-
-
Notifications
You must be signed in to change notification settings - Fork 15.3k
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
[WIP] pythonPackages.cvxopt: fix transient test failure #42918
Conversation
Another try: |
No attempt on x86_64-darwin (full log) The following builds were skipped because they don't evaluate on x86_64-darwin: sage.lib Partial log (click to expand)
|
No attempt on x86_64-darwin (full log) The following builds were skipped because they don't evaluate on x86_64-darwin: sage Partial log (click to expand)
|
No attempt on aarch64-linux (full log) The following builds were skipped because they don't evaluate on aarch64-linux: sage Partial log (click to expand)
|
Success on x86_64-linux (full log) Attempted: sage.lib Partial log (click to expand)
|
Failure on x86_64-linux (full log) Attempted: sage Partial log (click to expand)
|
Good to know Apparently the cause for the error made no sense because it was not actually the cause of the error (see sage ofBorg log). It did fix it locally for me though, so probably some transient error. I'll have to investigate. |
Failure on aarch64-linux (full log) Attempted: sage.lib Partial log (click to expand)
|
Quick update: This is a transient test failure of cvxopt and apparently was introduced by (my) commit 156e910, "pythonPackages.cvxopt: 1.1.7 -> 1.1.9" in which I also added glpk, gsl and fftw support. One of those is probably responsible, maybe cvxopt/cvxopt#120 is relevant (I basically only read the title). I don't have any time to continue right now. So in case anybody else feels like it, go ahead. |
This took way longer to track down than I'd like to admit. Turns out using openblas for lapack causes the error. Reported upstream here: cvxopt/cvxopt#122 Depending on how fast (and if) upstream fixes the issue I'll backport the fix or switch to liblapack. I'd rather backport the switch though, because IIRC liblapack depends on atlas which takes forever to build. |
It looks like it might take a while to fix upstream. Switching to liblapack for now. @GrahamcOfBorg build python2.pkgs.cvxopt |
Failure on aarch64-linux (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Failure on aarch64-linux (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
Success on x86_64-linux (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Success on x86_64-linux (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
See cvxopt/cvxopt#122 for more info.
@GrahamcOfBorg build python2.pkgs.cvxopt |
Success on x86_64-linux (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
Oh damn forgot to push first. |
Failure on aarch64-linux (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Failure on aarch64-linux (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
Success on x86_64-linux (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Success on x86_64-linux (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
Failure on aarch64-linux (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Failure on aarch64-linux (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
Success on x86_64-linux (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Looks like that doesn't actually fix the problem... Edit: The test error is very enlightening
|
Failure on x86_64-darwin (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
Failure on x86_64-darwin (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Failure on x86_64-darwin (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
Failure on x86_64-darwin (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Failure on x86_64-darwin (full log) Attempted: python2.pkgs.cvxopt Partial log (click to expand)
|
Failure on x86_64-darwin (full log) Attempted: python3.pkgs.cvxopt Partial log (click to expand)
|
Turns out neither sage nor cvxopt are at fault but openblas. #43234 |
Motivation for this change
For some reason after the openblas 3.0.1 update
(c8ed322), cvxopt fails to build with
the pinned glpk. Removing that dependency override fixes that. Its not
very pretty because now sage depends on 2 different glpk versions. But
it should be enough until the next glpk update, which will make the glpk
pinning unnecessary (https://trac.sagemath.org/ticket/24824).
I have really no clue how a minor version update in openblas made the build (test) of cvxopt fail, but who understands computers anymore anyhow...
Things done
sandbox
innix.conf
on non-NixOS)nix-shell -p nox --run "nox-review wip"
./result/bin/
)nix path-info -S
before and after)