------- Additional Comments From schwab at suse dot de 2004-09-29 08:16 ------- IMHO "differ in sign" would be wrong. It's not the sign (plus or minus) that matters, but whether the sign bit exists at all. Even the C standard talks about signedness of types (6.2.5#8).
-- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17720
