Add skeleton webrtc::SessionDescription and webrtc::MediaDescription classes.

webrtc/api/ortc/mediadescription.h

Index: webrtc/api/ortc/mediadescription.h
diff --git a/webrtc/api/ortc/mediadescription.h b/webrtc/api/ortc/mediadescription.h
new file mode 100644
index 0000000000000000000000000000000000000000..ac246cf8098bf461a6144542c3cca3cd06886284
--- /dev/null
+++ b/webrtc/api/ortc/mediadescription.h
@@ -0,0 +1,36 @@
+/*
+ *  Copyright 2017 The WebRTC project authors. All Rights Reserved.
+ *
+ *  Use of this source code is governed by a BSD-style license
+ *  that can be found in the LICENSE file in the root of the source
+ *  tree. An additional intellectual property rights grant can be found
+ *  in the file PATENTS.  All contributing project authors may
+ *  be found in the AUTHORS file in the root of the source tree.
+ */
+
+#ifndef WEBRTC_API_ORTC_MEDIADESCRIPTION_H_
+#define WEBRTC_API_ORTC_MEDIADESCRIPTION_H_
+
+#include <string>
+
+#include "webrtc/base/optional.h"
+
+namespace webrtc {
+
+// A structured representation of a media description within an SDP message.

[Taylor Brandstetter, 2017/03/11 00:24:29]

nit: I'd say "session description" instead of "message", here and in the other
header file. "message" could be interpreted as meaning SIP messages.

[Zhi Huang, 2017/03/11 01:20:38]

Done.

+class MediaDescription {
+ public:
+  explicit MediaDescription(std::string mid) : mid_(mid) {}

[Taylor Brandstetter, 2017/03/11 00:24:29]

Should be mid_(std::move(mid)) to ensure an extra copy isn't made.

[Zhi Huang, 2017/03/11 01:20:39]

Oh, I missed this one.

+
+  ~MediaDescription() {}
+
+  rtc::Optional<std::string> mid() const { return mid_; }

[Taylor Brandstetter, 2017/03/11 00:24:29]

Could add a comment that this is the "mid" attribute defined in RFC5888.

[Zhi Huang, 2017/03/11 01:20:39]

Done.

+  void set_mid(const rtc::Optional<std::string>& mid) { mid_ = mid; }

[Taylor Brandstetter, 2017/03/11 00:24:29]

Same comment as in previous CL: I think this method can just take a string for
convenience, since there shouldn't be a need to "unset" the mid.

Also, maybe we don't need to use Optional at all? An empty string could be
interpreted as being "unset".

[Zhi Huang, 2017/03/11 01:20:38]

Both ways sound good to me. Maybe it would make it more clear that this field is
optional by using rtc::Optional()?

[Taylor Brandstetter, 2017/03/11 01:37:06]

That's a good argument. I could go either way. Might as well go with "Optional"
since that's what we already have.

+
+ private:
+  rtc::Optional<std::string> mid_;
+};
+
+}  // namespace webrtc
+
+#endif  // WEBRTC_API_ORTC_MEDIADESCRIPTION_H_