Conversation
|
Basic evals checks fail in pytest. Bisected to setuptools 60.9.0 bump, so I stuck with 60.8.2 for now and reported the test failures upstream. pytest-dev/pytest#9688 |
435c548 to
19b5fba
Compare
|
Where did the pip update go? |
|
Force pushed it away when I downgraded from 60.9.0 to 60.8.2, it wasn't really relevant and can be added back again later. Want to get one eval so we see what breaks. |
5788b0e to
0b0f61d
Compare
0b0f61d to
bf5b343
Compare
|
Have built quite a lot of packages with this on macos 10.15 and am now only getting pre-existing failures. |
The library was renamed and this package didn't follow … yet?
Including pytest-xdist without configuring it has no effect...
setuptools 60.x (re-)introduces a bundled distutils, which we need to patch with C++ support in the same way as we do with the stdlib's distutils (mostly to placate clang on macos)
This is the tool that manages our certificate store.
|
The state of things right now is looking pretty good. Will rebase now, to drop the portpicker commit that conflicts. |
|
Thanks @risicle for coming up with that patch. That was a big chunk of the heavy-lifting to get this PR merged! |
|
Bisect claims that 5c09870 |
|
Looks like we're quite behind on scikit-build anyway. |
|
In fact the next "Python updates" includes a bump that may well fix it ada1d70 |
|
I see two different tests failing with that update, maybe we should just disable those. |
|
This pull request has been mentioned on NixOS Discourse. There might be relevant details there: |
This cycle is about creating a cheap baseline evaluation for the expected setuptools breakage to have a comparison against master.
https://hydra.nixos.org/eval/1744066
In the meantime it has been rebased and retargeted onto staging.
This is closely related to #155516 and we will start updating or removing packages that rely on this feature.
Previous python-unstable: #154969