Front page | perl.perl5.porters |
Postings from June 2012
IO::Socket::IP in core?
Thread Next
From:
Ricardo Signes
Date:
June 25, 2012 07:14
Subject:
IO::Socket::IP in core?
Message ID:
20120625141420.GA5237@cancer.codesimply.com
IO::Socket::IP is floating around the edges of the core distribution, poking
at the cell membrane and looking for a way in. Let's talk about it.
I'm not advocating for a teeny teeny tiny core, but I think a smaller one is
good. I don't think we want two big complicated libraries for doing the same
thing in core, for example. IO::Socket::IP seems promising because it says
it's a drop-in replacement for IO::Socket. That means that IO::Socket could
drop out, or (realistically) become an alias for or wrapper around
IO::Socket::IP.
If it can't ever "go away" in this way, this would suggest to me that
IO::Socket::IP is *not* a drop-in replacement for IO::Socket, and I'd be
unexcited to bring it into core without being able to eventually eject its
older cousin.
So, how can we prove that it's a good drop-in? Well, one place to start is by
making IO::Socket into an empty subclass of ::IP, right? Then we can run the
core test suite on a nice boring platform (to start!) to see what breaks.
Thoughts?
--
rjbs
Thread Next
-
IO::Socket::IP in core?
by Ricardo Signes