--On Montag, 21. Januar 2002 18:32 -0500 Sam Ruby <[EMAIL PROTECTED]> wrote:
> Andy Clark wrote:
>>
>> We're trying to have a production release of Xerces2 by the
>> end of this month.
>
> IMHO, that would be premature. Take a look at
> http://nagoya.apache.org/~rubys/gump/xerces2/index.html .
>
> Also, the xml-security subproject unit tests which pass with Xerces 1
> cause failures inside the parser with Xerces 2.
>
> http://nagoya.apache.org/~rubys/gump/xerces2/xml-security.html
> http://nagoya.apache.org/~rubys/gump/base/xml-security.html
Sam,
actually, these Xerces2 failures depend on the enabled parser-features and
on the input values. These tests run through the canonicalization test
suite to canonicalize arbitrary XML documents (which itself are the test
vectors from the c14n spec [1]). These examples include 'strange' XML
snippets with partly defined DTDs just to define a default attr for one of
many elements and things like this.
test31 and test31withComments fail:
http://www.w3.org/TR/xml-c14n#Example-OutsideDoc
test35 fails:
http://www.w3.org/TR/xml-c14n#Example-Entities
The three interop tests
IAIKTest.test_coreFeatures_signatures_anonymousReferenceSignature
IAIKTest.test_coreFeatures_signatures_manifestSignature_manifest
IAIKTest.test_coreFeatures_signatures_signatureTypesSignature
fail for similar reasons.
What I need from the parser (I had to search quite long for Xerces1 to find
sufficient settings) is that it does validation as much as possible ;-))
This means that if my DTD only defines one Attribute but I have many
different Elements and Attributes in the XML instance, well, he validates
that Attribute and is silent on all the other things that can't be
validated.
Regards,
Christian
[1] http://www.w3.org/TR/xml-c14n
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
