Conversation
acdanila
changed the title
ccuetom
reviewed
Aug 14, 2026
ccuetom
left a comment
ccuetom
left a comment
Contributor
There was a problem hiding this comment.
Looks good.
Please, add the testing call to the GitHub actions so they are triggered when needed. If testing dolphinx is required, then that should also be installed in the actions (maybe through the optional requirements or some other requirements file?)
Author
|
@ccuetom, dealing with the envs and adding dolfinx will require a bit more thought, I don't have a clear solution as of now, but I'm sure a clean one exists. dolfinx and some other packages need to be included into the environment.yml file to be used by conda. for now, I've added a pytest gh action that doesn't use dolfinx, so 3 of the tests are skipped |
…bility and env.yaml files for dolfinx builds
ccuetom
reviewed
Sep 4, 2026
added facet_tag addressibility
The two-step install left a broken environment on a fresh Linux box in three separate ways, each silent until something much later failed. The MPI implementation has to be pinned before the base environment goes in. On Linux the base pulls mpich through hdf5, and the overlay's openmpi pin then cannot be satisfied at all; on macOS the base pulls no MPI, which is why this never showed up here. The PETSc scalar type does not survive a later solve. A build-string spec that lives only in a yml is forgotten by the next one, and an ordinary "conda install python=3.11" is enough to swap the complex build for the real one, since conda-forge builds real by default and nothing else asks otherwise. That produced an environment named -complex that reported float64. h5py has to be named in the overlay even though the base already has it. Pinning fenics-dolfinx to 0.9.0 holds hdf5 at 1.14, and an h5py built against a newer one fails on import with "Not a datatype". A package the overlay does not name is one the solver leaves alone. conda-meta/pinned is honoured by every later solve, so all three go there, written before any package is installed. scripts/create_fem_env.sh does the whole thing in one pass and checks the Python version, both scalar types, the hdf5 pairing and the MPI implementation before installing anything on top, since each of those has now failed silently at least once. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
sbatch copies the script into /var/spool before running it, so BASH_SOURCE points at the copy and the checkout cannot be found from it. Fall back to the submit directory, and name the fix in the error when neither works. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
From a directory that merely contains one named stride, import stride resolves to that directory as a namespace package rather than to the installed one, and the suite dies with "cannot import name Geometry from stride", location unknown. Which directory sbatch happens to run in should not decide that. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.