W3C home > Mailing lists > Public > www-svg@w3.org > December 2005

Re: [SVGMobile12] SVGT12-307: setAttributeNS() raising exceptions on newly-created elements

From: Dean Jackson <dino@w3.org>
Date: Wed, 14 Dec 2005 20:40:13 +1100
Message-Id: <D7B15AA8-2CFC-4562-A014-9E116F2564A6@w3.org>
Cc: www-svg@w3.org
To: Ian Hickson <ian@hixie.ch>


On 14/12/2005, at 3:32 PM, Ian Hickson wrote:

> On Wed, 14 Dec 2005, Dean Jackson wrote:
>>>
>>> Please edit the spec so that it can't be misinterpreted to read that
>>> the setAttributeNS() method in DOM 3 Core should also raise that
>>> exception in the given case.
>>
>> Given the use case that you mention below, we think the best  
>> solution is
>> to remove the exception throwing behaviour. Obviously there are
>> situations that require a script to modify the version and  
>> baseProfile
>> attributes via the DOM, or uDOM.
>>
>> Would that satisfy your request?
>
> Sure, that would be great.

Good.

>
>> The next thing to specify is what does it mean when you change those
>> attributes on an actively displayed/interpreted document.
>> We intend to add wording that specifies no change in behaviour is
>> required from an implementation.
>
> That seems fine. (It isn't really clear to me what a UA is supposed  
> to do
> with these attributes anyway -- it seems that a UA that completely  
> ignores
> them is conformant, so saying that changes should similarly not  
> have any
> requirements seems quite reasonable.)

With our relaxed error processing rules, it is only identification.
In the case where an SVG Tiny UA gets content marked as Full, it is able
to display a warning. See section C.4.

>
>
>> This is similar to the behaviour of dynamically inserting a <script>
>> element into HTML or SVG. The DOM is updated but the processing  
>> doesn't
>> change.
>
> Inserting an <html:script> into a document would cause it to be  
> executed
> (assuming it hasn't been executed before), but mutating a <script>
> elements's contents would indeed not have any effect, which I  
> assume is
> what you meant.

Yes, that's what I meant. Thanks.

Dean
Received on Wednesday, 14 December 2005 09:40:33 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:32 GMT