Improving the fake clock and using it to fix a flaky STUN timeout test.

webrtc/base/timeutils.h

Index: webrtc/base/timeutils.h
diff --git a/webrtc/base/timeutils.h b/webrtc/base/timeutils.h
index 78ebacee38df5087737de039fa20a6e8c485c9f6..113793ade3c81567a317ad7a3669bc08d3aa72c9 100644
--- a/webrtc/base/timeutils.h
+++ b/webrtc/base/timeutils.h
@@ -39,8 +39,10 @@ class ClockInterface {
 
 // Sets the global source of time. This is useful mainly for unit tests.
 //
-// Does not transfer ownership of the clock.
-// SetClock(nullptr) should be called before the ClockInterface is deleted.
+// Returns the previously set ClockInterface, or nullptr if none is set.
+//
+// Does not transfer ownership of the clock. SetClockForTesting(nullptr)
+// should be called before the ClockInterface is deleted.
 //
 // This method is not thread-safe; it should only be used when no other thread
 // is running (for example, at the start/end of a unit test, or start/end of
@@ -49,7 +51,7 @@ class ClockInterface {
 // TODO(deadbeef): Instead of having functions that access this global

[nisse-webrtc, 2016/06/08 09:47:30]

I don't think it's a good idea to pass around the clock everywhere (even if
there's some code using webrtc::Clock doing that).

Except maybe if we can cut down a lot on the number of places where the clock is
read. And in many cases, I think passing the current time as an argument
downstream makes more sense than passing down a clock object.

 // ClockInterface, we may want to pass the ClockInterface into everything
 // that uses it, eliminating the need for a global variable and this function.
-void SetClock(ClockInterface* clock);
+ClockInterface* SetClockForTesting(ClockInterface* clock);
 
 // Returns the current time in milliseconds in 32 bits.
 uint32_t Time32();