On Jan 22, 2009, at 11:22 PM, Ovid wrote:
Well, since the thread was about Eric's method of eliminating
'plan', after you made your explanation, I think it was clear that
the proposal would require, as Eric suggested, an alteration to core
TAP:
ok 1
ok 2
ok 1
ok 2
ok 3
ok 4
1..4
ok 3
1..3
With the above, the nested TAP is essentially a 'no_plan' stream
which is summarized as 'ok 3'. I *think* this is what Eric was
referring to. This does, to an extent, merely push the plan/no_plan
issue down a level (how do I know that 4 tests in the nested TAP
were the tests which needed to be run?), but it's an interesting
idea. (I deliberately included plans to show how the idea could
still match our current TAP discussion and not be too far "out
there").
Right, and now it's no longer implied. :-)
David