Migrate WaitableEvent to enum-based constructor in chrome/

chrome/browser/net/predictor.cc

Index: chrome/browser/net/predictor.cc
diff --git a/chrome/browser/net/predictor.cc b/chrome/browser/net/predictor.cc
index 4904ec17df74b39a597568dc356f7dc0ebe1bef6..f04d2d0ef3c9450a923aa1c9f8cb2808f77d09f3 100644
--- a/chrome/browser/net/predictor.cc
+++ b/chrome/browser/net/predictor.cc
@@ -794,7 +794,9 @@ void Predictor::SaveStateForNextStartupAndTrim() {
   if (!CanPreresolveAndPreconnect())
     return;
 
-  base::WaitableEvent completion(true, false);
+  base::WaitableEvent completion(
+      base::WaitableEvent::ResetPolicy::MANUAL,

[Nico, 2016/06/04 01:48:18]

if this must be an enum class i feel it shouldn't have been a nested one. then
these would be half as verbose.

[gab, 2016/06/04 11:45:20]

We thought about that [1] but decided to go this route because every namespace
you remove you have to makeup for with a more specific enum name, i.e.
base::ResetPolicy is too generic and say base::WaitableEventResetPolicy is
merely saving a ::.

[1] https://crbug.com/612843#c11

[Nico, 2016/06/04 16:33:56]

I submit that if every step along the way seems to have made sense but the end
state is base::WaitableEvent::InitialState::NOT_SIGNALED, then one of the steps
along the way wasn't the right step, even if it felt right at the time ;-)

[gab, 2016/06/04 19:03:22]

Hmm, do you have a specific counterproposal? I'm happy to change it if you have
something we all agree is better. 

+      base::WaitableEvent::InitialState::NOT_SIGNALED);
 
   ListPrefUpdate update_startup_list(user_prefs_,
                                      prefs::kDnsPrefetchingStartupList);