[Telemetry] Convert long running user stories to the System Health format.

tools/perf/benchmarks/battor.py

 # Copyright 2016 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 from core import perf_benchmark
 from telemetry.timeline import chrome_trace_category_filter
 from telemetry.web_perf import timeline_based_measurement
 import page_sets
 from telemetry import benchmark
 
@@ skipping 112 matching lines @@
 @benchmark.Enabled('mac')
 class BattOrTrivialPages(_BattOrBenchmark):
 
   def CreateStorySet(self, options):
     # We want it to wait for 30 seconds to be comparable to legacy power tests.
     return page_sets.MacGpuTrivialPagesStorySet(wait_in_seconds=30)
 
   @classmethod
   def Name(cls):
     return 'battor.trivial_pages'
+

[petrcermak, 2016/08/25 09:57:19]

nit: there should be 2 blank lines between top-level definitions.

[rnephew (Reviews Here), 2016/08/25 15:28:57]

Charlies CL at https://codereview.chromium.org/2239053003/ will cause this new
pageset to run with battor tests, so I will be deleting these benchmarks.

+class DesktopBattOrLongRunning(_BattOrBenchmark):

[rnephew (Reviews Here), 2016/08/24 17:50:45]

If my understanding of how the system_health.memory_* tests work is correct, the
new pages should automatically be picked up by it correct?

If that is the case I can get rid of these battor tests (or keep them, depending
on popular consensus). I just wanted to make sure at least once benchmark is
running the new pages.

[petrcermak, 2016/08/25 09:57:18]

All benchmarks that don't specify |case| or have case='long_running' will
automatically pick these up. As far as I can tell, BattOr benchmarks have
case='load'
(https://cs.chromium.org/chromium/src/tools/perf/benchmarks/battor.py?l=63), so
will NOT pick up your stories. Rather than define another benchmark, I suggest
you either:

1) Change battor benchmarks to use ALL SH stories (not just the loading ones) by
removing "case='load'".
2) Modify the SystemHealthStorySet constructor to accept a |cases| list of
strings (instead of just a |case| string) and change the BattOr benchmark ot
pass cases=['load', 'long_running'].

[petrcermak, 2016/08/25 09:57:18]

If for some reason you decide to spin up a separate benchmark, then please pull
all the common code (all three methods, CreateStorySet,
ShouldTearDownStateAfterEachStoryRun, Name) into an abstract superclass.

+  PLATFORM = 'desktop'
+
+  def CreateStorySet(self, options):
+    return page_sets.SystemHealthStorySet(platform=self.PLATFORM,
+                                          case='long_running')
+
+  @classmethod
+  def ShouldTearDownStateAfterEachStoryRun(cls):
+    return True
+
+  @classmethod
+  def Name(cls):
+    return 'battor.long_running_%s' % cls.PLATFORM
+
+
+class MobileBattOrLongRunning(_BattOrBenchmark):
+  PLATFORM = 'mobile'
+
+  def CreateStorySet(self, options):
+    return page_sets.SystemHealthStorySet(platform=self.PLATFORM,
+                                          case='long_running')
+
+  @classmethod
+  def ShouldTearDownStateAfterEachStoryRun(cls):
+    return True
+
+  @classmethod
+  def Name(cls):
+    return 'battor.long_running_%s' % cls.PLATFORM