Makes a handful of DataSources not replace the existing

chrome/browser/sessions/session_restore_uitest.cc

Index: chrome/browser/sessions/session_restore_uitest.cc
diff --git a/chrome/browser/sessions/session_restore_uitest.cc b/chrome/browser/sessions/session_restore_uitest.cc
index d79efa5337a50b5d82452b611f04e8189c82d0d0..496f67195fbc177524453f0b7121a5864b8ddef7 100644
--- a/chrome/browser/sessions/session_restore_uitest.cc
+++ b/chrome/browser/sessions/session_restore_uitest.cc
@@ -450,9 +450,7 @@ TEST_F(SessionRestoreUITest, TwoWindowsCloseOneRestoreOnlyOne) {
 // process-per-site and process-per-site-instance, because we treat the new tab
 // as a special case in process-per-site-instance so that it only ever uses one
 // process.)
-//
-// Flaky: http://code.google.com/p/chromium/issues/detail?id=52022
-TEST_F(SessionRestoreUITest, FLAKY_ShareProcessesOnRestore) {
+TEST_F(SessionRestoreUITest, ShareProcessesOnRestore) {

[Evan Martin, 2011/02/23 00:17:32]

Did you mean to include this?

[sky, 2011/02/23 00:38:57]

Yes. All the changes I've done to ChromeURLDataManager (and URLRequest,
ProcessIterator, process launcher and automation classes) are so this test isn't
flakey! Yowza!

   if (ProxyLauncher::in_process_renderer()) {
     // No point in running this test in single process mode.
     return;