[Webkit-unassigned] [Bug 82251] [chromium] Layers should should know their visibility outside their content bounds

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=82251





--- Comment #10 from Adrienne Walker <enne at google.com>  2012-03-27 15:42:54 PST ---
(In reply to comment #6)

> So, it seems like we should use that as a replacement for visibleRect then no? It would also have the advantage of not having the axis-aligned bounding box applied, in case we wanted the exact transformed rect.

I would need to be convinced that this could be a replacement for visibleRect.  It seems to me that we'd need to add a bunch of intersect(contentBounds()) calls elsewhere.

(In reply to comment #7)

> I think Enne's suggestion is different than this (which is effectively the same as the unclipped visible rect, right Enne?).. Just without the need to compute/store it when we compute the visible rect.

Yes (and yes).

(In reply to comment #8)

> Sorry, one last question.. Are render surfaces always clipped to the viewport? Ideally what we actually want is the viewport transformed into content space, so we can expand that, which will expand into the content space even if the viewport was initially outside of it.

I'm not sure that you really want the viewport.  What if a surface masks to bounds and something is animating inside of that? That's why I was thinking clip rect and not viewport.

(In reply to comment #9)

> Simplified idea: visibility(Tile T) = distance outside target surface * visibility(closest tile in target surface)

That's along the lines of what I was thinking too.  If you project the target surface's clip rect into a FloatQuad in content space, you can make a prioritization function based on distance from tile rect to that projected clip rect, with the assumption that tiles that are closer to the clip rect will be visible first.  You might additionally be able to add in animation velocity as a second check and do a separating axes test to make sure that the tile rect will ever intersect the clip rect.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.