Work around exceptions when the 'high' value for a histogram is not present. This happens when the …

tools/telemetry/telemetry/value/histogram_util.py

Index: tools/telemetry/telemetry/value/histogram_util.py
diff --git a/tools/telemetry/telemetry/value/histogram_util.py b/tools/telemetry/telemetry/value/histogram_util.py
index bb6ff363d22943b0c8fde15340ceca46a431ae4c..99542bec43a1d6e3501d9be397ce20b790f2cc0e 100644
--- a/tools/telemetry/telemetry/value/histogram_util.py
+++ b/tools/telemetry/telemetry/value/histogram_util.py
@@ -16,6 +16,21 @@ BROWSER_HISTOGRAM = 'browser_histogram'
 RENDERER_HISTOGRAM = 'renderer_histogram'
 
 
+def GetHistogramBucketsFromJson(histogram_json):
+  return GetHistogramBucketsFromRawValue(json.loads(histogram_json))
+
+
+def GetHistogramBucketsFromRawValue(raw_value):
+  buckets = raw_value.get('buckets', [])
+  if buckets:
+    # If there are values greater than the maximum allowable for the histogram,
+    # the highest bucket will have a 'low': maxvalue entry in the dict but no
+    # 'high' entry. Code often assumes the 'high' value will always be present,
+    # and uses it to get bucket mean. So default it to the same value as low.
+    buckets[-1].setdefault('high', buckets[-1]['low'])
+  return buckets
+
+
 def CustomizeBrowserOptions(options):
   """Allows histogram collection."""
   options.AppendExtraBrowserArgs(['--enable-stats-collection-bindings'])
@@ -27,9 +42,10 @@ def SubtractHistogram(histogram_json, start_histogram_json):
   Both parameters and the returned result are json serializations.
   """
   start_histogram = json.loads(start_histogram_json)
+  start_histogram_buckets = GetHistogramBucketsFromRawValue(start_histogram)
   # It's ok if the start histogram is empty (we had no data, maybe even no
   # histogram at all, at the start of the test).
-  if 'buckets' not in start_histogram:
+  if not start_histogram_buckets:
     return histogram_json
 
   histogram = json.loads(histogram_json)
@@ -39,16 +55,16 @@ def SubtractHistogram(histogram_json, start_histogram_json):
         'Trying to compare histograms from different processes (%d and %d)'
         % (start_histogram['pid'], histogram['pid']))
 
-  start_histogram_buckets = dict()
-  for b in start_histogram['buckets']:
-    start_histogram_buckets[b['low']] = b['count']
+  start_histogram_bucket_counts = dict()
+  for b in start_histogram_buckets:
+    start_histogram_bucket_counts[b['low']] = b['count']
 
   new_buckets = []
-  for b in histogram['buckets']:
+  for b in GetHistogramBucketsFromRawValue(histogram):
     new_bucket = b
     low = b['low']
-    if low in start_histogram_buckets:
-      new_bucket['count'] = b['count'] - start_histogram_buckets[low]
+    if low in start_histogram_bucket_counts:
+      new_bucket['count'] = b['count'] - start_histogram_bucket_counts[low]
       if new_bucket['count'] < 0:
         logging.error('Histogram subtraction error, starting histogram most '
                       'probably invalid.')
@@ -72,8 +88,7 @@ def AddHistograms(histogram_jsons):
 
   buckets = collections.defaultdict(int)
   for histogram_json in histogram_jsons:
-    h = json.loads(histogram_json)
-    for b in h['buckets']:
+    for b in GetHistogramBucketsFromJson(histogram_json):
       key = (b['low'], b['high'])

[sullivan, 2015/02/20 16:43:19]

This line would be buggy without a fix in histogram_utils, too.

       buckets[key] += b['count']