MacViews: Implement NativeViewHostMac (take 3)

MacViews: Implement NativeViewHostMac (take 3)

This is done in a comparable way to NativeViewHostAura, but using a
ReparentNSView helper function which reparents an NSView to a
NativeWidgetMac's BridgedContentView (or to `nil`).

Earlier attempts tried to use NativeWidgetPrivate::ReparentNativeView()
which turns out to be a bad idea, since it's really designed to reparent
windows (aura::Windows or HWNDs). Mac's child window support is
insufficient for what toolkit-views expects, and WebContents on Mac are
managed by an NSView in any case, not an NSWindow.

Along with logic currently in render_widget_host_view_mac.mm and
web_contents_view_mac.mm, reparenting renderer views needs to be done in
the native view hierarchy, not the window hierarchy.

This CL no longer tries to enable the current
native_view_host_unittest.cc (tests there assume ReparentNativeView is
used), but adds native_view_host_mac_unittest.mm to give test coverage,
and to inspect properties at the Cocoa layer. Common parts of the 3
_unittest.* files are refactored into a NativeViewHostTestBase class.

Some other views_unittests are also made passing by implementing a
NativeViewHost::CreateWrapper that doesn't just return NULL:
4xFocusTraverslTest and NativeWidgetTest.GetTopLevelNativeWidget2

Before: 455 tests run 22 tests failed 21 tests crashed. After: 458 tests
run 22 tests failed 16 tests crashed.

BUG=378134, 415024
Committed: https://crrev.com/5ee3db9894c69bd53f19c92b38a39c4a345e8ec0
Cr-Commit-Position: refs/heads/master@{#295618}

[tapted, 2014-09-17 07:06:13]

Patchset #5 (id:80001) has been deleted

[tapted, 2014-09-17 12:24:51]

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

[tapted, 2014-09-17 12:24:52]

Hi Andre - could you take an initial look? (note there's a recent compile
regression in webview.cc that I'll need to fix too - http://crrev.com/574943002/
- still looking at alternatives to just adding #ifdefs..)

[Andre, 2014-09-17 20:57:14]

lgtm

https://codereview.chromium.org/530933002/diff/140001/ui/views/controls/nativ...
File ui/views/controls/native/native_view_host_mac.mm (right):

https://codereview.chromium.org/530933002/diff/140001/ui/views/controls/nativ...
ui/views/controls/native/native_view_host_mac.mm:119: [[host_->native_view()
window] makeFirstResponder:host_->native_view()];
From -[NSWindow makeFirstResponder:] documentation:
 If this method is called explicitly, first send acceptsFirstResponder to
responder, and do not call makeFirstResponder: if acceptsFirstResponder returns
NO.

https://codereview.chromium.org/530933002/diff/140001/ui/views/controls/nativ...
File ui/views/controls/native/native_view_host_mac_unittest.mm (right):

https://codereview.chromium.org/530933002/diff/140001/ui/views/controls/nativ...
ui/views/controls/native/native_view_host_mac_unittest.mm:77:
EXPECT_TRUE(NSEqualRects(NSZeroRect, [native_view_ frame]));
NSIsEmptyRect()?

[tapted, 2014-09-18 08:00:03]

Patchset #8 (id:160001) has been deleted

[tapted, 2014-09-18 08:21:07]

Patchset #9 (id:200001) has been deleted

[tapted, 2014-09-18 08:34:13]

Patchset #9 (id:220001) has been deleted

[tapted, 2014-09-18 08:34:25]

Patchset #9 (id:240001) has been deleted

[tapted, 2014-09-18 08:37:17]

Patchset #9 (id:260001) has been deleted

[tapted, 2014-09-18 08:40:14]

https://codereview.chromium.org/530933002/diff/140001/ui/views/controls/nativ...
File ui/views/controls/native/native_view_host_mac.mm (right):

https://codereview.chromium.org/530933002/diff/140001/ui/views/controls/nativ...
ui/views/controls/native/native_view_host_mac.mm:119: [[host_->native_view()
window] makeFirstResponder:host_->native_view()];
On 2014/09/17 20:57:13, Andre wrote:
> From -[NSWindow makeFirstResponder:] documentation:
>  If this method is called explicitly, first send acceptsFirstResponder to
> responder, and do not call makeFirstResponder: if acceptsFirstResponder
returns
> NO.

Done.

https://codereview.chromium.org/530933002/diff/140001/ui/views/controls/nativ...
File ui/views/controls/native/native_view_host_mac_unittest.mm (right):

https://codereview.chromium.org/530933002/diff/140001/ui/views/controls/nativ...
ui/views/controls/native/native_view_host_mac_unittest.mm:77:
EXPECT_TRUE(NSEqualRects(NSZeroRect, [native_view_ frame]));
On 2014/09/17 20:57:13, Andre wrote:
> NSIsEmptyRect()?

I grok the documentation for that as `width <= 0 || height <= 0`, so
NSEqualRects is a stronger check I think.

[tapted, 2014-09-18 08:42:39]

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

[tapted, 2014-09-18 08:42:40]

+sky - please take a look. You can diff against patchset 6 to see the code you
reviewed earlier (http://crrev.com/509943002 [closed]) - thanks!

[sky, 2014-09-18 12:58:29]

LGTM

[tapted, 2014-09-18 23:04:54]

The CQ bit was checked by tapted@chromium.org

[commit-bot: I haz the power, 2014-09-18 23:06:36]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/530933002/280001

[commit-bot: I haz the power, 2014-09-19 00:12:33]

Committed patchset #9 (id:280001) as 4dbb8443c28af960c0d778a506d7387afccceb01

[commit-bot: I haz the power, 2014-09-19 00:13:07]

Patchset 9 (id:??) landed as
https://crrev.com/5ee3db9894c69bd53f19c92b38a39c4a345e8ec0
Cr-Commit-Position: refs/heads/master@{#295618}