[Webkit-unassigned] [Bug 89828] Optimize RenderGeometryMap mappings gathering

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=89828





--- Comment #6 from Antti Koivisto <koivisto at iki.fi>  2012-06-24 16:13:28 PST ---
(In reply to comment #4)
> Doesn't this need to check canUseConvertToLayerCoords() on all layers between layer and ancestorLayer? There's no guarantee that ancestorLayer is an immediate ancestor.

Sounds likely. Can you think of a way to test that case? The patch is logic pretty well tested due to 

    ASSERT(enclosingIntRect(rendererMappedResult) == enclosingIntRect(FloatQuad(result).boundingBox()));

in RenderGeometryMap::absoluteRect() and we don't hit the case in layout tests. 

> I'd like to keep versions that are renderer-based, so we can call them from absoluteRects()/absoluteQuads() at some point (we already know of cases where these are a bottleneck).

We don't usually leave around dead code. It is trivial to add back later. Are you planning to use it soon?

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