| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 00:17 | |
| - [zuul/zuul-registry] 904533: Replace rehash with sha256 https://review.opendev.org/c/zuul/zuul-registry/+/904533 | ||
| - [zuul/zuul-registry] 904534: Add logging to filesystem driver https://review.opendev.org/c/zuul/zuul-registry/+/904534 | ||
| - [zuul/zuul-registry] 904535: Don't upload empty final chunks https://review.opendev.org/c/zuul/zuul-registry/+/904535 | ||
| - [zuul/zuul-registry] 904536: Add s3 support https://review.opendev.org/c/zuul/zuul-registry/+/904536 | ||
| -@gerrit:opendev.org- Bernhard Berg proposed wip on behalf of Lukas Kranz: [zuul/zuul-jobs] 887917: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 10:21 | |
| @jim:acmegating.com | fungi: does https://github.com/cloudtools/sha256/pull/5 look right to you? (that wfm in a python3.11-bullseye container) | 15:44 |
|---|---|---|
| @jim:acmegating.com | fungi: oh your latest comment makes me think that because what they currently do is run an sdist and upload that to pypi, the problem is probably just that their c code was generated from a too-old cython. and that probably they (or anyone) could just run an sdist with a newer cython and upload that to fix. my pull request just does that on every install. more future-proof but less static. :| | 15:49 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 15:49 | |
| - [zuul/zuul-registry] 904533: Replace rehash with sha256 https://review.opendev.org/c/zuul/zuul-registry/+/904533 | ||
| - [zuul/zuul-registry] 904534: Add logging to filesystem driver https://review.opendev.org/c/zuul/zuul-registry/+/904534 | ||
| - [zuul/zuul-registry] 904535: Don't upload empty final chunks https://review.opendev.org/c/zuul/zuul-registry/+/904535 | ||
| - [zuul/zuul-registry] 904536: Add s3 support https://review.opendev.org/c/zuul/zuul-registry/+/904536 | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 15:58 | |
| - [zuul/zuul-registry] 904533: Replace rehash with sha256 https://review.opendev.org/c/zuul/zuul-registry/+/904533 | ||
| - [zuul/zuul-registry] 904534: Add logging to filesystem driver https://review.opendev.org/c/zuul/zuul-registry/+/904534 | ||
| - [zuul/zuul-registry] 904535: Don't upload empty final chunks https://review.opendev.org/c/zuul/zuul-registry/+/904535 | ||
| - [zuul/zuul-registry] 904536: Add s3 support https://review.opendev.org/c/zuul/zuul-registry/+/904536 | ||
| @fungicide:matrix.org | > <@jim:acmegating.com> fungi: oh your latest comment makes me think that because what they currently do is run an sdist and upload that to pypi, the problem is probably just that their c code was generated from a too-old cython. and that probably they (or anyone) could just run an sdist with a newer cython and upload that to fix. my pull request just does that on every install. more future-proof but less static. :| | 16:25 |
| yep, sorry, i completely missed that the sdist on pypi contained the results of the cythonize step, so was just done with a too-old version of cython. latest pr lgtm | ||
| @fungicide:matrix.org | also the latest test results on your change seem to bear that out | 16:26 |
| @jim:acmegating.com | fungi: https://zuul.opendev.org/t/zuul/build/18b1961942fb4adeb1447b0af63a0c62/log/job-output.txt#2251 | 16:31 |
| @jim:acmegating.com | fungi: the way i read that is that our magic wheel caching doesn't work for git+https entries in requirements files... | 16:32 |
| @jim:acmegating.com | like... it built the wheel in the build phase, but then wanted to build it again during install (and we don't have git in the install image) | 16:32 |
| @jim:acmegating.com | i guess easy fix is to just install git there too for now... but it'd be cool if that worked with the assemble script/two phase build | 16:33 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 16:35 | |
| - [zuul/zuul-registry] 904533: Replace rehash with sha256 https://review.opendev.org/c/zuul/zuul-registry/+/904533 | ||
| - [zuul/zuul-registry] 904534: Add logging to filesystem driver https://review.opendev.org/c/zuul/zuul-registry/+/904534 | ||
| - [zuul/zuul-registry] 904535: Don't upload empty final chunks https://review.opendev.org/c/zuul/zuul-registry/+/904535 | ||
| - [zuul/zuul-registry] 904536: Add s3 support https://review.opendev.org/c/zuul/zuul-registry/+/904536 | ||
| @fungicide:matrix.org | is that a separate build layer that gets discarded>? | 16:35 |
| @jim:acmegating.com | yep | 16:35 |
| @fungicide:matrix.org | * is that a separate build layer that gets discarded? | 16:35 |
| @jim:acmegating.com | we do wheel builds, then copy over the cache layer, then install them again | 16:35 |
| @jim:acmegating.com | i'm guessing that pip can't be sure that a pointer to a git repo is still the same version... | 16:36 |
| @fungicide:matrix.org | but yeah, i suppose the alternative would be to explicitly run wheel and add the result to the cache, then `sed` the requirements file to match or something? | 16:36 |
| @jim:acmegating.com | (maybe if we used the @hash approach, it would?) | 16:36 |
| @fungicide:matrix.org | seems messy | 16:36 |
| @jim:acmegating.com | i mean, that's how we've built every python image for years :) | 16:37 |
| @jim:acmegating.com | so that we don't need to install build deps in the final images | 16:37 |
| @fungicide:matrix.org | oh, right i mean inline editing the requirements file to match seems messy | 16:37 |
| @jim:acmegating.com | oh heh | 16:38 |
| @jim:acmegating.com | i'm going to try the hash approach in parallel, see if that fixes it | 16:38 |
| @fungicide:matrix.org | worth a shot, sure. i'd be inclined to not put a lot of effort into elegantly solving this if it's just until we get a pull request merged/new release of sha256 | 16:39 |
| @jim:acmegating.com | same | 16:39 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul-registry] 904586: DNM test git hash https://review.opendev.org/c/zuul/zuul-registry/+/904586 | 16:42 | |
| -@gerrit:opendev.org- Ahmon Dancy proposed: [zuul/zuul] 899425: gitlabconnection.py: Handle 404 on unapprove https://review.opendev.org/c/zuul/zuul/+/899425 | 16:48 | |
| @fungicide:matrix.org | corvus: unfortunately, the hash notation doesn't seem to have made any difference, it still tried to git clone | 17:17 |
| @jim:acmegating.com | yep :( also, there's more weirdness from our two-phase build | 17:20 |
| @jim:acmegating.com | https://zuul.opendev.org/t/zuul/build/9c017d56bff44d5b83de75398b840d94/log/job-output.txt#2350 | 17:21 |
| @jim:acmegating.com | but on the plus side; there's movement on the PR | 17:21 |
| @jim:acmegating.com | the way i'm reading that error is that cython isn't getting stashed into the wheel cache for the second phase | 17:22 |
| @jim:acmegating.com | that might be a real problem for us | 17:22 |
| @jim:acmegating.com | probably fixable most simply by adding cython as a run-time dependency to zuul-registry | 17:24 |
| @jim:acmegating.com | oh, and if i'm right about that, this may not be necessary once the PR merges too | 17:25 |
| @fungicide:matrix.org | yeah, because we won't be running cython to build the sha256 wheel anyway | 17:26 |
| @fungicide:matrix.org | that happens at sdist generation | 17:26 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 17:26 | |
| - [zuul/zuul-registry] 904533: Replace rehash with sha256 https://review.opendev.org/c/zuul/zuul-registry/+/904533 | ||
| - [zuul/zuul-registry] 904534: Add logging to filesystem driver https://review.opendev.org/c/zuul/zuul-registry/+/904534 | ||
| - [zuul/zuul-registry] 904535: Don't upload empty final chunks https://review.opendev.org/c/zuul/zuul-registry/+/904535 | ||
| - [zuul/zuul-registry] 904536: Add s3 support https://review.opendev.org/c/zuul/zuul-registry/+/904536 | ||
| -@gerrit:opendev.org- Zuul merged on behalf of Ahmon Dancy: [zuul/zuul] 899425: gitlabconnection.py: Handle 404 on unapprove https://review.opendev.org/c/zuul/zuul/+/899425 | 18:06 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 19:01 | |
| - [zuul/zuul-registry] 904533: Replace rehash with sha256 https://review.opendev.org/c/zuul/zuul-registry/+/904533 | ||
| - [zuul/zuul-registry] 904534: Add logging to filesystem driver https://review.opendev.org/c/zuul/zuul-registry/+/904534 | ||
| - [zuul/zuul-registry] 904535: Don't upload empty final chunks https://review.opendev.org/c/zuul/zuul-registry/+/904535 | ||
| - [zuul/zuul-registry] 904536: Add s3 support https://review.opendev.org/c/zuul/zuul-registry/+/904536 | ||
| @clarkb:matrix.org | corvus: maybe also switch the docker image to bookworm instead of bullseye to show that openssl 3.x isnt' a problem with the new lib? | 19:23 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul-registry] 904590: Upgrade to bookworm image https://review.opendev.org/c/zuul/zuul-registry/+/904590 | 19:38 | |
| @jim:acmegating.com | Clark: ^ yep (also i've run it locally so i know it works) | 19:38 |
| @fungicide:matrix.org | entire series lgtm | 19:40 |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: | 20:01 | |
| - [zuul/zuul-registry] 904533: Replace rehash with sha256 https://review.opendev.org/c/zuul/zuul-registry/+/904533 | ||
| - [zuul/zuul-registry] 904534: Add logging to filesystem driver https://review.opendev.org/c/zuul/zuul-registry/+/904534 | ||
| - [zuul/zuul-registry] 904535: Don't upload empty final chunks https://review.opendev.org/c/zuul/zuul-registry/+/904535 | ||
| @fungicide:matrix.org | bookworm image change is passing now too | 20:09 |
| @fungicide:matrix.org | * bookworm image change is passing too | 20:09 |
| @clarkb:matrix.org | I'll take a look at the other two changes after lunch | 20:39 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 21:01 | |
| - [zuul/zuul] 904592: Avoid unecessary join with provides table https://review.opendev.org/c/zuul/zuul/+/904592 | ||
| - [zuul/zuul] 904593: Add an index to speed build time estimation https://review.opendev.org/c/zuul/zuul/+/904593 | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 21:07 | |
| - [zuul/zuul] 899766: Add a build-times web endpoint https://review.opendev.org/c/zuul/zuul/+/899766 | ||
| - [zuul/zuul] 899767: Add runtime graphs to web https://review.opendev.org/c/zuul/zuul/+/899767 | ||
| @clarkb:matrix.org | corvus: left a comment about some s3 behavior I figured you might want to check before approving the change. But I think it is ok to approve it if there isn't a reason to set the isdir flag like we do with swift | 21:08 |
| @jim:acmegating.com | Clark: thx, replied. i'm leaning toward approving and updating later; let me know what you think. | 21:20 |
| @clarkb:matrix.org | ya that makes sense given it is used by the pruning which alsio needs works | 21:21 |
| @clarkb:matrix.org | * ya that makes sense given it is used by the pruning which also needs work | 21:21 |
| @clarkb:matrix.org | I've gone ahead and approved it | 21:28 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 904595: Use getBuildTimes for build time estimator https://review.opendev.org/c/zuul/zuul/+/904595 | 21:30 | |
| @jim:acmegating.com | kk, thx | 21:30 |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul-registry] 904536: Add s3 support https://review.opendev.org/c/zuul/zuul-registry/+/904536 | 21:41 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul-registry] 904590: Upgrade to bookworm image https://review.opendev.org/c/zuul/zuul-registry/+/904590 | 22:13 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!