| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 893180: Connection health: add gerrit info https://review.opendev.org/c/zuul/zuul/+/893180 | 09:51 | |
| -@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 893180: Connection health: add gerrit info https://review.opendev.org/c/zuul/zuul/+/893180 | 13:19 | |
| -@gerrit:opendev.org- Benjamin Schanzel proposed: [zuul/nodepool] 893375: Kubernetes driver: allow setting dynamic k8s labels https://review.opendev.org/c/zuul/nodepool/+/893375 | 13:57 | |
| @mhuin:matrix.org | Anybody got an idea about failing zuul-nox-docs tests since today? this change https://review.opendev.org/c/zuul/zuul/+/892985 failed in gate because of it. Seems to affect master as well | 14:58 |
|---|---|---|
| @fungicide:matrix.org | `autodoc: failed to import class 'web.LogStreamHandler' from module 'zuul'; the following exception was raised: PyO3 modules may only be initialized once per interpreter process` | 15:12 |
| @clarkb:matrix.org | I suspect ^ needs to be fixed first, but afterwards if we land https://review.opendev.org/c/zuul/zuul/+/892696 and https://review.opendev.org/c/zuul/nodepool/+/892697 we can get zuul and nodepool running on bookworm in OpenDev over the weekend (which then serves as a form of beta test for everyone downstream) | 15:30 |
| @clarkb:matrix.org | https://pyo3.rs/main/migration.html seems related to the pyo3 thing | 15:31 |
| @clarkb:matrix.org | did something add a new rust dependency maybe? | 15:33 |
| @clarkb:matrix.org | https://pyo3.rs/main/migration.html#from-016-to-017 is the specific migration that raises that error and that is older | 15:33 |
| @jim:acmegating.com | the crypto stuff was rusting, right? | 15:33 |
| @clarkb:matrix.org | yes, but it has for a very long time I think. But maybe they changed their linkage library | 15:34 |
| @clarkb:matrix.org | LogStreamHandler is a subclass of ws4py.websocket.WebSocket. I bet that is where the rust linkage is | 15:37 |
| @clarkb:matrix.org | hrm maybe not ws4py hasn't had a release in quite some time | 15:38 |
| @fungicide:matrix.org | setuptools-rust latest release was tuesday of last week (2023-08-22) | 15:39 |
| @fungicide:matrix.org | but zuul-nox-docs only broke between 13:38z (last success) and 23:03z (first failure) yesterday | 15:41 |
| @fungicide:matrix.org | so roughly a 9.5-hour window there | 15:41 |
| @clarkb:matrix.org | can compare package differences between those | 15:42 |
| @fungicide:matrix.org | most recent pyo3 release was a month ago too: https://github.com/PyO3/pyo3/tags | 15:43 |
| @clarkb:matrix.org | I wonder if dropping eithe rthe pyclass or autodoc call against the class will workaround it. Currently trying ot test that theory locally | 15:45 |
| @clarkb:matrix.org | I want to reproduce locally first though | 15:45 |
| @jim:acmegating.com | sphinx 7.2.4 -> 7.2.5 is the culprit | 15:48 |
| @jim:acmegating.com | #11645: Fix a regression preventing autodoc from importing modules within packages that make use of if typing.TYPE_CHECKING: to guard circular imports needed by type checkers. Patch by Matt Wozniski. | 15:49 |
| @jim:acmegating.com | looks suspciously related to the py03 thing | 15:49 |
| @clarkb:matrix.org | if I remove the logstreamhandler autodoc the error moves to zuulweb. So cherrypy must be doing rust things? | 15:49 |
| @jim:acmegating.com | Clark: https://paste.opendev.org/show/bffjJPhzXpkbth4pKurA/ | 15:50 |
| @fungicide:matrix.org | in the initial pip install, the other thing i see changing is virtualenv 20.24.3->20.24.4 | 15:50 |
| @clarkb:matrix.org | corvus: oh how do you get the full list of warnings? Because ya I'm finding each one I remove leads to another popping up | 15:51 |
| @clarkb:matrix.org | that is unlikely to be a viable short term workaround then | 15:51 |
| @jim:acmegating.com | `sphinx-build -E -d doc/build/doctrees -b html doc/source/ doc/build/html` | 15:51 |
| @jim:acmegating.com | https://github.com/sphinx-doc/sphinx/issues/11662 | 15:53 |
| @fungicide:matrix.org | smoking gun | 15:53 |
| @jim:acmegating.com | based on the state of discussion in that bug, i think we should pin sphinx with a link to it. i'll do that. | 15:54 |
| @fungicide:matrix.org | sgtm, thanks! | 15:54 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 893411: Avoid sphinx 7.2.5 https://review.opendev.org/c/zuul/zuul/+/893411 | 15:56 | |
| @jim:acmegating.com | did that as a != in the hope they'll magically fix it and release soon :) | 15:56 |
| @clarkb:matrix.org | I'm surprised that removing a module from sys.modules is ever sufficient to actually unload it... | 15:57 |
| @jim:acmegating.com | yeah... like in the python dictionary sense, yes... in the sense of having run machine code... not so much, eh? :) | 15:58 |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 893411: Avoid sphinx 7.2.5 https://review.opendev.org/c/zuul/zuul/+/893411 | 17:01 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 892985: Add failure-output option to zuul-nox-py311 job https://review.opendev.org/c/zuul/zuul/+/892985 | 17:20 | |
| @jpew:matrix.org | Would someone mind a peek at https://review.opendev.org/q/topic:decline-from-quota ? | 18:40 |
| @jpew:matrix.org | Also we found https://review.opendev.org/c/zuul/nodepool/+/892805 useful for debugging purposes | 18:41 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!