 |
 |
Joerg Stiller writes: > Does the Fortran 95 standard define how the sign of an > integer number is to be handled within the bit manipulation > functions? No. Indeed, it specifically says that "the interpretation of bits in such objects is processor dependent". The "such objects" in question are those with the top bit set. The main point is that this allows integers to be implemented as 1's complement, 2's complement, or even sign-magnitude. The Fortran standard does not force that level of implementation detail on a processor. -- Richard Maine [log in to unmask] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%