In the OP case, where the final intend is to run an external command, the optimum solution would probably be to fork a new process and exec the command. | [reply] [Watch: Dir/Any] |
forking, and then calling system to run the command, means he's forking twice for every command. Maybe 3 times depending what's in $command.
You can achieve the goal without threads on win32 like this:
#! perl -slw
use strict;
my @dates = 1 .. 10;
my @pids;
(my $cmd = <<'CMD' ) =~ s[\s+][ ]g;
-wle"print qq[task $ARGV[0] starting];
sleep rand( 10 );
print qq[task $ARGV[ 0 ] finishing];"
CMD
while( my $date = pop @dates ) {
push @pids, system 1, 'perl.exe', "$cmd $date";
if( @pids >= 5 ) {
my $kid = wait;
@pids = grep $_ != $kid, @pids;
}
}
waitpid $_, 0 for @pids;
Since you need to load a new executable, why not just start it in a new process rather than duplicating the current one first?
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
| [reply] [Watch: Dir/Any] [d/l] [select] |
Since you need to load a new executable, why not just start it in a new process rather than duplicating the current one first?
Well, I was thinking about Unix systems where calling fork & exec is the standard (and sometimes only) way to start a new process. On Windows, you are right, it will be quite inefficient.
| [reply] [Watch: Dir/Any] |