[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] Failed assertion with the --nofuseki option
From: |
Gunnar Farneback |
Subject: |
Re: [gnugo-devel] Failed assertion with the --nofuseki option |
Date: |
Mon, 05 May 2003 17:53:43 +0200 |
User-agent: |
EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/20.7 (sparc-sun-solaris2.7) (with unibyte mode) |
St?phane wrote:
> It seems that GnuGo sulks when asked to generate a first move
> as black without its belovced fuseki database :-) The following
> command :
>
> gnugo --quiet --nofuseki --color white
>
> gives the following output :
>
> [...]
>
> ***assertion failure:
> genmove.c:647 - dragon2 != 0 near PASS***
Thanks for finding this bug. It's fixed by the appended patch.
By the way, --fuseki completely turns off generation of fuseki moves
and is mainly useful if you have a local move generation problem (e.g.
a life and death problem in one corner of an otherwise empty board)
and don't want to have a move generated in some empty corner, even if
it should happen to be bigger. On an empty board, GNU Go passes with
--nofuseki.
To only turn off the fuseki database, use --nofusekidb.
/Gunnar
Index: engine/genmove.c
===================================================================
RCS file: /cvsroot/gnugo/gnugo/engine/genmove.c,v
retrieving revision 1.70
diff -u -r1.70 genmove.c
--- engine/genmove.c 21 Apr 2003 21:40:35 -0000 1.70
+++ engine/genmove.c 5 May 2003 15:51:54 -0000
@@ -643,6 +643,9 @@
int found_one = 0;
int other = OTHER_COLOR(color);
+ if (stones_on_board(BLACK | WHITE) == 0)
+ return 0;
+
gg_assert(dragon2 != NULL);
for (pos = BOARDMIN; pos < BOARDMAX; pos++) {