I now pushed the patches to fix these problems to my repository at
winboard.nl .
When testing with the PGN file you gave, I still noticed something
weird, though: The moves after the Rook leaves e1 (so 24... Rhe8 25.
Re1) do not seem to appear in book.
"h.g.muller" via XBoard and WinBoard bugs schreef op 26-10-2025 11:27:
Ah, I see what causes the problem. The long-algebraic for Rh1 produced
by move_to_string already is in error, and produces e1g1. The culprit
is a patch aimed to catch non-compliant notation for castling as King
takes own Rook (used in UCI Chess960 implementation) in normal Chess. I
copied this routine from the public-domain code for probing Polyglot
books. At that time I had already noticed that the patch was no good,
and written a remark about that in a comment:
// correct FRC-style castlings in variant normal.
// [HGM] This is buggy code! e1h1 could very well be a normal R or
Q move.
if(!strcmp(move_s,"e1h1")){
safeStrCpy(move_s,"e1g1", 6);
}else if(!strcmp(move_s,"e1a1")){
safeStrCpy(move_s,"e1c1", 6);
}else if(!strcmp(move_s,"e8h8")){
safeStrCpy(move_s,"e8g8", 6);
}else if(!strcmp(move_s,"e8a8")){
safeStrCpy(move_s,"e8c8", 6);
}
But then I apparently forgot all about it. What is needed here is first
a test to see if we are in variant normal. And then whether there is a
King on the from-square. If these conditions are not both satisfied,
this code should not be executed. I guess we have gotten away with this
for so long because in genuine opening lines there will almost never be
a Rook on e1/e8. And even if there is, it is very unlikely that anyone
moved it to the corner square.
Alan Bennet schreef op 26-10-2025 07:12:
Hi,
Thanks again for replying. I have a new issue with "add next move",
but first to tidy up the previous email.
On 10/25/25 08:36, h.g.muller" via XBoard and WinBoard bugs wrote:
It would not be so easy to implement adequate error reporting /
warning here.
Okay. I don't _really_ need it, because if I make a mistake I can fix
it.
... Perhaps hand-creating a book is a feature that just should not be
used at 3 am. ;-)
Indeed!
... a PGN with a single game is immediately loaded in the memory
buffer as the current game, but when there are more games in the file
a Game List is created,
Expected behavior then. I can just make sure there is more than one
game.
## ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ##
New issue. For the following move list I am having trouble adding the
move 24.Rh1 to the book. It keeps adding 24.Rg1 instead. Or it's doing
something else. Very spooky, just in time for Halloween.
1.e4 d5 2.exd5 Qxd5 3.Nc3 Qa5 4.d4 Nf6 5.Nf3 c6 6.Bd2 Bf5 7.Bc4 e6
8.Qe2 Bb4 9.O-O-O Nbd7 10.a3 Bxc3 11.Bxc3 Qc7 12.Ne5 Nd5 13.Bd2 b5
14.Bb3 Nxe5 15.dxe5 O-O-O 16.g4 Bg6 17.f4 h5 18.h3 a6 19.Qf3 f5
20.exf6 gxf6 21.Rhe1 hxg4 22.hxg4 e5 23.f5 Bf7
User debug session below. Note that I am using the xboard feature that
allows me to add the same move more than once. :)
--
Alan
$ xboard -v
xboard version 4.9.1
configure options: prefix=/usr; datarootdir=/usr/share;
datadir=/usr/share; gamedatadir=/usr/share/games/xboard;
desktopdir=/usr/share/applications; mimedir=/usr/share/mime/packages;
iconsdir=/usr/share/icons/hicolor/48x48/apps;
svgiconsdir=/usr/share/icons/hicolor/scalable/apps;
infodir=/usr/share/info; sysconfigdir=; update_mimedb=yes; NLS=yes;
GTK=yes; Xaw3d=no; Xaw=no; ptys=pipes; zippy=yes; sigint=yes
$ grep -e "-bookDepth" ~/.xboardrc
-bookDepth 51
$ cat ~/Documents/tmp/Rf1-Rg1-Rh1.pgn
[White "Rf1"]
[Result "1/2-1/2"]
1. e4 d5 2. exd5 Qxd5 3. Nc3 Qa5 4. d4 Nf6 5. Nf3 c6 6. Bd2 Bf5
7. Bc4 e6 8. Qe2 Bb4 9. O-O-O Nbd7 10. a3 Bxc3 11. Bxc3 Qc7 12.
Ne5 Nd5 13. Bd2 b5 14. Bb3 Nxe5 15. dxe5 O-O-O 16. g4 Bg6 17. f4
h5 18. h3 a6 19. Qf3 f5 20. exf6 gxf6 21. Rhe1 hxg4 22. hxg4 e5
23. f5 Bf7
24. Rf1 Rhe8 25. Rfe1
1/2-1/2
[White "Rg1"]
[Result "1/2-1/2"]
1. e4 d5 2. exd5 Qxd5 3. Nc3 Qa5 4. d4 Nf6 5. Nf3 c6 6. Bd2 Bf5
7. Bc4 e6 8. Qe2 Bb4 9. O-O-O Nbd7 10. a3 Bxc3 11. Bxc3 Qc7 12.
Ne5 Nd5 13. Bd2 b5 14. Bb3 Nxe5 15. dxe5 O-O-O 16. g4 Bg6 17. f4
h5 18. h3 a6 19. Qf3 f5 20. exf6 gxf6 21. Rhe1 hxg4 22. hxg4 e5
23. f5 Bf7
24. Rg1 Rhe8 25. Rge1
1/2-1/2
[White "Rh1"]
[Result "1/2-1/2"]
1. e4 d5 2. exd5 Qxd5 3. Nc3 Qa5 4. d4 Nf6 5. Nf3 c6 6. Bd2 Bf5
7. Bc4 e6 8. Qe2 Bb4 9. O-O-O Nbd7 10. a3 Bxc3 11. Bxc3 Qc7 12.
Ne5 Nd5 13. Bd2 b5 14. Bb3 Nxe5 15. dxe5 O-O-O 16. g4 Bg6 17. f4
h5 18. h3 a6 19. Qf3 f5 20. exf6 gxf6 21. Rhe1 hxg4 22. hxg4 e5
23. f5 Bf7
24. Rh1 Rhe8 25. Rhe1
1/2-1/2
$ rm ~/chesspgbook/book.bin
$ xboard -ncp &
[1] 29001
GUI | File | Load Game | Recently Used | Rf1-Rg1-Rh1.pgn
GUI | File | Save Games as Book
GUI | File | Quit
$ hexdump -ve '1/1 "%.02x" 1/1 "%.02x" 1/1 "%.02x" 1/1 "%.02x" 1/1
"%.02x" 1/1 "%.02x" 1/1 "%.02x" 1/1 "%.02x " 1/1 "%.02x" 1/1 "%.02x "
1/1 "%.02x" 1/1 "%.02x " 1/1 "%.02x" 1/1 "%.02x" 1/1 "%.02x" 1/1 "%02x
" "\n"' ~/chesspgbook/book.bin | grep -F "612d49983aa49672"
612d49983aa49672 0105 0001 00010001 ## Rf1
612d49983aa49672 0106 0001 00010001 ## Rg1
612d49983aa49672 0107 0001 00010001 ## Rh1
[1]+ Done xboard -ncp
$ xboard -ncp &
[1] 29061
GUI | Edit | Edit Book
GUI | right-click through the book until 23...Bf7 is played
33.3% 1 Rf1 {1/1}
33.3% 1 Rg1 {1/1}
33.3% 1 Rg1 {1/1} ## where is Rh1 ??
GUI | add next move | drag Re1-h1
33.3% 1 Rf1 {1/1}
33.3% 1 Rg1 {1/1}
33.3% 1 Rg1 {1/1}
0.0% 1 Rh1 ## so far so good
GUI | commit changes
25.0% 1 Rf1 {1/1}
25.0% 1 Rg1 {1/1}
25.0% 1 Rg1 {1/1}
25.0% 1 Rg1 ## what happened ??
GUI | File | Quit
$ hexdump -ve '1/1 "%.02x" 1/1 "%.02x" 1/1 "%.02x" 1/1 "%.02x" 1/1
"%.02x" 1/1 "%.02x" 1/1 "%.02x" 1/1 "%.02x " 1/1 "%.02x" 1/1 "%.02x "
1/1 "%.02x" 1/1 "%.02x " 1/1 "%.02x" 1/1 "%.02x" 1/1 "%.02x" 1/1 "%02x
" "\n"' ~/chesspgbook/book.bin | grep -F "612d49983aa49672"
612d49983aa49672 0105 0001 00010001 ## Rf1
612d49983aa49672 0106 0001 00010001 ## Rg1
612d49983aa49672 0106 0001 00010001 ## Rg1
612d49983aa49672 0107 0001 00000000 ## Rh1 spooky
[1]+ Done xboard -ncp
$
