gnugo-devel
[Top][All Lists]
Advanced

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

Re: [gnugo-devel] problem with score feature of 3.2


From: Gunnar Farnebäck
Subject: Re: [gnugo-devel] problem with score feature of 3.2
Date: Fri, 21 May 2004 05:21:03 +0200
User-agent: EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/21.3 (sparc-sun-solaris2.9) MULE/5.0 (SAKAKI)

Adam wrote:
> Ok. I stopped being lazy and tested it with 3.4. 10 of the games now
> seem to work properly! But 2 of them still have trouble. I've
> enclosed them.
> 
> Once again the issue is that gnugo seems to assume a non-integer
> komi and yet ends up with an integer result.

This problem becomes clearer if an explicit komi is added to the game
record, e.g. KM[0]. I've debugged the first example and came up with
the appended patch. I leave it for Dan to check that the patch is
appropriate and for Adam to test whether it's effective also for the
second example.

- ignore inessential critical dragons in in dilate_erode()

/Gunnar

Index: engine/score.c
===================================================================
RCS file: /cvsroot/gnugo/gnugo/engine/score.c,v
retrieving revision 1.23
diff -u -r1.23 score.c
--- engine/score.c      24 Jan 2004 04:04:56 -0000      1.23
+++ engine/score.c      21 May 2004 03:09:28 -0000
@@ -68,7 +68,8 @@
     if (!ON_BOARD(ii))
       continue;
 
-    if (board[ii] && dragon[ii].status == CRITICAL)
+    if (board[ii] && dragon[ii].status == CRITICAL
+       && DRAGON2(ii).safety != INESSENTIAL)
       critical_found = 1;
     if (board[ii] == WHITE && !captured_territory(ii, color))
       gb[ii] = 128;




reply via email to

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