Re: Re: [GENERAL] random() function produces wrong range

Stephan Szabo wrote:
>
> On Tue, 1 Aug 2000, Tom Lane wrote:
>
> > Stephan Szabo <sszabo@megazone23.bigpanda.com> writes:
> > > Actually, on my machines, both man pages for rand() and random() say
> > > they return values between 0 and RAND_MAX (whether that's true or not
> > > is another matter).  In my case RAND_MAX==INT_MAX so the change wouldn't
> > > be a problem, but it might be problematic on some of the 64 bit machines.
> >
> > Oh, that's interesting.  What platform do you use?  If RAND_MAX applies
> > to random() on some machines that'd probably explain why the code is
> > written like it is.  But on my box (HPUX) the rand() function is old
> > and crufty and considerably different from random().
>
> That's from a pair of linux boxes, although checking on a FreeBSD box a
> friend has, his boxes man pages show the range as explicitly 0 to 2^31-1
> as your box does.

On my SCO 5.0.4 box, rand() from the man page...

The rand function uses a multiplicative congruential random-number
generator
  with period 2^32 that returns successive pseudo-random numbers in the
range
  from 0 to (2^15)-1.


The following functions define the semantics of the functions rand and
srand.

     static unsigned long int next = 1;
     int rand()
     {
          next = next * 1103515245 + 12345;
          return ((unsigned int)(next/65536) % 32768);
     }
     void srand(seed)
     unsigned int seed;
     {
          next = seed;
     }


--
Dave Smith
Candata Systems Ltd.
(416) 493-9020
dave@candata.com