Fix recent null pointer error regression in the flakiness dashboard
Created attachment 146377 [details] Patch
Comment on attachment 146377 [details] Patch looks fine to me, but my JS-fu isn't terribly strong. Is it possible for something in the test to throw an exception that would cause doXHR not to be reassigned back to oldDoXHR? i.e. does that need to wrapped in a try/finally like you would in python?
(In reply to comment #2) > (From update of attachment 146377 [details]) > looks fine to me, but my JS-fu isn't terribly strong. Is it possible for something in the test to throw an exception that would cause doXHR not to be reassigned back to oldDoXHR? i.e. does that need to wrapped in a try/finally like you would in python? yes. good idea.
Committed r119754: <http://trac.webkit.org/changeset/119754>