Re: lynx-dev dotless IP-address in URLs

[Klaus Weide]

On Sun, 22 Aug 1999, Gisle Vanem wrote:

> I've noted some web-owners use dot-less IP-adresses in their URLs for
> some reason. To make it harder for non-Windows users to connect maybe?

To hide.

I've yet to see any site to use this for a legitimate reason.  I bet here
are none (sites or reasons).

> Do you think it could be a job for Lynx to try to decode these to a
> "dot address" before connecting.

No, leave it to the name resolver where it belongs.  The only form of IP
address defiend for ULS is the dotted quad.  Therefore everything else
should be fed to gethostbyname() not inet_addr().  Some resolver libraries
may interpret such strings as dot-less IP-adresses - it's up to them to
be broken.  If anything, http://NNNNNN/ and ftp://NNNNNN/ etc. should
stand for host NNNNNN in the local domain (or search list as configured),
digit-only hostnames _are_ valid for DNS after all.

> I don't know if other besides Winsock's gethostbyname() or inet_addr()
> that behave this way, but I'm sure it's not an universally accepted
> form.
> 
>         URL: http://3480473412/leadinc/

A ping or traceroute should show you the "real" address, if you don't
feel like decoding it by hand.  Have a look some at some of those pages
Not worth adding a single line of lynx code for those sites, who after
after all have _intentionally broken_ addresses, IMNSHO.

This has been discussed before; sorry, no pointers.


   Klaus