Harry Putnam wrote:
>
> "John W. Krahn" <[EMAIL PROTECTED]> writes:
>
> > open FILE, '>file' or die "Cannot open file: $!";
> > my $pid = open DUMP, 'tcpdump -v -ieth0 |' or die "Cannot open pipe from
> > tcpdump: $!";
> > print FILE while <DUMP>;
> > close DUMP or die "Cannot close pipe from tcpdump: $!";
> > close FILE;
> >
> > print "Pid was $pid\n";
>
> Looking at this, I'm guessing one doesn't get to know the pid until
> its all over. (I can't try it out right now). Which in this case
> would not be usefull.
You can print the PID before the loop starts:
open FILE, '>file' or die "Cannot open file: $!";
my $pid = open DUMP, 'tcpdump -v -ieth0 |' or die "Cannot open pipe from
tcpdump: $!";
print "Pid is $pid\n";
print FILE while <DUMP>;
close DUMP or die "Cannot close pipe from tcpdump: $!";
close FILE;
> This thing needs to run until at the 4hr mark syslog sends a SIGHUP
> and rotates the ouput files. All that is taken care of by syslog.
>
> But the tcpdump process runs continuously. Only stopped by outside
> forces and then restarted.
>
> There is probably some standard way to put those pid numbers in
> /var/run but I don't know it.
use File::Basename;
my $pid_file = '/var/run/' . basename( $0 ) . '.pid';
{ open my $PID, '>', $pid_file and print $PID "$pid\n" };
John
--
use Perl;
program
fulfillment
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
<http://learn.perl.org/> <http://learn.perl.org/first-response>
