Per wrote: > I probably broke this, and hopefullt fixed it with my September 1 > "PATCH to maybe_yank_clinit". (I'll verify this after I re-build.) I > can check it in, but I'd prefer Alex to take a look first, as I don't > know the reasons for the old logic. I don't think so. I had this patch applied to my tree when I created the test case for this bug. AG