Measuring data use of different ModelTypes in Sync Service.

net/url_request/data_use_measurement.cc

Index: net/url_request/data_use_measurement.cc
diff --git a/net/url_request/data_use_measurement.cc b/net/url_request/data_use_measurement.cc
index d27cc1b397e084f9ea9d8f086d01ffba4a6c2426..2b965d266deb9c1650ad1936ababf80cca80b011 100644
--- a/net/url_request/data_use_measurement.cc
+++ b/net/url_request/data_use_measurement.cc
@@ -135,6 +135,14 @@ void ReportDataUsage(const std::string& service_name,
 }
 
 }  // namespace
+void SparseDataUseReport(const std::string& histogram_name,

[bengr, 2015/08/07 17:13:37]

Add a blank line above.

[amohammadkhan, 2015/08/07 18:52:30]

Done.

+                         int value,
+                         int count) {
+  if (count > 0)
+    UMA_HISTOGRAM_SPARSE_SLOWLY_WITH_VALUE(histogram_name, value, count);
+  else
+    LOG(ERROR) << "Ali1 zero or negative count " << count;

[bengr, 2015/08/07 17:13:37]

What is Ali1?

[amohammadkhan, 2015/08/07 18:52:30]

Sorry, basically this LOG(ERROR) was for testing to see if we have any negative
or zero values and it should be removed.

But beside that I have a question. Because at first I changed
Data_use_measurement to add a function for measuring Sync data use. But I had to
include it (net/url_request/data_use_measurement.h) in Sync files that I wasn't
sure it is Okay or not. So I added my Macro to those files and didn't use
function of data_use_measurement. So I wanted to ask you which approach is
better and remove the code for the other one. Because now I have the code for
both of approaches (The added function in data_use_measurement and the Macro in
Sync files)

+}
 
 void DataUseReport(const std::string& service_name,
                    const net::URLFetcher* fetcher) {