if (-s "$imagefile" < 700_000_000 and -e "$imagefile") { print '<script language="JavaScript"> var x=window.confirm("Is there a CD/DVD in the drive?" +) if (!x) window.location="http://hostname/cgi-bin/swish.cgi" </script>'; print "\n\nISO not too big and exists. \nBurning the CDR +W.\n\n" + ."\nPlease be patient, this will take a while...\n\n"; system "sudo cdrecord -v -eject dev=$device $imagefile"; + + } ( $? == 0 ) ? &success($date, $time, $hostname, $logdir, $to, $from) : &failure($date, $time, $hostname, $logdir, $to, $from);
Hmm. That looks like mixed success to me. What happens if the test at the top of the if statement fails? Then $? will be 0 (unless it was fortuitously still set from a previous failed child process), even though cdrecord was never attempted.
One technique for simplifying the code that requires a lot of error checking is to stick everything in an eval, use dies liberally within the eval to "raise exceptions", and pick through the mess afterwards; something like:
eval { foo() or die "foo failed"; bar() or die "bar bombed"; baz() or die "baz blew chunks"; # etc. }; # don't forget that semicolon! handle_error($@) if $@; # life is good... # ...and in some dark corner of your code sub handle_error { my $msg = shift; if ( $msg =~ /foo/ ) { # deal with foo's error; } elsif ($msg =~ /bar/ ) { # deal with bar's error } # etc. else { # deal with unexpected error } }
the lowliest monk
In reply to Re^4: Control Structures
by tlm
in thread Control Structures
by ghenry
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |