Wraparound on multiply overflow

[Paul Graham <pgraham@cadence.com>]

In C it is defined that the result of multiplying two unsigned integers
is the mathematical result modulo the value (int_max + 1), where int_max
is the maximum unsigned integer.  I'm interested in porting to Ada  a
random number generator which relies on this rule.  The code looks like:

    unsigned int X[10];
    ...
    X[0] = seed;
    for (i = 1; i <= 10; i++)
        X[i] = X[i-1] * 12345;

The idea is that a table of random looking numbers is created by
repeated multiplication of a seed by some constant.  This sequence is
guaranteed to be the same across conforming C implementations, because
of  the modulo rule I mentioned above.

Can this code be written in Ada?  A few iterations of the loop would
result in a range error or overflow.  I could suppress checks, but would
the result of an overflowing multiply be defined in that case?  Is there
any other way of emulating C's behavior for this program?  I know that
Ada has modular types, but these don't work with multiply, do they?

Paul