[whatwg] IE6 vs XHR vs XMLNS

[Henri Sivonen]

On Feb 14, 2008, at 20:16, Julian Reschke wrote:

> Henri Sivonen wrote:
>> On Feb 14, 2008, at 19:45, Julian Reschke wrote:
>>> IE6 XHR is part of MSXML, which deals just fine with XMLNS, except  
>>> for the absence of some DOM L2 methods.
>>

>> The practical effect of the lack of Level 2 methods in IE6 and the
>
> The only problem in IE6 I'm aware of is the lack of createElementNS  
> and createAttributeNS, for which there's a simple workaround.

Well, those methods are rather central to synthesizing namespaced DOM  
trees.

>> brokenness* of the XHR XML serializer in older WebKit is that using  
>> namespaces with cross-browser XHR is at least hard and using Level  
>> 1 DOM APIs with namespaceless XML 1.0 (< 5th ed.) is the practical  
>> single-codepath cross-browser solution.
>> * http://bugs.webkit.org/show_bug.cgi?id=5262
>
> XMLSerializer is neither part of DOM level 2 nor XHR, right?

void send(in Document data); is part of XHR and causes a serializer to  
run.

It isn't clear to me how you want the editor to take what you are  
saying into account in a WHATWG deliverable. The relevance of my IRC  
remarks was that unnamespaced XML is alive in XHR APIs due to browser  
problems and, therefore, we can't unconditionally make  
createElement(foo) equivalent to createElementNS("http://www.w3.org/1999/xhtml 
", foo) in DOM5.

-- 
Henri Sivonen
hsivonen at iki.fi
http://hsivonen.iki.fi/