Add a PrintTo function for rtc::Optional to aid with testing.

webrtc/base/optional.cc

 /*
  *  Copyright 2016 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.
  */
 
 #include "webrtc/base/optional.h"
 
+#include <iomanip>
+
 namespace rtc {
 namespace optional_internal {
 
 #if RTC_HAS_ASAN
 
 void* FunctionThatDoesNothingImpl(void* x) { return x; }
 
 #endif
 
+void OptionalPrintObjectBytes(const unsigned char* bytes,
+                              size_t size,
+                              std::ostream* os) {
+  *os << "<optional with " << size << "-byte object [";
+  for (size_t i = 0; i != size; ++i) {
+    *os << (i == 0 ? "" : ((i & 1) ? "-" : " "));
+    *os << std::hex << std::setw(2) << std::setfill('0')
+        << static_cast<int>(bytes[i]);
+  }
+  *os << "]>";
+}

[kwiberg-webrtc, 2017/02/17 13:51:22]

This is just for testing, so you should probably inline it (making tests bigger,
but production binaries smaller).

[ossu, 2017/02/17 14:02:19]

I'd rather keep this cruft as isolated as possible. It's only <iomanip> but it'd
be nice to avoid including that in everything that uses Optional, if possible. I
don't think the size increase is noticeable.

[tommi, 2017/02/17 14:08:55]

My concern is that if we get to the point that we can do builds that are free of
iostream, then just one place that pulls it in, will be enough to make a
significant difference.  If the argument is "it's just a drop in the ocean",
then the counter argument is "death by a thousand cuts" :)

Logging is very costly as it is, so we have our work cut out to address that and
there are indeed situations where we've had to throw out all logging in order to
be able to reach performance targets.
If we add more things to battle, then we make the battle harder for ourselves in
the end.
So, whatever we can do to make utilities like these unavailable to production
code, are appreciated. (if we could do what Karl suggests, make testing easier
while making production code smaller, then that would be great)

+
 }  // namespace optional_internal
+
 }  // namespace rtc