Timeouts in conda-based tests

[tobiasdiez]

Steps To Reproduce

Recently, a few tests in CI always fail due to timeout. This came out of nowhere with no apparent changes: no changes to the underlying os image, no changes to the conda packages and no code changes during this time period.

All of these failures seem to be related to cysignals/gap interaction.

I have no idea what changed and might trigger these issues.

Expected Behavior

Pass

Actual Behavior

Fail:

sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/doc/en/thematic_tutorials/group_theory.rst  # Timed out
sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/sage/combinat/designs/incidence_structures.py  # Timed out
sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/sage/graphs/generic_graph.py  # Timed out
sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/sage/groups/finitely_presented_named.py  # Timed out
sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/sage/groups/libgap_wrapper.pyx  # Timed out
sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/sage/groups/perm_gps/permgroup.py  # Timed out
sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/sage/groups/perm_gps/permgroup_morphism.py  # Timed out
sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/sage/topology/simplicial_complex.py  # Timed out
sage -t --warn-long 5.0 --random-seed=223128428498463667388028472981004216973 src/sage/topology/simplicial_complex_examples.py  # Timed out

Additional Information

Good: https://github.com/sagemath/sage/actions/runs/12560678326/job/35018373971
Bad: https://github.com/sagemath/sage/actions/runs/12572431063/job/35044413316?pr=39242

Environment

CI

Checklist

• I have searched the existing issues for a bug report that matches the one I want to file, without success.
• I have read the documentation and troubleshoot guide

[tobiasdiez]

Strange: it works again...

[user202729]

 python3 /.../sage/src/bin/sage-runtests --warn-long 5.0 --random-seed=113808444557570614600856870960997183849 src/sage/topolog
 ├─ python3 /.../sage/src/bin/sage-cleaner
 └─ python3 /.../sage/src/bin/sage-runtests --warn-long 5.0 --random-seed=113808444557570614600856870960997183849 src/sage/topo
    └─ wget --quiet -O - http://www.math.rwth-aachen.de/~mfer/mfertoc_SHA.json

maybe the culprit is unreliable network. Why is it downloading from some server anyway?

[user202729]

https://docs.gap-system.org/pkg/atlasrep/doc/chap4.html#X7C3293A98577EE68

[user202729]

plan:

• wait until the server is back up
• run all the tests locally or on CI to see what gets downloaded
• zip the files that the program download (i.e. those needed to run the tests) and put it somewhere reliable
• make the CI download it in advance

we may want to vendor the files for users as well.

[cxzhong]

I also can not visit the website. Maybe we need to skip this or change a reliable upstream.

[user202729]

also that test-long works because it uses an incremental docker image. I guess if the site is down permanently, we have a copy in the docker image.

[cxzhong]

Now, it is OK again. @dimpase Can you let it be in the distribution of GAP?