Most CPAN modules have tests in either test.pl or the t/*.t files. Take a look at cpan.org: here are the tests for CGI.pm for example.

update: other links that might be of interest:

A Test::MockObject Illustrated Example

the Test::Unit documentation

Unit Testing @ extremeperl.org

update2: also take a look at

The Joy of Test

the Test::Unit documentation

Even though I'm biased I'd recommend avoiding Test::Unit and using Test::Class instead. Unlike the former Test::Class is maintained (by me :-) and plays with all the other Test::Builder based Perl test modules.

Also check out Test::Tutorial.

It's quite simple. All code starts with a specification (even if just in your head) and some documentation (eg the POD for your module). You write tests to ensure that everything in the spec and the documentation is true. Additionally, once you have written your code you should be aware of where any weak points in it are - where, for example, your off-by-one errors might be lurking - and so you would write extra tests to exercise those bits of the code more thoroughly. It is very important to test and document both for success and for failure. If your docs say that "the age parameter tells how old the person is" then you probably want to test (and document!) what happens when a user enters -1, 1000, and "grapefruit", as well as for reasonable values.

And when you change the code, you of course have updated the docs too, and you add tests to ensure that the updated docs are still true.

Searching around the monestry will find you a lot of useful stuff like Perl Testing and Quality Assurance and What goes in a test suite?. There are also lots of useful delicious links under perl+testing.

package X::CSV;
use strict;
use warnings;
use Text::CSV;


=head1 NAME
    X::CSV - parse Comma Seperated Value files


=head1 SYNOPSIS

    #create CSV parse
    my $csvParser = X::CSV->new( data_file => $file_path);

    eval{
       my $data = $csvParser->parse();
    };

    if($@){
       die($@);
    }

sub parse(){
    my $self = shift
    my (@parsed, @csv_source);


    #open the csv file
    open(CSV_SOURCE , "<" . $self->data_file) or die "...";
    ....
    ....
}
[download]

#!/usr/bin/perl
use strict;
use warnings;

use X::CSV;


#Number of tests
use Test::More tests =>3 ;

#test 1 & 2
BEGIN { use_ok( 'X::CSV' ); }
   require_ok( 'X::CSV' ) ;

#parse the test.csv
my $parser = X::CSV->new(
                         data_file => $ENV{TEST_ROOT}.'/t/test.csv'
                         );

my $result = $parser->parse();

my $length = @{$result};

#test 3
#We have an csv file with 3 rows
is($length,'3','We have an sheet with 3 rows');
[download]