Hi Robert, On Thu, 2008-12-18 at 01:36 +0000, Robert Lougher wrote: > This should now be fixed in CVS. A missing prototype means long longs > are being interpreted incorrectly on 32-bit systems, leading to a > wrong timeout value (hence the harness is not killing the hung runner > process, and thus appearing to hang itself). Thanks Robert. Things go smoothly again on builder now! We might want to look into the reported failures though: http://lists.gnu.org/archive/html/classpath-testresults/2008-12/msg00074.html The list is a little messed up though, because it was missing the original results files from the times that I forced the run to end early. But it looks like there are some regressions in there. Cheers, Mark
