Re: Parsing media queries

On Dec 10, 2007, at 15:41, Anne van Kesteren wrote:

> On Mon, 19 Nov 2007 01:32:01 +0100, L. David Baron  
> <dbaron@dbaron.org> wrote:
>> On Wednesday 2007-11-14 18:01 +0200, Henri Sivonen wrote:
>>> I agree that it doesn't make sense to allow escapes. After all,  
>>> the spec
>>> are Basic Latin identifiers--not arbitrary strings.
>>
>> Have you tested what existing implementations do with character
>> escapes?  (Assuming that those implementations support character
>> escapes in CSS.)  In both media queries in HTML and in CSS?

I hadn't, although I probably should have. My comments about media  
query parsing were based on trying to implement parsing from the spec.

> http://tc.labs.opera.com/mediaqueries/parsing/022.htm tests it for  
> HTML. Safari 3 supports escaping and Opera 9.5 doesn't.

Thanks. So there's a backwards compat case for not allowing them.

>> The implementation that I'd started to write for Mozilla would have
>> accepted character escapes, although it wouldn't be too much extra
>> work to disallow them in one or both modes.  But I expect
>> implementors may default to accepting them.
>
> I personally was sort of convinced by the argument that HTML  
> entities do not work in CSS either.


I can see why escapes would be allowed for on the basis of media  
queries being CSS and CSS allowing escapes. However, I also think that  
there is no *need* for the CSS escapes in media queries in HTML so for  
someone like me who isn't reusing bits of a full CSS parser, the  
escapes are work with no real use case besides passing test suites.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Monday, 10 December 2007 14:37:35 UTC