filter the content by hooking in

Out of curiosity I looked into this a bit, and it turns out that hacking/hooking into Template::Parser (via Template::Directive, Template::Grammar, or even Parser.yp) is difficult, because it looks like Template::Parser::_parse drops the original source text and doesn't pass it into the handlers. But for a first step, all that's needed are the tokens, which can be provided by Template::Parser::split_text... but careful with the following, I haven't tested with a lot of different cases yet to see if there might be token types this doesn't handle.

#!/usr/bin/env perl
use warnings;
use strict;
use Data::Dump qw/dd pp/;
use Template::Parser;

my $text = <<'END';
before
[% # this is a comment to the end of line
   foo = 'bar'
%]

<p>bw, bliako</p>

[%# placing the '#' immediately inside the directive
    tag comments out the entire directive
%]
[% outside %]
after
END

my $parser = Template::Parser->new();
my $tokens = $parser->split_text($text);
#dd $tokens; # Debug

my $o = '';
for (my $i=0; $i<@$tokens; $i++) {
    if (ref $tokens->[$i]) {
        my $text = $tokens->[$i][0];
        #dd $text; # Debug
        $o .= "[% $text %]";
    }
    elsif ($tokens->[$i] eq 'TEXT') {
        my $text = $tokens->[++$i];
        #dd $text; # Debug
        $o .= $text;
    }
    else { die pp($i,$tokens->[$i]) }
}

print $o;

__END__

before
[% # this is a comment to the end of line
   foo = 'bar' %]

<p>bw, bliako</p>


[% outside %]
after
[download]

Great, i think you just found a reply for Tidy for Template Toolkit Files =)

Anyway ... please stop doing things "out of curiosity" and concentrate on running Perl in the browser! ;-)

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery FootballPerl is like chess, only without the dice}

some 2-second later edits below...

Thanks, your code's great. To do what LanX proposed looked to me too scary(=switch to another task and read their manuals, feel free to downvote my human ingredients). And using regex's is too fragile without knowing the full spec of TT, i.e. are nested comments allowed and how to deal with [% and [%#] and [% # inside strings as Corion said. So let TT do it seems the right way to me.

Edit2: To be fair to the regex solutions: it was me who asked for a regex in the first place.

thanks though that looks a big investment. plus I have my own semantics for language switching on top of TT so If I let TT parse that it may freak out ...

> have my own semantics ... it may freak out

If TT may freak out, how do you expect us to provide a regex doing it ? :)

Good luck! ;)

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery FootballPerl is like chess, only without the dice}