full 16 bit addressing support for 16 bit targets

Jim Wilson wilson@cygnus.com
Tue Sep 30 22:36:00 GMT 1997


It seems curious that you have code in make_signed_type to max precision
at 2 * HOST_BITS_PER_WIDE_INT, but you do not have such a limit in the
nearly identical code in make_unsigned_type.  Shouldn't it be in both
places?

Jim



More information about the Gcc mailing list