Don't allow updating tabs to javascript URLs without host

chrome/test/data/extensions/api_test/README.txt

 This directory contains extensions which are unit tests for the extension API.
 These tests are written using the extension API test framework, which allows
 us to do end-to-end testing of extension API in a browser_test.  The general way
 these tests work is to run code in an extension when they're loaded and to post
 a pass or fail notification back up to the C++ unit test which then reports the
 success or failure.  In the common case, the extension runs many subtests and
 then reports up a single pass or fail.  This case is made easy by the test
 framework.
 
 To write a new test:
@@ skipping 25 matching lines @@
     chrome.bookmarks.getTree(chrome.test.callbackPass(function(results) {
       chrome.test.assertTrue(compareTrees(results, expected),
                              "getTree() result != expected");
     }));
   },
 
   function get() {
     chrome.bookmarks.get("1", chrome.test.callbackPass(function(results) {
       chrome.test.assertTrue(compareNode(results[0], expected[0].children[0]));
     }));
-    chrome.bookmarks.get("42", chrome.test.callbackFail(function(){}, "Can't find bookmark for id."));
+    chrome.bookmarks.get("42", chrome.test.callbackFail("Can't find bookmark for id."));
   },
 
   function getArray() {
     chrome.bookmarks.get(["1", "2"], chrome.test.callbackPass(function(results) {
       chrome.test.assertTrue(compareNode(results[0], expected[0].children[0]),
                              "get() result != expected");
       chrome.test.assertTrue(compareNode(results[1], expected[0].children[1]),
                              "get() result != expected");
     }));
   }
 ]);
 
 // compareNode and compareTrees are helper functions that the bookmarks test
 // uses for convenience.  They're not really relevant to the framework itself.
 
-Note that chrome.test.callbackFail takes a second argument which is the error
-message that it expects to get when the callback fails (chrome.extension.lastError.message).
+Note that chrome.test.callbackFail takes an argument which is the error message
+that it expects to get when the callback fails
+(chrome.extension.lastError.message).
 
 Here's what the output of this test might look like:
 [==========] Running 1 test from 1 test case.
 [----------] Global test environment set-up.
 [----------] 1 test from ExtensionApiTest
 [ RUN      ] ExtensionApiTest.Bookmarks
 ( RUN      ) getTree
 (  SUCCESS )
 ( RUN      ) get
 (  SUCCESS )
 ( RUN      ) getArray
 (  SUCCESS )
 Got EXTENSION_TEST_PASSED notification.
 [       OK ] ExtensionApiTest.DISABLED_Bookmarks (2472 ms)
 [----------] 1 test from ExtensionApiTest (2475 ms total)
 
 [----------] Global test environment tear-down
 [==========] 1 test from 1 test case ran. (2482 ms total)
 [  PASSED  ] 1 test.
 1 test run
 0 test failed
 
 Note the RUN/SUCCESS messages in () - these are the subtests that are run in
 the extension itself.  Anything printed with chrome.test.log() will also display
 in stdout of the browser test (and hence in the buildbot output for that test).