These tests reveal that both functions assume that the target is initially positive zero if it is not identical with one of the operands, and fail in various ways if it is not, or if it is uninitialized.
Unlike assert(3), which uses abort(3), this has no other side effects (before raising SIGABRT) than an fprintf() call. The test framework will catch the SIGABRT, report that the test case failed, and proceed with the next case.
POSIX requires <stdio.h> and <unistd.h> to define off_t and ssize_t like <sys/types.h> does, but not all platforms respect that. Play it safe by always including <sys/types.h>.
Instead of having libcryb-test provide main() and assume that the test program defines t_prepare() and t_cleanup(), have libcryb-test provide a t_main() function which the test program calls with pointers to its prepare and cleanup functions.
Add several subtraction test cases and modify some others to better
exercise the carry logic, which is far more complex for subtraction than
for addition.
Add test cases where the target is the same MPI as one or both operands.
be better to systematically pad MPIs used in all tests with garbage.
Greatly expand the number of mpi_cmp() tests to cover all possible
combinations of same or different sign, same or different msb and same
or different magnitude.
Add a t_compare_mpi() predicate which checks that two MPIs have the same
value (but not necessarily the exact same representation).
Clean up and improve existing tests.
