Automatic link detection fails miserably on urls without http:// prefix

The regexp used in automatic link detection is flawed.

It fails to detect addresses without http:// prefix which are probably not in .com or .org zone, will not be marked as a link, and it thinks that anything before .org and until a space is a link, thus if I have a Markdown tag [WHATWG Web Site|whatwg.org], the link will point at "Web Site]%28whatwg.org".