Uh oh!
There was an error while loading. Please reload this page.
Reorganise tests - #367
Conversation
glesur
commented
Mar 19, 2026
- separate tests of specific IO routines from physics tests (XDMF and restart dumps)
- add all IO tests to github actions
- add Sedov to the test suite.
- make a dedicated venv in github-actions for pydefix
…way via testme.json + handle single variant defined in testme.json
There was a problem hiding this comment.
Pull request overview
Reorganizes the test suite by separating IO-focused tests (restart dumps, XDMF, Pydefix) from physics tests, and updates CI to run the expanded test matrix (including Sedov).
Changes:
- Moved restart validation out of the OrszagTang3D physics test into a dedicated IO “dump” test.
- Added new IO tests for XDMF output and Pydefix, and wired them into GitHub Actions.
- Added Sedov blast wave to CI and adjusted some test tolerances/output cadence.
Reviewed changes
Copilot reviewed 23 out of 24 changed files in this pull request and generated 5 comments.
Show a summary per file
| File | Description |
|---|---|
| test/MHD/OrszagTang3D/testme.py | Removes restart validation from the physics regression test. |
| test/MHD/OrszagTang3D/setup.cpp | Removes restart-analysis logic from OrszagTang3D setup. |
| test/MHD/LinearWaveTest/testme.py | Adjusts tolerance for the linear wave test. |
| test/MHD/LinearWaveTest/idefix-entropy.ini | Increases log cadence for the entropy variant. |
| test/IO/xdmf/testme.py | Adds a new IO test driver intended to validate XDMF output. |
| test/IO/xdmf/setup.cpp | Adds minimal setup to generate a representative MHD state for XDMF dumps. |
| test/IO/xdmf/idefix.ini | Adds a config that enables XDMF + dump output for the new test. |
| test/IO/xdmf/definitions.hpp | Adds build definitions for the new XDMF IO test. |
| test/IO/pydefix/testme.py | Adds a new IO test covering Pydefix dump generation/regression. |
| test/IO/pydefix/python_requirements.txt | Adjusts pybind11 minimum version and adds a comment. |
| test/IO/pydefix/idefix.ini | Adjusts output cadence and enables dmp output for Pydefix test. |
| test/IO/dump/testme.py | Adds a dedicated restart-dump IO test driver. |
| test/IO/dump/setup.cpp | Adds restart validation analysis (moved from OrszagTang3D). |
| test/IO/dump/definitions.hpp | Adds build definitions for the new dump IO test. |
| test/IO/dump/CMakeLists.txt | Enables MHD property for the IO dump test target. |
| test/HD/SedovBlastWave/idefix.ini | Removes XDMF output from Sedov and tweaks output formatting. |
| test/HD/SedovBlastWave/CMakeLists.txt | Adds an MHD property line (in addition to HDF5). |
| src/setup.cpp | Suppresses default Setup warning when built with Python enabled. |
| src/pydefix.cpp | Marks some arrays as maybe_unused and prints Python runtime details at startup. |
| src/output/xdmf.cpp | Adds maybe_unused annotations to avoid warnings in some build modes. |
| reference | Updates submodule reference SHA. |
| CMakeLists.txt | Adds stricter HDF5 parallel checks under MPI and adds Python3/pybind11 CMake path setup. |
| .github/workflows/idefix-ci-jobs.yml | Adds Sedov + IO test job and includes LinearWaveTest in CI. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
### Changed - fixed a bug that could lead to diverging results around the spherical axis in non-ideal MHD in 2.5D and 3D (#356) - refactor of the MPI exchange routines and boundary routines to avoid buildup of roundoff errors at domain faces/edges that could lead to the sudden burst of div(B) or incoherences between MPI sub-domains (#357) - fixed a bug that could lead to compilations error when targetting AMD APUs (#359) - fixed a bug that led to the generation of incorrect subviews in 2.5D with vector_potential enabled (#362) - fixed a bug that could lead to memory corruption when using the UCT_HLLD emf reconstruction scheme and DIMENSIONS < COMPONENTS (#363) - reorganise the test to separate specific IO tests from physics tests (#367) ### Added - magnetic vector potential is now accessible from pydefix when enabled (#361) - use ccache in the test suite to reduce the runtime of continuous integration (#364) - automatically detects pybind11 path with cmake when using pydefix (#367) - skeleton to run the test suite with pytest (#366) --------------------------------------------------------------------------- Co-authored-by: Clément Robert <cr52@protonmail.com> Co-authored-by: Sébastien Valat <sebastien.valat.dev@orange.fr>