If you have a look here you can check out the code of a link checker I wrote. Naturally it needs to crawl an entire site and correctly handle relative links. As you will see it takes lesss than half a dozen lines to do this.

Update

Actually looking at that code there is a subtle bug in it. Here is a complete sub that I will use to modify that code (one day :-)

print rel2abs( 'http://foo.com/aa/bb/cc', './../.././bar.htm' );

sub rel2abs {
    my ($root, $link) = @_;
    $link =~ s|^\s*/||;                 # trim link
    $root .= '/' unless $root =~ m|/$|; # ensure trailing /
    # move back up tree in response to ../ and ignore ./
    while ( $link =~ s|^(\.?\./)|| ) {
      next if $1 eq './';  # effectively just delete ./
        # trim dirs from root only until we can't go any further!
        $root =~ s|[^/]+/$|| unless $root =~ m|http://[^/]+/$|i;
    }
  return $root.$link;
}
[download]

cheers

tachyon

s&&rsenoyhcatreve&&&s&n.+t&"$'$`$\"$\&"&ee&&y&srve&&d&&print