[Webkit-unassigned] [Bug 91537] [Chromium] Textures drawn during occlusion are incorrectly re-used when unoccluded.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=91537





--- Comment #10 from zlieber at chromium.org  2012-07-19 10:49:02 PST ---
(In reply to comment #9)
> Thanks for all of the explanation.  That makes a lot of sense.  :)
> 
> Can you also convince me that != is the correct operation here rather than  intersection(targetUnoccluded, screenUnoccluded) == targetUnoccluded?
> 
> The case I'm thinking of is a surface that's rotated some non-zero amount.  It has no external occlusion, and it has internal occlusion that covers exactly half of it.  Now you're considering a layer that takes up the whole bounds of the surface.  There's no screen space occlusion because the layer is not rectilinear in screen space.  There's target space occlusion of half the layer.  However, since there's no external occlusion, you'd want to set hasOcclusionFromOutsideTargetSurface to false, but a simple inequality would return true.

The assumption that I had is that targetUnoccluded will always be bigger than screenUnoccluded and include it entirely. If that assumption is correct, the two conditions are equivalent, so checking for their intersection is equivalent to checking for their equality.

In your example, wouldn't we have a very small screenUnoccluded resulting from bounding rect of target occlusion?

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