[Mac] Avoid "adding unknown subview" warning.

chrome/browser/ui/cocoa/tabs/tab_strip_controller.mm

Index: chrome/browser/ui/cocoa/tabs/tab_strip_controller.mm
diff --git a/chrome/browser/ui/cocoa/tabs/tab_strip_controller.mm b/chrome/browser/ui/cocoa/tabs/tab_strip_controller.mm
index 626d742e101675067ae37c08e24de353f009093a..42cdd3c2e2627e3e4a2121b4cb545b48fcfa705e 100644
--- a/chrome/browser/ui/cocoa/tabs/tab_strip_controller.mm
+++ b/chrome/browser/ui/cocoa/tabs/tab_strip_controller.mm
@@ -229,6 +229,13 @@ CGFloat FlipXInView(NSView* view, CGFloat width, CGFloat x) {
   return NO;
 }
 
+// Called by AppKit to check if dragging this view should move the window.
+// NSButton overrides this method in the same way so dragging window buttons
+// has no effect.
+- (NSRect)_opaqueRectForWindowMoveWhenInTitlebar {

[tapted, 2016/10/10 23:38:06]

nit: we typically forward-declare these - I guess it's only strictly necessary
when calling [super ..] but it helps document where the private methods are
coming from. E.g. in a `@interface NSView (PrivateAPI)` at the top of the file.
(and we haven't started putting private APIs into sdk_forward_declarations, so I
guess we should copy it in tab_view.mm as well, [or give it its own header..]).

[Eugene But (OOO till 7-30), 2016/10/11 00:33:10]

Done.

+ return [self bounds];

[tapted, 2016/10/10 23:38:06]

Maybe comment here what the base class returns? (I'm guessing NSZeroRect?) --
i.e. say why [self bounds] is the right thing to convince AppKit not to move the
window

[Eugene But (OOO till 7-30), 2016/10/11 00:33:10]

Done in predeclaration.

+}
+
 - (id)initWithFrame:(NSRect)frameRect
          controller:(TabStripController*)controller {
   if ((self = [super initWithFrame:frameRect])) {