Re: URI references [was: XML 1.0 Conformance Test Results

[james anderson <james.anderson@setf.de>]

Richard Tobin wrote:
> 
> > i had understood the file scheme to always require a host
> 
> Definitely not - RFC1738 says:
> 
>    As a special case, <host> can be the string "localhost" or the empty
>    string; this is interpreted as `the machine from which the URL is
>    being interpreted'.
> 

I wasn't referring to this.

> It does not authorize omitting the // too as in file:/dev/null.

This is the issue.

>                                                                 On
> the other hand, the example(s) you quote from the test suite uses it as
> the system id for a notation, which is not required to be dereferenced
> by the XML parser and whose use is implementation defined.  The same
> goes for the use of "scheme:" and "d:", which appear in the system IDs
> of unparsed entities.
> 

Which is to say, that any constraints on the the content of a system
identifier should not be taken to apply to those identifiers which
belong to an unparsed entity?