[System-Keyboard-Lock] Forward navigator functions to RenderFrameHost

content/browser/keyboard_lock/keyboard_lock_service_impl.cc

Index: content/browser/keyboard_lock/keyboard_lock_service_impl.cc
diff --git a/content/browser/keyboard_lock/keyboard_lock_service_impl.cc b/content/browser/keyboard_lock/keyboard_lock_service_impl.cc
new file mode 100644
index 0000000000000000000000000000000000000000..2db29c7e992abf2bfcfe81847fbcbb51e04ba4a7
--- /dev/null
+++ b/content/browser/keyboard_lock/keyboard_lock_service_impl.cc
@@ -0,0 +1,35 @@
+// Copyright 2017 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "content/browser/keyboard_lock/keyboard_lock_service_impl.h"
+
+#include "content/public/browser/render_frame_host.h"
+
+namespace content {
+
+KeyboardLockServiceImpl::KeyboardLockServiceImpl(
+    blink::mojom::KeyboardLockServiceRequest request)
+    : binding_(this, std::move(request)) {}

[dcheng, 2017/04/26 13:27:17]

Nit: #include <utility>

[Hzj_jie, 2017/04/26 22:05:56]

Done.

+
+KeyboardLockServiceImpl::~KeyboardLockServiceImpl() = default;
+
+// static
+void KeyboardLockServiceImpl::CreateMojoService(
+    RenderFrameHost* render_frame_host,
+    blink::mojom::KeyboardLockServiceRequest request) {
+  new KeyboardLockServiceImpl(std::move(request));

[dcheng, 2017/04/26 13:27:17]

Is this leaked? The binding isn't a strong binding, and I see no delete.

[Hzj_jie, 2017/04/26 22:05:56]

I do not think so: the object is owned by binding_.

[dcheng, 2017/04/27 02:53:10]

Talked offline and we agree that this needs to be cleaned up.

I would just recommend using mojo::MakeStrongBinding() for now; we can change it
later if it doesn't make sense as the implementation is filled out.

[Hzj_jie, 2017/04/28 00:10:31]

Done.

+}
+
+void KeyboardLockServiceImpl::RequestKeyLock(
+    const std::vector<std::string>& key_codes,
+    const RequestKeyLockCallback& callback) {
+  // TODO(zijiehe): Implementation required.

[dcheng, 2017/04/26 13:27:17]

For IPC reviews, it's generally beneficial to include the actual service
implementation--otherwise, it's hard to evaluate for the IPC security review. Is
there any possibility of doing that here?

[Hzj_jie, 2017/04/26 22:05:56]

I'd only say we still have bunch of changes needed to support this callback,
which is not likely to be able to be done in this change.

[dcheng, 2017/04/27 02:53:10]

Please include an IPC security reviewer on the followup CLs then.

[Hzj_jie, 2017/04/28 00:10:31]

Done.

+  callback.Run(true, std::string());
+}
+
+void KeyboardLockServiceImpl::CancelKeyLock() {
+  // TODO(zijiehe): Implementation required.
+}
+
+}  // namespace content