Re: keyref not finding all bad data elements at once
Hi Guillaume, First of all, this implementation behavior could not be a bug with a XML Schema engine, as XSD specification doesn't recommend anything about details of the error reporting, upon XML Schema validation failure. To my opinion, at the least XSD implementations are expected to report validation success or failure (i.e, a kind of 'true' or 'false' validation outcome) after a XML schema validation episode. How validation error messages are constructed, and how much of error reporting is performed by schema engines, is implementation dependent. In this particular case, reporting all instances of identity-constraint violations may be a nice to have feature (but IMHO, it's still debatable, whether this is a best design in this case), but not doing so, doesn't make Xerces non-compliant to XSD spec :) It seems to me, that seeing all of identity-constraint violations if error instances are as less as 2-3, may be good to see. But if let's say we make the error list unbounded, producing all the error instances (imagine this list size to be say, 500!) could be unpleasant to see. On Fri, May 21, 2010 at 2:43 PM, Guillaume Deshors guilla...@deshors.net wrote: Hi I am facing exactly the issue reported by this oxygen user : http://www.oxygenxml.com/forum/topic2869.html?sid=a4b044cdcc3c06ae07059f924156f6d3 When using several keyrefs causing errors, Xerces reports only the first and then gives up. I would really like to have Xerces report all foreign key errors, like I've seen with the saxon or libXml parsers. Is there a way to do that ? I've done a few hours of search and I've read that it is a fatal error to Xerces and that's why it stops. I've tried to set continue-after-fatal-error to true but it didn't change anything ; anyway I've read that it's not recommended to set this option in a production environment. I'm joining a sample XML and XSD files that show my problem. In my example, each ASSREGARD references a ASSFILDO, I've created two bad ASSREGARD that reference non-existent ASSFILDO. Then Xerces reports only : Description: [Xerces] cvc-identity-constraint.4.3: Key 'FK_ASSFILDOREGARD' with value 'non-existent1' not found for identity constraint of element 'FeatureCollection'. URL: http://www.w3.org/TR/xmlschema-1/#cvc-identity-constraint Instead of this I would like it to report both errors, on non-existent1 and non-existent2. Thank you very much for any support ! Regards, Guillaume Here are the two files : keyrefDemo.xml ?xml version=1.0 encoding=UTF-8? glml:FeatureCollection xmlns:gml=http://www.opengis.net/gml; xmlns:glml=http://www.grandlyon.com/glml; xmlns:xsi=http://www.w3.org/2001/XMLSchema-instance; xsi:schemaLocation=http://www.grandlyon.com/glml ./keyrefDemo.xsd glml:ASSFILDO glml:IID_IDENTFILDO28532/glml:IID_IDENTFILDO /glml:ASSFILDO glml:ASSREGARD glml:IID_IDENTREGARD23797/glml:IID_IDENTREGARD glml:IID_ASSFILDO28532/glml:IID_ASSFILDO /glml:ASSREGARD glml:ASSREGARD glml:IID_IDENTREGARDbad1/glml:IID_IDENTREGARD glml:IID_ASSFILDOnon-existent1/glml:IID_ASSFILDO /glml:ASSREGARD glml:ASSREGARD glml:IID_IDENTREGARDbad2/glml:IID_IDENTREGARD glml:IID_ASSFILDOnon-existent2/glml:IID_ASSFILDO /glml:ASSREGARD /glml:FeatureCollection - keyrefDemo.xsd ?xml version=1.0 encoding=UTF-8 ? schema targetNamespace=http://www.grandlyon.com/glml; xmlns:glml=http://www.grandlyon.com/glml; xmlns:xlink=http://www.w3.org/1999/xlink; xmlns=http://www.w3.org/2001/XMLSchema; elementFormDefault=qualified version=0.9 !-- element -- element name=FeatureCollection complexType sequence element name=ASSFILDO maxOccurs=unbounded complexType sequence element name=IID_IDENTFILDO type=string / /sequence /complexType /element element name=ASSREGARD maxOccurs=unbounded complexType sequence element name=IID_IDENTREGARD type=string / element name=IID_ASSFILDO type=string / /sequence /complexType /element /sequence /complexType key name=PK_ASSFILDO selector xpath=glml:ASSFILDO/ field xpath=glml:IID_IDENTFILDO/ /key key name=PK_ASSREGARD selector xpath=glml:ASSREGARD/ field xpath=glml:IID_IDENTREGARD/ /key keyref name=FK_ASSFILDOREGARD refer=glml:PK_ASSFILDO selector xpath=glml:ASSREGARD/ field xpath=glml:IID_ASSFILDO/ /keyref /element /schema -- Regards, Mukul Gandhi
Re: keyref not finding all bad data elements at once
Hi Guillaume,
Which API are you using and did you register an error handler?
The default error handler for a JAXP Validator is defined as:?
class DraconianErrorHandler implements ErrorHandler {
public void fatalError( SAXParseException e ) throws SAXException {?
throw e;
}
public void error( SAXParseException e ) throws SAXException {
throw e;
}
public void warning( SAXParseException e ) throws SAXException {
// noop
}
}
If you're using this API and didn't supply your own error handler, the
validator will always stop on the first error.
Thanks.
[1]
http://xerces.apache.org/xerces2-j/javadocs/api/javax/xml/validation/Validator.html#setErrorHandler
(org.xml.sax.ErrorHandler)
Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: mrgla...@ca.ibm.com
E-mail: mrgla...@apache.org
guillaume.desh...@gmail.com wrote on 05/21/2010 05:13:25 AM:
Hi
I am facing exactly the issue reported by this oxygen user : http://
www.oxygenxml.com/forum/topic2869.html?sid=a4b044cdcc3c06ae07059f924156f6d3
When using several keyrefs causing errors, Xerces reports only the
first and then gives up.
I would really like to have Xerces report all foreign key errors,
like I've seen with the saxon or libXml parsers. Is there a way to do
that ?
I've done a few hours of search and I've read that it is a fatal
error to Xerces and that's why it stops. I've tried to set continue-
after-fatal-error to true but it didn't change anything ; anyway
I've read that it's not recommended to set this option in a
production environment.
I'm joining a sample XML and XSD files that show my problem. In my
example, each ASSREGARD references a ASSFILDO, I've created two bad
ASSREGARD that reference non-existent ASSFILDO. Then Xerces reports
only :
Description: [Xerces] cvc-identity-constraint.4.3: Key
'FK_ASSFILDOREGARD' with value 'non-existent1' not found for
identity constraint of element 'FeatureCollection'.
URL: http://www.w3.org/TR/xmlschema-1/#cvc-identity-constraint
Instead of this I would like it to report both errors, on non-
existent1 and non-existent2. Thank you very much for any support !
Regards, Guillaume
Here are the two files :
keyrefDemo.xml
?xml version=1.0 encoding=UTF-8?
glml:FeatureCollection
xmlns:gml=http://www.opengis.net/gml;
xmlns:glml=http://www.grandlyon.com/glml;
xmlns:xsi=http://www.w3.org/2001/XMLSchema-instance;
xsi:schemaLocation=http://www.grandlyon.com/glml ./keyrefDemo.xsd
glml:ASSFILDO
glml:IID_IDENTFILDO28532/glml:IID_IDENTFILDO
/glml:ASSFILDO
glml:ASSREGARD
glml:IID_IDENTREGARD23797/glml:IID_IDENTREGARD
glml:IID_ASSFILDO28532/glml:IID_ASSFILDO
/glml:ASSREGARD
glml:ASSREGARD
glml:IID_IDENTREGARDbad1/glml:IID_IDENTREGARD
glml:IID_ASSFILDOnon-existent1/glml:IID_ASSFILDO
/glml:ASSREGARD
glml:ASSREGARD
glml:IID_IDENTREGARDbad2/glml:IID_IDENTREGARD
glml:IID_ASSFILDOnon-existent2/glml:IID_ASSFILDO
/glml:ASSREGARD
/glml:FeatureCollection
- keyrefDemo.xsd
?xml version=1.0 encoding=UTF-8 ?
schema targetNamespace=http://www.grandlyon.com/glml; xmlns:glml=
http://www.grandlyon.com/glml;
xmlns:xlink=http://www.w3.org/1999/xlink;
xmlns=http://www.w3.org/2001/XMLSchema;
elementFormDefault=qualified version=0.9
!-- element --
element name=FeatureCollection
complexType
sequence
element name=ASSFILDO maxOccurs=unbounded
complexType
sequence
element name=IID_IDENTFILDO
type=string /
/sequence
/complexType
/element
element name=ASSREGARD maxOccurs=unbounded
complexType
sequence
element name=IID_IDENTREGARD
type=string /
element name=IID_ASSFILDO type=string /
/sequence
/complexType
/element
/sequence
/complexType
key name=PK_ASSFILDO
selector xpath=glml:ASSFILDO/
field xpath=glml:IID_IDENTFILDO/
/key
key name=PK_ASSREGARD
selector xpath=glml:ASSREGARD/
field xpath=glml:IID_IDENTREGARD/
/key
keyref name=FK_ASSFILDOREGARD refer=glml:PK_ASSFILDO
selector xpath=glml:ASSREGARD/
field xpath=glml:IID_ASSFILDO/
/keyref
/element
/schema
Re: keyref not finding all bad data elements at once
Hi Mukul Thanks for your answer. Understand me correctly, I'm not reporting this as a possible bug ; this behavior is understandable but I hoped there would be a way to nevertheless get all occurrences of this error. In my case it is necessaray, let me explain why : the system we're building validates the same XML three times against three different XSD, which are basically the same but with a growing severity ; each one contains more checks than the previous. The last XSD contains validations that are functionnaly only considered warnings for our application, and that's where foreign keys lie. The warnings can be ignored to process the step anyway, but it's definitely not identical to the user if there's one warning or one hundred. And in that case I would be glad that it's unpleasant to see ! I'll check the other answer, thank you again. 2010/5/21 Mukul Gandhi muk...@apache.org Hi Guillaume, First of all, this implementation behavior could not be a bug with a XML Schema engine, as XSD specification doesn't recommend anything about details of the error reporting, upon XML Schema validation failure. To my opinion, at the least XSD implementations are expected to report validation success or failure (i.e, a kind of 'true' or 'false' validation outcome) after a XML schema validation episode. How validation error messages are constructed, and how much of error reporting is performed by schema engines, is implementation dependent. In this particular case, reporting all instances of identity-constraint violations may be a nice to have feature (but IMHO, it's still debatable, whether this is a best design in this case), but not doing so, doesn't make Xerces non-compliant to XSD spec :) It seems to me, that seeing all of identity-constraint violations if error instances are as less as 2-3, may be good to see. But if let's say we make the error list unbounded, producing all the error instances (imagine this list size to be say, 500!) could be unpleasant to see.
Re: keyref not finding all bad data elements at once
Hi
Yes that's what I did, though it's not directly the class Validator class
that's directly involved. I used a SAXParser on which I set a custom
DefaultHandler, like this :
SAXParserFactory spf = SAXParserFactory.newInstance();
spf.setNamespaceAware(true);
spf.setValidating(true);
SAXParser sp = spf.newSAXParser();
sp.setProperty(JAXP_SCHEMA_LANGUAGE, W3C_XML_SCHEMA);
sp.setProperty(JAXP_SCHEMA_SOURCE, file: + xsdUri);
sp.parse(xml, new MessageContentHandler(type));
Note : I've checked in the code of SaxParser (
http://kickjava.com/src/javax/xml/parsers/SAXParser.java.htm ) and it
affects the HandlerBase as the ErrorHandler.
Here is my custom class MessageContentHandler (part of it) :
private class MessageContentHandler extends DefaultHandler {
private String type;
private int niveau = 0;
private int numBalise = 0;
public MessageContentHandler(String type) {
super();
this.type = type;
}
public void warning(SAXParseException exception) throws SAXException
{
creerAlerte(exception.getLocalizedMessage(), type,
getNumBalise());
}
public void fatalError(SAXParseException exception) throws
SAXException {
creerAlerte(exception.getLocalizedMessage(), type,
getNumBalise());
}
public void error(SAXParseException exception) throws SAXException {
creerAlerte(exception.getLocalizedMessage(), type,
getNumBalise());
}
(snip)
Please note that the method creerAlerte doesn't throw any exception, it
just registers the messages.
Hoping that you have any hint...
Regards,
Guillaume
2010/5/21 Michael Glavassevich mrgla...@ca.ibm.com
Hi Guillaume,
Which API are you using and did you register an error handler?
The default error handler for a JAXP Validator is defined as:
class DraconianErrorHandler implements ErrorHandler {
public void fatalError( SAXParseException e ) throws SAXException {
throw e;
}
public void error( SAXParseException e ) throws SAXException {
throw e;
}
public void warning( SAXParseException e ) throws SAXException {
// noop
}
}
If you're using this API and didn't supply your own error handler, the
validator will always stop on the first error.
Thanks.
[1]
http://xerces.apache.org/xerces2-j/javadocs/api/javax/xml/validation/Validator.html#setErrorHandler(org.xml.sax.ErrorHandler)http://xerces.apache.org/xerces2-j/javadocs/api/javax/xml/validation/Validator.html#setErrorHandler%28org.xml.sax.ErrorHandler%29
Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: mrgla...@ca.ibm.com
E-mail: mrgla...@apache.org
guillaume.desh...@gmail.com wrote on 05/21/2010 05:13:25 AM:
Re: keyref not finding all bad data elements at once
You're right. Only one error message is reported. It's been a very long
time (years) since I've looked at identity constraints in Xerces. Couldn't
tell you why it was implemented that way but that seems to be the
behaviour. It isn't wrong. Xerces would be producing the correct PSVI (i.e.
same error code applies for all the keyref errors). It's just less helpful
for troubleshooting.
Thanks.
Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: mrgla...@ca.ibm.com
E-mail: mrgla...@apache.org
guillaume.desh...@gmail.com wrote on 05/21/2010 12:48:15 PM:
Hi
Yes that's what I did, though it's not directly the class Validator
class that's directly involved. I used a SAXParser on which I set a
custom DefaultHandler, like this :
SAXParserFactory spf = SAXParserFactory.newInstance();
spf.setNamespaceAware(true);
spf.setValidating(true);
SAXParser sp = spf.newSAXParser();
sp.setProperty(JAXP_SCHEMA_LANGUAGE, W3C_XML_SCHEMA);
sp.setProperty(JAXP_SCHEMA_SOURCE, file: + xsdUri);
sp.parse(xml, new MessageContentHandler(type));
Note : I've checked in the code of SaxParser ( http://kickjava.com/
src/javax/xml/parsers/SAXParser.java.htm ) and it affects the
HandlerBase as the ErrorHandler.
Here is my custom class MessageContentHandler (part of it) :
private class MessageContentHandler extends DefaultHandler {
private String type;
private int niveau = 0;
private int numBalise = 0;
public MessageContentHandler(String type) {
super();
this.type = type;
}
public void warning(SAXParseException exception) throws
SAXException {
creerAlerte(exception.getLocalizedMessage(), type,
getNumBalise());
}
public void fatalError(SAXParseException exception) throws
SAXException {
creerAlerte(exception.getLocalizedMessage(), type,
getNumBalise());
}
public void error(SAXParseException exception) throws
SAXException {
creerAlerte(exception.getLocalizedMessage(), type,
getNumBalise());
}
(snip)
Please note that the method creerAlerte doesn't throw any
exception, it just registers the messages.
Hoping that you have any hint...
Regards,
Guillaume
2010/5/21 Michael Glavassevich mrgla...@ca.ibm.com
Hi Guillaume,
Which API are you using and did you register an error handler?
The default error handler for a JAXP Validator is defined as:
class DraconianErrorHandler implements ErrorHandler {
public void fatalError( SAXParseException e ) throws SAXException {
throw e;
}
public void error( SAXParseException e ) throws SAXException {
throw e;
}
public void warning( SAXParseException e ) throws SAXException {
// noop
}
}
If you're using this API and didn't supply your own error handler,
the validator will always stop on the first error.
Thanks.
[1] http://xerces.apache.org/xerces2-j/javadocs/api/javax/xml/
validation/Validator.html#setErrorHandler(org.xml.sax.ErrorHandler)
Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: mrgla...@ca.ibm.com
E-mail: mrgla...@apache.org
guillaume.desh...@gmail.com wrote on 05/21/2010 05:13:25 AM:
