Perf tests that measure time to load libraries

chrome/chrome_tests.gypi

 # Copyright 2013 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.
 {
   'variables' : {
     'chromedriver_unittest_sources': [
       'test/chromedriver/capabilities_unittest.cc',
       'test/chromedriver/chrome/chrome_finder_unittest.cc',
       'test/chromedriver/chrome/console_logger_unittest.cc',
       'test/chromedriver/chrome/device_manager_unittest.cc',
@@ skipping 2815 matching lines @@
       # Required for WebRTC PyAuto tests.
       'target_name': 'webrtc_test_tools',
       'type': 'none',
       'dependencies': [
         'pyautolib',
         '../third_party/libjingle/libjingle.gyp:peerconnection_server',
         '../third_party/webrtc/tools/tools.gyp:frame_analyzer',
         '../third_party/webrtc/tools/tools.gyp:rgba_to_i420_converter',
       ],
     },  # target 'webrtc_test_tools'
+    {
+      # Executable to measure time to load libraries.
+      'target_name': 'load_library_perf_tests',

[sky, 2014/03/01 00:09:21]

Is there a reason this needs to be in its own target?

[shadi, 2014/03/01 00:35:22]

Similar to other perf tests targets. Main reason is to be able to run this perf
test on widevine bots for example. There was a discussion with tonyg@ in the
related CL (https://codereview.chromium.org/151283003/) about how we should
combine perf test targets all together... 

+      'type': '<(gtest_target_type)',
+      'dependencies': [
+        '../base/base.gyp:test_support_perf',
+        '../testing/gtest.gyp:gtest',
+        '../testing/perf/perf_test.gyp:*',
+        '../third_party/widevine/cdm/widevine_cdm.gyp:widevine_cdm_version_h',
+      ],
+      'sources': [
+        'browser/load_library_perf_test.cc',
+      ],
+      'include_dirs': [
+        '..',
+        '<(SHARED_INTERMEDIATE_DIR)',
+      ],
+    },  # target 'load_library_perf_tests'
   ],
   'conditions': [
     ['OS=="mac"', {
       'targets': [
         {
           # This is the mac equivalent of the security_tests target below. It
           # generates a framework bundle which bundles tests to be run in a
           # renderer process. The test code is built as a framework so it can be
           # run in the context of a renderer without shipping the code to end
           # users.
@@ skipping 386 matching lines @@
             'utility',
           ],
           'sources': [
             'tools/service_discovery_sniffer/service_discovery_sniffer.h',
             'tools/service_discovery_sniffer/service_discovery_sniffer.cc',
           ],
         }]
     }],
   ],  # 'conditions'
 }