Depends what you mean by "uniquely"? Ie. Is it the internal handle you're interested in; or where it is pointing:

C:\test>perl -mstrict -wle"print \*STDERR; open STDERR, '>', 'junk.err
+' or die $!; print \*STDERR;"
GLOB(0x1da110)
GLOB(0x1da110)
[download]

The value of the internal file handle should suffice. To give some background, I am writing a simple mock tcp server, using IO::Socket::INET and IO::Select, to use in system testing. It is not finished yet, but as a minimum I'd like to at least display the socket ids in a trace so I can tell which client is sending what. Here are some (incomplete) code snippets to give you a feel:

my $listener = IO::Socket::INET->new(
   LocalPort => $port,
   Proto     => 'tcp',
   Listen    => 5,
   ReuseAddr => 1,
) or die "error: IO::Socket::INET: $@";
my $selector = IO::Select->new($listener);
while ( my @ready = $selector->can_read() )  {
   for my $client (@ready) {
      if ( $client == $listener ) {
         my $new_conn = $listener->accept();
         $selector->add($new_conn);
         my $fh_hex = sprintf '0x%x', $new_conn;
         print "Accepted new connection ($fh_hex)\n";
# ...
sub recv_client {
   my $client = shift;
   my $fh_hex = sprintf '0x%x', $client;
   print "Recv from client ($fh_hex):\n";
   # ... display data received ...
}
[download]

Update: Thanks to BrowserUk's excellent tip I am now using:

my $fh_hex   = sprintf '0x%x', $new_conn;
my $peerhost = $new_conn->peerhost();
my $peerport = $new_conn->peerport();
my $peeraddr = $new_conn->peeraddr();
my $peerhostfull = gethostbyaddr($peeraddr, AF_INET) || "Cannot resolv
+e";
print "Accepted new connection $fh_hex ($peerhost:$peerport,host=$peer
+hostfull)\n";
[download]

which is much better.

Be aware that handles do get reused. Once a client disconnects, the next connect might reuse that handle.

For uniqueness, you might consider combining the output from IO::Socket's PeerHost() and PeerPort() methods: xxx.xxx.xxx.xxx:nnnn.

Also, I don't see a lot of value in sprintfing the handle, over the default stringification 'GLOB(0xHHHHHH)"?

---

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'

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". I knew I was on the right track :)

In the absence of evidence, opinion is indistinguishable from prejudice.

$unique{$new_conn}++;

$unique_id = "$client $unique{$client}";
[download]

glob plus counter per glob

Hi eyepopslikeamosquito,

There's also fileno, but those numbers are re-used. I don't have much experience with fileno, so perhaps a monk who does can say more, but I imagine that if you closely keep track of when filehandles are closed, you could still use fileno to tell different connections apart.

open my $fh1, '<', '/tmp/foo' or die $!;
open my $fh2, '<', '/tmp/bar' or die $!;
print "fh1: ", fileno($fh1), "\n";
print "fh2: ", fileno($fh2), "\n";
__END__
fh1: 3
fh2: 4
[download]

Otherwise, I agree with BrowserUk: a good way to identify network connections is by properties like remote IP + remote port.

A while back I tried out POE. Its API takes quite a bit of getting used to, but once you've got the hang of it, it works well. Since each "session" has a unique ID, keeping track of clients is easy. The Cookbook is really helpful in getting started.

Hope this helps,
-- Hauke D