Refactor lifetime management of UserMediaController

third_party/WebKit/Source/web/WebLocalFrameImpl.cpp

 /*
  * Copyright (C) 2009 Google Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
  *     * Redistributions in binary form must reproduce the above
@@ skipping 1406 matching lines @@
 }
 
 WebLocalFrameImpl::WebLocalFrameImpl(WebTreeScopeType scope, WebFrameClient* client)
     : WebLocalFrame(scope)
     , m_frameLoaderClientImpl(FrameLoaderClientImpl::create(this))
     , m_frameWidget(0)
     , m_client(client)
     , m_autofillClient(0)
     , m_contentSettingsClient(0)
     , m_inputEventsScaleFactorForEmulation(1)
-    , m_userMediaClientImpl(this)
     , m_webDevToolsFrontend(0)
     , m_selfKeepAlive(this)
 {
     frameCount++;
 }
 
 WebLocalFrameImpl::WebLocalFrameImpl(WebRemoteFrame* oldWebFrame, WebFrameClient* client)
     : WebLocalFrameImpl(oldWebFrame->inShadowTree() ? WebTreeScopeType::Shadow : WebTreeScopeType::Document, client)
 {
 }
@@ skipping 23 matching lines @@
     m_frame = frame;
 
     // FIXME: we shouldn't add overhead to every frame by registering these objects when they're not used.
     if (!m_frame)
         return;
 
     if (m_client)
         providePushControllerTo(*m_frame, m_client->pushClient());
 
     provideNotificationPermissionClientTo(*m_frame, NotificationPermissionClientImpl::create());
-    provideUserMediaTo(*m_frame, &m_userMediaClientImpl);
+    provideUserMediaTo(*m_frame, UserMediaClientImpl::create(m_client->userMediaClient()));

[dcheng, 2016/08/17 02:32:47]

Can m_client be null here? It's hard to tell.

(Personally, I'm not sure if we should allow WebLocalFrame to be created with a
null client, it seems to lead to more problems than it solves...)

[haraken, 2016/08/17 02:41:27]

Yeah, as far as I look at line 1466 and line 1480 etc, m_client can be null.

However, m_client can be null only after WebLocalFrameImpl::close() is called?
If that's the case, we won't need to instantiate these supplement objects in the
first place...

I'll look into more details.

     provideIndexedDBClientTo(*m_frame, IndexedDBClientImpl::create());
     provideLocalFileSystemTo(*m_frame, LocalFileSystemClient::create());
     provideNavigatorContentUtilsTo(*m_frame, NavigatorContentUtilsClientImpl::create(this));
 
     bool enableWebBluetooth = RuntimeEnabledFeatures::webBluetoothEnabled();
 #if OS(CHROMEOS) || OS(ANDROID) || OS(MACOSX)
     enableWebBluetooth = true;
 #endif
     if (enableWebBluetooth)
         BluetoothSupplement::provideTo(*m_frame, m_client ? m_client->bluetooth() : nullptr);
@@ skipping 666 matching lines @@
 {
     frame()->loader().forceSandboxFlags(static_cast<SandboxFlags>(flags));
 }
 
 void WebLocalFrameImpl::clearActiveFindMatch()
 {
     ensureTextFinder().clearActiveFindMatch();
 }
 
 } // namespace blink