Tim, as an XML implementor, this is the reason why we need a test
validation suite for XML parsers. That's the best way to ensure that
there's compatibility between the parsers people are writing. I know I
sound like a broken record on this, but we're doing our own test bed, but
that's only guesswork, based on our interpretation of the spec. You and Eve
and a few others *know* what the intention of the spec is.
I'm not sure if the idea is clear... Back when I was a CS grad student,
writing a student compiler, we had a series of test cases that our compiler
had to handle. If we didn't run all the tests we couldn't hand it in.
That's what I'm asking for. A definitive set of test cases that I could
hand off to my engineers and say "make sure we can handle all these cases."
At 05:49 PM 8/5/98 -0700, you wrote:
>At 01:01 AM 8/6/98 +0100, Peter Flynn wrote:
>>ANY means "any element markup", not "anything". If you want the element type
>>to contain both character data and element markup, I think it needs to say
>NO!!!!!!!!!!!!!!! Any means anything. Having said that, this is an
>irritant in the spec. If you read between the lines at the front
>of section 3., it is inescapable that nothing rules out mixed content
>in ANY... but the spec could have avoided immense confusion if it just
>said "ANY means ANYTHING". Once again, the annotated spec at xml.com
>tries to patch this hole. -Tim