Re: Re: Re: Re: Activeperl 5.6 fork() doesn't friggin work

[jtx]

What I have been able to do, though it is a cheap hack and will probably have to be re-written if I want this to be a serious program, is taken your idea of using a while loop and somewhat modifying the server/client. This is the modified client, and it works, though like I said, I think it's pretty cheesy:

#!/usr/bin/perl

use IO::Socket;

# Make our connections
my $ns1 = new IO::Socket::INET (PeerAddr  => 'ns1',
                                PeerPort  => 1200,
                                Proto     => 'tcp'
                               ) or die "Couldn't create socket: $!\n"
+;

my $line;
my $count = 0;

my $domain = 'hello.com';
my $ipaddress = '192.168.0.1';

while ($line ne 'CAIO') {

   if ($count == 0) {
      print $ns1 "HELO\n";
      $count++;
   } elsif ($count == 1) {
      print $ns1 "dom:$domain\n";
      $count++;
   } elsif ($count == 2) {
      print $ns1 "ipaddr:$ipaddress\n";
      $count++;
   } elsif ($count == 3) {
      print $ns1 "addzone\n";
      $count++;
   }

   sleep(1);
   chomp($line = <$ns1>);
   print "$line\n";
}

close($ns1);

As you might have gathered, the server is a name server, and this program was created to add zones via sockets. It's irrelevant to me at this point whether or not BIND supports this natively - This is just a good way for me to practice using Sockets and TK, and make my boss happy while I write programs that will eventually end up being my replacement :)

[premchai21]

Or:

my (@cmds) = ("HELO", "dom:$domain", "ipaddr:$ipaddress", "addzone");

while (@cmds)
{
    print $ns1 shift(@cmds), "\n";
    $line = <$ns1>;          # don't need the sleep;
                             # the read should block until
                             # a line is available.

    defined($line) or last;  # quit if connection closed
                             # (could die instead)

    print $line;             # don't need \n now, $line
                             # not chomped

    ($line =~ /^CAIO$/) and last;
                             # $ can handle trailing
                             # newline, with no
                             # significant speed penalty
                             # (should be optimized by
                             #  regex engine)
}