![](https://secure.gravatar.com/avatar/93000ba847c7071e4a7a2fdf975df825.jpg?s=120&d=mm&r=g)
Hi, I’m trying to use the automatic pytest discovery in mmtbx and I notice that there are a few “libtbx-style” tests, namely "$D/conformation_dependent_library/test_cdl.py", "$D/conformation_dependent_library/test_cdl_esd.py", “$D/conformation_dependent_library/test_rdl.py", "$D/conformation_dependent_library/test_hpdl.py”, that follow the “pytest” naming convention, and are thus picked up by the pytest.discover() function. They then error because pytest cannot run them. There are — as far as I can tell — normal libtbx tests (they are registered in run_tests.py). Is there any reason they do not follow the typical naming convention? Changing their filenames and modifying run_tests.py accordingly seems to lead to the same behaviour when running tests. Is there any reason they should not be renamed from “test_*.py” to “tst_*.py”? Thanks, Nick