[
Lists Home |
Date Index |
Thread Index
]
Aha! A fairly simple, yet elusive answer. Where does it say that
validateOnParse (default true) only does a lax validation? That could have
saved me a lot of trouble.
Anyway, thank you very much for the examples and quick reply! My faith in
MSXML is restored (for the moment...).
Strangely, though, I can't get a decent error message by using your
"validate()" trick. That is, I get one of two messages, depending on whether
I lack the correct namespace or having correct namespace+invalid data.
The ones I get are:
"Validate failed because the root element had no associated DTD/schema."
(-1072897500)
"Validate failed because the document does not contain exactly one root
node." (-1072897501)
These messages are obvious enough, even the latter one, because trying to
output the XMLDoc.xml property when having an invalid document yields only
an empty string. Thus, the document does of course lack a root node. It's
somewhat strange though, giving just that message, and not one where it says
exactly WHAT is wrong, and not only describing the symptoms of the error.
This means that for proper error logging, I have to combine the two
approaches, by both checking the parseError.errorCode after load, and if
that's OK, also perform the validate() and see if that fails. That way I can
log any schema violations with a decent error message, providing the
document uses the correct namespace, and finally trapping the cases where
the namespace itself is incorrect.
Case closed, I'd presume.
Mvh / Kind regards,
Morten Grøtan
morten.grotan@electricfarm.no
Utvikler, MCSD
Electric Farm ASA, Rosenkrantz gt. 21, NO-0160 OSLO, Norway,
http://www.electricfarm.no/
mob: +47 92 88 59 72, tel: +47 48 00 09 99, fax: +47 23 10 30 10
-----Original Message-----
From: Dare Obasanjo [mailto:dareo@microsoft.com]
Sent: 30. september 2002 11:04
To: Morten Grøtan; xml-dev@lists.xml.org
Subject: RE: [xml-dev] Validating documents against XML Schema with
different namespace using MSXML
Calling the validate( ) method on the document is the preferred way to use
MSXML to validate a document since it does strict validation while loading
with validateOnParse set to true does lax validation.
I could not reproduce the results of your scenarios using MSXML 4.0 SP1 when
I called validate() directly. Below are the code, XML instance and schema I
used
var xmlDoc = new ActiveXObject("Msxml2.DOMDocument.4.0");
var xsdCache = new ActiveXObject("Msxml2.XMLSchemaCache.4.0");
xsdCache.add("", "example.xsd");
xmlDoc.async = false;
xmlDoc.schemas = xsdCache;
xmlDoc.load("example.xml");
var err = xmlDoc.validate();
if (err.errorCode == 0){
WScript.Echo("Document is valid");
}else{
WScript.Echo("Validation error:" + err.reason);
}
<?xml version="1.0" ?>
<xs:schema
xmlns:xs="http://www.w3.org/2001/XMLSchema";>
<xs:element name="A">
<xs:complexType mixed="true">
<xs:choice>
<xs:element name="B" />
<xs:element name="C" />
</xs:choice>
</xs:complexType>
</xs:element>
</xs:schema>
<?xml version="1.0"?>
<A xmlns="http://www.example.com"; >
foo <C /> foo
</A>
-----Original Message-----
From: Morten Grøtan [mailto:Morten.Grotan@electricfarm.no]
Sent: Mon 9/30/2002 1:29 AM
To: XML Dev List (xml-dev@lists.xml.org)
Cc:
Subject: [xml-dev] Validating documents against XML Schema with
different namespace using MSXML
Hi,
I've only just started using XML Schemas with the MSXML parser (4.0
SP1), and have trouble grasping a few baseline things.
The real-life task I set out to solve with XML Schemas was to be
able to validate documents prior to importing to a production database. The
documents are received as part of an automatic collector service, hence I
don't have 100% control of what kind of files are received. I want to be
able to only import 100% valid documents, according to my schema, and avoid
having tons of custom validation logic within my import program.
Making the schema itself for a given instance of a typical valid
document has proved to be no problem, and I am also able to enforce the
constraints specified in my schema. So far so good. But what happens is that
I purposely place an invalid document among the other valid ones, to see
what happens.
The real problem, I've found, is in the use of namespaces (what a
surprise, having read lots of different articles about namespaces
lately....). I have defined both a default and a target namespace in my
schema and everything works just fine as long as I have the same namespace
in my instance document. But as soon as there is the slightest mismatch in
namespaces, the document is NOT validated against the schema, hence
rendering my validation mechanism utterly useless.
I've sketched out 4 scenarios below, using the following
terminology:
xsd: An XML Schema
xml: An XML instance document
xmlns: Default namespace
targetns: XML Schema target namespace
Scenario 1:
xsd: no xmlns and no targetns
xml: no xmlns
uses xsd and correctly says invalid xml is invalid
Scenario 2:
xsd: no xmlns and no targetns
xml: xmlns set to something else than targetns
does NOT use xsd and incorrectly says invalid xml is valid
Scenario 3:
xsd: xmlns and targetns
xml: xmlns
uses xsd and correctly says invalid xml is invalid
Scenario 4:
xsd: xmlns and targetns
xml: no xmlns or xmlns set to something else than targetns
does NOT use xsd and incorrectly says invalid xml is valid
The way I've implemented this is using the MSXML2.XMLSchemaCache.4.0
object to add a schema, associating it with the namespace defined in the
schema, and then setting the "schemas" property of the XMLDom object
(MSXML2.DOMDocument.4.0) to point at the XMLSchemaCache object. Then I just
call the load method of the XMLDom object, and check the parseError object
of the XMLDom object for any errors.
So, why does this happen? Shouldn't MSXML be able to find out that
my instance document does not share the same namespace as my schema, hence
invalidating the document? What is the purpose of the whole XML Schema
shebang elsewise? If you can't be sure that invalid documents are in fact
invalidated, without writing tons of validation code yourself, why even
bother using a schema? Needless to say, I'm slightly disillusioned right
now, and possibly soon will be having a problem with my boss, trying to
explain why I've wasted so much time on this schema thing without producing
a useful validation mechanism.....
As for the XMlSchemaCache object, I'm aware that you can add several
schemas into it before attaching it to the XMLDom object, hence using all
the schemas to validate the instance document. Admittedly it can seem hard
to see which (if any) schema should be responsible for invalidating the
document, but in my opinion ANY of them should be able to do that, as long
as they can't be validated against the document. So, if at least one of the
added schemas fail, the whole validation should fail. Or am I totally naïve?
Mvh / Kind regards,
Morten Grøtan
morten.grotan@electricfarm.no
Utvikler, MCSD
Electric Farm ASA, Rosenkrantz gt. 21, NO-0160 OSLO, Norway,
http://www.electricfarm.no/ <http://www.electricfarm.no/>
mob: +47 92 88 59 72, tel: +47 48 00 09 99, fax: +47 23 10 30 10
|
