hiddenSymmetries
Spec compilation fails with python 3.13
For the error message take a look at https://github.com/hiddenSymmetries/simsopt/actions/runs/16472784292/job/46566116701
Also with python 3.12. The build system of SPEC needs to be updated and was never adapted to the deprecation of distutils. SPEC still doesn't have a pyproject.toml.
The build system is still basically what you wrote a few years ago. I am getting a little better at fixing this, but I will need some time. Any help or pointers would be much appreciated,.
@smiet, I thought you switched SPEC build system to meson. We need pyoculus also to be updated. I noticed you had a PyOculus fork with meson build system. Can we switch to that?
@mbkumar Update on this seems to be that spec compilation is now failing in the CI with python 3.9 as well (although python 3.10, 3.11 worked?) https://github.com/hiddenSymmetries/simsopt/actions/runs/18782919008/job/53593584365?pr=559
Python 3.9 reached EOL. So we are not going to focus on 3.9. I will start working on the fix for SPEC compilation.
Bharat Medasani
On Mon, Oct 27, 2025 at 11:20 AM Alan Kaptanoglu @.***> wrote:
akaptano left a comment (hiddenSymmetries/simsopt#542) https://github.com/hiddenSymmetries/simsopt/issues/542#issuecomment-3451846668
@mbkumar https://github.com/mbkumar Update on this seems to be that spec compilation is now failing in the CI with python 3.9 as well (although python 3.10, 3.11 worked?) https://github.com/hiddenSymmetries/simsopt/actions/runs/18782919008/job/53593584365?pr=559
— Reply to this email directly, view it on GitHub https://github.com/hiddenSymmetries/simsopt/issues/542#issuecomment-3451846668, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA62VEAEUXJ5HTUEPN55CE33ZYZ5HAVCNFSM6AAAAACCGG3SQOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTINJRHA2DMNRWHA . You are receiving this because you were mentioned.Message ID: @.***>
In that case, mind if I submit a PR that updates the CI to run things without python 3.9?
It's already in the pipeline.
Bharat Medasani
On Tue, Oct 28, 2025, 5:08 PM Alan Kaptanoglu @.***> wrote:
akaptano left a comment (hiddenSymmetries/simsopt#542) https://github.com/hiddenSymmetries/simsopt/issues/542#issuecomment-3458489681
In that case, mind if I submit a PR that updates the CI to run things without python 3.9?
— Reply to this email directly, view it on GitHub https://github.com/hiddenSymmetries/simsopt/issues/542#issuecomment-3458489681, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA62VEDZTSPM2LVIJSXZROL3Z7LOBAVCNFSM6AAAAACCGG3SQOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTINJYGQ4DSNRYGE . You are receiving this because you were mentioned.Message ID: @.***>
I believe the other runners are failing because numpy is no longer installed in the container (or changes to the deprecating setuptools are no longer picking this up as a dependency). I have a branch on SPEC repo (move_to_toml_3 ) that is almost there, but I cannot figure out exactly how to get all the files to copy correctly... I am working with a EuroFusion developer on this, but if you see the solution @mbkumar, this might be a better starting point.
I started a parallel branch in SPEC yesterday trying to solve the same issue, but you guys are far ahead of me. Some of the changes you made to the f2py compilation process are not clear to me. Perhaps we should meet, and you can walk me through the changes.
Bharat Medasani
On Thu, Oct 30, 2025 at 6:17 AM Chris Smiet @.***> wrote:
smiet left a comment (hiddenSymmetries/simsopt#542) https://github.com/hiddenSymmetries/simsopt/issues/542#issuecomment-3467166272
I believe the other runners are failing because numpy is no longer installed in the container (or changes to the deprecating setuptools are no longer picking this up as a dependency). I have a branch on SPEC repo (move_to_toml_3 https://github.com/PrincetonUniversity/SPEC/tree/move_to_toml_3 ) that is almost there, but I cannot figure out exactly how to get all the files to copy correctly... I am working with a EuroFusion developer on this, but if you see the solution @mbkumar https://github.com/mbkumar, this might be a better starting point.
— Reply to this email directly, view it on GitHub https://github.com/hiddenSymmetries/simsopt/issues/542#issuecomment-3467166272, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA62VEBKFW24MS4UW7MF4HL32HQS7AVCNFSM6AAAAACCGG3SQOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTINRXGE3DMMRXGI . You are receiving this because you were mentioned.Message ID: @.***>