Force composited opacity to imply flattening |
||||
Issue descriptionRight now, browsers implement a quirk in which composited opacity of preserves-3d descendants applies before flattening into the ancestor. This causes unfortunate complexity in effect property trees. We should force flattening before applying composited opacity, if it is web compatible. Related design doc: https://docs.google.com/document/d/1ImVe52dXZnaTWnmBThSgOvYsWqrTqDr6XlXbbnqypMg/edit?pli=1 Assigning to myself for the next step of reaching out to representatives of other browsers to see if they would be ok with making this change along with us.
,
May 26 2016
Use counter implemented in: https://codereview.chromium.org/1987283003
,
Jun 6 2016
Data from beta channel, 7-day aggregation: OpacityWithPreserve3DQuirk hits 0.006% of the time. Stable channel, 7-day aggregation: 0.0006%
,
Jun 6 2016
,
Jun 10 2016
trchen: we have 3 LGTMs to ship this. Could you do the honors?
,
Jun 13 2016
Yes sure! I will put up a CL today. Thanks for pushing this through!
,
Sep 27 2016
Why do you guys think this change is fine from a 3D-developer perspective? Applying opacity and having a 3D object's vertices flattened doesn't seem intuitive from a 3D-developer perspective. It seems to me this only makes it easy to implement, not that it is actually correct behavior. I proposed a `blend-3d` idea at https://lists.w3.org/Archives/Public/public-fx/2016JulSep/0077.html Another possibility could be to have an `alpha` property that is for use with 3D content, whereas `opacity` would be unchanged from what it is in Chrome 53, being mostly useful for 2D content.
,
Sep 13
Now opacity other than 1.0 will force transform-style:flat. |
||||
►
Sign in to add a comment |
||||
Comment 1 by petermayo@chromium.org
, May 18 2016