On Thu 19 Feb 2009, Geoffrey Young wrote:
> I've love if this particular aspect had a better solution, though
> threads really complicates things...
child_terminate is forbidden with a threaded MPM. And I think it is
rightly so.
void mpxs_Apache2__RequestRec_child_terminate(pTHX_ request_rec *r)
> And then we have $r->child_terminate which simply calls exit() at
> C-level in a request pool cleanup. So, no perl-level cleanup at all is
> done.
fwiw, I've never been happy with the implementation of child_terminate()
in 2.0 - in mp1 it was a nice, graceful way of cleaning up a child
proces
Hi,
why do we not call perl's cleanup at child exit?
Here my setup:
package My::Test;
use strict;
use warnings;
sub new {
my (undef, $data)=...@_;
my $inst=bless \$data;
warn "$$: a new instance is born: $inst ($$inst)\n";
$inst;
}
sub DESTROY {warn "$$: DESTROY $_[0] (${$_[0]})\n"}
$M
