From: Gennadiy Rozental (rogeeff_at_[hidden])
Date: 2002-07-18 21:03:04
> > > Yeah, testing this library is difficult. I'd be interested in
> > > discussing what an improved test would look like.
Not that I thrust myself, but naybe new Test Library could help?
> > Me too. First thing I would do (and have done) is to split the test
> > the same ways the headers are splitted - one test per trait. Somehow it
> > makes you willing to put more work in testing each particular trait :).
You may create one test case per trait. No need to create myriads of files.
> Yes, of course. The big problem is how to deal with the expected failures.
New BTL support number of expected failures on per test case basis.
> I want to know exactly which sub-checks in a test are expected to
> fail/succeed on a given compiler.
It could be done.
Right now you are using STATIC_ASSERTSs which does not seems correct since
it cause compilation error if fails, while we want runtime one.
I would be happy to help with any testing issues.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk