[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
Re: [xml-dev] RELAX NG equivalent of <xsd:any processContents="lax">?
- From: "Radu Cernuta" <radu.cernuta@gmail.com>
- To: "noah_mendelsohn@us.ibm.com" <noah_mendelsohn@us.ibm.com>
- Date: Tue, 6 Feb 2007 23:38:25 +0100
2007/2/6, noah_mendelsohn@us.ibm.com <noah_mendelsohn@us.ibm.com>:
Thanks, that makes sense. I'm still a bit rusty on RELAX NG, but I
>think< what you've got says:
1) The root can be anything except interesting elements.
2) The children can by anything at all to arbitrary depth.
3) Interesting elements among the children are validated.
4) The schemas for those interesting children are pointed to in that final
choice.
That is correct.
Do I have that right? If I understand this, then if you want to validate
the container itself, that's a separate validation. Not a big deal, but a
slight difference in style I'd think compared to W3C Schemas, in which
container and containee are validated together yielding one net validity
result, and the composition of the schema pieces (SOAP schema and purchase
order schema) is controlled on the command line or validator invocation
API. Anyway, I appreciate the tutorial. I'm convinced that in practice
people can pretty much do what they need for container validation with
either the RELAX NG or W3C Schema approaches, but the nuances are a bit
different. Thanks!
You are welcome,
Radu
--------------------------------------
Noah Mendelsohn
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------
"Radu Cernuta" <
radu.cernuta@gmail.com>
02/06/2007 04:58 PM
To: "noah_mendelsohn@us.ibm.com" <noah_mendelsohn@us.ibm.com
>
cc: "Jeff Lowery" <jlowery@myrio.com>, xml-dev@lists.xml.org
Subject: Re: [xml-dev] RELAX NG equivalent of <xsd:any
processContents="lax">?
Hi Noah,
It is not necessary to modify the existing official XML Schema provided
for SOAP by W3C. I think the purpose of the example in the tutorial was to
emphasize on the necessity to avoid ambiguity when using <anyName>, not to
show the one and only way of doing things.
For example, it is possible to write a RELAX NG schema that would validate
interesting elements (such as purchaseOrder in your example) within
whatever container vocabulary(soap in your example), except the vocabulary
of the interesting elements; the interesting elements can be defined in
this schema or another one:
<?xml version="1.0"?>
<grammar xmlns="http://relaxng.org/ns/structure/1.0">
<start>
<ref name="rootElement"></ref>
</start>
<define name="rootElement">
<element>
<anyName>
<except>
<nsName ns="
http://interestingElements.com"></nsName>
</except>
</anyName>
<ref name="children"></ref>
</element>
</define>
<define name="children">
<zeroOrMore>
<element>
<anyName>
<except>
<nsName ns="
http://interestingElements.com
"></nsName>
</except>
</anyName>
<ref name="children"></ref>
</element>
</zeroOrMore>
<optional>
<ref name="interestingElements"></ref>
</optional>
</define>
<define name="interestingElements">
<choice>
<externalRef href=""interestingElement1.rng"/>
<externalRef href=""interestingElement2.rng"
"/>
</choice>
</define>
</grammar>
The document can be validated against the official SOAP schema for the
container structure and against the other schema for the content. If one
of the validations fails, the document is not valid.
Greetings,
Radu Cernuta
2007/2/6, noah_mendelsohn@us.ibm.com <
noah_mendelsohn@us.ibm.com>:
I'm not a RELAX NG expert, but I am interested in this question, so maybe
your or other experts can provide a bit more detail. What I see at
http://relaxng.org/tutorial-20011203.html#IDAFLZR gives some of what I
think people look for in W3C Schema lax validation, but not quite all of
it. The difference I think I see is in the case where the attribute that
does have a declaration (xml:space in the example at the link) gets its
declaration from a separate schema document (in W3C Schema terms), perhaps
for a different targetNamespace. The RELAX NG solution shown seems to
require that a list of the elements to be validated be provided with the
<anyName> wildcard in an <except> element.
Consider a SOAP envelope:
<soap:envelope>
<soap:body>
<po:purchaseOrder>
...
</po:purchaseOrder>
</soap:body>
</soap:envelope>
There are a variety of ways to acheive different effects using W3C XML
Schema, but the schema that's officially published by W3C for the envelope
has the body marked <any processContents="lax"/>. If someone comes along
and writes a completely separate schema document describing the
po:purchaseOrder, and passes both schema documents in on the API call to
the validator, the envelope and the purchaseOrder will be validated
together, and the overall envelope will be marked invalid if either the
envelope markup or the purchaseOrder is in error. As I understand the
solution proposed above, it would involve modifying the W3C-provided
envelope schema to list <po:purchaseOrder> in an <except>.
Does RELAX NG have another approach, or is this considered a non-goal?
Certainly one always has the option of validating the container and the
payload separately. Thanks!
Noah
--------------------------------------
Noah Mendelsohn
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------
"Radu Cernuta" <
radu.cernuta@gmail.com>
02/06/2007 06:49 AM
To: "Jeff Lowery" <jlowery@myrio.com>
cc:
xml-dev@lists.xml.org, (bcc: Noah
Mendelsohn/Cambridge/IBM)
Subject: Re: [xml-dev] RELAX NG equivalent of <xsd:any
processContents="lax">?
For lax validation in RELAX NG you could take a look at the RELAX NG
Tutorial by James Clark and Murata Makoto. Section 11 (Name Classes)
handles this issue.
http://relaxng.org/tutorial-20011203.html#IDAFLZR
Greetings,
Radu Cernuta
2007/2/6, Jeff Lowery <jlowery@myrio.com>:
This has got to be a FAQ, but no luck with googling:
Trying to find a shorthand method to validate content using lax
constraints in RELAX NG. I want something like this:
#=================
namespace foo = "http://www.w3.org/foo"
qux =
element baz {empty} # some element
bar =
element bar { anyLax* } #allow baz, bar, anyOtherElement
anyOtherElement =
element * - foo:* { # any elements not in this namespace
attribute * { text }*,
(text
| anyOtherElement)*
}
anyLax = (
grammar {*} #all patterns
)*
#===========
Don't work, though.
Only examples I've seen enumerate all the grammars by name in the
schema. Doesn't appear there is wildcard of that type.
_______________________________________________________________________
XML-DEV is a publicly archived, unmoderated list hosted by OASIS
to support XML implementation and development. To minimize
spam in the archives, you must subscribe before posting.
[Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
Or unsubscribe:
xml-dev-unsubscribe@lists.xml.org
subscribe: xml-dev-subscribe@lists.xml.org
List archive: http://lists.xml.org/archives/xml-dev/
List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]