Notify GCMAppHandlers when the store is reset, so they clear cached IDs

chrome/browser/chromeos/policy/heartbeat_scheduler.cc

Index: chrome/browser/chromeos/policy/heartbeat_scheduler.cc
diff --git a/chrome/browser/chromeos/policy/heartbeat_scheduler.cc b/chrome/browser/chromeos/policy/heartbeat_scheduler.cc
index 6782c70d408e8dd6e693c53d8c05ebc86071abc2..92148858acc2264e9c900a5102d17335be369f45 100644
--- a/chrome/browser/chromeos/policy/heartbeat_scheduler.cc
+++ b/chrome/browser/chromeos/policy/heartbeat_scheduler.cc
@@ -424,6 +424,11 @@ void HeartbeatScheduler::ShutdownHandler() {
   NOTREACHED() << "HeartbeatScheduler should be destroyed before GCMDriver";
 }
 
+void HeartbeatScheduler::OnStoreReset() {
+  // Nothing to do here since HeartbeatScheduler doesn't locally persist
+  // registration ID, instead it just re-registers.

[Peter Beverloo, 2016/11/03 17:28:56]

registration_id_?

[Peter Beverloo, 2016/11/03 17:28:56]

What about CloudPolicyClient::UpdateGcmId()? There's a TODO on line 331 that
says we send the IDs to the server. It also says we re-send it each time anyway.

Up to the OWNERS, but what about mimicking that TODO?

[johnme, 2016/11/08 19:22:41]

Done.

[johnme, 2016/11/08 19:22:41]

Done - I now call ShutdownGCM.

+}
+
 void HeartbeatScheduler::OnMessage(const std::string& app_id,
                                    const gcm::IncomingMessage& message) {
   // Should never be called because we don't get any incoming messages