reorganize tests

[ranocha]

TODO:

• On macOS/Windows, only run MPI tests, threaded tests, and exercise important or “fragile” binary dependencies (like P4est)
  • Keep 2d_mpi
  • update 2d_threaded
• Decide what to do with misc
• Distribute parts approximately uniformly (by run time)
• Update required status checks

Closes #673; closes #794

[sloede]

The following consensus was reached at today's Trixi meeting:

• On macOS/Windows, only run MPI tests, threaded tests, and exercise important or “fragile” binary dependencies (like P4est)
• Use only Linux tests for coverage reports to speed up the process of adding new tests
• Try to configure CI to run macOS/Windows tests only on ready-for-review PRs to increase turnover and reduce pressure on concurrent jobs when still in draft phase. This can (maybe) be achieved by looking at the github.event.pull_request.draft == true/false property

[ranocha]

• Try to configure CI to run macOS/Windows tests only on ready-for-review PRs to increase turnover and reduce pressure on concurrent jobs when still in draft phase. This can (maybe) be achieved by looking at the github.event.pull_request.draft == true/false property

Could you please open a new issue with this?

[ranocha]

We need to update the required status checks before merging this.

[sloede]

While going through it, I realize that you removed many RestartCallbacks. Are all mesh-solver combinations still checked for restarting properly? And what was your motivation (just curious)?

[ranocha]

While going through it, I realize that you removed many RestartCallbacks. Are all mesh-solver combinations still checked for restarting properly?

They should be - I tried to include at least one example using a restart callback.

And what was your motivation (just curious)?

The restart callbacks

• are just annoying locally in many elixirs since they produce unnecessary IO
• take CI time that we can save easily

[ranocha]

We might think about merging threaded and mpi jobs on each OS

[sloede]

Overall LGTM, with one proposal for more consistency (of course 😎) and the request to not remove the LSA ICs/BCs, plus a few minor questions/remarks.

[sloede]

One thing I forgot: I think it would be good to adapt the relevant section in the docs as well. That is, fix the obsolete file name and maybe add a sentence or two about where to add new tests, such that everything remains roughly balanced?

[ranocha]

One thing I forgot: I think it would be good to adapt the relevant section in the docs as well. That is, fix the obsolete file name and maybe add a sentence or two about where to add new tests, such that everything remains roughly balanced?

Good point - I'll do that

[sloede]

Thanks a lot for taking care of this mundane but extremely crucial piece of infrastructure work!

[codecov]

Codecov Report

Merging #809 (6158da9) into main (6491e0c) will decrease coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #809      +/-   ##
==========================================
- Coverage   92.75%   92.72%   -0.02%     
==========================================
  Files         188      188              
  Lines       17398    17399       +1     
==========================================
- Hits        16136    16133       -3     
- Misses       1262     1266       +4     

Flag	Coverage Δ
unittests	92.72% <100.00%> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...dgsem/elixir_advection_amr_solution_independent.jl	100.00% <ø> (ø)
...xamples/p4est_3d_dgsem/elixir_euler_free_stream.jl	100.00% <ø> (ø)
...s/structured_3d_dgsem/elixir_euler_source_terms.jl	100.00% <ø> (ø)
...ree_2d_dgsem/elixir_advection_amr_coarsen_twice.jl	100.00% <ø> (ø)
...tree_2d_dgsem/elixir_advection_amr_refine_twice.jl	100.00% <ø> (ø)
...dgsem/elixir_advection_amr_solution_independent.jl	100.00% <ø> (ø)
...ree_2d_dgsem/elixir_advection_amr_visualization.jl	100.00% <ø> (ø)
examples/tree_2d_dgsem/elixir_lbm_couette.jl	80.00% <ø> (ø)
src/solvers/dgsem_p4est/dg_2d.jl	99.02% <100.00%> (+<0.01%)	⬆️
src/callbacks_step/save_restart.jl	84.00% <0.00%> (-8.00%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6491e0c...6158da9. Read the comment docs.