Primarily what I'm wondering is if there is already a good way to do this, since I'm not volunteering to implement it if there isn't.
Still, I don't suppose it hurts to mention what I would like. For polynomial rings, it would be a function along the lines of makeSomeRing(predicate) where predicate is a function mapping rings to true/false, and makeSomeRing returns some ring that makes pred evaluate to true and throws an exception if it can't find any such ring. Then Sage could detect that this method has been called and re-run the test where the method returns a different ring each time. Or it could just be run once by default for some preferred ring and many times given some option to do more exhaustive testing. At the most simple, makeSomeRing could just try pred on a fixed list of rings. Add a newly written ring to the list as you add it to Sage, and you get the new ring tested in every situation where the old rings are already tested. So the intent is not to construct random rings, though the scheme could be extended to support that as well. This scheme applies to just about anything you might wish your test to be parameterized by, it's just that rings is what is relevant to me for the Frobby interface. Testing the output could be handled by constructing the correct answer in the same ring, or by converting at the end to some fixed ring. If this were adopted, probably there should be a rule that the first test does not use this, so that the tests remain useful as documentation. -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
