Re: Updated filters specification

On Apr 28, 2011, at 12:28 AM, Erik Dahlstrom wrote:

> On Wed, 27 Apr 2011 06:05:13 +0200, Vincent Hardy <vhardy@adobe.com> wrote:
> 
>> 
>> On Apr 26, 2011, at 8:55 PM, Robert O'Callahan wrote:
>> 
>> On Tue, Apr 26, 2011 at 7:20 PM, Erik Dahlstrom  
>> <ed@opera.com<mailto:ed@opera.com>> wrote:
>> Deprecating (or dropping) 'enable-background' would essentially mean  
>> that BackgroundImage and BackgroundAlpha would always generate a  
>> transparent black result in the context of filters (I presume this is  
>> what the browsers that don't support enable-background already do, and  
>> it follows the error handling that is defined for when there was no  
>> "enable-background:new" parent element). Typically that means that  
>> things still render without errors but probably not as intended by the  
>> author.
>> 
>> Why can't we define BackgroundImage and BackgroundAlpha in a way that  
>> continues to work in the absence of enable-background? I think we can,  
>> even with GPU-accelerated rendering. In the worst case you can detect  
>> use of BackgroundImage/BackgroundAlpha and automatically infer the  
>> equivalent of enable-background for the ancestor elements.
>> 
>> I think it would be difficult to get the exact equivalent of the current  
>> behavior because currently, enable-background can be on the element's  
>> parent, grand-parent, or any other ancestor.
>> 
>> Vincent.
> 
> That's true, but in practice I think that most files with  
> 'enable-background' has had that attribute set only on the root svg  
> element.

Yes. Are you suggesting to then default to having BackgroundImage be the root svg canvas's current, accumulated rendering?

> That accounts for most of the Illustrator files I've seen anyway.  
> Btw does Illustrator always add 'enable-background' there, even when it's  
> not used by any filters?

I'll have to check on that.

Thanks,
Vincent.