W3C home > Mailing lists > Public > xmlschema-dev@w3.org > March 2009

Re: invalid within XML-Spy, .NET and Xerces parsers but valid with the W3C validator

From: Kevin Braun <kbraun@obj-sys.com>
Date: Thu, 12 Mar 2009 09:43:24 -0400
Message-ID: <49B9117C.2090606@obj-sys.com>
To: Robert Melskens <robert.melskens@gmail.com>
CC: xmlschema-dev@w3.org
In addition to what Michael already said, here's a little additional 
explanation that points you toward where you run afoul of the spec...

The error Xerces gives is: rcase-Recurse.2: There is not a complete 
functional mapping between the particles.

Xerces seems to be following 2.2 at 
http://www.w3.org/TR/2004/REC-xmlschema-1-20041028/structures.html#cos-particle-restrict. 

It removes the "pointless" sequence, so that comp1A1 becomes:
   <xs:complexType name="complA1">
       <xs:complexContent>
           <xs:restriction base="complA">
               <xs:sequence minOccurs="0">
                       <xs:element name="B" minOccurs="0"/>
                       <xs:element name="C" minOccurs="0"/>
                       <xs:element name="D" minOccurs="0"/>
                       <xs:element name="E" minOccurs="0"/>
               </xs:sequence>
           </xs:restriction>
       </xs:complexContent>
   </xs:complexType>

However, after doing this, applying the rules for a sequence particle 
restricting a sequence particle 
(http://www.w3.org/TR/2004/REC-xmlschema-1-20041028/structures.html#rcase-Recurse) 
leads to the noted failure (element B would map to the original nested 
sequence, elements C..E don't map to anything).

Kevin

On 3/11/2009 4:39 AM, Robert Melskens wrote:
> Hi,
>  
> With the attached XML-Schema, a simplified version of one of our 
> schema's, we had some problems within XML-Spy.
> The schema contains a content-model, which is judged by XML-Spy as 
> invalid.
> I know how to solve this invalidness, but that solution conflicts with 
> the desired model.
>
> I wasn''t able to get a grip on the real source of this 'invalid' 
> message by XML-Spy.
> Why is it invalid to have a minOccurs of 1 on the sequence element? 
> So I decided to try another validator, it might give me more clues.
> I tried '_http://www.w3.org/2001/03/webdata/xsv_' enad surprisingly 
> the schema was judged here as valid.
>
> I concluded XML-Spy would probably have a bug because I put more trust 
> in a W3C validator.
>
> So I sent a bug-report to Altova. A moment ago they answered me. They 
> did test the schema also on a .Net and the Xerces paser and also those 
> parsers reported an error.
> Does this mean the W3C validator has a bug? If so, can you answer the 
> question above. The XML-Schema recommendation doesn't help me much.
>
> I attached the mentioned schema.
>
> Greetings,
>
> Robert Melskens
>
> *Atos Origin Nederland B.V.*
>
Received on Thursday, 12 March 2009 13:44:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 11 January 2011 00:15:11 GMT