specific warnings they produced were being suppressed anyway. The truncation warnings
that were suppressed are now enabled, and the few that still occur ought to be fixed.
The only warnings suppressed now are the "negating an unsigned is still unsigned", the
validity of which Chris does not accept, and the "implicit conversion of <type> to bool
performance warning". Making the conversion explicit won't make it run any faster and
this is an easier way to shut the compiler up.
llvm-svn: 19031
Jeff Cohen