The few unit test failures were different
from before. So, I will assume they all are intermittent failures and no
new regressions. I would encourage teams to open bugs about the failing
tests, as I'm sure it indicates _some_ problem, even if intermittent, and
even if worse on "busy machine". But, I'll promote this
build soon, unless I hear otherwise. (Oh, and I did misunderstand what
I was seeing in the map files, and Raghu told me the versioning issue on
the test bundle would be expected, and he'll fix after M5).