Implement nextHopProtocol in PerformanceResourceTiming and PerformanceNavigationTiming.

third_party/WebKit/public/platform/WebURLResponse.h

 /*
  * Copyright (C) 2009 Google Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
  *     * Redistributions in binary form must reproduce the above
@@ skipping 247 matching lines @@
   // This indicates the location of a downloaded response if the
   // WebURLRequest had the downloadToFile flag set to true. This file path
   // remains valid for the lifetime of the WebURLLoader used to create it.
   BLINK_PLATFORM_EXPORT WebString downloadFilePath() const;
   BLINK_PLATFORM_EXPORT void setDownloadFilePath(const WebString&);
 
   // Remote IP address of the socket which fetched this resource.
   BLINK_PLATFORM_EXPORT WebString remoteIPAddress() const;
   BLINK_PLATFORM_EXPORT void setRemoteIPAddress(const WebString&);
 
+  // ALPN negotiated protocol of the socket which fetched this resource.
+  BLINK_PLATFORM_EXPORT WebString alpnNegotiatedProtocol() const;

[jochen (gone - plz use gerrit), 2017/04/18 11:37:58]

it seems that these aren't strings, but registry controlled  values. Should this
be an enum?

[panicker, 2017/04/18 17:01:06]

Ditto for this (see other comment).

[panicker, 2017/04/18 21:43:50]

While we could use net::HttpResponseInfo::ConnectionInfo for connectionInfo
there is no enum for ALPN in net/ i.e. at the source.
Adding something here is outside the scope of this change; although I could work
with net/ folks on a follow up CL if you think it necessary. 

+  BLINK_PLATFORM_EXPORT void setALPNNegotiatedProtocol(const WebString&);
+
+  // Information about the type of connection used to fetch this resource.
+  BLINK_PLATFORM_EXPORT WebString connectionInfo() const;

[jochen (gone - plz use gerrit), 2017/04/18 11:37:58]

same here, is that a free-form value, or really some enum?

[panicker, 2017/04/18 17:01:06]

There's a lot of internal and deprecated values here that the net/ team is
constantly evolving, so I'm not sure it makes sense to bake it into public
interface:
https://cs.chromium.org/chromium/src/net/http/http_response_info.h?type=cs&q=...

[jochen (gone - plz use gerrit), 2017/04/18 18:21:11]

The interface is not really public - it used to be an interface to WebKit but
now it's just historical baggage

[panicker, 2017/04/18 21:43:50]

I'm open to suggestions:
while we could propagate enum around, we cannot access net/http from Blink
(Webkit/Source/core) right?
Which means maintaining another version of
net::HttpResponseInfo::ConnectionInfoToString ?
https://cs.chromium.org/chromium/src/net/http/http_response_info.cc?q=net::Ht...

[jochen (gone - plz use gerrit), 2017/04/19 15:01:25]

platform/loader already depends on net/ and public/web also has a long list of
dependencies, so I'd propose to just add a DEPS entry to net/http and use that

[kinuko, 2017/04/20 03:37:08]

Yes, it's fine to make platform/loader code depend on net/.

+  BLINK_PLATFORM_EXPORT void setConnectionInfo(const WebString&);
+
   // Remote port number of the socket which fetched this resource.
   BLINK_PLATFORM_EXPORT unsigned short remotePort() const;
   BLINK_PLATFORM_EXPORT void setRemotePort(unsigned short);
 
   // Original size of the response before decompression.
   BLINK_PLATFORM_EXPORT void setEncodedDataLength(long long);
 
   // Original size of the response body before decompression.
   BLINK_PLATFORM_EXPORT void addToEncodedBodyLength(long long);
 
@@ skipping 26 matching lines @@
   // instance it contains.
   std::unique_ptr<ResourceResponseContainer> m_ownedResourceResponse;
 
   // Should never be null.
   ResourceResponse* m_resourceResponse;
 };
 
 }  // namespace blink
 
 #endif