Re: [Bug-gnubg] "SetMatchID cannot handle positions where a double has been offered"

[Christian Anthon]

Hi Ingo,

the problem is that gnubg gets confused if the take/pass moverecord
isn't preceded by a a double move record. It doesn't matter for you
since you don't continue playing, but rather just enter a new gnubgid
every time. The fix isn't optimal, but for a normal user it doesn't
make too much of a difference as one can easily work around the
problem. If you open a bug report I'll get around to fixing it at some
point.

For now the easiest work-around for you is probably to enter the
gnubgid prior to the double. The output returned will contain the cube
action for the take side as well. Or even easier, just remove the
warning. I believe that it should work for your setup.

Christian.

On Wed, Aug 5, 2009 at 5:36 AM, Ingo Macherius<address@hidden> wrote:
> Hi,
>
> I'm debugging a playing FIBS bot which has problems with a gnubg change by 
> Christitan done on January 11th in play.c. This bug showed after the gbubg 
> binary was updated to a recent  gnubg CVS version.
>
> The bot does attach gnubg to STDOUT and sends batch commands to it, parsing 
> the output. For cube decisions, the pseudo code is:
>
> String fibsboard = ... ;
> String posid = calculateGnubgPositionID(fibsboard);
> String matchid = calculateGnubgMatchID(fibsboard);
> String[] hint = sendToGnubg(new String[] {
>        "set gnubgid " + posid + ":" + matchid",
>        "hint" };
> if (acceptDouble(hint)) {
>        ...
>
>
> Here is an excerpt from the logs:
> ...
>>FIBS: 
>>board:You:inim:3:0:0:0:-1:-1:-1:0:0:5:0:3:0:0:0:-4:5:0:0:1:0:0:-5:-3:0:0:0:1:0:1:0:0:0:0:1:1:1:0:-1:1:25:0:0:0:0:0:4:0:0:0
>>FIBS:
>>FIBS: inim doubles. Type 'accept' or 'reject'.
> <FIBS: board
>>FIBS: 
>>board:You:inim:3:0:0:0:-1:-1:-1:0:0:5:0:3:0:0:0:-4:5:0:0:1:0:0:-5:-3:0:0:0:1:0:1:0:0:0:0:1:0:0:1:-1:1:25:0:0:0:0:0:4:0:0:0
> <gnubg: set gnubgid cB/wACrgc/ARIA:cBFgAAAAAAAA
>>gnubgI'm sorry, but SetMatchID cannot handle positions where a double has 
>>been offered
> <gnubg: hint 1
>>gnubg: 
>>board:opp:gnubg:3:0:0:0:-1:0:0:0:3:5:0:0:-1:0:0:-5:4:0:0:0:-3:0:-5:0:0:1:1:1:0:-1:3:3:3:3:1:1:1:0:1:-1:0:25:0:0:0:0:0:0:0:0
> <gnubg: Setting GNUBG id cB/wACrgc/ARIA:cBFgAAAAAAAA
> <gnubg:     1. Cubeful 2-ply    8/5(2) 6/3*(2)               MWC:  55.34%
> ...
>
> Note the "I'm sorry, but SetMatchID cannot handle positions where a double 
> has been offered", which breaks the approach so far used by (not only this) 
> bot. Please note courtesy of threading the log entries are slightly out of 
> order, but I guess still clear.
>
> The patch causing the bot to fail is:
> http://cvs.savannah.gnu.org/viewvc/gnubg/gnubg/play.c?r1=1.347&r2=1.348
>
> Questions:
> - What is the rationale of the patch, it seems that bots played very well 
> with the old code (going to 2050 and above) ?
> - Is there a workaround I can do other than reverting to a pre-Januaray 11th 
> version?
>
> Thx in advance,
> Ingo
>
>
>
> _______________________________________________
> Bug-gnubg mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/bug-gnubg
>