You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, there are no published tests for scene_rdl2 Python bindings. These are somewhat incomplete and buggy in places. As part of a larger project, I think it would be worthwhile to expose scene_rdl2 to Python as comprehensively as possible. But before doing that, I think the first step is to build up a test suite so that the existing bindings can be tested, fixes for existing bugs can be verified, new bindings can be validated, and regressions can be caught over time.
I propose using Python's built-in unittest library - it works fine, and it doesn't require any extra dependencies. Tests would be added to tests/mod/python/py_scene_rdl2 alongside the existing tests/lib. The new tests will be integrated into the CTest unit tests as a new test, scenerdl2_py_scene_rdl2_tests, so that these run as part of the standard unit label tests. It would also depend on the scenerdl2_scene_rdl2_tests target, in order to use the DSOs that are generated there for testing (fake / test-only SceneObjects).
The tests themselves would mirror the C++ scene_rdl2 tests, as much as is appropriate in Python. For the most part, the file structure should mirror the C++ tests as well, to make it easy to jump back and forth and compare (e.g. TestAttribute.cc would have a paired test_attribute.py).
I"m happy to take on this work if there's no concerns about the approach.
Activity