From: "Patrick R. Michaud" <[email protected]>
Date: Sun, 6 Dec 2009 00:01:24 -0600
TT #1091 and DEPRECATED.pod claim that label-based exception
handlers are eligible to be removed in 2.1. I think label-based
exception handlers are necessary; but I'm willing to be proven wrong.
If a handler "should be" a Sub, then presumably a closure should also be
allowed. In that case, you could create a closure that contained the
loop_handler body, calling continuations instead of using "goto". But
not only is that a much more heavy-weight solution, it doesn't avoid the
inferior runloop issue, either.
Furthermore, if there is an inferior runloop problem with using a
Continuation as a handler, then there is also a problem with invoking a
Continuation from a handler. And if that is so, then handler code can't
make a nonlocal exit, which seems unacceptable in any case.
But if a handler can be a Closure or a Sub, and if handlers should be
allowed to make nonlocal exits, then I fail to see the advantage of
eliminating the Continuation case. It should be possible to use all
classes of invokable objects interchangeably, IMHO -- here as elsewhere.
So I am puzzled, too.
OK, back to lurking. ;-}
-- Bob Rogers
http://www.rgrjr.com/
_______________________________________________
http://lists.parrot.org/mailman/listinfo/parrot-dev
