Re: testcases for optional axes?

Kay, Michael wrote:

> I don't have (or know of) any tests that don't depend on an XSLT 
> harness, but if you write your tests as XPath 1.0 expressions then you 
> could verify the results by putting the same expressions through an XSLT 
> processor or a free-standing XPath processor.

That's a good idea.  Though I'm worried about tricky corner cases I
may not have thought of.

So far I've written 7 little tests for the additional axes, and they
at least provide a sanity test.  (If anyone is curious, they're in the
Kawa CVS repository in gnu/xquery/testsuite/ along with
a bunch of other XQuery tests.)
	--Per Bothner

Received on Monday, 15 September 2003 15:31:13 UTC