- From: Rik Cabanier <cabanier@gmail.com>
- Date: Thu, 14 Mar 2013 11:21:32 -0700
- To: Simon Fraser <smfr@me.com>
- Cc: "L. David Baron" <dbaron@dbaron.org>, Dirk Schulze <dschulze@adobe.com>, "robert@ocallahan.org" <robert@ocallahan.org>, "public-fx@w3.org" <public-fx@w3.org>
- Message-ID: <CAGN7qDCg_T60Qah3GZuxyTQP1fXuu4tid6hYZ4ghMwu1Mhvr-Q@mail.gmail.com>
On Thu, Mar 14, 2013 at 11:04 AM, Simon Fraser <smfr@me.com> wrote: > > On Mar 14, 2013, at 10:43 AM, Rik Cabanier <cabanier@gmail.com> wrote: > > On Thu, Mar 14, 2013 at 10:21 AM, L. David Baron <dbaron@dbaron.org>wrote: > >> On Thursday 2013-03-14 09:55 -0700, Rik Cabanier wrote: >> > On Thu, Mar 14, 2013 at 9:19 AM, L. David Baron <dbaron@dbaron.org> >> wrote: >> > >> > > On Thursday 2013-03-14 05:44 -0700, Dirk Schulze wrote: >> > > > No, not WebKit's rules. And I do not think that we want to specify >> > > > buffering. A behavior in situations like scrolling for blending >> > > > should be specified and browser need to follow. I see that this >> > > > can be challenging but would be most desireabale. After all, >> > > > scrolling should not affect the browser experience of the user on >> > > > the visual side - especially for blending. >> > > >> > > As I said in http://dbaron.org/log/20130306-compositing-blending , I >> > > think there's a lot less to specify and a lot less to drive towards >> > > interoperability if compositing and blending operations are limited >> > > to things that create stacking contexts. This limitation would be >> > > present if background-blend-mode and background-composite are >> > > dropped, which I think should be done. >> > >> > >> > David, >> > the exact same issue will happen if blending applies to elements (in >> which >> > case stacking contexts are created). >> > For instance, an element with blending that is a child of an element >> that >> > uses fixed positioning will render differently today in FF and WK. >> > >> > Dropping background-blend-mode will not solve this problem. >> >> Dropping background-blend-mode simplifies it a lot, because you only >> have to consider elements that form stacking contexts when >> addressing it. > > > Maybe I'm not seeing the problem that makes background-blend-mode harder. > What would be the difference between: > > <div style="position: fixed"> > > <img src="ducky.png" style="mix-blend-mode: multiply"> > > and > > <div style="position: > fixed;backgound:url('ducky.png');background-blend-mode:multiply"> > > Both will render differently because 'position:fixed' creates an offscreen > buffer in WK but not in FF > > > I'm confused about background-blend-mode, and the spec < > https://dvcs.w3.org/hg/FXTF/rawfile/tip/compositing/index.html#background-blend-mode> > is not helping. > > Does it blend with anything in other elements, or only between the > background images/background color of the element to which it's applied? > > The spec's use of "the background images at a lower z-index." is confusing > here. Is that "the background images of elements at a lower z-index", or > "later background images of the element"? > > If it's only within elements and not between them, I don't see why it > matters whether position:fixed creates a group. > > >From the spec: Each background image will blend with the element's background and the background images at a lower z-index So, it will blend between elements. David Baron previously proposed to group background images so they only blended with each other and not the backdrop. (This would necessitate creating a new off screen buffer.) This is certainly viable, but not as powerful. Since we have to solve this problem for regular elements anyway, I don't think it helps to group them.
Received on Thursday, 14 March 2013 18:22:03 UTC