Instrument serial API code to make testing easier.

chrome/browser/extensions/api/serial/serial_api.h

Index: chrome/browser/extensions/api/serial/serial_api.h
diff --git a/chrome/browser/extensions/api/serial/serial_api.h b/chrome/browser/extensions/api/serial/serial_api.h
index 1d31ba652eb1d0282a4a5cb4072c3a784821b447..52ff92824153c8bbf9f140650d060b6c185ea918 100644
--- a/chrome/browser/extensions/api/serial/serial_api.h
+++ b/chrome/browser/extensions/api/serial/serial_api.h
@@ -16,6 +16,7 @@
 namespace extensions {
 
 class APIResourceEventNotifier;
+class SerialConnection;
 
 extern const char kConnectionIdKey[];
 
@@ -49,6 +50,13 @@ class SerialOpenFunction : public AsyncAPIFunction {
   virtual void Work() OVERRIDE;
   virtual bool Respond() OVERRIDE;
 
+  // Overrideable for testing.

[Mihai Parparita -not on Chrome, 2012/07/11 00:45:45]

FWIW, the delegate pattern that Antony and I mentioned would be to move these
two methods into a Delegate inner class, and to default to an implementation
that returned real connections, but to allow it to be swapped out with one that
returns mocks.

The advantage of that over this is that it's more obvious what methods need to
be implemented for testing, instead of a seemingly random subset of all the
virtual methods.

+  virtual SerialConnection* CreateSerialConnection(
+      const std::string& port,
+      int bitrate,
+      APIResourceEventNotifier* event_notifier);
+  virtual bool DoesPortExist(const std::string& port);
+
  private:
   scoped_ptr<api::experimental_serial::Open::Params> params_;
   int src_id_;