Script3.pl would be run if script2.pl died.

To prevent it, use

script1.pl
script2.pl && script3.pl
[download]

Script2 will always be run regardless of how script1 exits. Script3 will only be be run if script2 succeeds -- Ie. returns 0, which it won't if it dies or exits with any value other than 0.

You can also do

script1 || script2;
[download]

In which case, script2 will only be run if script1 fails.

P:\test>perl -e"exit 1" && echo succeeded || echo failed
failed

P:\test>perl -e"exit 0" && echo succeeded || echo failed
succeeded

P:\test>perl -e"die" && echo succeeded || echo failed
Died at -e line 1.
failed
[download]

You can also group commands

(prog1 && prog1) || (prog3 || prog4)
[download]

The thing about system is that it returns the exit value of the system command. System commands return 0 for success. So you really want to say system "blah" and die as opposed to system "blah" or die.

Win32::Process might work better than system()

system "blah"and die ...

Yes. Exactly.

The following code examples may be of use.

retcode.pl

#! /usr/bin/perl -w
# retcode.pl
# Sets the errorlevel passed as first argument
use strict;

my $retcode = $ARGV[0] || 0;
print "Return code = $retcode\n";
exit $retcode;
[download]

returncode.cmd

@echo off
rem A Command file that uses conditional processing.

echo Executing First Script
retcode.pl 0

if errorlevel 1 goto end_label

echo Executing Second Script
retcode.pl 0

:end_label
echo Done!
[download]

Play around with the return codes sent to the script and the name of the first scrip to see that the script will jump to the label if the script returns an error value greater than 1 or the script itself is not found.

inman

Something I've done in the past (admittedly for cron jobs but could still be applied here) is have each script set a fail flag (a zero byte size file for example) that would tell the next script to exit without running.

In script2.pl:

if ($you_do_not_want_script3_to_run) {
   # create zero byte size file...
   open(FAILFLAG, ">/tmp/script2.failflag") || do {
      # handle the error here
   };
   close (FAILFLAG);
}
[download]

Then in script3.pl:

my $script2_failflag = "/tmp/script2.failflag";
if (-e $script2_failflag) {
   # script3.pl should not contine any further
   exit();
}
[download]

Why on earth wouldn't you just check the exit status of the scripts you're executing?

------------
:Wq
Not an editor command: Wq

Well, as I mentioned, I have used fail flags for cron jobs in the past (rules out checking the exit status) and just thought I'd bring it up as something else the OP could try if so inclined. You know, TIMTOWTDI and all that :-)

-- vek --