You're right. One problem is that opening paren, which doesn't have a match; another one is that I mistyped ")+" instead of "+) -- otherwise we capture one or more character next to a pipe ... that oughta learn me =)

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

It's funny, in browsing this site and answering questions, you often pick up litle snippets that you want to include in some of your code somewhere. I used the regex that I modified to optionally allow "http://".

Then I got to thinking - what if you want to type "https://" at the beginning?

In my regex - it would output the links as:
http://https://www.whatever.com
Which is obviously no good.

I fiddled around with it, but then got stuck and just worked it out (I think):

$text =~  s#\[(http[s]?://)?([^\|]+)\|([^\]]+)\]#<a href="$1$2">$3
+</a>#g;
[download]

That'll work, but I do want to point out one thing: you don't need a character class for a single character. :) $text =~  s#\[(https?://)?([^\|]+)\|([^\]]+)\]#<a href="$1$2">$3</a>#g;