no-caret-repaint-in-non-content-editable-element.html is the only test that explicitly disables caret browsing. ... if (window.layoutTestController) layoutTestController.overridePreference("WebKitEnableCaretBrowsing", false); ... However, this is a useless statement, since all DRT that support testing caret browsing (gtk, qt anc chromium) have it reset by default to FALSE before each test execution. On Mac and Windows, DRT that do not support testing caret browsing, it is obviously also FALSE by default. I propose removing it. Patch coming ...
Created attachment 76961 [details] patch v1 (r74349, r=xan)
Comment on attachment 76961 [details] patch v1 (r74349, r=xan) Makes sense to me.
Comment on attachment 76961 [details] patch v1 (r74349, r=xan) Clearing flags on attachment: 76961 Committed r74349: <http://trac.webkit.org/changeset/74349>
After talking to Daniel Bates, the original author of this test, we both agreed that the statement has no affect as DRTs are today (since caret browsing it reset before each test execution). However, explicitly setting caret browsing to DISABLED ensures that the test will run under the circumstance it needs to catch the bug: caret browsing is DISABLED. We decided to put this code back in, and he will add a more descriptive comment saying why it should be there. See bug 51372