Just as an experiment I hacked at the sl_prioritized sub directly like
so....
sub sl_prioritized {
my ($kernel, $heap, $priority, $msg) = @_[KERNEL, HEAP, ARG0, ARG1]; my
$now = time();
#$heap->{send_time} = $now if $heap->{send_time} < $now;
if (@{$heap->{send_queue}}) {
my $i = @{$heap->{send_queue}};
$i-- while ($i and $priority <
$heap->{send_queue}->[$i-1]->[MSG_PRI]); splice(
@{$heap->{send_queue}}, $i, 0,
[ $priority, # MSG_PRI
$msg, # MSG_TEXT
]
);
#} elsif ( $heap->{dont_flood} and
# $heap->{send_time} - $now >= 10 or not defined
# $heap->{socket} # ) {
# push( @{$heap->{send_queue}},
# [ $priority, # MSG_PRI
# $msg, # MSG_TEXT
# ]
# );
# $kernel->delay( sl_delayed => $heap->{send_time} - $now - 10 );
} else {
warn ">>> $msg\n" if $heap->{debug};
#$heap->{send_time} += 2 + length($msg) / 120;
$heap->{socket}->put($msg);
}
}
The results were the same as before 20secs for 14 short lines. :) hmmm I
suspect that another layer of "throttling" is taking place in
POE::Wheel::ReadWrite ? On the other hand I could just be dense...
Mike Tremaine
[EMAIL PROTECTED]
