Re: [Bug-gnubg] Re-evaluating moves doesn't update assessment

[Jim Segrave]

On Wed 31 Jul 2002 (09:55 +0100), Ian Shaw wrote:
> 
> Windows build 020730
 
> In the annotation window, changing the evaluation settings to a higher play
> and re-evaluating does not update GnuBg's assessment of the move i.e.
> Doubtful, Bad, Very Bad.
> 
>     GNU Backgammon  Position ID: 7GwHAyDo3gYDIA
>                     Match ID   : UYkyAQAAAAAA
>     +24-23-22-21-20-19------18-17-16-15-14-13-+  O: StrangeLuck
>     | X     O  O     O |   | O  O           X |  0 points
>     |       O  O     O |   | O  O           X |  
>     |          O       |   |    O             |  
>     |                  |   |                  |  
>     |                  |   |                  |  
>     |                  |BAR|                  |v 9 point match
>     |                  |   |                  |  
>     |                X |   |                  |  
>     |             X  X |   |                  |  
>     |             X  X |   | X  X           O |  Rolled 54
>     | O        X  X  X |   | X  X           O |  0 points
>     +-1--2--3--4--5--6-------7--8--9-10-11-12-+  X: ian (Cube: 2)
> 
> 
>     1. Cubeful 0-ply    13/4                         Eq.:  +0.404
>        0.575 0.154 0.004 - 0.425 0.104 0.003
>         0-ply cubeful [expert]
>     2. Cubeful 0-ply    24/15                        Eq.:  +0.363 ( -0.042)
>        0.589 0.094 0.002 - 0.411 0.126 0.002
>         0-ply cubeful [expert]
> 
> Marked "doubtful"
> 
>     1. Cubeful 2-ply    13/4                         Eq.:  +0.410
>        0.577 0.149 0.005 - 0.423 0.105 0.004
>         2-ply cubeful 100% speed, 16 cand., 1 tol. [world class++]
>     2. Cubeful 2-ply    24/15                        Eq.:  +0.409 ( -0.000)
>        0.593 0.100 0.002 - 0.407 0.130 0.004
>         2-ply cubeful 100% speed, 16 cand., 1 tol. [world class++]
> 
> Remained "doubtful"

I just tried this under a Unix version from last night. The window
updates correctly (doubtful under expert evaluation, no commentary
under world class++), the same equities being shown.

But something awful did happen:

Clicking Help, commands or simply entering Help as text freezes
gnubg. It appears to be in an endless loop doing GTKHelpAdd of
"set level for tutor warnings". 

Joern changed the way the commands for tutor mode work and it looks
like he's created a loop in the help structure. I think the error is
a lack of a terminating entry in gnubg.c, acSetTutorSkill[] is missing
it's terminator:

--- gnubg.c.orig        Wed Jul 31 11:49:13 2002
+++ gnubg.c     Wed Jul 31 11:49:40 2002
@@ -16,7 +16,7 @@
  * along with this program; if not, write to the Free Software
  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  02111-1307  USA
  *
- * $Id: gnubg.c,v 1.261 2002/07/29 18:55:27 thyssen Exp $
+ * $Id: gnubg.c $
  */
 
 #include "config.h"
@@ -918,6 +918,7 @@
     NULL, NULL },
   { "verybad", CommandSetTutorSkillVeryBad, N_("Warn about `very bad'
   play"),
     NULL, NULL },
+  { NULL, NULL, NULL, NULL, NULL }
 }, acSetTutor[] = {
   { "mode", CommandSetTutorMode, N_("Give advice on possible
   errors"),
     szONOFF, &cOnOff },


-- 
Jim Segrave           address@hidden