Closed mahays0 closed 10 years ago
As far as I can see looks good to me. I would probably have preferred us not to have to make such wide changes but if we have to to make the test run I guess there is no choice.
Likewise from me. I think backporting those DOH fixes will make it more reliable in general to run regression tests against 1.7, which is a good thing.
PS: One other thing that bothers me is that IIUC you are breaking the dojo/util code between 2d90107 and 2a75c9648d5dd10218cda85f3fafa24fc3db70bf. That stops people from using git bisect
.
Can you squash 2a75c9648d5dd10218cda85f3fafa24fc3db70bf into changeset 2d90107?
I merged those aspects of the commits as Bill suggested. We want to make sure to pull this at the same time as the dojo/dojo PR because some of the _internal robot APIs were refactored.
I've merged all the changes mentioned in this PR, although not taken directly from this PR since the check in comments here (ex: 1875d6c0e9c0970ac66f98ec0512176a30138b08) are incorrect.
There were substantial infrastructure and robot changes to DOH in 1.8. They are necessary to get the tests running smoothly in 1.7 under IE11. A lot of the 1.8 code is promise-based so I simplified the code to use legacy deferral code where possible.