HTML::Mason does not depend on mod_perl, but it can use it ;-)

Here's your example rendered through Mason:

#!/usr/bin/perl
use strict;
use warnings;
use HTML::Mason;

sub foo { $_[0] + $_[1] }
my $template = << 'END';
This is some line.
This is a line with some expression: <% foo(3,5) %>
% for my $s qw(this that) {
This is <% $s %> line
% }
This is the end
END

my $output;
my $interp = HTML::Mason::Interp->new( out_method => \$output,
                                       in_package => __PACKAGE__,
                                     );
my $comp = $interp->make_component( comp_source => $template );
$interp->exec($comp);

print $output;
[download]

You mentioned you don't care too much about the syntax, that's why I preferred do not show you how to override Mason's default syntax. Check out http://masonhq.com/docs/manual/ if you're interested.

Thank you for providing the example!!!

-- 
Ronald Fischer <ynnor@mm.st>

Also it looks as if Mason assumes to be embedded in HTML, though perhaps it will work with other text files too...

Most template systems ignore the type of document they are embedded in, mostly it's relevant for choosing the delimiters of the interpreted parts, and for stuff like escaping (which is always optional).

I have no experience with Mason myself, but I'd guess it's similar there (and I'm quite sure there's also a way to use Mason wihout mod_perl. Just look at the test cases, it has to be tested somehow).