Perl for Web Site Management

Making the Script Smarter

Once we know about it, it’s an easy matter to modify the original data file to fix the Best Glue Products listing. But it would be nice if we could make the &parse_exhibitor subroutine smart enough to spot those sorts of problems on its own. By adding some sanity checks and having the script output appropriate error messages if they fail, we can make the process of policing the data file for these sorts of errors much easier.

For example, &parse_exhibitor uses the following lines to parse the company’s name:

        if ($line_count == 1) {
            $co_name = $line;

What if that line somehow ended up being blank? The code in the main body of our script shouldn’t allow that to happen, but it doesn’t hurt to program defensively. Let’s modify those lines to look like this:

        if ($line_count == 1) {
            unless ($co_name = $line) {
                warn <<"EOF";
line_count=1, but got a false co_name. skipping exhibitor.
($exhibit_file line number $.)
EOF
                return;
            }

By placing the assignment to the $co_name variable inside a logical test, we can test the value being assigned for truth. In the event that a false company name is assigned, we issue a warning, including the name of the file we were parsing ($exhibit_file) and the special variable $., which gives the line number of the filehandle most recently read from. (This will help us locate the problematic listing in $exhibit_file.) Then we use the return function, which causes a subroutine to stop whatever it’s doing and returns program execution ...