Re: URI Reference creation

Mike Brown wrote:
> Sebastian Pipping wrote:
>> Yes, but shouldn't the host "A" become lowercase and "%7b" and
>> "%7d" become uppercase?
> In our API, the different types of normalization are performed by separate 
> functions. The test cases for the percent-encoding normalization are only for 
> the NormalizePercentEncoding function that does what's described in section
> of the spec. There is a separate function and test data for the 
> NormalizeCase function corresponding to sec.

I see. Should have figured that out myself, sorry.

>> I meant that maybe "a/b///c" and "a/b/c" are different
>> for some applications and therefore fixing it would
>> not work for them.
> You have origin URI foo and target URI bar, and one or the other contains 
> empty segments. The function gives you relative ref baz, which works (gets you 
> from the origin to the target) regardless of how many empty segments there 
> are. *Not* fixing it means you might be producing relative refs that, when 
> resolved against the origin, result in something other than the target URI.

Not sure what you mean. Can you give a concrete example?


Received on Wednesday, 1 August 2007 01:24:32 UTC