Re: [filters] notes to CSS filters

Am 23.09.2011 um 08:25 schrieb Robert O'Callahan:

> On Fri, Sep 23, 2011 at 6:20 PM, Dirk Schulze <vbs85@gmx.de> wrote:
> Am 23.09.2011 um 08:08 schrieb Robert O'Callahan:
>> On Fri, Sep 23, 2011 at 5:02 PM, Dirk Schulze <vbs85@gmx.de> wrote:
>> But that is exactly the point. If we don't have a filter region, the filter effect would be unbound! Right now just the filter region bounds lightning effects. What else can clip the visible area?
>> 
>> The viewport? :-)
> 
> I am sure you remember the ECMA-cloud [1]. On WebKit we did not have an intelligent auto clipping algorithm from the beginning. And blurring all elements on this site took more than a minute (1:30 to 2 IIRC). Imagine this page would use lightning effects instead of blur effects. We would need even more time.
> 
> Yeah, I fixed that bug in Gecko too :-).
> 
> If there really is nothing bounding the lighting effect, then it will cover the viewport, and since that's probably not what the author wants, they'll fix it by clipping the element somehow or combining the lighting image with some kind of mask, presumably bounded. Hopefully whatever they use to restrict the visibility of the lighting can be used by the browser to limit how much of the lighting image needs to be computed.


You trust web developers a lot ;) At the end this could cause a browser to be unresponsive. But viewport would at least bound filters, thats true.

Dirk