Collect separate V8 histograms for Top10 non-Google sites

content/renderer/render_thread_impl.cc

Index: content/renderer/render_thread_impl.cc
diff --git a/content/renderer/render_thread_impl.cc b/content/renderer/render_thread_impl.cc
index 5fe2984fd116ddc99da82aac920e70bef0775dce..551f59134b639eafd368fd2fe36dbb469d459a68 100644
--- a/content/renderer/render_thread_impl.cc
+++ b/content/renderer/render_thread_impl.cc
@@ -22,6 +22,7 @@
 #include "base/single_thread_task_runner.h"
 #include "base/strings/string16.h"
 #include "base/strings/string_number_conversions.h"
+#include "base/strings/string_split.h"
 #include "base/strings/string_tokenizer.h"
 #include "base/strings/sys_string_conversions.h"
 #include "base/strings/utf_string_conversions.h"
@@ -288,6 +289,28 @@ std::string HostToCustomHistogramSuffix(const std::string& host) {
     return ".inbox";
   if (host == "www.youtube.com")
     return ".youtube";
+
+  // The Top10 sites have different TLD and/or subdomains
+  // depending on the localization
+   if (host == "sina.com.cn")
+    return ".top10";
+
+  std::vector<base::StringPiece> host_tokens = base::SplitStringPiece(
+    host, ".", base::TRIM_WHITESPACE, base::SPLIT_WANT_NONEMPTY);
+
+  if (host_tokens.size() == 3) {
+    if ((host_tokens[1] == "facebook") ||
+        (host_tokens[1] == "baidu") ||
+        (host_tokens[1] == "yahoo") ||
+        (host_tokens[1] == "amazon") ||
+        (host_tokens[1] == "qq") ||
+        (host_tokens[1] == "twitter") ||
+        (host_tokens[1] == "taobao") ||
+        (host_tokens[1] == "live") ||
+        (host_tokens[1] == "linkedin"))

[jochen (gone - plz use gerrit), 2015/10/21 13:17:57]

according to http://www.alexa.com/topsites linkedin is #11 (excluding google
sites), but wikipedia is on the list

I'm not sure we should ignore the TLDs, I also don't understand why you require
three components?

what about .co.uk? if you really want to filter out the tld, use
net::registry_controlled_domains::GetDomainAndRegistry() to extract this
information.

I would vote for just hardcoding the names.

also, please add { } for a multi-line if

[Michael Hablich, 2015/10/22 13:16:35]

You are right: Removed linkedin, added wikipedia

Using GetDomainAndRegistry now, thanks

I still against hardcoding the names because just have a look at wikipedia:
de.wikipedia.org, jp.wikipedia.org and the reverse is also happening like in
Amazon's case: amazon.com, amazon.de. I think we do not only want US-centric
data.

+      return ".top10";
+  }
+
   return std::string();
 }