Re: Styling conflict

On Thu, 18 Nov 2004, Chris Lilley wrote:
>> 
>> It is tested in the CSS3 Selectors test suite (and has been for some 
>> years, Chris).
> 
> Yes, I happen to know that. On the other hand, the casual reader might 
> conclude from
> 
>    http://www.w3.org/TR/css3-selectors/#Tests
> 
> that there is no such test suite.

Yeah, we should republish that spec with updated links.


>> However, we currently do not have an SVG version of the CSS3 Selectors 
>> test suite, so testing of this feature in an SVG context has not been 
>> done by the CSSWG.
> 
> Right. Although SVG 1.1 and 1.2 use CSS2 selectors, not CSS3 ones.

I'm not sure what that means. SVG 1.1 and 1.2 might require support for 
selectors from CSS2 in order for an SVG UA to claim conformance, but 
nothing stops an SVG UA from implementing selectors from CSS3 as well (as, 
indeed, Mozilla does).


>> Note that the change in the specificity of style="" from 1,0,0 to 
>> 1,0,0,0 was made after examining numerous implementations, including a 
>> large number of SVG UAs. The overwhelming majority of UAs at the time 
>> implemented what CSS2.1 now says.
> 
> Sure, the issue isn't the change as such (although I consider removing 
> the ability to restyle content to be unfortunate) but rather how to 
> reference it.

Just reference CSS2.1.


> Do you have the test results from the large number of SVG UAs available? 
> it would be helpful to tie that statement to specific UAs and specific 
> revisions, and to a specific test.

Not off hand. This was years ago now. Check w3c-css-wg archives.

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

Received on Thursday, 18 November 2004 23:00:22 UTC