Is this best for multiline parsing?

Hello,

  I'm working on Mail::Audit so that I can pull out file attachment
names from it.  To do so requires that you look at the headers for a
"boundary" marker to figure out what seperates different attachments.
  This is my first attempt at using RecDescent and would like to know if
this looks like the best way to do it or if someone would offer some
style tips.
  My main inspiration in this script was Damian's post of how to parse a
command seperated over multiple lines.  Except that one was a little
easier than this as it had a line continuation marker at the end.
Looking through some mail headers they don't do this, rather just
indenting the next line a couple of spaces.
  My next question is how to continue on with the rest of the file.  I
don't really care what the file attachment is beyond its listed filename
and what lines it takes up in the ->body() reference to an array returned
by Mail::Audit.  I'm thinking about doing a while() loop till I get to
the boundary marker and then sending the next non-blank lines through
another RecDescent parse.  I don't want to have to tell RecDescent to
look through thousands of lines of a mail file when it is not needed.
  I realize that most of this could be done without RecDescent but
thought I would give it a try.  Maybe I'm a masochist :)

Chris

----------cut----------cut----------cut----------cut----------
#!/usr/bin/perl
 
use Data::Dumper;
use Parse::RecDescent;
use strict;
 
my ($result, $parser, $grammar, $header);
 
$header = getdata();
 
$grammar = join '', <DATA>;
$parser = Parse::RecDescent->new($grammar)
  or die "Bad grammar '$grammar'\n";
 
$result = $parser->headerinfo($header);
 
print Dumper($result);
 
sub getdata {
  my $header = <<'EOF';
Delivered-To: [email protected]
Content-Type: Multipart/Mixed;
  charset="iso-8859-1";
  boundary="------------Boundary-00=_2YQQSZM4Z8EP8VA9F3K8"
From: Joe Blow <[email protected]>
EOF
  return $header;
}
 
__DATA__
  headerinfo:
    parameter(s)
    {  my %return;
       foreach (@{$item{parameter}}) {
         push @{$return{$_->[0]}}, $_->[1];
       }
       { \%return } }
  parameter: key ':' value
      { [ $item{key}, $item{value} ] }
  key:
    /^([^:]+)/ { $1 }
  value:
    multiline_value | singleline_value
    singleline_value:
      /(.*)\n/ { $1 }
    multiline_value:
      /^(.*?)\s*\n(\s+.*\n)+/
        { $item[-1] =~ s/^\s+/ /gm;
          $item[-1] =~ s/\s*\n//gm;
          { $item[-1] } }