On Thu, Aug 21, 2003 at 02:38:03PM +0100, Fergal Daly wrote: > 1 problem here is that you have not indicated anywhere how many times you will > be extending the plan
Some sort of *optional* way to specify that would be nice. > > The former has the large disadvange of requiring all subtests to be aware > > of test-wide state. And I can't think of any major advantages, though it > > is a bit more readable. > > The other advantage is the ability to do > > 1..4 > ok 1 > plan 1..2 > ok 2.1 > ok 2.2 > ok 2 # plan was ok > ok 3 > ok 4 > > That is, extend again before you finish the current extension. > > You could allow extensions at any time but then you lose the ability to know > if you ran 4 + 2 tests or 5 + 1, Not if you introduce an end tag (though I'd rather not). 1..4 ok 1 1..2 ok 1 ok 2 2..1 ok 2 ok 3 ok 4 4..1 Though now the 'no_plan' style in a subplan gets confusing. We might have to change the no_plan style so that it has to produce some sort of header. It might be literally "1..N". -- Michael G Schwern [EMAIL PROTECTED] http://www.pobox.com/~schwern/ Cheating is often more efficient. - Seven of Nine