Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Comet vs Crafty via winboard... a bug hunt

Author: Robert Hyatt

Date: 13:11:07 05/23/98

Go up one level in this thread


On May 23, 1998 at 09:18:02, Stefan Meyer-Kahlen wrote:

>
>On May 22, 1998 at 21:08:57, Robert Hyatt wrote:
>
>>
>>I received a debug file from KK, and it has some interesting stuff
>>in it.  I'm going to clip the relevant part and include it below with
>>comments inserted and indented to separate them from the text.  Note
>>that Crafty is "first" (playing black) and Comet is "second" playing
>>white...
>>
>>Sending to first: d4d5
>>Received from first: Hint: O-O
>>Received from first:     5    -42     111  14055  11. ... Qf6 12. dxe6
>>Qxe6 13. Qxe6+ fxe6 14. Bc4 Kf7
>>Received from first:     5    -42     191  23318  11. ... Qf6 12. dxe6
>>Qxe6 13. Qxe6+ fxe6 14. Bc4 Kf7
>>Received from first:     6    -67     330  41590  11. ... Qf6 12. Bg4
>>Nh6 13. dxe6 Nxg4 14. exd7+ Kxd7
>>Received from first:     6    -37     364  47921  11. ... exd5!!
>>Received from first:     6    -36     443  61004  11. ... exd5 12. Qxd5
>>Ne7 13. Qb7 Qc8 14. Qxc8+ Nxc8 15. O-O
>>Received from first:     6    -36     546  74996  11. ... exd5 12. Qxd5
>>Ne7 13. Qb7 Qc8 14. Qxc8+ Nxc8 15. O-O
>>Received from first: 11. ... e6d5
>>
>>   up to here, all looks good.  Crafty played exd5, after searching to
>>   a depth of 6 plies, and is expecting Qxd5 as a reply.  It sent the
>>   proper string to winboard "11. ... e6d5" which was the move it wanted
>>   to play.
>>
>>
>>
>>Interrupting second
>>Sending to second: time 19182otim 14399
>>Sending to second: e6d5
>>Received from second: 11. e6d5
>>
>>   this doesn't look so good.  Comet echoed crafty's move.  I don't
>>   know that this causes a problem, but look below...
>>
>>
>>Received from first: Hint: Qxd5
>>Received from second:  83 [4] :   b3d5  g8e7  d5d6  d8b6
>>Received from second:  85 [5] :   b3d5  g8e7  d5b7  d8c8  b7e4
>>Received from first:     6    -84     310  22401  12. ... Ne7 13. Qd3
>>Nf6 14. Qxd8+ Kxd8 15. O-O Ke8 16. Rd1
>>Interrupting first
>>Sending to first: time 14545
>>otim 18899
>>Sending to first: e6d5
>>
>>   that is bad.  the move Crafty played was sent right back to crafty
>>   as comet's move.  Crafty rejected it, properly, and winboard probably
>>   popped up an illegal move window at this point...
>>
>>
>>Received from second:  85 [5] :   b3d5  g8e7  d5b7  d8c8  b7e4
>>Received from first: illegal move.
>>Received from first:     6    -60    1151  85897  12. ... Ngf6 13. Qc6
>>a5 14. O-O Rc8 15. Qf3
>>Interrupting first
>>Sending to first: time 14600
>>otim 18399
>>Sending to first: e6d5
>>
>>
>>   winboard sends it again, since it was rejected.  Crafty rejects it
>>   a second time.  with the same illegal move message.
>>
>>
>>Received from first:     6    -60    1192  88346  12. ... Ngf6 13. Qc6
>>a5 14. O-O Rc8 15. Qf3
>>Received from first: illegal move.
>>Received from second:  87 [6] :   b3d5  g8f6  d5g2  a8a7  b1d2  d8a8
>>Received from second:  87 [6] :   b3d5  g8f6  d5g2  a8a7  b1d2  d8a8
>>Received from second: 12. ... b3d5
>>
>>
>>   now Comet sends the real move it wants to play.
>>
>>
>>who is wrong?  I don't know.  Either this is a problem caused by the
>>echo of the opponent's move by Comet (which is not needed) or it is a
>>bug in winboard.  I'll send a note to Tim Mann to let him look...  but
>>clearly something is wrong...
>
>It is bug in Winboard. If you send a string containing "..." (like in
>your pv) it will be misinterpreted. Tim Mann knows about it and has
>already fixed the problem.
>
>Stefan



talked with Tim last night, in fact...  I converted to a more readable
PV format several months ago, producing analysis that fits the PGN
standard a little better.. and the "23. ... h3 etc" format that is
correct for PGN is confusing winboard.  This can be turned off with
"display nomovenum" or by leaving off the "show thinking"...

And Tim said 3.6.8beta will have better move parsing...  soon...



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.