This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Both of these functions test new-each-time="no", but this is a hint to the processor. I propose to fix to allow the alternate outcome(s).
I disagree. The new-each-time attribute is not a hint, it has prescriptive semantics which these tests are designed to test.
Oh, you are right, I see it now (I was still in the "mode" of the idenity-sensitive hint of previous draft, I think). I see: <quote> DeterministicFO31 functions: as the term is defined in [Functions and Operators 3.1], these offer a guarantee that when a function is called repeatedly with the same arguments, it returns the same results. A classic example is the docFO30 function, which offers the guarantee that doc($X) is doc($X): that is, two calls supplying the same URI return the same node. </quote> which is set with new-each-time="no" (which seems to me that a better word would be new-each-time="never"). So we allow this to blow up if caching is not enough. I have inadvertently already pushed the changes, I will revert them back.
> I will revert them back. Not done yet, apparently. Assigned and on todo-list as discussed at F2F Prague 2nd day.
Done