Eclipse Community Forums: XML Schema Definition (XSD)

Help

Home

Home » Archived » XML Schema Definition (XSD) » QName validation(QName attributes are not validated)

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

QName validation [message #545223]

Wed, 07 July 2010 08:59

Tsvetan Stoyanov

Messages: 10
Registered: July 2009

Junior Member

Hi all,

I noticed that the literal values of QName attributes are not validated.
Here's an example:

In one schema:

<simpleType name="My;SimpleType" .../>

In another schema:

<element name="e1" type="sch1:My;SimpleType"/>

Notice that the value of the name and type attributes is not correct - it contains ;.

As expected the validation of the first schema fails because of the incorrect value, but the second schema validation is OK - even if the type attribute value is incorrect the type itself is resolved.

Is this an expected behavior? Shouldn't be the literal attribute value the one that must be validated first?

[Updated on: Wed, 07 July 2010 09:00]

Report message to a moderator

Re: QName validation [message #545324 is a reply to message #545223]

Wed, 07 July 2010 12:53

Ed Merks

Messages: 33137
Registered: July 2009

Senior Member

I'm not sure what's to be gained from repeating that error on each use?
A single error at the point of definition would seem more useful;
resolving the name properly regardless of the ill formed nature of its
characters helps with fixing such an error...

Tsvetan Stoyanov wrote:
> Hi all,
>
> I noticed that the literal values of QName attributes are not validated.
> Here's an example:
>
> In one schema:
> <simpleType name="My;SimpleType" .../>
>
> In another schema:
> <element name="e1" type="sch1:My;SimpleType"/>
>
> Notice that the value of the name and type attributes is not correct -
> it contains ;.
>
> As expected the validation of the first schema fails because of the
> incorrect value, but the second schema validation is OK - even if the
> type attribute value is incorrect the type itself is resolved.
>
> Is this an expected behavior? Shouldn't be the literal attribute value
> the one that must be validated first?
>

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

Report message to a moderator

Re: QName validation [message #545348 is a reply to message #545324]

Wed, 07 July 2010 13:41

Tsvetan Stoyanov

Messages: 10
Registered: July 2009

Junior Member

Hi Ed,

I think these are two errors - one for the definition and one for the usage. And hiding (skipping) the usage error makes the whole schema valid, which in actually not true.
In a hypothetic UI editor, the second schema will look as valid since the type is resolved and no validation errors are reported, but actually the attribute value remains invalid in respect to its schema.

Best Regards,
Tsvetan

Report message to a moderator

Re: QName validation [message #545385 is a reply to message #545348]

Wed, 07 July 2010 15:08

Ed Merks

Messages: 33137
Registered: July 2009

Senior Member

Tsvetan,

What does Xerces do in this situation? I suppose we should try to do
much the same thing.

Tsvetan Stoyanov wrote:
> Hi Ed,
>
> I think these are two errors - one for the definition and one for the
> usage. And hiding (skipping) the usage error makes the whole schema
> valid, which in actually not true.
> In a hypothetic UI editor, the second schema will look as valid since
> the type is resolved and no validation errors are reported, but
> actually the attribute value remains invalid in respect to its schema.
>
> Best Regards,
> Tsvetan

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

Report message to a moderator