Re: [CSSOM] CSSRule.NAMESPACE_RULE

On Fri, 12 Mar 2010 19:14:48 +0100, Simon Fraser <smfr@me.com> wrote:
> On Mar 12, 2010, at 10:05 AM, Anne van Kesteren wrote:
>> On Fri, 12 Mar 2010 19:01:12 +0100, Simon Fraser <smfr@me.com> wrote:
>>> WebKit already uses 8 and 9 for animation keyframe-related rules:
>>>
>>>        const unsigned short WEBKIT_KEYFRAMES_RULE = 8;
>>>        const unsigned short WEBKIT_KEYFRAME_RULE = 9;
>>
>> The specification already states that private extensions should be  
>> outside the range 0-1000. That range is reserved for the CSS WG.
>
> At what point during the standardization process should we switch from  
> considering these as private extensions to values that are approved by  
> the CSS WG?

Probably at the point where you rename them from WEBKIT_ to something  
else? For prefixed properties this usually happens at the CR phase. I'm  
happy with settling constants before then if it is at least somewhat clear  
the proposal is going to make it.


>>> We're playing a game of whack-a-mole here. We really need partition  
>>> out numeric ranges somehow.
>>
>> The specification contains some suggestions for private extension  
>> ranges.
>
> It seems that the real problem here is that we use raw numeric values  
> for CSSRulel types. If rule types were identified by a string label, or  
> if we could force authors to always use the constant values on CSSRule  
> (maybe by some trick like making the type opaque, if IDL allows that),  
> then we wouldn't have this problem.

That would be inconsistent with Node... Also, isn't it a bit too late for  
that?


-- 
Anne van Kesteren
http://annevankesteren.nl/

Received on Friday, 12 March 2010 18:21:51 UTC