[Webkit-unassigned] [Bug 76274] [chromium] Refactor canvas, plugin, and video drawing to be more data-driven

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=76274





--- Comment #12 from Adrienne Walker <enne at google.com>  2012-01-17 09:28:00 PST ---
(In reply to comment #11)

> This test currently relies on some culling code which I think is going to behave differently once this refactor is complete.
> 
> CCLayerImpl::appendQuads is called for all layers, but CCLayerImpl::draw was not called for all layers.
> 
> I see two options here:
> 
> 1. Modify the test to expect CCLayerImpl::appendQuads to always be called.
>    This essentially undoes https://bugs.webkit.org/show_bug.cgi?id=75783
> 2. Test if the layer's quads are actually drawn.
> 
> Which approach do you think I should take?

This is more complicated than I expected.  As a part of refactoring all these layer types, I'm hoping to remove CCLayerImpl::draw entirely.  And, on top of that, this test stretches across way too many layers of code.  It goes all the way from layer tree => render surface lists => quads => layer drawing code => graphics context for testing.  Ack.

For what it's worth, I would expect appendQuads to always be called for all of those layers.  The change from bug 75783 is not worth preserving, because culling is well-tested (and better-tested) elsewhere.  I'd rather do that than test the drawing of quads in this test.

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