Maybe I should go back to implementing my own.

There is no need, nor any point to implementing it as a module.

It takes exactly 5 lines of code:

#! perl -slw
use strict;
use threads;
use Thread::Queue;

sub worker {
    my $Q = shift;
    my $tid = threads->tid;
    print "Thread $tid started";
    while( defined( my $work = $Q->dequeue ) ) {
        print "Thread $tid processing workitem $work";
        sleep 1;
    }
    print "Thread $tid done";
}

our $WORKERS //= 10;
our $ITEMS   //= 30;

my $Q = new Thread::Queue;                                    ## 1
my @pool = map threads->new( \&worker, $Q ), 1 .. $WORKERS; ## 2
$Q->enqueue( 1 .. $ITEMS );                                  ## 3
$Q->enqueue( (undef) x $WORKERS );                           ## 4
$_->join for @pool;                                           ## 5

__END__
C:\test>tqpool -WORKERS=3 -ITEMS=10
Thread 1 started
Thread 2 started
Thread 2 processing workitem 1
Thread 3 started
Thread 3 processing workitem 3
Thread 1 processing workitem 2
Thread 2 processing workitem 4
Thread 3 processing workitem 5
Thread 1 processing workitem 6
Thread 2 processing workitem 7
Thread 3 processing workitem 8
Thread 1 processing workitem 9
Thread 2 processing workitem 10
Thread 3 done
Thread 1 done
Thread 2 done
[download]