Implement WiFi Display elementary stream packetizer.

extensions/renderer/api/display_source/wifi_display/wifi_display_stream_packet_part.h

Index: extensions/renderer/api/display_source/wifi_display/wifi_display_stream_packet_part.h
diff --git a/extensions/renderer/api/display_source/wifi_display/wifi_display_stream_packet_part.h b/extensions/renderer/api/display_source/wifi_display/wifi_display_stream_packet_part.h
new file mode 100644
index 0000000000000000000000000000000000000000..097c0b8c32be99a22c8927ace2761bb75b5c5ae4
--- /dev/null
+++ b/extensions/renderer/api/display_source/wifi_display/wifi_display_stream_packet_part.h
@@ -0,0 +1,46 @@
+// Copyright 2015 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.
+
+#ifndef EXTENSIONS_RENDERER_API_DISPLAY_SOURCE_WIFI_DISPLAY_WIFI_DISPLAY_STREAM_PACKET_PART_H_
+#define EXTENSIONS_RENDERER_API_DISPLAY_SOURCE_WIFI_DISPLAY_WIFI_DISPLAY_STREAM_PACKET_PART_H_
+
+#include <stddef.h>
+#include <stdint.h>
+
+#include "base/macros.h"
+
+namespace extensions {
+
+// WiFi Display elementary stream unit packetization consists of multiple
+// packetization phases. During those phases, unit buffer bytes are not
+// modified but only wrapped in packets.
+// This class allows different kind of WiFi Display stream packets to refer to
+// unit buffer bytes without copying them.
+class WiFiDisplayStreamPacketPart {
+ public:
+  WiFiDisplayStreamPacketPart(const uint8_t* data, size_t size)
+      : data_(data), size_(size) {}

[asargent_no_longer_on_chrome, 2016/03/16 22:47:58]

nit: inlining constructors is discouraged

http://dev.chromium.org/developers/coding-style/cpp-dos-and-donts#TOC-Stop-in...

(here and below)

[e_hakkinen, 2016/03/17 15:29:26]

But since there is no inheritance, no explicit destructor and only a few POD
members, isn't it OK to inline?

[asargent_no_longer_on_chrome, 2016/03/17 18:22:59]

Ok, since it has only POD members it should be alright to leave as you have it,
assuming that this will actually be performance sensitive code. (The only danger
I see is that in some future refactor, we might add some non-POD data members
and forget to de-inline the constructor). 

+  template <size_t N>
+  explicit WiFiDisplayStreamPacketPart(const uint8_t (&data)[N])
+      : WiFiDisplayStreamPacketPart(data, N) {}
+  // Allow copy elision. Should never be really called.
+  WiFiDisplayStreamPacketPart(WiFiDisplayStreamPacketPart&& other)
+      : WiFiDisplayStreamPacketPart(other.data(), other.size()) {}

[asargent_no_longer_on_chrome, 2016/03/16 22:47:58]

I may be misunderstanding what you're trying to accomplish here, but do you want
to use Function Suppression, ie "= delete", if this should never be called? 

https://groups.google.com/a/chromium.org/forum/#!topic/chromium-dev/i1o7-RNRnMs

[e_hakkinen, 2016/03/17 15:29:26]

It should never be called but it should still be available so that compiler is
allowed to elide the calls and do return value optimization.
But maybe it is more clear to move it to WiFiDisplayElementaryStreamPacket there
move constructibility is needed.

[asargent_no_longer_on_chrome, 2016/03/17 18:22:59]

ok

+
+  const uint8_t* begin() const { return data(); }
+  const uint8_t* data() const { return data_; }
+  bool empty() const { return size() == 0u; }
+  const uint8_t* end() const { return data() + size(); }

[asargent_no_longer_on_chrome, 2016/03/16 22:47:58]

consider not inlining this function

http://dev.chromium.org/developers/coding-style/cpp-dos-and-donts#TOC-Stop-in...

+  size_t size() const { return size_; }
+
+ private:
+  const uint8_t* const data_;
+  const size_t size_;
+
+  DISALLOW_COPY_AND_ASSIGN(WiFiDisplayStreamPacketPart);
+};
+
+}  // namespace extensions
+
+#endif  // EXTENSIONS_RENDERER_API_DISPLAY_SOURCE_WIFI_DISPLAY_WIFI_DISPLAY_STREAM_PACKET_BASE_H_