Add immediate variants of doc.write page load metrics

chrome/browser/page_load_metrics/page_load_metrics_browsertest.cc

Index: chrome/browser/page_load_metrics/page_load_metrics_browsertest.cc
diff --git a/chrome/browser/page_load_metrics/page_load_metrics_browsertest.cc b/chrome/browser/page_load_metrics/page_load_metrics_browsertest.cc
index 711d340d1e3dcb71ca2df96639dce116b808a824..34c2faea3bae3e1fdb2a228efb537c8d4fb36d5a 100644
--- a/chrome/browser/page_load_metrics/page_load_metrics_browsertest.cc
+++ b/chrome/browser/page_load_metrics/page_load_metrics_browsertest.cc
@@ -136,15 +136,14 @@ IN_PROC_BROWSER_TEST_F(MetricsWebContentsObserverBrowserTest,
   histogram_tester_.ExpectTotalCount(
       internal::kHistogramDocWriteBlockParseStartToFirstContentfulPaint, 1);
 
-  histogram_tester_.ExpectTotalCount(
-      internal::kHistogramDocWriteBlockReloadCount, 1);
-
   ui_test_utils::NavigateToURL(browser(),
                                embedded_test_server()->GetURL("/title2.html"));
 
   histogram_tester_.ExpectTotalCount(
       internal::kHistogramDocWriteBlockParseStartToFirstContentfulPaint, 1);
 
+  // Testing this histogram's value only in the end else the test will be

[Bryan McQuade, 2016/06/17 13:24:03]

yeah, i had encountered similar issues with timing races in this test. i think
you can remove this comment. it's useful for the code reviewer but i don't know
it'll be useful for readers of the code once this patch is committed.

+  // dependent on IPC timing.
   histogram_tester_.ExpectTotalCount(
       internal::kHistogramDocWriteBlockReloadCount, 2);
 }