Fix to random.m

[Thomas Charles CONWAY]

Hi

Here's a simple fix for the poor behaviour of random.m reported in
the mercury-users mailing list.

-- 
Thomas Conway <conway at cs.mu.oz.au>
Nail here [] for new monitor.  )O+

library/random.m:
	Change the coefficients for the linear-congruential generator.
	They now assume only 31 bits (unsigned) rather than 32. This
	should prevent the defective behaviour reported on the
	mercury-users mailing list.

cvs diff: Diffing .
Index: random.m
===================================================================
RCS file: /home/staff/zs/imp/mercury/library/random.m,v
retrieving revision 1.15
diff -u -r1.15 random.m
--- random.m	1998/01/23 12:33:30	1.15
+++ random.m	1998/10/28 04:11:58
@@ -12,6 +12,8 @@
 % uses a threaded random-number supply. It could be made non-unique, but
 % since each thread returns the same list of random numbers, in the interests
 % of safety, it is declared with (backtrackable) unique modes.
+% The coefficients used in the implementation were taken from Numerical
+% Recipes in C (Press et al), and are originally due to Knuth.
 %
 %---------------------------------------------------------------------------%
 
@@ -64,7 +66,7 @@
 :- pred random__params(int, int, int).	% a, c, m
 :- mode random__params(out, out, out) is det.
 
-random__params(2416, 374441, 1771875).
+random__params(9301, 49297, 233280).
 
 random__init(I0, RS) :-
 	copy(I0, RS).