Because of Window's low resolution timer, the failing test's time targets were not met. By giving the test a little more time, the failure was resolved.
Created attachment 259475 [details] the fix.
Comment on attachment 259475 [details] the fix. I wonder if we can figure out some kind of threshold to use to support Windows in future tests. Thank you for figuring this out! r=me.
Comment on attachment 259475 [details] the fix. Clearing flags on attachment: 259475 Committed r188689: <http://trac.webkit.org/changeset/188689>
All reviewed patches have been landed. Closing bug.