I think it could be useful to add some property-based tests to resolvelib, to stress test some of the basic properties of the resolver. Some ideas I had for useful tests:
- If the resolver succeeds, the resolver result will satisfy the supplied requirements.
- Given a set of requirements, adding an extra one that contradicts one in the set will cause a resolver failure.
The key will be setting up "strategies" that generate sets of candidates and requirements.
I'm mainly adding this here as a reminder, so that I don't forget to look at this at some point in the future.
I think it could be useful to add some property-based tests to resolvelib, to stress test some of the basic properties of the resolver. Some ideas I had for useful tests:
The key will be setting up "strategies" that generate sets of candidates and requirements.
I'm mainly adding this here as a reminder, so that I don't forget to look at this at some point in the future.