Re: [css-compositing] blending and inline SVG

On Mon, May 20, 2013 at 4:27 PM, Dirk Schulze <dschulze@adobe.com> wrote:

>
> On May 20, 2013, at 4:05 PM, Rik Cabanier <cabanier@gmail.com> wrote:
>
> >
> >
> > On Mon, May 20, 2013 at 3:42 PM, Dirk Schulze <dschulze@adobe.com>
> wrote:
> >
> > On May 20, 2013, at 3:39 PM, Rik Cabanier <cabanier@gmail.com> wrote:
> >
> > >
> > >
> > > On Mon, May 20, 2013 at 1:55 PM, Dirk Schulze <dschulze@adobe.com>
> wrote:
> > >
> > > On May 20, 2013, at 11:16 AM, Rik Cabanier <cabanier@gmail.com> wrote:
> > >
> > > > All,
> > > >
> > > > I was editing the chapter on isolation [1] and wondered if inline
> svg should happen in an isolated group. (So the <svg> tag would establish a
> new group/stacking context)
> > > > It seems that it would be very hard to implement if this was not the
> case.
> > > >
> > > > Is everyone that inline SVG is always isolated?
> > > >
> > > > We also need to discuss what other constructs in SVG create
> isolation. The current filter spec assumes that nothing does, but that
> doesn't correspond with reality.
> > >
> > > The first question is how inline SVG cooperates with HTML in general.
> We did not specify that anywhere to my knowledge. In Blink and WebKit
> inline SVG elements are handled as replacement elements, same as <img>,
> <video> or <canvas>. It would make sense to not treat inline SVG elements
> different from the other elements for these two engines. However, I would
> like to understand where you see the technical difference to other
> "graphical" HTML elements like <div> or <p>.
> > >
> > > I'm unsure what you are asking.
> > > Browser are indeed treating svg as a canvas (and not as a change from
> the css box model to the svg drawing model) so it makes sense to have the
> content isolated.
> >
> > I was more revering to the sentence "It seems that it would be very hard
> to implement if this was not the case.". The question is why it would be
> harder for SVG. Treating SVG as replacement element is not necessarily a
> problem for not having an isolation group IMO.
> >
> > Not neccesarily, but as you know, the graphics engines that the browser
> rely on, are not set up to deal with non-isolated groups.
> > Do browsers create a new context when they see an inline SVG?
>
> Blink and WebKit do not create a new context at this point and the content
> could blend with it's backdrop at this point IIRC. I don't think that this
> is much different from Firefox.
>

Interesting!
So, it seems that WebKit and Blink would have to go out of their way to
make inline SVG isolated. FWIW I think <svg> should not cause isolation.
Does anyone know how Firefox works? How about IE?


>
> >
> >
> > The question is although: Should <video> always create an isolation
> group?
> >
> > Does it matter? There is no visible content inside the video tag.
>
>