Steve landed the analysis in bug 1294915 (\o/) but there are still 7 hazards, and actually 9 as of a few minutes ago. We have a threshold value to prevent us from adding more, but we should drive that number to zero. We have some dependent bugs about this already, which I've assigned to Manish for easy fixes or triaging.
We should try to fix all of Stylo's static analysis hazards before we enable Stylo by default on Nightly.
I think I've fixed all of the actual bugs in the code (and updated the whitelist as much as I can.) The remaining ones seem to be cases where the current analysis is inadequate. Steve, can you take it from here?