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.

--dave