Front page | perl.qa |
Postings from January 2008
--- Paul Johnson <paul@pjcj.net> wrote: > I don't want it on all the time. I don't want to change anything in > my > code to do it. I don't (yet) care that it doesn't run my cleanup > code > or whatever. I just want to easily find out what the first error > was, > and then have a short edit/test cycle until the bug is fixed. Then > I'll > go back to plain "make test" as usual. > > Anyway, thanks for making this happen. Schwern's agreed that if a somewhat clean solution can be found, he'll add it. Aristotle's suggestion of dieing on the next ok() seems like the easiest to implement. The problem is this: is 3, 4, 'whee!'; is factorial(123456), $some_num, '... be patient ...'; If we have 'die on fail' for the first test, the arguments of the second test will still be evaluated, but we'll die before the test is executed (what I wouldn't give for lazy evaluation ...). You'll get most of what you want, but the "next arguments are evaluated" caveat is important. I think the interface should have three ways of being used: use Test::More 'no_plan', 'fail'; That means it's always on. use Test::More 'no_plan'; is 3, 4, 'whee!'; die_on_fail(); is factorial(123456), $some_num, '... be patient ...'; The first failure won't trigger the die. DIE_ON_FAIL=1 prove ... That let's people have global control (and is easy to bind to keys in editors). Cheers, Ovid -- Buy the book - http://www.oreilly.com/catalog/perlhks/ Perl and CGI - http://users.easystreet.com/ovid/cgi_course/ Personal blog - http://publius-ovidius.livejournal.com/ Tech blog - http://use.perl.org/~Ovid/journal/Thread Previous