Running individual random tests might still be useful for finding crashes or assertion failures, but there is no need to run the styloVsGecko visual comparison tests if the results are random.
MozReview-Commit-ID: Brz6zf25lCO
--HG--
extra : rebase_source : 74b453e952da2710f069afc6b8d8017a4ec021dd
extra : source : 95dc8a8a752d1a1ac05e026115aff319fc27e772