| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 825668: Skip flaky disconnect tests https://review.opendev.org/c/zuul/zuul/+/825668 | 00:25 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 825788: Fix update of mutable ZKObject attributes https://review.opendev.org/c/zuul/zuul/+/825788 | 13:12 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 825788: Fix update of mutable ZKObject attributes https://review.opendev.org/c/zuul/zuul/+/825788 | 13:25 | |
| @fungicide:matrix.org | all's quiet in opendevland today... no reports of new problems from the apac/emea crowds, and people in the americas are mostly awake now | 14:48 |
|---|---|---|
| @fungicide:matrix.org | looking good for a 4.12.0 | 14:48 |
| @jim:acmegating.com | fungi: thanks! swest found https://review.opendev.org/825788 overnight... i think that may be worth another iteration... i'm looking into adding tests for that now and maybe an alternate fix | 14:57 |
| @fungicide:matrix.org | ahh, yep thanks for the heads up | 15:25 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 825880: Use hash to de-duplicate ZKObject writes https://review.opendev.org/c/zuul/zuul/+/825880 | 21:04 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 825880: Use hash to de-duplicate ZKObject writes https://review.opendev.org/c/zuul/zuul/+/825880 | 21:04 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 825880: Use hash to de-duplicate ZKObject writes https://review.opendev.org/c/zuul/zuul/+/825880 | 21:06 | |
| @fungicide:matrix.org | looks like some after-dinner reading | 21:07 |
| @jim:acmegating.com | swest tobiash Clark fungi ^ there's an alternate fix with a test.... i think that's worth holding 4.12 for | 21:07 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 824482: Add "zuul delete-pipeline-state" command https://review.opendev.org/c/zuul/zuul/+/824482 | 21:53 | |
| @jim:acmegating.com | the google search results look *a lot* better today (at least, in my google bubble) | 21:57 |
| @clarkb:matrix.org | I can take a look in maybe 45 minutes to an hour? Errand running is taking longer than anticipated | 22:03 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 22:04 | |
| - [zuul/zuul] 825022: Add a model API version https://review.opendev.org/c/zuul/zuul/+/825022 | ||
| - [zuul/zuul] 825023: DNM: Example model upgrade change https://review.opendev.org/c/zuul/zuul/+/825023 | ||
| @jim:acmegating.com | no worries. if it gets approved any time today, i can roll opendev over the weekend to try to get 4.12 out early next week | 22:04 |
| @clarkb:matrix.org | corvus for the bugfix if we are hashing the serialized version how is that different when comparing hashes vs serialized data? Is it that we are looking at two serialized versions that are the same because they are both the live versions in zuul and not what is written to the db? | 23:21 |
| @jim:acmegating.com | Clark: basically the sequence is: 1) create zkobject with mutable attribute and write it to zk. 2) mutate attribute. 3) serialize #1 4) update attribute 5) serialize #2 6) #1 == #2 so don't update zk. | 23:23 |
| @jim:acmegating.com | Clark: so the key fix is that we're moving the base of the comparison from step 3 to step 1. that way we catch the mutation in step 2. | 23:24 |
| @clarkb:matrix.org | gotcha | 23:24 |
| @clarkb:matrix.org | I think that is what I was tryingto ask about with my second question so seems I was on the right track | 23:24 |
| @jim:acmegating.com | Clark: we could have kept that as serializing, but then we'd be story a duplicate of all data; thus the switch to hash | 23:25 |
| @jim:acmegating.com | * Clark: we could have kept that as serializing, but then we'd be storing a duplicate of all data; thus the switch to hash | 23:25 |
| @clarkb:matrix.org | just trying to understand the test then I'll approve it | 23:31 |
| @clarkb:matrix.org | corvus: in the case of the circular dep zuul enqueues them together as a single item? | 23:32 |
| @jim:acmegating.com | Clark: two items, but they're enqueued at the same time. they both have the same changes | 23:36 |
| @jim:acmegating.com | (doing that makes more sense as a cross-repo dependency of course :) | 23:37 |
| @jim:acmegating.com | they also have the same bundle; so the test is making sure we have coverage of the bundle being written correctly; without the bugfix it's missing one of the changes and throws exceptions | 23:39 |
| @clarkb:matrix.org | ok so we start with 2 builds for Z. Then wait until for that to finish up and get out of the way. Then we get 4 builds. Two each for A and B. Why does releasing builds[0] end up being sufficient at the end there for all the other jobs to complete. Wouldn't we also need to release builds[1] which corresponds to job2? | 23:40 |
| @jim:acmegating.com | they abort | 23:40 |
| @clarkb:matrix.org | (and possibly release the builds for the second set of job1 and job2) | 23:40 |
| @clarkb:matrix.org | oh because it is dependency between them | 23:41 |
| @clarkb:matrix.org | and job1 failed | 23:41 |
| @jim:acmegating.com | since build[0] fails yep | 23:41 |
| @clarkb:matrix.org | got it. I'm glad we don't use circular dependencies in opendev :) this makes the brain melt a bit | 23:41 |
| @jim:acmegating.com | yeah i have some brain slag :) | 23:42 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!