EulerMB::Coroutine - http://www.math2.org/eulermb/pod/EulerMB/Coroutine.html
========== SYNOPSIS
use EulerMB::Coroutine qw(:all);
sub gen_func
{
my $x = 0;
return coroutine {
COBEGIN();
while($x < 10) {
YIELD($x);
$x++;
}
}->wrap();
}my $func = &gen_func();
while(my @x = $func->()) {
print "$x[0]\n";
}
# prints 0..9
==========It is inspired by the Coro distribution (http://search.cpan.org/~mlehmann/Coro), particularly Coro::Cont (BTW, is the use of the term "continuation" correct in that module?), as well as coroutines in other languages (e.g. Lua). I wrote this specifically for another lexing/parsing module:
http://www.math2.org/eulermb/pod/EulerMB/Content.html
and it did the job for that application. You may refer to it for examples. I had reasonable success using Coro instead, but I found it segfaulted if I wasn't careful, it had some problems with threading-enabled Perl, and I didn't want my module dependent upon something that might make deployment more difficult.
My question is whether it is worth renaming this EulerMB::Coroutine module to something more generic and putting it onto CPAN. Also, I'm open to any ideas to making it better. It relies on a bit of source filtering, which some people don't like, but the source filtering is fairly tame, and I don't know of any better way to do coroutines reliably and conveniently in Perl without modifying Perl itself. There is also a number of "Limitations and Caveats" as given in the POD. Some type of symmetric coroutine might in implemented in a similar manner as well.
best regards, -david manura, dm.list[-at-]math2.org
