Input data is structured in a way where changing one test could effect others.
Also test outputs are sometimes placed in the current working directory, and sometimes in a temporary directory. These files may or may not be deleted in different combinations when tests fail or pass.
It would be preferable to structure the inputs and expected outputs for each test in a consistent and separated directory structure for each test, nested by test class and test function so that out of order execution of tests does not effect the results.
Additionally when a test function within a test class fails, the resulting observed files should be left in a folder within that structure so they can be inspected, and only deleted if the test function passes.
This, and directory path naming and path construction in code should all be handled consistently across test classes.
Input data is structured in a way where changing one test could effect others.
Also test outputs are sometimes placed in the current working directory, and sometimes in a temporary directory. These files may or may not be deleted in different combinations when tests fail or pass.
It would be preferable to structure the inputs and expected outputs for each test in a consistent and separated directory structure for each test, nested by test class and test function so that out of order execution of tests does not effect the results.
Additionally when a test function within a test class fails, the resulting observed files should be left in a folder within that structure so they can be inspected, and only deleted if the test function passes.
This, and directory path naming and path construction in code should all be handled consistently across test classes.