Re: [Backgrounds/Borders] What to do when a border-image fails to load

Brad Kemper wrote:
> On Nov 19, 2009, at 12:13 PM, fantasai <fantasai.lists@inkedblade.net> 
> wrote:
> 
>>
>> I've added
>>  # The area outside the curve of the border edge does not accept
>>  # mouse events on behalf of the element.
>> to the definition of 'border-radius' and
>>  # [portions of the border-image outside the border box] are
>>  # invisible to mouse events and do not capture such events on
>>  # behalf of the element.
>> to the definition of 'border-image-outset'. Please let me know if
>> this is acceptable.
> 
> That's understandable, and worded in such a way that it covers the 
> combination of those two properties too.
> 
> One question: were we set on using "do not" instead of "may not"?

Yes. That was the whole "note/recommend/require" discussion. :)

> Could there, for instance, be a mobile device that is able to draw curves, but 
> only capture renctangular areas fir mouse events (or touch events).

There could be, but I'd say we can wait and see if someone actually
has this problem and complains about the spec being too stringent.
It doesn't seem very likely to me, given that mouse events are, afaik,
given as screen coordinates and the layout engine is responsible for
figuring out which element they hit.

> Or would a UA be non-complient if it did register clicks on border-image 
> pixels outside the border-box, based on the alpha mask?

Yes, that UA would be non-compliant. I think it's actually more
important that we get consistent behavior across UAs here than that
we allow a UA to come up with some fancy heuristics.

In CSS4 we should be able to allow more corner shapes than just
elliptical curves, so the author will be able to tailor the active
area pretty close to the shape they want in most cases. And the
pointer-events spec might be able to add some extra controls.

~fantasai

Received on Thursday, 19 November 2009 21:46:30 UTC