Re: Parsing: CDATA section processing

On Sun, 6 Apr 2008, John-Mark Bell wrote:
> The current specification of how the tokeniser decides whether to 
> process CDATA sections or treat them as bogus comments relies upon the 
> tokeniser having access to the treebuilder's insertion mode and the 
> current node in the stack of open elements. I can't see any reason why 
> there is a need for this that couldn't be equally well served by a flag 
> controlled by the treebuilder (much like the existing content model 
> flag). Therefore, I propose the following:
> Introduce a new tokeniser flag "processCDATASection", which is initially
> clear.
> [...]

Implementing it as you describe is completely acceptable, but I'm not 
confident that I'd be able to catch all the ways in which you can jump in 
and out of states to set the flag correctly. The flag would only be a 
proxy for the condition in the spec now anyway -- it seems better to just 
be explicit in the spec about what the spec is really doing.

Ian Hickson               U+1047E                )\._.,--....,'``.    fL       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Saturday, 24 May 2008 09:20:06 UTC