Closes #185 by fixing the browsers we test the polyfll itself against - the rest are obviously covered by WPT and what we lack here would be a simple test switch otherwise that isn't very useful
(the rationale for which is that the polyfill has to make approximations with the tools we had available, like removing a tabindex attribute, and restoring it later - and it tests those approximations which of course won't be true if you're using the new 'primitive' which can juggle those internally for you)
Closes #185 by fixing the browsers we test the polyfll itself against - the rest are obviously covered by WPT and what we lack here would be a simple test switch otherwise that isn't very useful
(the rationale for which is that the polyfill has to make approximations with the tools we had available, like removing a tabindex attribute, and restoring it later - and it tests those approximations which of course won't be true if you're using the new 'primitive' which can juggle those internally for you)