Failing nightly regression tests with
what appear to be either build or environment problems (all builds fail).
Anyone looking into the problem or can at least say what the problem
is? (Chris? Eliot?) Given current problems with time zones + other bugs
in 2.3.4 that have been long since fixed I really want to make a 2.3.5
either tomorrow or Friday. I'm willing to do that if OS X is still
broken since that's been an unresolved issue for almost 2 months; I'm less
happy about making a release where Linux/PPC is broken as well since that
platform was working as recently as 3/31.