On Fri, Aug 02, 2002 at 04:41:01PM -0700, [EMAIL PROTECTED] wrote:
> First hurdle:
>
> if (pid <= 0) {
> /* Sorry folks, we don't presently implement rooting around for
> the first child we can find, and we definitely don't want to
> pass a pid of -1 to $getjpi, where it is a wildcard operation.
> */
> set_errno(ENOTSUP);
> return -1;
> }
>
> This doesn't go well with "while ($pid = waitpid(-1, WNOHANG))" in
> POE::Kernel.
>
> It's just a Blocking World out there!
I'm confused. Why would passing -1 to $getjpi (I'm assuming it's
syscall-esque) be a bad idea?
For VMS we could go with a $SIG{CHLD} handler and hope Perl doesn't
dump core. That should be a safe course (rhymes with cores) of action
in 5.8 and beyond.
I would rather improve Perl than POE here, though. If a VMS
programmer can work it out in C, it'll be faster and increase
portability for Perl programs on that system.
-- Rocco Caputo / [EMAIL PROTECTED] / poe.perl.org / poe.sf.net
