Do not dynamically call setLayer on RenderWidgetHostViewCocoa

Do not dynamically call setLayer on RenderWidgetHostViewCocoa

Instead, call it once at creation, and then call addSublayer
to cover up the layer when content comes in.

Remove the atomic add-layer-and-remove-old-one functions,
since the new arrangement allows adding the new layer before
removing the old one.

Change places that used [self layer] to explicitly call out to the
software and compositing layers.

The reason for this change is that it is not safe to dynamically
call setLayer on NSViews because of the following undocumented
feature of CoreAnimation:

Ordinarily, the NSView hierarchy and the CALayer hierarchy
match. That is, the ordering of a NSView's subviews' CALayers is
the same as the ordering of the NSView's CALayer's sublayers.
This gets completely broken when you call setLayer on a NSView.
The new CALayer will be appended to the end of the NSView's
superview's CALayer's sublayer array, even if that NSView wasn't at
the end of its superview's subview array.

BUG=348490
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=255886

[ccameron, 2014-03-06 22:31:49]

Patch is simpler than the explanation, promise!

[Ken Russell (switch to Gerrit), 2014-03-07 00:29:56]

https://codereview.chromium.org/188873005/diff/20001/content/browser/renderer...
File content/browser/renderer_host/render_widget_host_view_mac.mm (right):

https://codereview.chromium.org/188873005/diff/20001/content/browser/renderer...
content/browser/renderer_host/render_widget_host_view_mac.mm:1481: if
(!CreateCompositedIOSurfaceLayer()) {
Who takes care of deleting the software layer in this case?

https://codereview.chromium.org/188873005/diff/20001/content/browser/renderer...
content/browser/renderer_host/render_widget_host_view_mac.mm:1987:
CreateSoftwareLayer();
Who takes care of deleting the composited IOSurface layer in this case?

[ccameron, 2014-03-07 07:30:00]

https://codereview.chromium.org/188873005/diff/20001/content/browser/renderer...
File content/browser/renderer_host/render_widget_host_view_mac.mm (right):

https://codereview.chromium.org/188873005/diff/20001/content/browser/renderer...
content/browser/renderer_host/render_widget_host_view_mac.mm:1481: if
(!CreateCompositedIOSurfaceLayer()) {
On 2014/03/07 00:29:56, Ken Russell wrote:
> Who takes care of deleting the software layer in this case?

I moved this to where the software frame and backing store are removed at [1] in
GotAcceleratedFrame.

Once we remove the non-CoreAnimation and legacy-software code (maybe in 2
weeks?), I may in-line the GotSoftware/AcceleratedFrame code.

https://codereview.chromium.org/188873005/diff/20001/content/browser/renderer...
content/browser/renderer_host/render_widget_host_view_mac.mm:1982:
DestroySoftwareLayer();
[1] from above.

https://codereview.chromium.org/188873005/diff/20001/content/browser/renderer...
content/browser/renderer_host/render_widget_host_view_mac.mm:1987:
CreateSoftwareLayer();
On 2014/03/07 00:29:56, Ken Russell wrote:
> Who takes care of deleting the composited IOSurface layer in this case?

This is done further down in the if (last_frame_was_accelerated_) block's
DestroyCompositedIOSurfaceAndLayer.

[Ken Russell (switch to Gerrit), 2014-03-07 20:52:44]

OK. Thanks for bearing with me. LGTM

[ccameron, 2014-03-08 01:25:52]

Thanks for being thorough!

[ccameron, 2014-03-08 01:25:55]

The CQ bit was checked by ccameron@chromium.org

[commit-bot: I haz the power, 2014-03-08 11:16:38]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ccameron@chromium.org/188873005/20001

[commit-bot: I haz the power, 2014-03-08 14:11:24]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-03-08 14:11:25]

Retried try job too often on mac_rel for step(s) telemetry_perf_unittests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=mac_rel&nu...

[ccameron, 2014-03-10 05:44:24]

The CQ bit was checked by ccameron@chromium.org

[commit-bot: I haz the power, 2014-03-10 05:44:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ccameron@chromium.org/188873005/20001

[commit-bot: I haz the power, 2014-03-10 06:36:00]

Change committed as 255886