| POE::Pipe::TwoWay - portable two-way pipe creation |
POE::Pipe::TwoWay - portable two-way pipe creation (works without POE)
my ($a_read, $a_write, $b_read, $b_write) = POE::Pipe::TwoWay->new(); die "couldn't create a pipe: $!" unless defined $a_read;
POE::Pipe::TwoWay makes unbuffered two-way pipes or it dies trying.
It can be more frugal with filehandles than two OneWay pipes when
socketpair() is available.
Pipes are troublesome beasts because the different pipe creation
methods have spotty support from one system to another. Some systems
have pipe(), others have socketfactory(), and still others have
neither.
POE::Pipe::TwoWay tries different ways to make a pipe in the hope that
one of them will succeed on any given platform. It tries them in
socketpair() -> pipe() -> IO::Socket::INET order. If socketpair() is
available, the two-way pipe will use half as many filehandles as two
one-way pipes.
So anyway, the syntax is pretty easy:
my ($a_read, $a_write, $b_read, $b_write) = POE::Pipe::TwoWay->new(); die "couldn't create a pipe: $!" unless defined $a_read;
And now you have an unbuffered pipe with two read/write sides, A and
B. Writing to $a_write passes data to $b_read, and writing to
$b_write passes data to $a_read.
my ($a_read, $a_write, $b_read, $b_write) = POE::Pipe::TwoWay->new();
It's possible to force POE::Pipe::TwoWay to use one of its underlying pipe methods. This was implemented for exercising each method in tests, but it's possibly useful for others.
However, forcing TwoWay's pipe method isn't documented because it's cheezy and likely to change. Use it at your own risk.
The INET domain socket method may block for up to 1s if it fails.
POE::Pipe::TwoWay is copyright 2000 by Rocco Caputo. All rights reserved. POE::Pipe::TwoWay is free software; you may redistribute it and/or modify it under the same terms as Perl itself.
| POE::Pipe::TwoWay - portable two-way pipe creation |