Touchpad scroll latching enabled for Mac behind flag.

cc/input/scroll_state_data.h

Index: cc/input/scroll_state_data.h
diff --git a/cc/input/scroll_state_data.h b/cc/input/scroll_state_data.h
index 90339649f5ffbd475d17f5b36421923e1cae812e..62a197a2d26edd3cea43eebcdad2ec577b1604e5 100644
--- a/cc/input/scroll_state_data.h
+++ b/cc/input/scroll_state_data.h
@@ -32,6 +32,9 @@ class CC_EXPORT ScrollStateData {
   bool is_beginning;
   bool is_in_inertial_phase;
   bool is_ending;
+  // True if the next ScrollBegin is in inertial_phase (is_beginning &&
+  // in_inertial_phase). May only be true if is_ending == true.
+  bool precedes_fling;

[aelias_OOO_until_Jul13, 2016/12/02 05:17:56]

It's confusing to add a special boolean to the scroll end meaning "don't
actually end".  I'd prefer having a single flingstart event, or perhaps no event
at all since the first update event with is_in_inertial_phase = true is
self-explanatory.

As for the scroll deltas that may be attached to ScrollEnd, I'm not sure whether
that ever actually happens, but if it does, then it's a bug to delay them for
100ms anyway.  So either send a ScrollUpdate message for them separately, or
confirm that it never happens and don't worry about it.

[sahel, 2016/12/13 19:14:23]

-Done.
-GSE don't have any scroll delta. I confirm that this case won't happen ever.

 
   bool should_propagate;
   bool from_user_input;