Re: How much background content gets blurred by feGaussianBlur

On Sun, Oct 30, 2011 at 11:20 AM, Boris Zbarsky <> wrote:
> On 10/30/11 12:00 PM, David Dailey wrote:
>> Firefox redraws the blurred ellipses dynamically; Chrome waits until
>> mouseup to redraw; Opera requires a refresh; ASV doesn’t understand the
>> relative coordinates and hence doesn’t redraw at all.
>> Is that something the spec addresses?
> I may be biased, but it seems to me that if the spec describes the static
> rendering and does not explicitly define how it changes dynamically then the
> correct conceptual behavior is that any time any of the inputs to the
> rendering algorithm changes the whole thing is rerendered.  This is the
> behavior that browsers should pretend is taking place (in reality, of
> course, the optimize away as much of the "redoing the work" aspect as they
> can).
> So unless the spec explicitly says otherwise, the Firefox behavior seems
> most correct to me, with WebKit's being maybe barely acceptable.

Agreed.  Our behavior isn't great, but it's understandable; Firefox's
behavior is the best.


Received on Sunday, 30 October 2011 18:40:26 UTC