Use subtests

[schwern]

This would all be a whole lot simpler if each call to ok ran all the tests as a subtest.

Right now, a Test::Proto test produces a single ok line. This isn't a very good indication of what's really going on. It also goes to some lengths to implement existing Test::More and Test::Builder functions itself so they only fail. Finally, only the first test fails which doesn't give the full story for debugging.

If ok ran validate in a subtest, then each test method (is_like and so on) can run as a normal test, succeeding or failing. The normal Test::More and Test::Builder calls can be used. And all test output and failures can be displayed.

[pdl]

I agree. It is issue #1 - I'm just getting there slowly so the work isn't in the master branch at the moment, nor applied to most of the prototype classes.

[pdl]

Closed as duplicate.