Re: regarding Testcase 5.1.a

 Hi Joern - I'm sorry to have carelessly phrased my response to your original email...and certainly did not mean to imply *you* were being picky...I was just having fun with the way you described the problem.

Thanks for your feedback!  Charlie
 
Charles Wiecha
Manager, Multichannel Web Interaction
IBM T.J. Watson Research Center
P.O. Box 704
Yorktown Heights, N.Y.  10598
Phone: (914) 784-6180, T/L 863-6180, Cell: (914) 320-2614
wiecha@us.ibm.com 

 -----Joern Turner <joernturner@me.com> wrote: -----

 =======================
 To: Charles F Wiecha/Watson/IBM@IBMUS
 From: Joern Turner <joernturner@me.com>
 Date: 12/16/2009 06:22PM 
 Cc: Joern Turner <joern.turner@googlemail.com>, "public-forms@w3.org" <public-forms@w3.org>
 Subject: Re: regarding Testcase 5.1.a
 =======================
   Charles,

On 16.12.2009, at 23:07, Charles F Wiecha <wiecha@us.ibm.com> wrote:

> Hi Joern -- Firefox is not so picky so I'm not sure if this version  
> has been updated as you suggest. There's now a namespace declaration  
> on the instance (and an extra one for good measure on the root  
> document though probably not needed there...).
>
Ok, thanks. It's not my private pickyness but that of Xerces ;) As  
this is probably one of the most used validating parsers i thought  
this alone may make an argument. We traced it down and found that  
xerces (2.9) would never validate a QName without the prefix being  
declared and for my personal opinion this is absolutely correct. - we  
surely didn't wanted to patch Xerxes if it's not really a bug.

But anyway the issue is resolved now.
>
> Lacking any comments to the contrary, we'll include this in the  
> updated test suite zip file shortly. Thanks, Charlie Wiecha
>
>
My collegue Tobi Krebs sent a comment which unfortunately didn't reach  
the list. We've found another error through these test cases. The  
xforms variant of nonNegativeInteger is missing in the schema linked  
by the Spec. Result was that the one of the forms never showed a valid  
value.  (the latter relates to the 5.2.x test cases)

Thanks for the effort you and the WG put into clarifying these issues.

Regards,

Joern
> (See attached file: 5.1.a.xhtml)
>
> Charles Wiecha
> Multichannel Web Interaction
> IBM T.J. Watson Research Center
> P.O. Box 704
> Yorktown Heights, N.Y. 10598
> Phone: (914) 784-6180, T/L 863-6180, Cell: (914) 282-3483
> wiecha@us.ibm.com
>
>
> <graycol.gif>Joern Turner ---12/07/2009 06:33:57 AM---When  
> activating the 'Valid Values' trigger our Validator (Xerces)  
> complains about the Node bound to
>
> <ecblank.gif>
> From:	<ecblank.gif>
> Joern Turner <joern.turner@googlemail.com>
> <ecblank.gif>
> To:	<ecblank.gif>
> public-forms@w3.org
> <ecblank.gif>
> Date:	<ecblank.gif>
> 12/07/2009 06:33 AM
> <ecblank.gif>
> Subject:	<ecblank.gif>
> regarding Testcase 5.1.a
> <ecblank.gif>
> Sent by:	<ecblank.gif>
> public-forms-request@w3.org
>
>
>
> When activating the 'Valid Values' trigger our Validator (Xerces)
> complains about the Node bound to QName as there's no namespace
> defined for 'my:myelement' and therefore the prefix 'my' cannot be
> resolved. The Node becomes invalid.
>
> Declaring a namespace with that prefix fixes that problem in our
> implementation. I think Xerces is right here as the QName validation
> does not only test for lexical correctness of the QName but also
> checks if there a namespace in scope for the given prefix.
>
> Regards,
>
> Joern
>
>
>
> <5.1.a.xhtml>

Received on Wednesday, 16 December 2009 23:34:37 UTC