> Do we want to address pointer signedness issues? I guess... if we have a prototype void foo(long *x); then we should probably fix code that does unsigned long y; foo(&y); unless it gets too messy.