[gnugo-devel] Bug report gnugo-3.3.8

[Markus Enzenberger]

Hi,

I recently tried to use the GnuGo 3.3.8 API for some experiments.
Maybe I will try to combine GnuGo and NeuroGo later using GnuGo for providing 
input features, if I can get it running fast and stable enough.

I implemented a class for syncronizing GnuGo's board with NeuroGo's board by 
using gnugo_play_move and gnugo_undo_move(1) and then calling 
gnugo_examine_position. Unfortunately GnuGo crashes after only about
150 play/undo cycles.

Here is some more information:

I initialize GnuGo with:

gg_srand(0); /* Undocumented in the GnuGo docs but necessary */
init_gnugo(32);
level = 0;

The assertion output is:

***assertion failure:
move_reasons.c:431 - k < 100 near PASS***


   A B C D E F G H J
 9 . . . . . . O . . 9
 8 . . . . . . . X . 8
 7 . . + . . . + . . 7
 6 . . . . . . . . . 6
 5 . . . . + . . . . 5
 4 . . . . . . . . . 4
 3 . . + . . . + . . 3
 2 . . . . . . . . . 2     WHITE has captured 0 stones
 1 . . . . . . . . . 1     BLACK has captured 0 stones
   A B C D E F G H J

gnugo 3.3.8 (seed 0): You stepped on a bug.

The debugger shows that k==100, is this bug related to the FIXME comment
following in move_reason.c:add_move_reason() or am I doing something wrong
with initializing or using the GnuGo API?

Thanks

- Markus