Load built in tts engine when launching tts settings page

We are typically extremely conservative when loading the built in tts engine. The only codepath we do this before was in TtsControllerImpl, in response to a speak call.

This change enables loading immediately after the tts settings page registers js message handlers in C++, where a non-null profile is available.

Test: on a fresh profile, navigate to settings -> accessibility ->  tts -> Google tts engine settings. Verify all voices show up.
Change-Id: Id2294f25f034da0a254d904c1325e0d8ba274d21
Reviewed-on: https://chromium-review.googlesource.com/1120948Reviewed-by: Hector Carmona <hcarmona@chromium.org>
Commit-Queue: David Tseng <dtseng@chromium.org>
Cr-Commit-Position: refs/heads/master@{#572091}

chrome/browser/ui/webui/settings/tts_handler.cc

@@ -30,6 +30,9 @@ void TtsHandler::HandleGetAllTtsVoiceData(const base::ListValue* args) {
 }
 
 void TtsHandler::HandleGetTtsExtensions(const base::ListValue* args) {
+  // Ensure the built in tts engine is loaded to be able to respond to messages.
+  TtsExtensionEngine::GetInstance()->LoadBuiltInTtsExtension(
+      Profile::FromWebUI(web_ui()));
   base::ListValue responses;
   Profile* profile = Profile::FromWebUI(web_ui());
   extensions::ExtensionRegistry* registry =