Re: [Bug 10] Round-tripping namespace decls in properties

Lisa Dusseault wrote:
> 
> On Dec 5, 2005, at 3:01 AM, bugzilla@soe.ucsc.edu wrote:
> 
>>
>> http://ietf.cse.ucsc.edu:8080/bugzilla/show_bug.cgi?id=10
>>
>> ------- Additional Comments From julian.reschke@greenbytes.de  
>> 2005-12-05 03:01 -------
>>
>>> - Does it make sense to say that on attributes, [attributes] must be
>>> preserved?
>>> I think this was a cut-and-paste error so I am leaving it out.
>>
>> It's a mistake. Please let's dicuss this change over here, and only 
>> integrate it
>> when done.
> 
> Ok; in what sense are there attributes on attributes? As far as I can 
> tell, Infoset talks about attributes having namespace name, local name, 
> prefix, normalized value, type, references, owner (like parent) and the 
> specified flag.

As I said: it was a mistake. Attributes do not have attributes.

>>> - Did you consider whether [references] ought to be preserved on 
>>> attributes?
>>> What's the consequence if they're not?
>>
>> As far as I can tell, this question is meaningless at it doesn't 
>> affect the
>> serialization.
> 
> Why is that?

I'm not sure how to answer that... What's relevant to us is what's going 
over the wire, thus the XML serialization. That part of the Infoset 
doesn't seem to be relevant here. Am I missing something?

Best regards, Julian

Received on Thursday, 8 December 2005 08:51:41 UTC