This can mean anywhere that you need to read it from a filehandle, or get it as a return from some other function. For example, you could have it in an actual file. You could put it in a __DATA__ or __END__ block at the end of your file and read it from the *DATA filehandle. You could retrieve it using LWP from a webserver somewhere, then parse it from that. As long as it isn't contained as text inside your actual perl code where the perl compiler will see it.

Here's an example:

#!/usr/bin/perl -w

use strict;
use Parse::RandGen::Regexp;

my $regexp = <DATA>;
chomp $regexp; # get rid of the line termination.
my $r = Parse::RandGen::Regexp->new(qr/$regexp/);
my $string = $r->pick(match=>1, captures=>{}); 
print("\$string: $string\n");
__END__
(?smi)^STOR\s[^\n]{100}
[download]

Here we're loading it from the __END__ section. For more info on how to use the DATA filehandle, see perldata. Otherwise, you can take the string, put it in a file, and then open/read/close that file as normal.

Shouldn't that be

(?smi:^STOR\s[^\n]{100})
[download]

You made me go back to perlre again - I knew I had looked it up when giving the original answer, but you made me second guess myself. ;-) Turns out, we're both right. Without the colon, the directive applies to the end of the current block (capturing or non-capturing parenthesis). With the colon, of course, it applies only to that non-capturing block.

Cool thanks. :-)