Isn't it bad style to require a particular order for the tests?  I
thought the idea was they were self-contained, so if you get 'test 55
fails' then you can fix the bug and run test 55 again.  Also, if one
test relies on an environment left around by the previous one, a
single failure in an early test can screw up the whole run.  That
probably doesn't matter, but if it is not possible to fail one test
and pass its successor, then they could be considered as a single
test.

OTOH, the tests in a distribution are not normally run in any other
way than by 'make test', so if it saves time to require a particular
order you can do that without inconveniencing the user.

OTOOH, it is quite simple to make a single test which does two or more
things, if you need them to happen in that order.  So there is no real
_need_ for MakeMaker or anything else to support requiring a
particular test order.

-- 
Ed Avis <[EMAIL PROTECTED]>


Reply via email to