[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnubg] non-random seeding

From: Jim Segrave
Subject: Re: [Bug-gnubg] non-random seeding
Date: Mon, 19 Jul 2004 20:44:17 +0200
User-agent: Mutt/1.4i

On Mon 19 Jul 2004 (08:06 +0000), Joern Thyssen wrote:
> On Sun, Jul 18, 2004 at 09:33:50PM +0200, Jim Segrave wrote
> > As I posted on GOL I think it's a bug. The call to InitRNG to set up
> > the rollout RNG in gnubg.c specicially tells it not to initialise the
> > seed. It then xor's the seed with a magic number.
> > 
> > Possible solutions:
> > 
> > 1) change the fSet paramter in the call to TRUE
> Committed.
> The call to InitRNG for the rollout seed was introduced in gnubg.c
> 1.109, and had fSet = FALSE!
> So this bug has been here all along. However, maybe your fix in
> gtkgame.c
> (http://savannah.gnu.org/cgi-bin/viewcvs/gnubg/gnubg/gtkgame.c.diff?r1=1.399&r2=1.400&sortby=date)
> is the reason why it pops up now?

I think it was simply an oversight when the code first went in. I
wasn't sure if there was another reason for not asking for the seed to
be set at this call, which is why I suggested two possible
fixes. I have a vague memory that there was a bug which made rollouts
interact with the normal dice rng which was fixed some time ago. I
think before then, rollouts were picking up whatever the dice rng was
using, so this problem would have been rare, having separated contexts
exposed it. I never noticed because I used to force the rollout seed
in .gnubgrc, so I could test the changes I was doing to rollout code -
I needed to be able to run identical rollouts lots of times. 

Jim Segrave           address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]