Repository navigation
CI: enable python dependencies caching - #18835
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.
|
|
Welcome to Astropy 👋 and congratulations on your first pull A project member will respond to you as soon as possible; in the If you feel that this pull request has not been responded to in a |
|
👋 Thank you for your draft pull request! Do you know that you can use |
|
As I suspected, this doesn't seem to work: with jobs competing for a single cache key, only one of them is able to save a cache, which in turn will be essentially useless on re-runs since the bulk of resources (mostly bandwidth) we're trying to save are taken up by platform-specific binaries (extensions from numpy, scipy ...), so, I'm not sure that we can cache anything efficiently with the current interface from OpenAstronomy/github-actions-workflows I'll report this upstream |
|
reported as OpenAstronomy/github-actions-workflows#324 |
3bffcc2 to
56c2f9e
Compare
de6d6e8 to
b02a091
Compare
|
Maybe possible: OpenAstronomy/github-actions-workflows#324 (comment) |
b02a091 to
e77ff5c
Compare
|
Indeed, and apparently I already did the work ! I just forgot to check the logs and undraft. I did now ! |
| macos: py312-test-alldeps | ||
| posargs: --durations=50 --run-slow | ||
| runs-on: macos-latest | ||
| posargs: --durations=50 |
There was a problem hiding this comment.
oh, this wasn't intentional. Thanks for ca(t)ching it !
|
I don't think the test failure caused by a failed file download is related, but with dependency caching it's hard to tell. |
|
Note that I'm intentionally not enabling cache for unstable CI, so we can trust the failure isn't related. |
e77ff5c to
71c7cd0
Compare
|
I want a couple of other devs to review as well and requested them. If you don't hear back in a week, feel free to ping me and then maybe we can just merge at that point. Thanks. |
|
Let's merge this, it seems to be working. If it all goes to hell the revert button is not far away. |
Description
context: https://pyfound.blogspot.com/2025/10/open-infrastructure-is-not-free-pypi.html
Caching is already supported by OpenAstronomy's reusable workflows, although I'm not 100% sure it'll work out of the box, because cache keys can only be supplied before the test matrix is even output, so it doesn't seem possible to set pet-job cache keys. Any way, let's give this a spin and see how it behaves in practice !
TODO: