Duh, Abigail, you know me for at least 10 years now. Do you really
think you have to explain prototypes to me? Same for the other people:
yes I do know *all* complications of changing it. And no, it is not
important for the requested simplification of POSIX.pm
CORE::* uses prototypes and POSIX.pm does not which is inconsequent. There
are about 50 functions from POSIX.pm which are one-to-one mapped to CORE
functions with the same name. Because of a missing prototype, they currently
behave differently. If "the perl porters" prefer that situation, then
let it be.
In my experience, all 5.* releases have broken at least some of
my modules. Adding deprecation messages, as example, do often hinder
me. For instance, I use a module which started to produce "defined %hash"
complaints. It's not my module which caused the complaint, but something
currently unmaintained. Took me hours to repair it cleanly.
Still, I think we should make those improvements with major releases.
* Abigail (abigail@abigail.be) [110130 19:39]:
> > In the common case you can. Of course, there are situations where it
> > does not work. A prototype of ($$) works as well as @_==2
>
> Eh, no.
> sub foo ($$) {...}
> @arg = (1, 2);
> foo (@arg);
>
> isn't the same as
> foo (1, 2);
--
MarkOv
------------------------------------------------------------------------
Mark Overmeer MSc MARKOV Solutions
Mark@Overmeer.net solutions@overmeer.net
http://Mark.Overmeer.net http://solutions.overmeer.net
Thread Previous
|
Thread Next