Problem with winboard 4.6.2

[Pedro Castro]

I played a tournament with multiple engines.

There are many engines that have problems playing with black, do not start the game, for example Eveann, Clarabit, chessKISS. These engines have no problem playing with earlier versions (winboard_x).

Here is a debug file not starting the game for Averno.

One problem seems to be that Winboard engine sends a move without waiting for the engine has sent the command "feature", then the engine may not recognize commands such as move or white or black command.

Code: Select all

recognized 'normal' (-1) as variant normal
recognized 'normal' (-1) as variant normal
WinBoard 4.6.2 + tnwb
Reset(1, 0) from gameMode 0
recognized 'normal' (-1) as variant normal
GameEnds(0, (null), 2)
shuffleOpenings = 0
StartChildProcess (dir=".\engines\toledo") "./tnwb.exe"
nice engine proc to 10
656 >first : xboard
protover 2
656 <first : feature myname="Toledo Nanochess Jan/11/2010" done=1
656 >first : accepted myname
656 >first : accepted done
672 >first : new
random
672 >first : level 0 3 2
672 >first : post
672 >first : hard
672 >first : easy
Impossible move , type = 0
Reset(0, 1) from gameMode 0
recognized 'normal' (-1) as variant normal
GameEnds(0, (null), 2)
672 >first : quit
shuffleOpenings = 0
StartChildProcess (dir=".\engines\toledo") "./tnwb.exe"
nice engine proc to 10
1203 >first : xboard
protover 2
1203 >first : new
random
1203 >first : level 0 3 2
1203 >first : post
1203 >first : hard
1203 >first : easy
Impossible move , type = 0
1203 >first : force
StartChildProcess (dir=".\engines\averno081") "./Averno081.exe" 081
nice engine proc to 10
1766 >second: xboard
protover 2
1766 <second: feature done=0
1766 >second: accepted done
1766 <first : feature myname="Toledo Nanochess Jan/11/2010" done=1
1766 >first : accepted myname
1766 >first : accepted done
1781 >second: new
random
1781 >second: level 0 3 2
1781 >second: post
1781 >second: hard
1781 >second: easy
Impossible move , type = 0
1781 >second: force
New game (1): Toledo Nanochess Jan/11/2010-Averno081 (w)
1797 >first : computer
1797 >second: computer
1797 >first : black
1797 >first : time 18000
1797 >first : otim 18000
1797 >first : white
book key = 463b96181691fc9c
book move field = 731
book hit = d2d4
1797 >first : force
1797 >first : d2d4
nps: w=-1, b=-1
1813 >second: time 18000
1813 >second: otim 18198
book hit = (NULL)
1813 >second: d2d4
1813 >second: black
1813 >second: go
silence
2047 <second: feature ping=1 setboard=1 playother=1 san=0 usermove=1 time=1 draw=0 sigint=0 sigterm=0 reuse=1 analyze=1 myname="Averno v0.81ai (081)" variants="normal" colors=0 ics=1 name=1 pause=0 done=1
Interrupting second
2047 >second: accepted ping
2047 >second: accepted setboard
2047 >second: accepted playother
2047 >second: accepted san
2047 >second: accepted usermove
2047 >second: accepted time
2047 >second: accepted draw
2047 >second: accepted sigint
2047 >second: accepted sigterm
2047 >second: accepted reuse
2047 >second: accepted analyze
2047 >second: accepted myname
2047 >second: accepted variants
2047 >second: accepted colors
2047 >second: accepted ics
2047 >second: accepted name
2047 >second: accepted pause
2047 >second: accepted done
2094 <second: Error (unknow command): d2d4
2094 <second: Error (unknow command): black
2094 <second: 0 0 0 0  <L> e2e4(978) <L> d2d4(958)
2094 <second: move e2e4
GameEnds(25, Xboard: Forfeit due to invalid move: e2e4 (e2e4) res=23, 4)
GE(25, Xboard: Forfeit due to invalid move: e2e4 (e2e4) res=23, 4) bare king k=16 color=0
Interrupting first
2110 >first : result 1-0 {Xboard: Forfeit due to invalid move: e2e4 (e2e4) res=23}
2110 >second: result 1-0 {Xboard: Forfeit due to invalid move: e2e4 (e2e4) res=23}
2110 >first : quit
2110 >second: quit
Saliendo: Partida Toledo Nanochess Jan/11/2010 vs. Averno v0.81ai (081): puntuación final 1-0-0
GameEnds(25, Xboard: Forfeit due to invalid move: e2e4 (e2e4) res=23, 2)


[H.G.Muller]

After a first glance at the debug file I get the impression that this could be a bug in the GUI-book code, rather than a difference between versions. Does this problem also occur when the GUI book is switched off? Winboard_x of course did not support a GUI book.

Toledo uses the GUI book here, and gets (of course) an immediate hit on the opening position. This hit is then sent to Averno before it is done starting up. But as the feature usermove=1 has not been received, the move is not prefixed by usermove, and Averno ignores it. Because of that it then thinks it has to play white after receiving go. IIRC I do wait for ping/pong to acknowledge when the second engine is ready, but in this case the engine has not even told WB that it does support ping. So I guess that code will be skipped here.

I will have a look at it.

[Pedro Castro]

I am repeating the tournament, but this time I'm using Winboard 4.5. (20110826)

Engines that have their own opening book using it.
Uci engines without opening book using default_book provided by Polyglot.
4 Xboard engines without opening book using default_book provided by the GUI; Betsabe, IACH, Roque and Toledo.

Now everything seems fine, although I have not played many games.

I have not tried yet with the latest version to disable the GUI book. But in the previus tournament Clarabit not started with black even his rival used own book, PolyGlot book or GUI book.