Friday, 2022-04-01

@clarkb:matrix.orgvlotorev: it is an optimization to ensure that every project has a check and gate pipeline I think00:01
@clarkb:matrix.orgi think if you drop that and explicitly add check and gate pipelines to all your projects it would work. I'm not sure yet if this is a bug in the regex project name handling00:02
@clarkb:matrix.orgI think it is a bug and we should swap the order of the if and else here https://opendev.org/zuul/zuul/src/branch/master/zuul/model.py#L7480-L749300:05
@clarkb:matrix.orgbasically we want to check if we can match with canonical names first since they are safest. Then if we don't match with canonical names (because the regex doesn't start with a hostname) try the loser match and check for conflicts00:06
@clarkb:matrix.orgI'll push that up00:06
@clarkb:matrix.orglooking at the code more closely it is possible that this condition is expected to be an error but I am not sure yet00:14
@clarkb:matrix.orgI'll work up what I think the fix is then people can chime in on gerrit00:14
@vlotorev:matrix.orgThanks. I'll just remove the .*.00:16
@vlotorev:matrix.org> <@clarkb:matrix.org> vlotorev: it is an optimization to ensure that every project has a check and gate pipeline I think00:18
What's be the benefit for enqueuing the change into check pipeline without jobs? Is it to check mergebility (when Zuul reports like 'failed to merge one of it's cross-repo depencency')?
@vlotorev:matrix.org> <@clarkb:matrix.org> vlotorev: it is an optimization to ensure that every project has a check and gate pipeline I think00:19
* What's the benefit for enqueuing the change into check pipeline without jobs? Is it to check mergibility (when Zuul reports like 'failed to merge one of it's cross-repo depencency')?
@clarkb:matrix.orgyes00:19
@clarkb:matrix.orgzuul won't try to merge anything if it doesn't need to enqueue the change iirc00:20
-@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 836108: Handle greedy project name regexes https://review.opendev.org/c/zuul/zuul/+/83610800:23
@vlotorev:matrix.org * What's the benefit for enqueuing the change into check pipeline without jobs? Is it to check mergeability (when Zuul reports like 'failed to merge one of it's cross-repo depencency')?00:34
-@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 834857: Fix bug in getting changed files https://review.opendev.org/c/zuul/zuul/+/83485705:29
-@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 834857: Fix bug in getting changed files https://review.opendev.org/c/zuul/zuul/+/83485705:31
-@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 834857: WIP: Fix bug in getting changed files https://review.opendev.org/c/zuul/zuul/+/83485706:58
@ekapoun1:matrix.org> <@ekapoun1:matrix.org> Hello, regarding circular dependencies - do they work across repos, or only within the same repo? We're testing the consequences of enabling it, but although we have a shared queue set up between two different repos, where both queue definitions say circular dependencies are allowed, Zuul says a dependency cycle is detected and gives a -1. Is there a tenant configuration to enable/disable circular dependencies as well?06:58
Just to reiterate, we were expecting it to work across repos, since circular dependencies within the same repo doesn't make much sense (you could just squash the commits into one).
-@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 834857: WIP: Fix bug in getting changed files https://review.opendev.org/c/zuul/zuul/+/83485707:25
-@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 834857: WIP: Fix bug in getting changed files https://review.opendev.org/c/zuul/zuul/+/83485707:47
-@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 834857: Fix bug in getting changed files https://review.opendev.org/c/zuul/zuul/+/83485708:13
-@gerrit:opendev.org- Dong Zhang marked as active: [zuul/zuul] 834857: Fix bug in getting changed files https://review.opendev.org/c/zuul/zuul/+/83485710:39
@dong.zhang:matrix.org> <@ekapoun1:matrix.org> Hello, regarding circular dependencies - do they work across repos, or only within the same repo? We're testing the consequences of enabling it, but although we have a shared queue set up between two different repos, where both queue definitions say circular dependencies are allowed, Zuul says a dependency cycle is detected and gives a -1. Is there a tenant configuration to enable/disable circular dependencies as well?12:17
The config is not in tenant but in queue, you need to config "allow-circular-dependencies: true" for the queue. See https://zuul-ci.org/docs/zuul/latest/config/queue.html#attr-queue.allow-circular-dependencies
@tobias.henkel:matrix.orgekapoun1: circular dependencies work cross repos but the repos must share the same queue13:10
@tobias.henkel:matrix.org * ekapoun1: circular dependencies work across repos but the repos must share the same queue13:10
-@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 835452: Test zuul-client dequeue-all https://review.opendev.org/c/zuul/zuul/+/83545214:43
-@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 835452: Test zuul-client dequeue-all https://review.opendev.org/c/zuul/zuul/+/83545215:17
-@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 835452: Test zuul-client dequeue-all https://review.opendev.org/c/zuul/zuul/+/83545215:53
@clarkb:matrix.orgHere is an April fools' joke this channel can probably get behind https://www.rfc-editor.org/rfc/rfc9225.html19:47
-@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 836108: Handle greedy project name regexes https://review.opendev.org/c/zuul/zuul/+/83610819:59
@clarkb:matrix.orgcorvus: ^ I think the release note I wrote there may be a bit long but not sure how to explain it with fewer words. Open to ideas if we think that is a problem. ++ to getting other reviewres like tobiash to take a look too20:00
@jim:acmegating.comwell, tobiash specifically because he wrote it :)20:03
@tobias.henkel:matrix.orgwhat did I write?20:03
@clarkb:matrix.orgthe project regex name handling20:04
@tobias.henkel:matrix.orgYeah, I remember20:04
@clarkb:matrix.orgvlotorev discovered an interesting behavior with it yesterday which is the ^.*$ with colliding short names is always an error. The above change tries to make it possible to deal with that but does introduce a change of behavior (a subtle one)20:05
@tobias.henkel:matrix.orgI need to wrap my head around why we even need to error out on ambigious matches20:13
@jim:acmegating.comyeah, i've been trying to come up with scenarios, and so far i can't think of a case where the change would be detrimental; nor can i think of a case where even allowing same-host duplicates would be detrimental.  maybe it was just the principle of least surprise20:33
@jpew:matrix.orgHmm, changing your keystore password is a good way to crash your whole zuul install :(21:21

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!