Re: Parsing of HTML fragments

Ian Hickson schrieb:
> On Tue, 24 Apr 2007, Dão Gottwald wrote:
>>> If this WG adopts the WHATWG draft, I imagine this would be handled by 
>>> the innerHTML parsing mode.
>> Um, not necessarily. A feed consumer doesn't have to use innerHTML to 
>> insert fragments into a given document. In fact, HTML fragments can be 
>> delivered as XML nodes instead of CDATA.
> 
> The "innerHTML parsing mode" isn't restricted to innerHTML; it's just a 
> way to parse fragments with a particular context. You can use the 
> algorithm described for innerHTML in the spec for other purposes, such as 
> this one.

Using the described algorithm is an option. Yet I don't think that 
specifies "how HTML fragments, such as those in Atom documents, are to 
be parsed". There are too many use cases and different valid approaches 
to handle them. A server / feed reader could inject fragments by 
building up the document with string operations and passing the whole 
thing to the user agent / HTML parser afterwards; the latter would never 
know that there were fragments. Heck, you don't even need a context; you 
could treat feed entries as isolated documents.

--Dao

Received on Wednesday, 25 April 2007 00:23:55 UTC