The Perl Debugger

Debugging is an annoying necessity in any language, whether it's debugging your own code or somebody else's that you've been given to make work on your system. Anything you can do to make debugging easier is a big win. Perl includes a command-line debugger that can make your debugging job considerably easier. This article covers the basics of the debugger and shows off a few tricks you may find useful.

An astounding number of bugs can be caught by Perl automatically by turning on warnings and strict at the beginning of your program. If your program includes the line use warnings; you can catch dozens of common errors, including variables used only once, which often are typos; scalar variables used before they are set; and redefined subroutines.

These diagnostic messages can be explained further by including the line use diagnostics;, which prints an explanation for each warning. Or, you can look up the explanations using man perldiag.

If your Perl version is older than 5.6, instead of use warnings; you have to use the -w option on the first line of the script, like this: #!/usr/bin/perl -w.

Finally, you can catch additional common errors with use strict;, which in effect, forbids a few unsafe programming shortcuts. The rules that use strict turns on are as follows:

As you can see, warnings and strict tighten up a few of Perl's features that can be used for good but also can be abused. These commands make debugging easier, because Perl catches these bugs for you.

You might ask at this point, what's wrong with debugging by scattering print statements in your code? Nothing is wrong with this debugging technique, but you have more power with the interactive debugger. You can examine all aspects of the program and environment, not only those you thought of when you ran the program, and you can see more clearly what the program actually does. Hopefully, by the end of this article you will agree that investing a little effort in learning the debugger pays off in saved time.

The debugger is run from the command line by passing Perl the -d option:

perl -d filename.pl

If you are debugging a CGI program written with CGI.pm, simply call it on the command line with the arguments you'd like to pass, along with -d:

perl -d filename.pl param=value param2=value

Instead of using the command line, you could use the Perl debugger as part of certain IDEs, such as GNU Emacs or Activestate Komodo, or from debugger GUI front ends, such as ddd or ptkdb. For space reasons, I discuss only the command line in this article, but the principles hold for a GUI debugger as well.

If you're using the command-line debugger, it is useful to have the Term::ReadLine module installed, which enables cursoring through the command history.

Here's an example program we use in this article. Copy the following to a file called sample.pl:

#!/usr/bin/perl

use warnings;
use strict;

my $name = "Pengu";

foreach (1..20) {
  &shout($name);
}

sub shout {
  my $name = shift;
  print "*** $name ***\n";
}