A test case for XML validation: a Cocoon sitemap

J.M. Vanel jmvanel@remove-this.free.fr
jeu jan 23 12:42:33 CET 2003

ABSTRACT
I tried several XML Schema validators (XSV, Xerces, Microsoft) 
on a work-in-progress XML Schema for Cocoon sitemap.
Some validators find some errors, some others find other errors, 
and I can not decide wether the Schema nor the document are correct.
But you might find the journey interesting.
I give all the recipies.
This is work-in-progress, comments welcome.

_____________________________________________________

I have a (corrected) XML Schema for Cocoon sitemap here:
http://jmvanel.free.fr/cocoon/sitemap.xsd

It has been taken from the CVS at apache.org:
http://cvs.apache.org/viewcvs.cgi/*checkout*/xml-cocoon2/src/documentation/xdocs/drafts/sitemap-2.1-draft.xsd?rev=HEAD&content-type=text/plain

I have a minimal Cocoon sitemap here:
http://jmvanel.free.fr/cocoon/exemple1/sitemap.xmap

I want to try XML Schema for my small Cocoon sitemap.

After an obvious Schema error (an inner complexType having
 a name line 330), sitemap.xsd passes validation by XSV, 
the tool made by Henry Thompson, XML Schema editor. 
XSV can be used by a Web form at:
http://www.w3.org/2001/03/webdata/xsv

Xerces gives on sitemap.xsd about 10 error messages that maybe are not relevant.
It didn't find the inner complexType having a name. 
But it also found an error that no other processor found: 
maxOccurs="0" (lines 205 and 206) 
(See ANNEXES for details on running Xerces). 
I don't know if it is formally erroneous, 
but it has no common sense to specify something to occur zero time,
better nothing at all.

_____________________________________________________

I continue the investigation with instance validation with XSV, 
the only one that aproves the corrected XML Schema for Cocoon sitemap 
(see below details about installing XSV on Linux).

python $PYTHONPATH/XSV/commandLine.py sitemap.xmap ~/jmvanel.free.fr/cocoon/sitemap.xsd 
<?xml version='1.0'?>
<xsv xmlns='http://www.w3.org/2000/05/xsv'
     docElt="{http://apache.org/cocoon/sitemap/2.0}sitemap"
     instanceAssessed="true" instanceErrors="0" rootType="[Anonymous]"
     schemaDocs="/home/jmv/jmvanel.free.fr/cocoon/sitemap.xsd" schemaErrors="2"
  target="/home/jmv/install/jakarta-tomcat-4.1.12-LE-jdk14/webapps/cocoon/test/sitemap.xmap"
     validation="strict" version="XSV 2.2-3 of 2003/01/20 12:21:00">
  <schemaDocAttempt URI="/home/jmv/jmvanel.free.fr/cocoon/sitemap.xsd"
                    outcome="success" source="command line"/>
  <schemaError char="5" line="397" phase="instance"
               resource="file:///home/jmv/jmvanel.free.fr/cocoon/sitemap.xsd">non-deterministic content model for type pipelineType: {http://apache.org/cocoon/sitemap/2.0}:select/{http://apache.org/cocoon/sitemap/2.0}:select</schemaError>
  <schemaError char="5" line="376" phase="instance"
               resource="file:///home/jmv/jmvanel.free.fr/cocoon/sitemap.xsd">non-deterministic content model for type matchType: {http://apache.org/cocoon/sitemap/2.0}:act/{http://apache.org/cocoon/sitemap/2.0}:act</schemaError>
</xsv>

It appears that on instance validation messages about the Schema alone are produced.
What does this mean ?
It needs further investigation, but it seems that parallel choices 
for matchType lead to what is considered "non-deterministic 
content model" by XSV. The element "act" can appear in several
 parallel choices.

It is possible that sitemap.xsd has to be completely rewriten.
The sitemap.xsd on my site has a few corrections that are without doubt necessary, but nothing more.
TO BE CONTINUED

_____________________________________________________

ANNEXES

http://www.gotdotnet.com/services/xsdvalidator/
GotDotNet XSD Schema Validator

Schema document:
NamespaceURI:
XML document:

Validation error:

The content model must be deterministic. Multiple definition of element 'http://apache.org/cocoon/sitemap/2.0:match' causes content model to become ambiguous. An error occurred at (243, 6).
 The content model must be deterministic. Multiple definition of element 'http://apache.org/cocoon/sitemap/2.0:match' causes content model to become ambiguous. An error occurred at (315, 6).
 The content model must be deterministic. Multiple definition of element 'http://apache.org/cocoon/sitemap/2.0:call' causes content model to become ambiguous. An error occurred at (359, 6).
 The content model must be deterministic. Multiple definition of element 'http://apache.org/cocoon/sitemap/2.0:match' causes content model to become ambiguous. An error occurred at (173, 6).
 The content model must be deterministic. Multiple definition of element 'http://apache.org/cocoon/sitemap/2.0:match' causes content model to become ambiguous. An error occurred at (376, 6).
 The content model must be deterministic. Multiple definition of element 'http://apache.org/cocoon/sitemap/2.0:call' causes content model to become ambiguous. An error occurred at (187, 6).
 The content model must be deterministic. Multiple definition of element 'http://apache.org/cocoon/sitemap/2.0:match' causes content model to become ambiguous. An error occurred at (397, 6).
 The content model must be deterministic. Multiple definition of element 'http://apache.org/cocoon/sitemap/2.0:match' causes content model to become ambiguous. An error occurred at (427, 18).
 Invalid particle derivation by restriction. An error occurred at (201, 6).
 
This page allows you to validate an XML documents against an W3C XML Schema (XSD) or an XML Data-Reduced (XDR) schema, verify the schemas are valid, or simply check the well-formedness of XML documents. Different actions are taken depends on the submitted information.
_________________________________________________________________________

Xerces in command line:

I hacked the sax.Counter sample shipped with Xerces to be able to validate
 against an external Schema (not in <?xml-styleshhet ). 
The source is here:
http://jmvanel.free.fr/xsd/Counter.java
The .class is also here in same directory. 
You can replace the original Counter.java in $XERCES_HOME/samples/sax , 
and download here the class from:
http://jmvanel.free.fr/xsd/Counter.class

Then put in your $HOME/bin a script "xmlvalid" :
java -classpath $XERCES_HOME/xercesImpl.jar:$XERCES_HOME/xmlParserAPIs.jar:$XERCES_HOME/samples sax.Counter $*

Use it this way:

[jmv@protea test]$ xmlvalid -esl http://apache.org/cocoon/sitemap/2.0  ~/contrats/thomasson/sitemap.xsd sitemap.xmap
[Error] sitemap.xsd:427:33: cos-nonambig: "http://apache.org/cocoon/sitemap/2.0":mount and "http://apache.org/cocoon/sitemap/2.0":mount (or elements from their substitution group) violate "Unique Particle Attribution".
[Error] sitemap.xsd:397:41: cos-nonambig: "http://apache.org/cocoon/sitemap/2.0":mount and "http://apache.org/cocoon/sitemap/2.0":mount (or elements from their substitution group) violate "Unique Particle Attribution".
[Error] sitemap.xsd:376:38: cos-nonambig: "http://apache.org/cocoon/sitemap/2.0":mount and "http://apache.org/cocoon/sitemap/2.0":mount (or elements from their substitution group) violate "Unique Particle Attribution".
[Error] sitemap.xsd:359:46: cos-nonambig: "http://apache.org/cocoon/sitemap/2.0":call and "http://apache.org/cocoon/sitemap/2.0":call (or elements from their substitution group) violate "Unique Particle Attribution".
[Error] sitemap.xsd:315:36: cos-nonambig: "http://apache.org/cocoon/sitemap/2.0":parameter and "http://apache.org/cocoon/sitemap/2.0":parameter (or elements from their substitution group) violate "Unique Particle Attribution".
[Error] sitemap.xsd:187:37: cos-nonambig: "http://apache.org/cocoon/sitemap/2.0":call and "http://apache.org/cocoon/sitemap/2.0":call (or elements from their substitution group) violate "Unique Particle Attribution".
[Error] sitemap.xsd:201:53: cos-particle-restrict.2: Forbidden particle restriction:  'choice:all,sequence,elt'.
[Error] sitemap.xsd:201:53: derivation-ok-restriction.5.3: Error for type 'noGenerationPipelineType'.  The particle of the type is not a valid restriction of the particle of the base.
[Error] sitemap.xsd:173:41: cos-nonambig: "http://apache.org/cocoon/sitemap/2.0":mount and "http://apache.org/cocoon/sitemap/2.0":mount (or elements from their substitution group) violate "Unique Particle Attribution".
[Error] sitemap.xsd:243:48: cos-nonambig: "http://apache.org/cocoon/sitemap/2.0":mount and "http://apache.org/cocoon/sitemap/2.0":mount (or elements from their substitution group) violate "Unique Particle Attribution".
sitemap.xmap: 6202 ms (21 elems, 21 attrs, 0 spaces, 251 chars)

_________________________________________________________________________

 Installing XSV on Linux

On RedHat 8.0, I downloaded these RPM as indicated on:
http://www.ltg.ed.ac.uk/%7Eht/xsv-status.html

and did:

rpm -Uvh XSV-2.2-3.noarch.rpm PyLTXML-1.3-2.i386.rpm

then I set the environnment variable:
PYTHONPATH=/usr/lib/python2.2/site-packages