cc: Move anti-aliasing decision to parent compositor.

cc: Move anti-aliasing decision to parent compositor.

This moves the final anti-aliasing decision to the renderer where
the exact tile transformation is known.

BUG=169163
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=176739

[enne (OOO), 2013-01-10 01:59:34]

lgtm

[commit-bot: I haz the power, 2013-01-10 14:50:59]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/reveman@chromium.org/11829047/1

[commit-bot: I haz the power, 2013-01-10 15:17:57]

Retried try job too often on ios_dbg_simulator for step(s) sync_unit_tests,
unit_tests

[Tom Hudson, 2013-01-10 15:36:54]

I think the comment in gl_renderer.cc:781-782 is out of date (as well has having
a grammar error), and the conditional in 783 is surprising (we antialias if the
quad is *not* rectilinear? that's not at all obvious) so deserves a comment.

[danakj, 2013-01-10 15:40:41]

On Thursday, January 10, 2013, wrote:

> I think the comment in gl_renderer.cc:781-782 is out of date (as well has
> having
> a grammar error), and the conditional in 783 is surprising (we antialias
> if the
> quad is *not* rectilinear? that's not at all obvious) so deserves a
> comment.


it means the quad is not axis aligned, so AA is needed. if its rectilinear
(axis aligned) why would you use AA?


>
https://codereview.chromium.**org/11829047/<https://codereview.chromium.org/1...
>

[reveman, 2013-01-10 19:50:10]

On 2013/01/10 15:40:41, danakj wrote:
> On Thursday, January 10, 2013, wrote:
> 
> > I think the comment in gl_renderer.cc:781-782 is out of date (as well has
> > having
> > a grammar error), and the conditional in 783 is surprising (we antialias
> > if the
> > quad is *not* rectilinear? that's not at all obvious) so deserves a
> > comment.
> 
> 
> it means the quad is not axis aligned, so AA is needed. if its rectilinear
> (axis aligned) why would you use AA?

hm, it also needs to snap to the pixel grid. ie, QuadF(RectF(0.5f, 0.5f, 10.0f,
10.0f)) still needs anti-aliasing. I swear I had such a check when I first wrote
this aa code but it looks like it might be gone now.

[reveman, 2013-01-11 16:58:34]

Removed old comment and added a TODO about problem with axis-aligned layers.
I'll fix that bug once solid color aa has landed as having that in place makes
it easy to verify correctness in gl_renderer_pixeltest.cc. Sounds good?

[Tom Hudson, 2013-01-14 09:49:23]

Sorry, didn't realize you were waiting for me; LGTM.

[commit-bot: I haz the power, 2013-01-14 17:35:03]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/reveman@chromium.org/11829047/16001

[commit-bot: I haz the power, 2013-01-14 22:24:05]

Change committed as 176739