Implemented BluetoothAdapterMac::AddDevices().

device/bluetooth/bluetooth_device.cc

 // Copyright (c) 2012 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 "device/bluetooth/bluetooth_device.h"
 
 #include <string>
 
 #include "base/utf_string_conversions.h"
 #include "device/bluetooth/bluetooth_utils.h"
 #include "grit/device_bluetooth_strings.h"
 #include "ui/base/l10n/l10n_util.h"
 
 namespace device {
 
 // static
 bool BluetoothDevice::IsUUIDValid(const std::string& uuid) {
   return !bluetooth_utils::CanonicalUuid(uuid).empty();
 }
 
-BluetoothDevice::BluetoothDevice()
-    : bluetooth_class_(0),
+BluetoothDevice::BluetoothDevice(const std::string& name,
+                                 const std::string& address,
+                                 uint32 bluetooth_class,
+                                 bool connected,
+                                 bool bonded)
+    : bluetooth_class_(bluetooth_class),
+      name_(name),
+      address_(address),
       visible_(false),
-      bonded_(false),
-      connected_(false),
+      bonded_(bonded),
+      connected_(connected),
       connectable_(true),
       connecting_(false) {

[keybuk, 2013/03/21 18:03:34]

What's the reason for adding these methods to the BluetoothDevice constructor?
That applies to all implementations

[youngki, 2013/03/21 18:43:28]

I think the member variables to BluetoothDevice should be initialized by
BluetoothDevice itself and not the subclasses. So far BluetoothDeviceWin and
BluetoothDeviceMac initialize the member variables in their constructors, so now
I think I should have BluetoothDevice constructor to initialize them instead of
having subclass constructors to do so.

[keybuk, 2013/03/21 18:46:25]

My concern is that these don't need to be member variables at all - a sub class
should chose to implement this differently.

In fact, the LE ChromeOS implementation just initializes BluetoothDeviceChromeOS
using a BluetoothDeviceClient::Properties* pointer, which it stores in the class
- and address() looks up data through that pointer, rather than using a member
variable.

[youngki, 2013/03/21 20:07:20]

Currently BluetoothDevice::address() is returning const std::string&. Do you
want this to be just const std::string instead of const std::string&?

Also if I follow what you are planning to do, I should make the following
methods pure virtual:

BluetoothDevice::address()
BluetoothDevice::GetName()
BluetoothDevice::IsBonded()
BluetoothDevice::IsConnected()

Let me know if this is what you want.

[keybuk, 2013/03/25 16:49:47]

It should probably just return std::string; and yes the others should be pure
virtual in this function

(I'm also ok with you not making that change btw, but if you're changing to pass
in all the property values in the constructor, you just "owned" the change ;p)

[youngki, 2013/03/25 17:52:03]

So according to the new design doc it's likely we should change the interface of
BluetoothAdapter and BluetoothDevice, so we probably have to change this file
again. In this case I am more inclined to just submit this change as is now,
then we can discuss the new interface of BluetoothDevice.

What do you think? :)

 }
 
 BluetoothDevice::~BluetoothDevice() {
 }
 
 const std::string& BluetoothDevice::address() const {
   return address_;
 }
 
 string16 BluetoothDevice::GetName() const {
@@ skipping 153 matching lines @@
   const BluetoothDevice::ServiceList& services = GetServices();
   for (BluetoothDevice::ServiceList::const_iterator iter = services.begin();
        iter != services.end();
        ++iter) {
     if (bluetooth_utils::CanonicalUuid(*iter) == canonical_uuid)
       return true;
   }
   return false;
 }
 
+void BluetoothDevice::SetVisible(bool visible) {
+  visible_ = visible;
+}
+
 }  // namespace device