Another dumb-mode PHP consumer demo

[Alan J Castonguay]

Parsing &#x0000; "&<> in 3 urls: Claimed Ident, openid.delegate, and
openid.server.

http://www.cs.uoregon.edu/~chowells/test.html redirects to
http://www.schtuff.com/?action=trust&cancel_to=http%3A%2F%2Fverselogic.net%2Fopenid%2Fguestbook.php%3Fopenid.mode%3Dcancel&dest=http%3A%2F%2Fwww.schtuff.com%2F%3Faction%3Dopenid_server%26openid.identity%3Dhttp%253A%252F%252Fusers.schtuff.com%252Fcarl%26openid.mode%3Dcheckid_setup%26openid.return_to%3Dhttp%253A%252F%252Fverselogic.net%252Fopenid%252Fguestbook.php%26openid.trust_root%3Dhttp%253A%252F%252Fverselogic.net%252F&identity=http%3A%2F%2Fusers.schtuff.com%2Fcarl&trust_root=http%3A%2F%2Fverselogic.net%2F

Thanks for the help hammering this. Seems I missed quite a few nuances
of the spec.

On 7/19/05, Carl Howells <chowells at janrain.com> wrote:
> Alan J Castonguay wrote:
> > The first was failing due to regex for the trailing slash using the
> > original claimed address rather than the semi-normalized (with http://
> > prefix). Fixed.
> >
> > The third was failing due to extra-required-whitespace regex in my
> > <link> tag parse. Changed the \s+ to \s*. Fixed.
> >
> > href="([^">]+)"\s* [/]?>#ix
> >
> > Any other broken regexen?
> 
> Kind of:  Try this one again:
> 
> http://www.cs.uoregon.edu/~chowells/test.html
> 
> Your parsing is breaking on (legal, and within spec) entities in the
> link tag.  It also needs to resolve them, after it correctly parses them
> out. :)
> 
> Carl
>