GTTF: Use a fresh TestingBrowserProcess for each test, part #1

This is going to be done in multiple steps,
because lots of tests depend on global TestingBrowserProcess.

However, the current solution is very error-prone,
and may lead to state being carried from one test
to another (via NotificationService), resulting in crashes.

In theory the ScopedTestingBrowserProcess could be made
empty for now, but re-initializing g_browser_process inside it
has two benefits:

1) It verifies that changing g_browser_process actually works.

2) It protects at least a subset of tests from carrying state.

BUG=61062
TEST=unit_tests, possibly more

Review URL: http://codereview.chromium.org/6478005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@74926 0039d316-1c4b-4281-b951-d872f2087c98

1 files changed, 5 insertions, 1 deletions

diff --git a/chrome/browser/automation/automation_provider_unittest.cc b/chrome/browser/automation/automation_provider_unittest.cc
index c931701..3631c44 100644
--- a/chrome/browser/automation/automation_provider_unittest.cc
+++ b/chrome/browser/automation/automation_provider_unittest.cc
@@ -4,6 +4,8 @@
 
 #include "chrome/browser/browser_thread.h"
 #include "chrome/browser/automation/chrome_frame_automation_provider.h"
+#include "chrome/test/testing_browser_process.h"
+#include "chrome/test/testing_browser_process_test.h"
 #include "ipc/ipc_message.h"
 #include "testing/gmock/include/gmock/gmock.h"
 #include "testing/gtest/include/gtest/gtest.h"
@@ -20,7 +22,9 @@ class MockChromeFrameAutomationProvider
                void (const IPC::Message& message));  // NOLINT
 };
 
-TEST(AutomationProviderTest, TestInvalidChromeFrameMessage) {
+typedef TestingBrowserProcessTest AutomationProviderTest;
+
+TEST_F(AutomationProviderTest, TestInvalidChromeFrameMessage) {
   MessageLoop message_loop;
   BrowserThread ui_thread(BrowserThread::UI, &message_loop);