MacViews: Ensure the RootView is in place before the first call to ReorderNativeViews

MacViews: Ensure the RootView is in place before the first call to ReorderNativeViews

This gets 3 views_unittests passing:
 - ViewTargeterTest.HitTestCallsOnView,
 - ViewTest.CanProcessEventsWithinSubtree,
 - ViewTest.GetEventHandlerForRect.

These were failing because the size of the RootView was being reset to
the Widget size when first added. This happened because NativeWidgetMac
was not ensuring the RootView was set up properly immediately after a
call to Widget::Init() for all Widget types (i.e. those without a
NonClientView).

When there _is_ a NonClientView, Widget::Init() calls SetContentsView()
which would eventually trickle down to
NativeWidgetMac::ReorderNativeViews(), which is where NativeWidgetMac
would set up the root view.

BUG=378134
Committed: https://crrev.com/f1d241a504fb4cdc18501ffb7b0369aa0135cc9d
Cr-Commit-Position: refs/heads/master@{#294296}

[tapted, 2014-09-10 06:43:18]

tapted@chromium.org changed reviewers:
+ sky@chromium.org

[tapted, 2014-09-10 06:43:19]

Hi sky - please take a look.

This is a piece from the earlier NativeWidgetHostMac CL (needed there to avoid
test regressions). I'm taking a different approach for that, and wanted to wrap
my head around this better.

Exactly why these particular tests were failing (and the added code fixes them)
was a bit subtle, so I've added a cross-platform test that captures the root
cause.

[sky, 2014-09-10 14:48:36]

LGTM

https://codereview.chromium.org/559873002/diff/1/ui/views/widget/widget_unitt...
File ui/views/widget/widget_unittest.cc (right):

https://codereview.chromium.org/559873002/diff/1/ui/views/widget/widget_unitt...
ui/views/widget/widget_unittest.cc:3106: const gfx::Rect kTestRect(0, 0, 500,
500);
nit: k is only for global or constants defined within a class, not functional
locals.

[tapted, 2014-09-11 00:17:16]

https://codereview.chromium.org/559873002/diff/1/ui/views/widget/widget_unitt...
File ui/views/widget/widget_unittest.cc (right):

https://codereview.chromium.org/559873002/diff/1/ui/views/widget/widget_unitt...
ui/views/widget/widget_unittest.cc:3106: const gfx::Rect kTestRect(0, 0, 500,
500);
On 2014/09/10 14:48:36, sky wrote:
> nit: k is only for global or constants defined within a class, not functional
> locals.

Done.

[tapted, 2014-09-11 00:17:21]

The CQ bit was checked by tapted@chromium.org

[commit-bot: I haz the power, 2014-09-11 00:19:27]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patchset/559873002/20001

[commit-bot: I haz the power, 2014-09-11 02:09:24]

Committed patchset #2 (id:20001) as 05d86a9adc1789af6ff9c694e54166827e21c4d0

[commit-bot: I haz the power, 2014-09-11 03:09:15]

Patchset 2 (id:??) landed as
https://crrev.com/f1d241a504fb4cdc18501ffb7b0369aa0135cc9d
Cr-Commit-Position: refs/heads/master@{#294296}