Hi, Jan Hubicka wrote: > this is variant I ended up re-testing and comitting today. This commit accidentally contained a change in tree-cfg.c that results in some additional fails, e.g. gcc.dg/pr18241-3.c (see http://gcc.gnu.org/ml/gcc-testresults/2009-08/msg02709.html ). I've reverted that in r151119. Ciao, Michael.