[CLUE-Tech] Strange URLs in a ColdFusion Site

Jed S. Baer thag at frii.com
Thu Apr 1 23:13:23 MST 2004


On Thu, 01 Apr 2004 20:53:17 -0700
David Anselmi <anselmi at anselmi.us> wrote:

> > I think it likely keeps working, as long as all URLs are absolute.
> 
> I don't think this is relevant.  All URLs sent to the server are 
> absolute.  The server doesn't have any idea what page you were on when 
> you clicked a link, and the browser doesn't have any idea what the CGI 
> code looks like.

But in dealing with relative URLs in a page, the browser constructs the
absolute URL using the existing page address, and the supplied URL. So how
does it determine where to cut the existing page address, and append the
string supplied in a relative URL? Until I saw these constructs, I'd have
said, "at the location of the last forward slash". And indeed, that's the
behavior I observe from the Gecko engine (I assume, as both Galeon and
Mozilla do this) when visiting the SSRN site.

OK, so there's at least one exception to this, which is when the relative
URL begins with a forward slash, and so the browser, instead of appending
it to the end of the existing path, appends it to the root of the existing
path. Determining what is the "root" can be difficult for URLs referring
to individual accounts on a hosting service, i.e. the
/server.tld/~username/ path construct. My ISP just switched to
/users.server.tld/username/. I haven't experimented with whether a "/page"
relative URL is interpreted relative to the the FQDN, or the username.

jed
-- 
http://s88369986.onlinehome.us/freedomsight/

... it is poor civic hygiene to install technologies that could someday
facilitate a police state. -- Bruce Schneier



More information about the clue-tech mailing list