Disable accelerated 2d canvas on the mac. The current implementation is based on SKIA and isn't ...

Disable accelerated 2d canvas on the mac.  The current implementation is based on SKIA and isn't yet available on this platform.
BUG=81678
TEST=verified that the mac now uses the non-accelerated path and canvas actually works!

[vangelis, 2011-05-31 03:11:54]

Please review when you get a chance.

[Avi (use Gerrit), 2011-05-31 12:07:15]

LGTM with fix.

http://codereview.chromium.org/7093002/diff/1/chrome/browser/tab_contents/ren...
File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):

http://codereview.chromium.org/7093002/diff/1/chrome/browser/tab_contents/ren...
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:322:
web_prefs.accelerated_2d_canvas_enabled = false;
Need a comment with a bug URL here.

[Justin Novosad, 2011-05-31 14:31:31]

I disagree with this fix.  The proper mechanism for excluding features on a per
platform basis is in feauture_override.gypi.  We already have a compile flag for
this feature in place: ENABLE(ACCELERATED_2D_CANVAS)
The problem is that the current logic in feauture_override.gypi turns on the
accelerated 2d canvas whenever accelerated compositing is on, which is not
necessary.  On the mac, we should be building with accelerated compositing only.

The change that is required on the WebKit side is to make sure that accelerated
compositing can be used without accelerated 2d canvas, which is currently not
the case because of unresolved symbols from TilingData.cpp.  To fix that, it is
just a matter of changing the #if at the top of that file.
Once that change is rolled into chromium, the change to feature_override.gypi
will not break the mac build.

On 2011/05/31 12:07:15, Avi wrote:
> LGTM with fix.
> 
>
http://codereview.chromium.org/7093002/diff/1/chrome/browser/tab_contents/ren...
> File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):
> 
>
http://codereview.chromium.org/7093002/diff/1/chrome/browser/tab_contents/ren...
> chrome/browser/tab_contents/render_view_host_delegate_helper.cc:322:
> web_prefs.accelerated_2d_canvas_enabled = false;
> Need a comment with a bug URL here.

[Avi (use Gerrit), 2011-05-31 14:49:38]

junov clearly is better-informed than I am. Please follow his advice.

[vangelis, 2011-05-31 16:27:27]

On 2011/05/31 14:49:38, Avi wrote:
> junov clearly is better-informed than I am. Please follow his advice.

Thanks, Justin.  I like your solution better too!  I'm closing this one since
you're already working on the GYP-based solution.