Hello

When developing new features and modifying how things are implemented it would be very convenient to have a set of test cases to go through to systematically being able to detect if changes introduces regressions.

There currently are some 'make check'-test cases, but it is quite tedious to write these kinds of checks for e.g. the tools.

If the author of e.g. a new tool would take some time to write a bunch of test cases it would help a lot for future developers that could use them to test wether their changes causes regressions.

I've just started thinking about this and I thought I'd bring this up to discussion on the list. There are a couple of questions in particular I feel needs discussion:

Is this a good idea? Where should the testcases be stored/managed? What format should they be in?
Does anyone know of good tools to help with these kinds of things?

I'm thinking we would store test-cases as simple version-controlled text documents in ./testcases or something? Perhaps doing that and then restructure things first when the need for a more strict structure arises?

Regards, Martin