Should be fixed in CVS.

This test (K2-SeqExprCast-421) was previously catalogued as expecting error FORG0001. It's not clear to me why it has changed. The test is

xs:anyURI("foo://")

and it seems to me that "foo://" is indeed an invalid URI: RFC3986 states:

When authority is not present, the path cannot begin with two slash characters ("//")

The rule is less clearly expressed in RFC2396 (which is what Schema Part 2 actually references) but I think it's still there. 

Ok, then http://www.w3.org/Bugs/Public/show_bug.cgi?id=5695 was resolved incorrectly. I changed back to the former.

Sorry, I think foo:// is valid.

Quoting the specification:

hier-part   = "//" authority path-abempty
                  / path-absolute
                  / path-rootless
                  / path-empty

   The scheme and path components are required, though the path may be
   empty (no characters).  When authority is present, the path must
   either be empty or begin with a slash ("/") character.  When
   authority is not present, the path cannot begin with two slash
   characters ("//").  These restrictions result in five different ABNF
   rules for a path (Section 3.3), only one of which will match any
   given URI reference.

   The following are two example URIs and their component parts:

         foo://example.com:8042/over/there?name=ferret#nose
         \_/   \______________/\_________/ \_________/ \__/
          |           |            |            |        |
       scheme     authority       path        query   fragment
          |   _____________________|__
         / \ /                        \
         urn:example:animal:ferret:nose

So this is saying that:


                   path-absolute
                   path-rootless
                   path-empty

may not begin with '//'.

path-abempty can begin with '//' even if the authority is the empty string - because the authority is still present.  Note that the empty string is a valid authority (matching the 'reg-name' production).

So foo:// is a valid URI with an empty authority and empty path.  The '//' here is _not_ part of the path.

Here's what Microsoft's .NET System.Uri class does with this...


$  New-object Uri('foo://')


AbsolutePath   : /
AbsoluteUri    : foo:///
Authority      :
Host           :
HostNameType   : Basic
IsDefaultPort  : True
IsFile         : False
IsLoopback     : True
IsUnc          : False
LocalPath      : /
PathAndQuery   : /
Port           : -1
Query          :
Fragment       :
Scheme         : foo
OriginalString : foo://
DnsSafeHost    :
IsAbsoluteUri  : True
Segments       : {/}
UserEscaped    : False
UserInfo       :

And this is what Java 5 (class java.net.URI) has to say about it:

new URI("foo://")

Exception in thread "main" java.net.URISyntaxException: Expected authority at index 6: foo://

These RFCs are a mess: intelligent implementors can interpret the phrase "when authority is present" in different ways. We're not trying to test URI class libraries, we should stick to cases that are uncontroversial.

Interesting!

Where can one get clarification on an RFC?

I agree, the test in CVS now accepts pretty much anything.

Splendid!  That'll kepp everyone happy :)