WebKit roll 99956:100033.

This will cause a ton of layout test failures on the Leopard
bots, but they'll be fixed with the next roll and this gets rid
of the big intl1 perf regression.

TBR=tonyg@chromium.org
BUG=none
TEST=none


Review URL: http://codereview.chromium.org/8540022

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@109754 0039d316-1c4b-4281-b951-d872f2087c98

DEPS

@@ -5,7 +5,7 @@ vars = {
   "sourceforge_url": "http://%(repo)s.svn.sourceforge.net/svnroot/%(repo)s",
   "webkit_trunk": "http://svn.webkit.org/repository/webkit/trunk",
   "nacl_trunk": "http://src.chromium.org/native_client/trunk",
-  "webkit_revision": "99956",
+  "webkit_revision": "100033",
   "chromium_git": "http://git.chromium.org/git",
   "swig_revision": "69281",
   "nacl_revision": "7141",

webkit/tools/layout_tests/test_expectations.txt

@@ -6,9 +6,3 @@
 // until the changes can be landed upstream.
 
 // PLEASE DO NOT LEAVE ENTRIES IN THIS FILE FOR MORE THAN 24 HRS.
-
-BUG_SENORBLANCO GPU : fast/canvas/canvas-transforms-fillRect-shadow.html = TEXT
-
-// Fixed ES5 bugs during the V8 roll, upstream in progress.
-BUGMSTARZINGER : fast/js/mozilla/strict/15.4.5.1.html = TEXT
-