Incorrect parsing of untagged numerical URLs

[#1]

If I post the untagged text http://192.168.1.1 it comes out as http://192.168.1.1 with the highlighting broken and the link pointing to http://192.0.0.168.
On the other hand:
If I post the tagged text [url]http://192.168.1.1[/url] it comes out as http://192.168.1.1 with highlighting correct and the link correct too.
Since the untagged text http://www.bbc.co.uk is highlighted and linked correctly it seems there's a bug in the parsing of uncoded numerical URLs.

[#2]

Seems the parsing is working if it is ended with a recognised TLD (correct me if I'm wrong)

[#3]

Sorry for the bluntness of my previous post; I was rushing out of the house at the time.
I meant to follow it on with a question to anyone who has seen the source to the forum software since I haven't used it, does automatic URL parsing occur using a conditional statement looking for a recognised TLD suffixed on the end?

[#4]

The first two octets may be assumed to be directory + domain, ie "www.plus". I'll download a copy of it and see if I can locate it.

[#5]

Yea I think that's the culprit. I have it downloaded now but don't have time to look at the code yet long enough to pinpoint any specific problem with it. Will look closer tomorrow and get back to you.

[#6]

I've had a look and I don't see it stripping numbers for any reason. However, to get around this happening, simply type the IP address as this: http://192.168.1.1/ . The trailing forwardslash will cause the link to be parsed correctly.