On Thursday 17 May 2007 18:30:08 chromatic wrote: > Sort of, but it also breaks test #15 in t/pmc/class.t. That's not your > fault though; the code fails to take into account role composition into > roles. I keep trying to fix it, but it gets a lot more complex than I > thought. > > I suspect what needs to happen is that Class#does() ought to loop through > all of the roles composed into the class, check if their names match the > given name, then call does() on each of them in turn.
That was the trick. I just fixed it in r18582. Thanks for tracking it down! -- c
