A problem with your regex, as it stands : you need to escape some of those square brackets, as those have special meaning within regexes.

Monolith-O, the syntax you're using is so close to HTML as it is that it doesn't seem to simplify things much. If you can change it, consider doing so. you could do something like what perlmonks does, and let [www.foo.com|the Foo site] stand in for <a href="http://www.foo.com">the Foo site</a>. A regex to match that would be:

$text =~ s#(\[([^|])+\|([^\]]+)]#<a href="http://$1">$2</a>#g;
[download]

Update that's wrong, as $code_or_die points out.

Sticking with the current setup, the first regex should be something like this (untested):

$text =~ s#\[link=(["'])(\S+)\1]([^[])+\[/link]#<a href="$2">$3</a>#g;
[download]

This says "when you find "link=" followed by a single or double quote, grab 1 or more characters that aren't a single or double quote (whichever matched) -- store that in $2 -- followed by the single or double quote, then grab one or more characters that aren't [ and store them in $3, finished up by [/link]. Replace that whole sequence with <a href="$2">$3</a>

Philosophy can be made out of anything. Or less -- Jerry A. Fodor