Re: URI Reference creation

Sebastian Pipping wrote:
> Mike Brown wrote:
> > Test cases are in 
> > http://cvs.4suite.org/viewcvs/4Suite/test/Lib/test_uri.py?view=markup.
> 
> ---------------------------------------------------------------------------
> I found a test case that seems strange to me.
> It's one of the cases in absolutize_test_cases:
> 
>    ('.//g', 'f:/a', 'f://g')
> 
> In the reference "g" is the name of a path
> while it becomes the host part in the resolved URI.
> Can this be correct?

That's Relative84 from Graham Klyne's tests at 
http://www.ninebynine.org/Software/HaskellUtils/Network/UriTest.xls

A bit of Googling reveals Rob Cameron brought up the issue before:
http://lists.w3.org/Archives/Public/uri/2003Jun/0037.html

Roy Fielding said it's not an issue:
http://lists.w3.org/Archives/Public/uri/2003Jun/0040.html

Rob demonstrated that it was, and suggested that the algorithm be
changed to better handle it:
http://lists.w3.org/Archives/Public/uri/2003Jun/0043.html

I don't see any evidence that it went any further than this.