I'm going to post a snip of the main loop for my Arcserve parser, mostly because I catch some of the weird things that pop up in our logs. Errors that start on a seperate line is one that tripped me up bad. These might help you if you run into them in your logs:

#%jobentry is defined outside of the while loop
        while (<LOG>) {
            next if /^\n/;
            my %line;
            my ($date, $time, $jobinfo, $jobtext);
            chomp;
            if (/^(E\d{4})/){
                $jobtext .= "[$1] ";
                $_ = substr $_, 6;
            }elsif (!/^\d{8}/){
                my $tmp_count = $count - 1;
                $job_entry[$tmp_count]{'text'} .= $_;
                next;
            }
            $date = substr $_, 0, 8;
            $time = substr $_, 9, 6;
            $jobinfo = substr $_, 16, 6;
            $jobtext .= substr $_, 23;
#o_trim is a little sub for taking out any straggling whitespace in th
+e job description
            $jobinfo = o_trim($jobinfo);
#Go on with rest of loop. I push the data into a hash, and then push t
+he hash into a database structure after I gather an entire job's wort
+h of entries.
               }
[download]

The entire code is tested and works, if quite quirky, as I was playing with some ideas about data structures when I did it. I'll post it if there is any interest.

----Asim, known to some as Woodrow.