Questions about Schema Support in Xerces 1.60

[Elisha Berns <e....@computer.org>]

Hi,

I have been attempting to use schema validation in Xerces 1.60 (C++ version
of course) and there are several things that seem to not be implemented or
implemented in a partial manner:

*	<import namespace=”…” … schemaLocation=”…”> doesn’t appear to work at all.
*	<include schemaLocation=”…”> appears to only work when the order of the
include statements is in the same order as the dependency order of the
declarations inside the file(s) included much in the same way as .h files
needs to be #included in the order of dependencies in C++.  If you include a
schema document that attempts to reference schema declarations that are
found in a document included later, the validator appears to not recognize
the declarations and the validation fails.
*	The validator doesn’t like complexType elements of complexContent in which
some of the member elements have a “minOccurs=0” attribute.  It seems to
expect the actual element in an instance document nevertheless.  So, the
minOccurs attribute isn’t implemented?

Can someone comment briefly on the status of these problems and where Xerces
C++ will be headed with the next release or bug-fixes?  Thanks for any help
you can be with these issues.

Yours truly,

Elisha Berns

e.berns@computer.org

RE: Questions about Schema Support in Xerces 1.60

[Elisha Berns <eb...@sbcglobal.net>]

Hi Khaled,

Thanks for the reply.  Here is the zip file with all of the schema files I
am using and the sample XML instance file and error logs.

I ran the test with two different command lines, and they produced different
errors.  For the first test the command line was:

DOMPrint.exe -v=auto  -s -f -e –n descriptionExample016.xml

See ErrorLog-1.txt for the results.

For the second test I ran:

DOMPrint.exe -v=auto -s -e –n descriptionExample016.xml

See ErrorLog-2.txt for the results.

The results in the first test regarding the ##any and <some schema item>
violating unique particle attribution wasn’t clear at all.  Perhaps if
possible you wouldn’t mind explaining the message in English.  The first
test always produces a stack overflow which isn’t clear either.  Can you see
why the –f “Full Schema Validation” would create a stack overflow?

The <import> directive specifically applies to the use of the xml:lang
attribute which doesn’t appear to be recognized or work yet.

Thanks for any help you can be on these matters.




-----Original Message-----
From: Khaled Noaman [mailto:knoaman@ca.ibm.com]
Sent: Wed, January 23, 2002 07:17
To: xerces-c-dev@xml.apache.org
Subject: Re: Questions about Schema Support in Xerces 1.60

Hi Elisha,
The C++ parser has support for the W3C schema language except for some
limitations as noted in http://xml.apache.org/xerces-c/schema.html.
o The problem with <include> directive is one of the limitations and will be
fixed in the next release.
o The <import> directive is supported. Would you please provide an example?
o The minOccurs/maxOccurs attributes are implemented.You may have found a
problem with complex types with elements having minOccurs = 0. A sample
schema would be helpful.
If you find any problems with the parser, you can use bugzilla (
http://nagoya.apache.org/bugzilla) to open a bug against the C++ parser.
Regards,
Khaled
Elisha Berns wrote:
Hi,
I have been attempting to use schema validation in Xerces 1.60 (C++ version
of course) and there are several things that seem to not be implemented or
implemented in a partial manner:
*           <import namespace=”…” … schemaLocation=”…”> doesn’t appear to
work at all.
*           <include schemaLocation=”…”> appears to only work when the order
of the include statements is in the same order as the dependency order of
the declarations inside the file(s) included much in the same way as .h
files needs to be #included in the order of dependencies in C++. If you
include a schema document that attempts to reference schema declarations
that are found in a document included later, the validator appears to not
recognize the declarations and the validation fails.
*           The validator doesn’t like complexType elements of
complexContent in which some of the member elements have a “minOccurs=0”
attribute.  It seems to expect the actual element in an instance document
nevertheless. So, the minOccurs attribute isn’t implemented?
Can someone comment briefly on the status of these problems and where Xerces
C++ will be headed with the next release or bug-fixes?Thanks for any help
you can be with these issues.
Yours truly,
Elisha Berns
e.berns@computer.org

Re: Questions about Schema Support in Xerces 1.60

[Khaled Noaman <kn...@ca.ibm.com>]

Hi Elisha,

The C++ parser has support for the W3C schema language except for some
limitations as noted in http://xml.apache.org/xerces-c/schema.html.

o The problem with <include> directive is one of the limitations and
will be fixed in the next release.
o The <import> directive is supported. Would you please provide an
example?
o The minOccurs/maxOccurs attributes are implemented.You may have found
a problem with complex types with elements having minOccurs = 0. A
sample schema would be helpful.

If you find any problems with the parser, you can use bugzilla
(http://nagoya.apache.org/bugzilla) to open a bug against the C++
parser.

Regards,
Khaled

Elisha Berns wrote:

> Hi,
>
> I have been attempting to use schema validation in Xerces 1.60 (C++
> version of course) and there are several things that seem to not be
> implemented or implemented in a partial manner:
>
>    * <import namespace=”…” … schemaLocation=”…”> doesn’t appear to
>      work at all.
>    * <include schemaLocation=”…”> appears to only work when the order
>      of the include statements is in the same order as the dependency
>      order of the declarations inside the file(s) included much in the
>      same way as .h files needs to be #included in the order of
>      dependencies in C++. If you include a schema document that
>      attempts to reference schema declarations that are found in a
>      document included later, the validator appears to not recognize
>      the declarations and the validation fails.
>    * The validator doesn’t like complexType elements of complexContent
>      in which some of the member elements have a “minOccurs=0”
>      attribute.  It seems to expect the actual element in an instance
>      document nevertheless. So, the minOccurs attribute isn’t
>      implemented?
>
> Can someone comment briefly on the status of these problems and where
> Xerces C++ will be headed with the next release or bug-fixes?Thanks
> for any help you can be with these issues.
>
> Yours truly,
>
> Elisha Berns
>
> e.berns@computer.org
>