Url regex is not quite right

[thestinger]

Url regex is not quite right

[vodik]

Oops, wrong bug.

[thestinger]

It includes matching for IPv6 hosts, and I think that's what leads to a lot of false positives.

[gtmanfred]

also fails with if they put ) at the end of it, it should match (with the )... http://pastebin.com/NFWnLnqq)

i will mess with this tonight

[thestinger]

Well, parentheses in URLs is perfectly legal so in theory it should always match those. However, then it no longer matches a URL in parentheses correctly (http://www.example.com/). That's easy to deal with though... at the very least this can be incrementally improved.

[vodik]

Honestly, I don't think we'll ever be able to do this 100% without moving to something slightly more powerful then regular expressions.

[thestinger]

This is pretty much as good as it can be now. I simply copied the most generic definition of a URL from the RFC, which will work for http, ftp, ssh, file, etc.

However, a valid url in the http scheme isn't necessarily on in the ssh scheme - that's not something can be dealt with because there are way too many schemes.