-
-
Notifications
You must be signed in to change notification settings - Fork 14k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
python3.pkgs.scipy: 1.10.1 -> 1.11.1 & more #239969
Conversation
It compiles, and I even added blas and lapack overriding support as requested at #206866 . Currently I test cross compilation of it using:
|
Tested both |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great work getting this going!
I'm of the opinion we should not be overriding the build phase. Adding pipBuildFlags
is a small change now. The downside of course is the mass rebuild making it a bit time consuming to test.
The |
Now using @FRidh or @SomeoneSerge if you'd approve the changes, I can squash some of the commits and make the history cleaner as well. |
Diff: scipy/scipy@v1.10.1...v1.11.1 Significant changes to expression: - Use `pypaBuildHook` to build it. - Use `fetchFromGitHub` instead of `fetchPypi`. - Support cross compilation using `pypaBuildHook` and `$mesonFlags`. - Add a `passthru.updateScript` that downloads and updates hashes for datasets needed for tests. - Cleanup: * Remove unneeded anymore `setupPyBuildFlags`. * Remove unnecessary `sed` command from preConfigure that fixed a Numpy issue in `setup.py` * Use `blas` and `lapack` directly, relying on them being overriden by the user in all of Python's packages' set - see commit 213d084 .
I rebased the git history a bit, with almost a single commit per package. If there are no objections, with your permissions I'd like to merge this today / tomorrow. |
I would like to cherry-pick this into python-updates unless there are review comments left we need to consider. |
Thanks, cherry-picked into #244135 |
This broke scipy on x86_64-darwin. PTAL. cc #245935 https://hydra.nixos.org/build/229710588 |
@mweinelt these links don't seem related to scipy, putting here a link that seems more relevant to me: https://hydra.nixos.org/eval/1797964?filter=scipy&compare=1797817&full=#tabs-now-fail |
A fix is possible at #246152 , will need to wait for ofborg to check it... |
Please recheck the build steps for the linked builds. They are very much related. https://hydra.nixos.org/log/cwpjczjxjav1kyq5p1ilav93rkl0gd3a-python3.10-scipy-1.11.1.drv |
I see. I'm still somewhat confused about the logs there (why is this aarch64 log not failing like the others?). But in any case, in #246152 as well, the build fails like you linked, and I have no idea how to fix that issue :/. It's also hard to debug this without a real darwin machine. Maybe contact upstream? |
I'm already paying attention here.
It looks like the problem is that
If you upgrade |
Description of changes
Update some dependencies of scipy, and added:
Things done
sandbox = true
set innix.conf
? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD"
. Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/
)