🚀 go-pugleaf

In article <352eml$r0g@wsdw05.win.tue.nl>,
Henny Wilbrink <wsdwhw@wsdw05.win.tue.nl> wrote:
>Try to compile gnuchess without the  -DAGING and -DKILLT defines in
>the definition of OPT in src/Makefile. This worked for me. I never
>bothered to find out which one of these two defines was causing the
>trouble.
>
>Henny Wilbrink
>wsdwhw@.win.tue.nl


This seems to have helped.  I've only run gnuchess a couple of times,
which isn't really enough to make sure, but it hasn't crashed like
before; it still tends to not make the time control.  There was a
fix to main.c mailed to me by Mike McGann which addresses this, but
which I haven't applied.  It follows the text of this article.

I removed the -DAGING and -DKILLT defines, as suggested by Henny, and
added -DBAREBONES.  It seems to be playing substantially stronger as
well. 

Thanks to everyone that responded,

Tony

>From: Mike McGann <mwm@hasler.ascom.ch>
>in pl71 there is a bug in the timing algorithm in main.c
>that would especially show up if the score turned since it
>would try to allocate more time.

>In main.c it should read:

>/* if I am doing really well use more time per move */
>  else if (me > him && tcompsum < topsum) increment = -1;
>  if (me > (him+60) ) increment = -1;
>  TimeControl.moves[computer] += increment;
>/* but dont let moves go below MINMOVES */
>  if (TimeControl.moves[computer] < MINMOVES )
>	TimeControl.moves[computer] = MINMOVES;
>}

-- 
Anibal Antonio Acero (Tony) (LLama) | Stone age or space age, man will be ask-
      a-acero@uchicago.edu          | ing the same question as his grandparents
___W(312)702-8214 H(312)752-5464____| before him and as his grandchildren after
him: "What star is that?" -- H.A. Rey, _The Stars, A New Way to See Them_