Added support for passing WindowOpenDisposition into BrowserInstantController::OpenURL() from the o…

chrome/browser/instant/instant_extended_browsertest.cc

Index: chrome/browser/instant/instant_extended_browsertest.cc
diff --git a/chrome/browser/instant/instant_extended_browsertest.cc b/chrome/browser/instant/instant_extended_browsertest.cc
index 08f8b4d094579c90ea12e3e047faf8820e4c4b55..52a9b3e1fb8ca6cb52a900d52432051cb5de1b9c 100644
--- a/chrome/browser/instant/instant_extended_browsertest.cc
+++ b/chrome/browser/instant/instant_extended_browsertest.cc
@@ -5,6 +5,7 @@
 #include "chrome/browser/instant/instant_loader.h"
 #include "chrome/browser/instant/instant_test_utils.h"
 #include "chrome/browser/ui/search/search.h"
+#include "chrome/browser/ui/tabs/tab_strip_model.h"
 #include "chrome/test/base/interactive_test_utils.h"
 #include "chrome/test/base/ui_test_utils.h"
 
@@ -80,3 +81,32 @@ IN_PROC_BROWSER_TEST_F(InstantExtendedTest, InputShowsOverlay) {
   EXPECT_TRUE(instant()->model()->mode().is_search_suggestions());
   EXPECT_EQ(preview_tab, instant()->GetPreviewContents());
 }
+
+// Test that middle clicking on a suggestion opens the result in a new tab.
+IN_PROC_BROWSER_TEST_F(
+    InstantExtendedTest, MiddleClickOnSuggestionOpensInNewTab) {
+  ASSERT_NO_FATAL_FAILURE(SetupInstant());
+  FocusOmniboxAndWaitForInstantSupport();
+
+  ASSERT_TRUE(browser()->tab_strip_model()->count() == 1);

[sreeram, 2013/02/06 22:00:50]

I'd just make this an EXPECT. Generally we use ASSERT when continuing with the
rest of the test will produce cascading crashes, say due to a null pointer or
such. If it's just a matter of initial assumptions/expectations being wrong, no
harm proceeding with the test. So:
EXPECT_EQ(1, ...);

[dougw, 2013/02/20 21:46:56]

Done.

+
+  // Typing in the omnibox should show the overlay.
+  SetOmniboxTextAndWaitForInstantToShow("test");
+  EXPECT_TRUE(instant()->model()->mode().is_search_suggestions());
+
+  EXPECT_TRUE(ExecuteScript("function Click() {"
+        "var eventObject = document.createEvent('MouseEvents');"
+        "eventObject.initMouseEvent('click');"
+        "document.getElementsByTagName('body')[0].dispatchEvent(eventObject);"
+      "}"
+      "document.body.addEventListener(\"click\", function() {"
+        "var nativeSuggestions = chrome.searchBox.nativeSuggestions;"
+        "if (nativeSuggestions.length > 0) {"
+          "chrome.searchBox.navigateContentWindow(nativeSuggestions[0].rid, 2);"

[sreeram, 2013/02/06 22:00:50]

I recommend that you navigate to a known URL, such as "chrome://version". So you
can then check that the newly added tab's URL is indeed that (and that we didn't
just open a tab and navigate to a random URL).

[dougw, 2013/02/20 21:46:56]

Done.

+        "}"
+      "});"
+      "Click();"));

[sreeram, 2013/02/06 22:00:50]

Why not actually click on the page using ui_test_utils::ClickOnView(browser(),
VIEW_ID_TAB_CONTAINER)? The reason why I insist on actually clicking on the page
is that it has a bunch of ripple effects - omnibox losing focus,
InstantController trying to commit the preview, etc. I'd like to ensure that the
navigation reliably succeeds amidst all that.

[dougw, 2013/02/20 21:46:56]

Done.

+
+  // Check that there are now two tabs.
+  EXPECT_EQ(2, browser()->tab_strip_model()->count());
+}