Eclipse Community Forums: XML Schema Definition (XSD)

Help

Home

Home » Archived » XML Schema Definition (XSD) » Problems with complexType(Problems with complexType)

Show: Today's Messages :: Show Polls :: Message Navigator

Problems with complexType [message #1062668]

Mon, 10 June 2013 10:14

Arjan Kok

Messages: 40
Registered: July 2009

Member

Hi,

I am working on a project that extends the org.eclipse.xsd module. There seem to be two problems regarding the handling of complexType.

1. The attributes of the complexContent element are lost when loading an xsd schema into the ecore model, and then deriving the xsd from the model again.
For example, when running XSDMainExample::loadAndPrint (in org.eclipse.xsd.example) for the xsd file example 1 (is attached), then the attributes of the complexContent element are lost. It looks like the attributes are not loaded into the model.

2. The complexContent element and restriction element are lost when the attribute base of the contained restriction element has value 'anyType'. See example 2.

Are these bugs or features? Should I enter a bugzilla, or is there a workaround?

Kind Regards,

Arjan Kok.

Attachment: example1.xsd
(Size: 0.84KB, Downloaded 295 times)
Attachment: example2.xsd
(Size: 0.81KB, Downloaded 318 times)

Report message to a moderator

Re: Problems with complexType [message #1062673 is a reply to message #1062668]

Mon, 10 June 2013 10:37

Ed Merks

Messages: 33140
Registered: July 2009

Senior Member

Arjan,

Comments below.

On 10/06/2013 12:14 PM, Arjan Kok wrote:
> Hi,
>
> I am working on a project that extends the org.eclipse.xsd module. There seem to be two problems regarding the handling of complexType.
>
>
> 1. The attributes of the complexContent element are lost when loading an xsd schema into the ecore model, and then deriving the xsd from the model again.
> For example, when running XSDMainExample::loadAndPrint (in org.eclipse.xsd.example) for the xsd file example 1 (is attached), then the attributes of the complexContent element are lost. It looks like the attributes are not loaded into the model.
For me that prints out

<?xml version="1.0" encoding="UTF-8"?>
<schema xmlns="http://www.w3.org/2001/XMLSchema"
xmlns:link="http://www.xbrl.org/2003/linkbase"
xmlns:xl="http://www.xbrl.org/2003/XLink">
<import namespace="http://www.xbrl.org/2003/XLink"
schemaLocation="http://www.xbrl.org/2003/xl-2003-12-31.xsd"/>
<import namespace="http://www.xbrl.org/2003/linkbase"
schemaLocation="http://www.xbrl.org/2003/xbrl-linkbase-2003-12-31.xsd"/>
<element name="reference" substitutionGroup="xl:resource">
<complexType mixed="true">
<complexContent>
<extension base="xl:resourceType">
<sequence>
<element maxOccurs="unbounded" minOccurs="0" ref="link:part"/>
</sequence>
</extension>
</complexContent>
</complexType>
</element>
</schema>

So you're specifically referring to the id attribute? I've not looked
closely, but the DOM itself should preserve that. They're not recorded
in the model. Are you discarding the DOM and regenerating it?
>
> 2. The complexContent element and restriction element are lost when the attribute base of the contained restriction element has value 'anyType'. See example 2.
The same issue here. What's serialized is semantically equivalent to
what's read in, so if you discard the DOM, you should expect this type
of thing.
>
> Are these bugs or features? Should I enter a bugzilla, or is there a workaround?
Are you discarding the DOM?
>
> Kind Regards,
>
> Arjan Kok.

Ed Merks
Professional Support: https://www.macromodeling.com/

Report message to a moderator

Re: Problems with complexType [message #1062687 is a reply to message #1062673]

Mon, 10 June 2013 11:43

Arjan Kok

Messages: 40
Registered: July 2009

Member

Hi Ed,

1) The complexContent element misses the attribute mixed = "true". Yes, I am discarding the DOM and regenerating the DOM from the model (that is what XSDMainExample::loadAndPrint does).
2) Yes, I am discarding the DOM. If the results are semantically equivalent then I am OK with it.

The goal is to work on the model level. Discarding the DOM reassures me that I don't miss information in the model.

Arjan.

[Updated on: Mon, 10 June 2013 11:47]

Report message to a moderator

Re: Problems with complexType [message #1062697 is a reply to message #1062687]

Mon, 10 June 2013 12:12

Ed Merks

Messages: 33140
Registered: July 2009

Senior Member

Arjan,

Comments below.

On 10/06/2013 1:43 PM, Arjan Kok wrote:
> Hi Ed,
>
>
> 1) The complexContent element misses the attribute mixed = "true".
That's redundant with respect to the mixed on the complex type itself
and is recorded only once in the semantic model and never serialized out
redundantly.
> Yes, I am discarding the DOM and regenerating the DOM from the model
> (that is what XSDMainExample::loadAndPrint does).
Yes as a demonstration of what's in the semantic model.
> 2) Yes, I am discarding the DOM.
>
> The goal is to work on the model level.
Yes that's supported. It incrementally updates the DOM.
> Discarding the DOM reassures me that I don't miss information in the
> model.
It does the opposite. It discards information not in the semantic
model, e.g., redundancies as well as things not in the semantic model as
documented in the XML Schema specification, such as non-schema namespace
attributes.
>
> Arjan.
>

Ed Merks
Professional Support: https://www.macromodeling.com/

Report message to a moderator

Re: Problems with complexType [message #1062703 is a reply to message #1062697]

Mon, 10 June 2013 12:30

Arjan Kok

Messages: 40
Registered: July 2009

Member

Hi Ed,

I understand your point regarding discarding redundancies. Thanks.

I changed the example (uploaded example 3), the complexContent element has now attribute mixed="false".

It is still missing in the output. The input DOM and output DOM are now not semantically equivalent, right?

PS Are there plans to support XML Schema 1.1?

Arjan.

Attachment: example3.xsd
(Size: 0.84KB, Downloaded 297 times)

Report message to a moderator

Re: Problems with complexType [message #1062801 is a reply to message #1062703]

Mon, 10 June 2013 19:23

Ed Merks

Messages: 33140
Registered: July 2009

Senior Member

Arjan,

Comments below.

On 10/06/2013 2:30 PM, Arjan Kok wrote:
> Hi Ed,
>
> I understand your point regarding discarding redundancies. Thanks.
>
> I changed the example (uploaded example 3), the complexContent element has now attribute mixed="false".
That sounds like a loop hole in the XML Schema specification. It has no
constraints about the redundant attributes being consistent.
>
> It is still missing in the output. The input DOM and output DOM are now not semantically equivalent, right?
I think it's not well specified.
>
> PS Are there plans to support XML Schema 1.1?
No, without funding there are no plans.
>
> Arjan.
>
>
>
>

Ed Merks
Professional Support: https://www.macromodeling.com/

Report message to a moderator