Determine prevalence of cache-control: no-store in main frame resources.

tools/metrics/histograms/histograms.xml

Index: tools/metrics/histograms/histograms.xml
diff --git a/tools/metrics/histograms/histograms.xml b/tools/metrics/histograms/histograms.xml
index 2e6cfb6d90c76cc3f15a908200ad9ee9e54e7c7a..3789bab816cce84ad45a9c58a7bbb99ed7dda095 100644
--- a/tools/metrics/histograms/histograms.xml
+++ b/tools/metrics/histograms/histograms.xml
@@ -13516,6 +13516,15 @@ Therefore, the affected-histogram name has to have at least one dot in it.
   </summary>
 </histogram>
 
+<histogram name="Net.MainFrameNoStore">
+  <owner>jkarlin@chromium.org</owner>
+  <summary>
+    After fetching a main frame no-store (1) is incremented if the response
+    headers contained &quot;cache-control: no-store&quot; else (0) is
+    incremented.

[Ilya Sherman, 2014/06/11 20:26:20]

Please use an enum to label the buckets, and make the <summary> more about what
the histogram is measuring as a whole.  A good way to write a histogram
<summary> is to answer two questions: "What question does this metric try to
answer?" and "When is this metric recorded, and what sorts of bias might that
lead to?".

[davidben, 2014/06/11 20:37:29]

(Oh yeah, sorry, when I suggested to use a boolean, I meant only in the metrics
gathering code, not histograms.xml. histograms.xml can still have labelled
buckets with a UMA_HISTOGRAM_BOOLEAN.)

[jkarlin, 2014/06/11 20:55:13]

Done.

[jkarlin, 2014/06/11 20:55:13]

Done.

+  </summary>
+</histogram>
+
 <histogram name="Net.MTPR_GetProxyForUrl_Thread_Wait_Time" units="milliseconds">
   <owner>Please list the metric's owners. Add more owner tags as needed.</owner>
   <summary>