * alexandrufazakas sighs | 10:38 | |
| juergbi | let me fix that | 10:38 |
| alexandrufazakas | yeah, I've just grepped for it | 10:39 |
| alexandrufazakas | Sorry about that :/ | 10:39 |
| juergbi | now it doesn't fail but it also doesn't seem to generate the file | 10:40 |
| juergbi | ah, no, it did generate it inside contrib, was expecting it at the top level | 10:40 |
| alexandrufazakas | Oh, yeah, it's in the same directory. It might've been a good idea to generate it at the root of the repository | 10:41 |
| juergbi | the link in CONTRIBUTING.rst also expects it at the root, afaict | 10:41 |
| gitlab-br-bot | juergbi opened MR !1510 (update_committers->master: Update COMMITTERS.rst file) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1510 | 10:43 |
| juergbi | the branch creation didn't quite work, it opened an empty MR | 10:45 |
| juergbi | but I've done it manually now | 10:45 |
| alexandrufazakas | Huh, the MR was created, but there's nothing? | 10:45 |
| alexandrufazakas | Ah | 10:45 |
| alexandrufazakas | Okay | 10:45 |
| alexandrufazakas | Yeah, it's not set up to work if there's no previous COMMITTERS.rst file in the repository | 10:45 |
| juergbi | it also doesn't create a branch on its own | 10:45 |
| juergbi | i.e., expects you already be on update_committers branch, afaict | 10:46 |
| juergbi | it should either create the branch or maybe simply check the current branch and report error if it's not the right branch | 10:46 |
| alexandrufazakas | I think I've tested it with no previous update_committers branch | 10:46 |
| alexandrufazakas | And it created it when pushing | 10:46 |
| juergbi | ah, it's supposed to create it only on the remote, not locally | 10:48 |
| juergbi | so it might indeed have been an issue just because it was the first time | 10:49 |
| alexandrufazakas | juergbi: yes | 10:50 |
| alexandrufazakas | So it's no need to checkout a new branch, clean up afterwards etc | 10:50 |
| *** qinusty has joined #buildstream | 11:11 | |
| *** qinusty is now known as qinusty_ | 11:13 | |
| *** qinusty_ is now known as qinusty | 11:54 | |
| tpollard | what would be the quickest way to check interactive failure handling? | 11:58 |
| persia | tpollard: How do you mean? Are you looking for something more complex than setting up a configuration for a known failure (e.g. a build element that calls `false` as the compilation command) and consuming that interactively? | 12:05 |
| tpollard | that would probably be fine, I realised when getting close to merging something that I'd have broke the handling | 12:05 |
| tpollard | but obviously the tests don't core it | 12:05 |
| tpollard | *cover | 12:05 |
| persia | If you need automation, expect(1) is a reasonable way to automate interactivity (remember to set your terminal parameters properly so that bst is being interactive). | 12:05 |
| persia | My general recommendation in such situations is to add a test that handles the case you've found, and then have the change you want to land (which no longer breaks the test) depend on the change with the new test. | 12:06 |
| tpollard | We can't add tests for it sadly | 12:07 |
| persia | But I don't know what needs to be set to make bst behave interactively in a test scenario: I know it checks a few things to try to be interactive for real users and non-interactive when launched from a script. | 12:07 |
| persia | Why not? | 12:07 |
| tpollard | But I think the easiest way will be to put nonsense in a working element and force it to break when I run it | 12:07 |
| tpollard | we can't test the interactive shell | 12:07 |
| persia | Why not? | 12:07 |
| tpollard | We do test that things fail though | 12:08 |
| tpollard | the test framework explicitly runs buildstream in non interactive | 12:08 |
| persia | Yes. You'd need to add any interactive tests to a separate CI job. | 12:08 |
| *** hergertme has quit IRC | 12:19 | |
| gitlab-br-bot | jennis opened (was WIP) MR !1403 (jennis/junctions_and_remotes->master: Revert default junction behaviour and add new junction config options) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1403 | 12:51 |
| jennis | To anyone interested, !1403 (changing how junction Elements interact with parent remotes) is now ready for review | 12:54 |
| jennis | :) | 12:54 |
| gitlab-br-bot | tpollard opened (was WIP) MR !1500 (tpollard/messageobject->master: Message rework) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1500 | 13:49 |
| gitlab-br-bot | tpollard approved MR !1510 (update_committers->master: Update COMMITTERS.rst file) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1510 | 14:07 |
| gitlab-br-bot | tpollard approved MR !1403 (jennis/junctions_and_remotes->master: Revert default junction behaviour and add new junction config options) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1403 | 14:31 |
| *** bochecha_ has joined #buildstream | 15:32 | |
| *** bochecha has quit IRC | 15:34 | |
| *** bochecha_ is now known as bochecha | 15:34 | |
| *** tristan has joined #buildstream | 15:34 | |
| *** bochecha_ has joined #buildstream | 15:51 | |
| *** bochecha has quit IRC | 15:54 | |
| *** bochecha_ is now known as bochecha | 15:54 | |
| *** jonathanmaw_ has quit IRC | 16:22 | |
| tpollard | thanks for the review comments benschubert | 16:31 |
| benschubert | tpollard: no problem, I'll give it a more thorough one tomorrow :) | 16:33 |
| tpollard | getting every corner case to work has been... fun | 16:33 |
| benschubert | I can imagine | 16:34 |
| *** bochecha has quit IRC | 16:48 | |
| *** hergertme has joined #buildstream | 17:20 | |
| *** phoenix has joined #buildstream | 18:33 | |
| *** phoenix has quit IRC | 18:54 | |
| *** cs-shadow has quit IRC | 20:03 | |
| *** traveltissues has joined #buildstream | 22:48 | |
| *** traveltissues has quit IRC | 22:52 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!