MsWin32 test divergence

Hi all,

	I’m comparing two cases:

Failure:
http://www.cpantesters.org/cpan/report/feecec6f-6c8a-1014-80c7-7ab93f003977 (gregor)

Success:
http://www.cpantesters.org/cpan/report/8173e2a8-6bf3-1014-af18-4cf4fbb2f7c2 (chorny)

It looks like there’s a divergence in build logic between these … is that by design? The failure case gives quite a bit more output and appears to go about its business a bit differently.

I’m wondering if the problem is with the tested module (CBOR::Free) or with the test environment.

Any pointers would be helpful. Thank you!

cheers,
-Felipe Gasper