Re: [CSS21] Status of defult (intrinsic) style sheet in UA

Lachlan Hunt wrote:
> However, it doesn't seem 
> to be defined how a UA should handle such rules in a UA stylesheet.

Correct.  But let's look at it another way.  What _is_ a UA stylesheet?  What 
does it mean to have a !important rule there?

> To test this, I tried the following rules in Firefox:

The fact that Firefox has an actual CSS file as a UA stylesheet is an 
implementation detail.  It could also have been done completely as a binary file 
(eg fastloaded version of ua.css) or even as part of the C++ code.

Modifying this file is equivalent to making changes to the C++ and recompiling 
the browser -- all you're doing is changing the rendering engine itself.

> I then restarted Firefox and was given red headings.  So, clearly 
> Mozilla is treating UA !important rules above all others, including user 
> !important stylesheets.

Yes, when Firefox needs a way to force the rendering engine to have a particular 
cascaded value for a property this is how it happens to do it.  That's subject 
to change without notice, of course.

> While it is unlikely that a vendor would put 
> !important rules in their stylesheet by default, the result should be 
> well defined in such cases.

Why?  What's the interoperability benefit?  Keep in mind that the reason there 
is a spec is to ensure interoperability.

-Boris

Received on Wednesday, 23 November 2005 02:18:06 UTC