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
scikits-odes: init at 2.4.0 #64195
scikits-odes: init at 2.4.0 #64195
Conversation
@GrahamcOfBorg build python3.pkgs.scikits-odes |
b9602e3
to
d4b0479
Compare
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.
This works for me on macOS 10.13.6 (17G65).
One test breaks on aarch64:
Any idea what could be the reason for that? |
I have never used the differential algebraic part of SUNDIALS. I think the only thing one can do is report it to the maintainers of scikits-odes. One of the problems I've had in the past is that if a matrix is selected at random and is close to being singular then tests will fail. Maybe it's something like that? |
I could reproduce it multiple times, and only on aarch64, so it might be an arch-specific issue.
|
Please move the numpy patch out. It causes this to be a mass-rebuild. |
I'd still like to eventually see this change land in master. What about dropping it from here, and file that change against staging? |
yes, open a separate PR for it |
87b3c53
to
2a37885
Compare
took the numpy patch out and rebased on latest master. |
@GrahamcOfBorg build python3.pkgs.scikits-odes |
This works
|
The nix derivation uses cmake to generate
@flokli if I run cmake outside of nix, I get a populated directory. How do I retain the equivalent directory in nix? |
I have this
but sadly don't get what I wanted
|
This is a bit better but still fails.
|
CMake can be convinced to use a working RUNPATH during |
liblapack needed to be built with shared library support ( |
@GrahamcOfBorg build python3.pkgs.scikits-odes |
0e679d8
to
65f600c
Compare
@GrahamcOfBorg build python2.pkgs.scikits-odes python3.pkgs.scikits-odes |
Any reason this can't be merged? |
Hm. It would have been nice if we heard back from LLNL/sundials about the patches we sent upstream, but their repo seems to be only code dumps of release snapshots so far, so I don't really see if and how patches are merged in 😢 I did not yet include the patch from LLNL/sundials#21 to remove test flakiness into this PR. It doesn't apply cleanly on I really hope LLNL/sundials is moving development to github soon as indicated on LLNL/sundials#8 (comment) - a sundials 3 branch containing a rebased patch would be really cool - nixpkgs policy is to not include patches into nixpkgs directly, but fetch them if possible… |
That's needed for scikits.odes. Upstream bug to migrate to sundials 4: bmcage/odes#98
This test fails on MacOS otherwise, due to slightly different math. Only add for sundials 4, it's not in sundials_3 yet.
The provided patch doesn't apply cleanly on sundials_3 (no SundialsTesting.cmake, SundialsAddTest.cmake containing trailing whitespaces inside context), so for now we just disable tests for sundials_3 - scikits-odes plans to move to sundials 4 anyways - see bmcage/odes#106.
I included the non-flakyness patches for sundials 4, and disabled tests for I also bumped scikits-odes to |
@GrahamcOfBorg build python.pkgs.scikits-odes python3.pkgs.scikits-odes sundials |
Motivation for this change
This adds
scikits.odes
, a scikit offering extra ode/dae solvers, as an extension to what is available in scipy.It's not yet compatible with sundials 4.x, so we add
sundials_3
, too.Things done
sandbox
innix.conf
on non-NixOS)nix-shell -p nix-review --run "nix-review wip"
./result/bin/
)nix path-info -S
before and after)cc @idontgetoutmuch