Create a run_multipage_benchmarks entry point for all benchmarks.

tools/perf/perf_tools/multipage_benchmark_runner.py

Index: tools/perf/perf_tools/multipage_benchmark_runner.py
diff --git a/tools/perf/perf_tools/multipage_benchmark_runner.py b/tools/perf/perf_tools/multipage_benchmark_runner.py
new file mode 100755
index 0000000000000000000000000000000000000000..2fcacdd430f548f913ded336e5edb2e96a637be7
--- /dev/null
+++ b/tools/perf/perf_tools/multipage_benchmark_runner.py
@@ -0,0 +1,76 @@
+#!/usr/bin/env python
+# Copyright (c) 2012 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.
+import csv
+import logging
+import os
+import sys
+
+from chrome_remote_control import browser_finder
+from chrome_remote_control import browser_options
+from chrome_remote_control import multi_page_benchmark
+from chrome_remote_control import page_runner
+from chrome_remote_control import page_set
+
+import perf_tools.first_paint_time_benchmark
+import perf_tools.scrolling_benchmark
+import perf_tools.skpicture_printer
+import perf_tools.texture_upload_benchmark
+
+# TODO(tonyg/nduca): Discover benchmarks automagically.
+_BENCHMARKS = {
+  'first_paint_time_benchmark':
+      perf_tools.first_paint_time_benchmark.FirstPaintTimeBenchmark,
+  'scrolling_benchmark':
+      perf_tools.scrolling_benchmark.ScrollingBenchmark,
+  'skpicture_printer':
+      perf_tools.skpicture_printer.SkPicturePrinter,
+  'texture_upload_benchmark':
+      perf_tools.texture_upload_benchmark.TextureUploadBenchmark
+}
+
+
+def Main():
+  """Turns a MultiPageBenchmark into a command-line program.
+
+  If args is not specified, sys.argv[1:] is used.
+  """
+  options = browser_options.BrowserOptions()
+  parser = options.CreateParser('%prog [options] <benchmark> <page_set>')
+  _, args = parser.parse_args()
+
+  if len(args) != 2 or args[0] not in _BENCHMARKS:
+    parser.print_usage()
+    import page_sets # pylint: disable=F0401
+    sys.stderr.write('Available benchmarks:\n%s\n\n' % ',\n'.join(

[dtu, 2012/10/16 23:53:44]

print >> sys.stderr, 'Available bench...'

[tonyg, 2012/10/17 00:13:44]

Done.

+        _BENCHMARKS.keys()))
+    sys.stderr.write('Available page_sets:\n%s\n\n' % ',\n'.join(
+        [os.path.relpath(f) for f in page_sets.GetAllPageSetFilenames()]))
+    sys.exit(1)
+
+  benchmark = _BENCHMARKS[args[0]]()
+  ps = page_set.PageSet.FromFile(args[1])
+
+  benchmark.AddOptions(parser)
+  _, args = parser.parse_args()
+
+  benchmark.CustomizeBrowserOptions(options)
+  possible_browser = browser_finder.FindBrowser(options)
+  if possible_browser == None:

[dtu, 2012/10/16 23:53:44]

if not possible_browser:

[tonyg, 2012/10/17 00:13:44]

Done.

+    sys.stderr.write(
+      'No browser found.\n' +
+      'Use --browser=list to figure out which are available.\n')
+    sys.exit(1)
+
+  results = multi_page_benchmark.CsvBenchmarkResults(csv.writer(sys.stdout))
+  with page_runner.PageRunner(ps) as runner:
+    runner.Run(options, possible_browser, benchmark, results)
+  # When using an exact executable, assume it is a reference build for the
+  # purpose of outputting the perf results.

[dtu, 2012/10/16 23:53:44]

Not sure about the validity of this assumption. What if you're comparing two
builds?

[tonyg, 2012/10/17 00:13:44]

It just changes the string that is printed in the results summary, so it doesn't
really hurt if this assumption breaks down. Anyway, this method was all just
copied from multi_page_benchmark.py, so it's not really the point of the patch.

+  results.PrintSummary(options.browser_executable and '_ref' or '')
+
+  if len(results.page_failures):
+    logging.warning('Failed pages: %s', '\n'.join(
+        [failure['page'].url for failure in results.page_failures]))
+  return max(255, len(results.page_failures))