EXP: make oldestdeps CI reproducible#19000
Conversation
|
Thank you for your contribution to Astropy! 🌌 This checklist is meant to remind the package maintainers who will review this pull request of some common things to look for.
|
oldestdeps\ CI reproducibleoldestdeps CI reproducible
b3e0e56 to
2cab93e
Compare
76971f9 to
7d5c792
Compare
7d5c792 to
84db076
Compare
mhvk
left a comment
There was a problem hiding this comment.
Super! This looks good to me, except that I think for our future selves it would be helpful to ensure the reason for things are explained in comments (see in-line).
84db076 to
0734e49
Compare
|
p.s. The one advantage of |
I've said this already but I'll repeat as many times as needed: this is only true if the global state of the package index (PyPI) doesn't change between your CI run and your local run, and if you're running on the same (or comparable) platform. |
|
In practice, it works reasonably well to reproduce test failures. Anyway, for another time. I'm sure I'm not the only one who basically copies the astropy/openastronomy setup, and then whenever things get too far out of whack, just makes updates that astropy did. Getting rid of |
|
There is also https://github.com/tox-dev/tox-uv to enable lock file support. |
IIUC the one thing we get from using it is ease to migrate between different CI systems, and considering the direction GitHub has been taking I wouldn't want to bet that another migration is implausible in the foreseeable future, so I'm definitely not proposing we "detox" the repo now of all times. Any way, I'm just blowing off accumulated steam, don't mind me !
Hum, we've been using tox-uv for a over a year now (and this PR is actually all about updating a parameter from tox-uv) so let's say I'm aware 😅 |
| # on oldestdeps, we let warnings be warnings (-Wdefault) instead of treated-as-errors, | ||
| # so we don't need to deal with filtering anything triggered from dependencies | ||
| # Other options replicate the static configuration from pyproject.toml, since they | ||
| # would otherwise be shaddowed by the env variable. |
There was a problem hiding this comment.
Ooops belated comment on a typo:
shaddowed -> shadowed
There was a problem hiding this comment.
thanks ! I can smuggle a typofix into my next PR updating this file, which I plan to open this weekend :)
There was a problem hiding this comment.
nevermind I dropped this plan (no actual need for a PR). I'll just fix this one typo but will skip CI
There was a problem hiding this comment.
hum actually if I skip CI the PR will just be unmergeable...
There was a problem hiding this comment.
Let's just let it be fixed by the next person who looks at it -- it is not like the text is incomprehensible!
Description
This includes #18955 and adds minimal changes to finalize #18782, including a workaround for #18992 (which I don't intend to get merged). This is meant as a demonstration of where the effort is at, and illustrates what's still missing to resolve #18782.