W3C home > Mailing lists > Public > xmlschema-dev@w3.org > April 2007

RE: Determinization in case of ambiguity or Should XML validators report on schemas that violate the UPA constraint?

From: Michael Kay <mike@saxonica.com>
Date: Mon, 16 Apr 2007 23:48:44 +0100
To: "'Shlomo Yona'" <S.Yona@F5.com>, <xmlschema-dev@w3.org>
Message-ID: <001201c78079$80df6270$4b0f210a@turtle>
Once I fix your typographical quotation marks to turn them into ASCII
quotation marks, Saxon reports:
 
Error at xsd:schema on line 1 of file:/c:/temp/test.xsd:
  The zero-length string is not a legal namespace URI
Error at xsd:element on line 11 of file:/c:/temp/test.xsd:
  Cannot reference schema components in namespace foo as it has not been
imported
  (more of the same)
Error at xsd:all on line 25 of file:/c:/temp/test.xsd:
  Element <xsd:all> is not allowed as a child of <xsd:sequence>
Error at xsd:all on line 25 of file:/c:/temp/test.xsd:
  Within <xs:all>, an <xs:element> must have @maxOccurs equal to 0 or 1
Error at xsd:all on line 25 of file:/c:/temp/test.xsd:
  Within <xs:all>, an <xs:element> must have @maxOccurs equal to 0 or 1
 
When I fix these problems Saxon actually reports both your schema and the
instance as valid. This is because Saxon doesn't detect a UPA in the
situation where you have two element particles that both refer to the same
element declaration - if Saxon can determine unambiguously which element
declaration an element refers to, it doesn't report any ambiguity. A known
minor departure from the spec.
 
Michael Kay
http://www.saxonica.com/


  _____  

From: xmlschema-dev-request@w3.org [mailto:xmlschema-dev-request@w3.org] On
Behalf Of Shlomo Yona
Sent: 16 April 2007 08:57
To: xmlschema-dev@w3.org
Subject: Determinization in case of ambiguity or Should XML validators
report on schemas that violate the UPA constraint?



Hello,

 

In the following example schema the element <b/> can appear either as part
of a sequence model group or as part of a choice model group or as an all
model group.

This means that the instance does not follow the Unique Particle Attribution
Constraint (see: "Section 3.8.6: Constraints on Model Group Schema
Components " and Appendix H in "XML Schema Part 1: Structures Second
Edition, W3C Proposed Edited Recommendation").

 

<?xml version="1.0" encoding="UTF-8"?>

<xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema" targetNamespace=""
xmlns="foo">

            <xsd:element name="a">

                        <xsd:complexType>

                                    <xsd:sequence minOccurs="0"
maxOccurs="2">

                                                <xsd:sequence minOccurs="0"
maxOccurs="2">

                                                            <xsd:element
ref="b" minOccurs="0" maxOccurs="2"/>

                                                            <xsd:element
ref="c" minOccurs="0" maxOccurs="2"/>

                                                </xsd:sequence>

                                                <xsd:choice minOccurs="0"
maxOccurs="2">

                                                            <xsd:element
ref="b" minOccurs="0" maxOccurs="2"/>

                                                            <xsd:element
ref="d" minOccurs="0" maxOccurs="2"/>

                                                </xsd:choice>

                                                <xsd:all minOccurs="0"
maxOccurs="1">

                                                            <xsd:element
ref="e" minOccurs="0" maxOccurs="2"/>

                                                            <xsd:element
ref="b" minOccurs="0" maxOccurs="2"/>

                                                </xsd:all>


                                    </xsd:sequence>

                        </xsd:complexType>

            </xsd:element>

            <xsd:element name="b">

                        <xsd:complexType/>

            </xsd:element>

            <xsd:element name="c">

                        <xsd:complexType/>

            </xsd:element>

            <xsd:element name="d">

                        <xsd:complexType/>

            </xsd:element>

            <xsd:element name="e">

                        <xsd:complexType/>

            </xsd:element>

</xsd:schema>

 

 

My question is: should the following example be valid against the schema or
not valid?

 

<a>

   <b/><c/><b/>

</a>

 

 

I see that different XML validators behave differently on this example.
However, I did not see them report an error in the schema.

 

Now, please consider the following schema:

 

<?xml version="1.0" encoding="UTF-8"?>

<xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema"
targetNamespace="bar" xmlns="foo">

            <xsd:element name="a">

                        <xsd:complexType>

                                    <xsd:sequence>

                                                <xsd:sequence minOccurs="0">

                                                            <xsd:element
name="b">

 
<xsd:complexType/>

                                                            </xsd:element>

                                                </xsd:sequence>

                                    </xsd:sequence>

                        </xsd:complexType>

            </xsd:element>

</xsd:schema>  

 

Instance:

 

<a/>

 

Is it valid?

 

 

See, the element b might not appear and that's OK if you ask the inner
sequence but the outer sequence needs to "happen" once (the minOccurs="1" is
implied).

 

So, how should I read this? Is the above single <a/> element instance valid
or not with this schema? Is such a schema legal or not? If the schema is not
legal, must a schema processor report an error?

 

Thanks.

 

Shlomo.

 
Received on Monday, 16 April 2007 22:50:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 11 January 2011 00:14:59 GMT