To chime in with the others, you shouldn't parse HTML by hand. There's so much possible variation that you can't possibly get it right in any reasonable timeframe with simple methods (like a couple regexen). Your problem that a tag may span multiple lines is just one of a myriad of cases. You may soon run across a case where the name you're looking for is two lines down from the start of the tag; then you'll have to readjust everything all over. Save yourself the headache of writing a parser and your users the headache of working with a simpleminded one, and just use the appropriate modules.

To answer your question though, it's simply a matter of keeping a copy of the current line around, and going through the file line by line.

my $buffer = <FH>;

my $line;
while(<FH>) {
    $buffer .= $line = $_;

    # ..
    # ..

    $buffer = $line;
}
[download]

Makeshifts last the longest.