Clint> Could be. Is the strace output below consistent with that Clint> bug? No. I don't see any calls to sigaltstack() and even if there were such a call, the bug triggered only under very specific circumstances. Clint> merulo% grep sigaltstack faked Binary file faked matches Clint> strace -e signal ./saved_acl2 and fakeroot strace -e signal Clint> ./saved_acl2 What does gdb's backtrace say when you get the first SEGV? --david