Re: [css3-background] does border-radius round the border-image ?

On Jan 13, 2009, at 11:45 AM, Brad Kemper wrote:

> I think that is much simpler. But its less specific about _which_  
> curve, and WebKit is currently clipping to the outside in all cases.  
> Can we add a little something and still keep it clear, perhaps as  
> follows?
>

WebKit's clipping here isn't really intentional.  I never thought  
about how background-clip would interact with border-radius.  Clipping  
to the inner border edge of the curve definitely sounds right to me  
when background-clip is padding-box.

> # Backgrounds, but not the border-image, are clipped to the curve.
> # Other effects that clip to the border or padding edge (such as
> # 'overflow') also must clip to the curve. The edge of the curve
> #   to use for clipping depends on the value of ‘background-clip’
> #  (outer edge for 'border-box', inner edge for ‘padding-box’). When
> #  the value of ‘background-clip’ is 'content-box', then the content
> #  box is clipped with corner radii equal to the 'border-radius'
> #  values, minus the padding values, with a floor of zero.
>
> That last part about the content-box is where we would expect to  
> find the curve in that case, as a logical extension of where the two  
> curves are in the other two ‘background-clip’ cases.
>
> By the way, neither WebKit or FireFox (Minefield) are currently  
> doing any clipping of the foreground when 'overfow' is hidden, as  
> this text says it should. I agree that they should do so, and maybe  
> that just hasn't yet been implemented but will be.
>
>

It will be.  I'm the one who proposed the overflow change in the first  
place. :)

For compatibility reasons we won't be changing our -webkit-****  
property implementations at all, so any changes we make will have to  
wait until this module is ready enough that we can implement the  
unprefixed versions of these properties that behave more like the spec.

dave
(hyatt@apple.com)

Received on Thursday, 15 January 2009 20:08:46 UTC