| @iwienand:matrix.org | corvus: https://review.opendev.org/c/zuul/zuul-jobs/+/871539 is a role for checking out git submodules. i'm just a bit unsure one this so would like some input. on one hand, it's simple enough to run the submodule pull as this does, and it seems like this is safe from a gating POV as you're pulling the references in the zuul checkout, which should be reproducible. on the other hand, the checkout from prepare-workspaces has the origin reset to /dev/null presumably to avoid pulling remotely in gate, etc. so if your project uses relative submodules (gerrit) you can't find them. | 00:43 |
|---|---|---|
| @iwienand:matrix.org | Michael Kelly has provided notes on this as i requested in the README. still something about it makes me hesitant on it, i feel like i'm missing something. whenever free, could you take a look? | 00:44 |
| @jim:acmegating.com | ianw: using submodules with zuul is very situation dependent. often what people want to do is incompatible with zuul. but there are some ways to use it for a vendored dependency that is not itself managed by zuul (in which case it behaves and breaks like any other dependency mechanism (lockfile, requirements, etc). Michael Kelly's role should work for that. there is also a way to use zuul with submodules that are themselves co-gated. that relies on gerrit's submodule subscription system. mordred and i set up gerrit's zuul for that. i don't have any objection to a role that runs "git submodule init" as long as it doesn't suggest that that is a correct or necessary thing to do in all cases. i do think that change should be updated not to reference the gerrit repos because it is definitely incorrect behavior for that specific submodule configuration. | 00:57 |
| @iwienand:matrix.org | ok, the gerrit thing i probably injected confusion into, because i asked for testing and the only thing we clone that i could really think of that had submodules was gerrit | 00:59 |
| @jim:acmegating.com | i think a synthetic test might be appropriate | 00:59 |
| @iwienand:matrix.org | what happens in that test is pretty much what we do to build the gerrit docker image in opendev | 01:01 |
| @iwienand:matrix.org | but yeah, one issue i had with that was now we need to keep the submodule list in sync in zuul-jobs which seems not right at all | 01:02 |
| @jim:acmegating.com | ianw: yep. that's why i say situation dependent. we're simply consumers of gerrit repos, and we're not gating any of the repos that either are or have submodules. so it works in that case. but it's not okay for gerrit itself. | 01:02 |
| @jim:acmegating.com | like, i don't even know if i could write a succinct list of rules for when a certain kind of submodule strategy is okay or not.. :/ | 01:03 |
| @jim:acmegating.com | 1) do not use submodules | 01:03 |
| @jim:acmegating.com | 2) ...? | 01:03 |
| @iwienand:matrix.org | yeah, i guess that was my issue, i was probably thinking a bit too big with this trying to cover every angle. just working for the simple case of non-relative submodules is probably enough | 01:07 |
| @iwienand:matrix.org | the other changes ontop of that are fairly similar for git-lfs, which is pretty much the same gist i guess | 01:08 |
| -@gerrit:opendev.org- Tobias Henkel proposed: [zuul/nodepool] 876881: Process paused requests first https://review.opendev.org/c/zuul/nodepool/+/876881 | 08:00 | |
| -@gerrit:opendev.org- Tobias Henkel proposed: [zuul/nodepool] 876881: Process paused requests first https://review.opendev.org/c/zuul/nodepool/+/876881 | 08:05 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/nodepool] 876943: Respect timeout when only declining requests https://review.opendev.org/c/zuul/nodepool/+/876943 | 08:10 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/nodepool] 876943: Respect timeout when not accepting requests https://review.opendev.org/c/zuul/nodepool/+/876943 | 08:11 | |
| -@gerrit:opendev.org- Tobias Henkel proposed: [zuul/nodepool] 876881: Process paused requests first https://review.opendev.org/c/zuul/nodepool/+/876881 | 09:14 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/nodepool] 876943: Respect timeout when not accepting requests https://review.opendev.org/c/zuul/nodepool/+/876943 | 09:33 | |
| -@gerrit:opendev.org- Benjamin Schanzel proposed on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/nodepool] 872256: Offload openstack delete api calls to an executor https://review.opendev.org/c/zuul/nodepool/+/872256 | 09:48 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/nodepool] 876943: Respect timeout when not accepting requests https://review.opendev.org/c/zuul/nodepool/+/876943 | 11:36 | |
| -@gerrit:opendev.org- Simon Westphahl proposed on behalf of Tobias Henkel: [zuul/nodepool] 876881: Process paused requests first https://review.opendev.org/c/zuul/nodepool/+/876881 | 11:36 | |
| @vlotorev:matrix.org | Speaking of submodules: I wrote a tool that maps zuul.projects canonical names to urls in .gitmodules and uses 'git submodule absorbgitdirs' if project canonical paths match. | 19:48 |
| It is combined with gerrit super-project subscription. | ||
| (Someone might find the tool useful). | ||
| @jpew:matrix.org | vlotorev: I'd be interested to see that | 21:32 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 21:34 | |
| - [zuul/nodepool] 877044: Don't update node request in assign handlers loop https://review.opendev.org/c/zuul/nodepool/+/877044 | ||
| - [zuul/nodepool] 877045: Make assignHandlers a generator https://review.opendev.org/c/zuul/nodepool/+/877045 | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed on behalf of Simon Westphahl: [zuul/nodepool] 876943: Respect timeout when not accepting requests https://review.opendev.org/c/zuul/nodepool/+/876943 | 21:34 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed on behalf of Tobias Henkel: [zuul/nodepool] 876881: Process paused requests first https://review.opendev.org/c/zuul/nodepool/+/876881 | 21:34 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/nodepool] 870430: Add a cache for image lists https://review.opendev.org/c/zuul/nodepool/+/870430 | 23:11 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 877048: Allow duplicate queue definitions on same project branches https://review.opendev.org/c/zuul/zuul/+/877048 | 23:30 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!