GraphicsLayer contentsVisible property miscomputed when using backing sharing

[Antti Koivisto]

The attached test case produces layer tree

(GraphicsLayer
  (anchor 0.00 0.00)
  (bounds 800.00 600.00)
  (deep color 1)
  (children 1
    (GraphicsLayer
      (bounds 800.00 600.00)
      (contentsOpaque 1)
      (deep color 1)
      (children 2
        (GraphicsLayer
          (position 8.00 8.00)
          (bounds 500.00 500.00)
          (contentsOpaque 1)
          (deep color 1)
        )
        (GraphicsLayer
          (position 30.00 30.00)
          (bounds 250.00 250.00)
          (drawsContent 1)
          (contentsVisible 0)
          (deep color 1)
        )
      )
    )
  )
)

The layer with (contentsVisible 0) has visible contents in backing sharing descendants.

[Antti Koivisto]

Created attachment 369729 [details]
test

[Antti Koivisto]

Maybe we can simply remove contentsVisible and userInteractionEnabled GraphicsLayer properties? Their purpose seems to be entirely subsumed by the event region.

[Simon Fraser (smfr)]

Does the event region control whether an overflow scroll is interactive? If so, then it sounds like we can remove these. But we need to make sure that the pointer-events and visibility properties are correctly reflected in event regions, and that event regions correctly disable interaction on UIScrollViews. More tests required!

[Antti Koivisto]

(In reply to Simon Fraser (smfr) from comment #3)
> Does the event region control whether an overflow scroll is interactive?

Yep.

> If
> so, then it sounds like we can remove these. But we need to make sure that
> the pointer-events and visibility properties are correctly reflected in
> event regions, and that event regions correctly disable interaction on
> UIScrollViews. More tests required!

Event regions do respect both visibility and pointer-events and has some testing for those too already.