If you use Net::FTP there is no need for any multitasking. Just supply Hash => \*STDOUT option to the constructor and the module will do it for you.

Hi,

Here you have an example:

#!/usr/bin/perl -w
$|++;

use strict;
use threads;
use threads::shared;

# Flag variable we use to stop the 2nd thread
my $complete : shared = 0;

# Create the thread
my $thread = threads->create(\&dothis);

# Code of main thread goes here
sleep 5;

# Change the flag variable so the thread ends
$complete = 1;

# Join the threads together
$thread->join();

# End of the program
exit;

sub dothis {
        until($complete) {
                print "#";
                sleep 1;
        }
        print "\n";
}
[download]

Take a look at threads.

But you can use fork and sending a kill signal to the child instead, I think it would be faster and more efficient.

For more complicated solution think about Event, Event::Lib or POE.

Update: This code is a example of threads, but of course for a simple stats of a file download is not necessary... as stated before. I supossed that you wanted a simple example to see how threads work in Perl.

Regards,

Multithreading is necessary when you need separate computations to happen at the same time (e.g. on a multi-cpu machine) (*). Multithreading is occasionally useful when you have to monitor multiple different input streams or monitor an input stream while doing some complex calculations in the background. (Some people just don't like writing state machines or using POE) The problem you mention sounds nothing like either of these cases.

To approach the problem you mentioned, it seems to that the matter is simple: pull down some of the file, see if you've pulled down enough bytes to write another '#' mark, print it if so, and repeat. Of course, if you're using Net::FTP to pull the file down (you aren't rewriting your own ftp protocol, are you?) then another poster already mentioned the option that'll do that for you. If you're running an external ftp program in the background, I'd suggest that you need to read perlipc and consider inter-process communication, and then check if your ftp program has the ability to print out hash marks as it's working (many do).

(*) Yes, I'm aware that multithreading per se isn't the only way to split computations across multiple CPUs; obviously one can run entirely separate processes.

--
@/=map{[/./g]}qw/.h_nJ Xapou cets krht ele_ r_ra/;
map{y/X_/\n /;print}map{pop@$_}@/for@/
[download]

But here are some parallel processing modules

Parallel::ForkManager

Proc::Queue

Sample codes are available in the docs

-SK

However you may not need multithreading, care to explain more what you want to do?

C.