Duck Chess

[hgm]

The patch that supports Duck Chess is quite recent, and I don't think it propagated any further than my own on-line repository. You can find it in the latest snapshot of the v4.9.x branch there ( http://hgm.nubati.net/cgi-bin/gitweb.cg ... ads/v4.9.x ).

To compile and install the source you download from there you would need the commands

./autogen.sh
./configure
make
sudo make install

[rgrosz789]

To compile and install the source you download from there you would need the commands

./autogen.sh
./configure
make
sudo make install

I downloaded and extracted your tar.gz file for xboard. When I ran ./autogen.sh, it failed to run autopoint.

I searched on github, and there are too many sources named autopoint - which one should I use?

[rgrosz789]

NM
I just used apt-get install autopoint.

[rgrosz789]

The patch that supports Duck Chess is quite recent, and I don't think it propagated any further than my own on-line repository. You can find it in the latest snapshot of the v4.9.x branch there ( http://hgm.nubati.net/cgi-bin/gitweb.cg ... ads/v4.9.x ).

To compile and install the source you download from there you would need the commands

./autogen.sh
./configure
make
sudo make install

Running ./configure returns an error 'no such file or directory'

After reading configure.ac, I installed autoconf. When I tried running 'autoconf ./configure.ac' I got lots of warnings that suggested running autoupdate. Not sure exactly what I should try next ...

[hgm]

That is bad. It must be some difference between distros. Problem is that I did not create this automake/autoconf stuff myself, and understand very little of it.

Since the problem seems to be that running ./autogen.sh does not produce a file ./configure for you, I have run "make dist" (after ./autogen.sh and ./configure) on my machine for the latest snapshot, to make a "distribution tar ball". Such a tar ball already contains the configure script, as the process involves running ./autogen.sh. This is the form in which XBoard sources are normally distributed. I uploaded it to

http://hgm.nubati.net/transfer/xboard-4.9.1.tar.gz .

It should be possible for you to build XBoard from that by running

./configure
make
sudo make install

If not, you could try to directly use a binary that I produced on my machine, although I am not sure such binaries can be transferred between distros. (They might dynamically link to libraries that are kept in different places; this is what ./configure would normally iron out.) Anyway, I uploaded it to

http://hgm.nubati.net/transfer/xboard.gz .

You could use that to replace an XBoard binary that was installed on your system by another method. You would then have to move the Duck image (which is new in this version) to the directory where XBoard expects its default piece images (something like /usr/local/share/games/xboard/...).

I hope one of these methods works.

[rgrosz789]

...
http://hgm.nubati.net/transfer/xboard-4.9.1.tar.gz .

It should be possible for you to build XBoard from that by running

./configure
make
sudo make install

If not, you could try to directly use a binary that I produced on my machine, although I am not sure such binaries can be transferred between distros. (They might dynamically link to libraries that are kept in different places; this is what ./configure would normally iron out.) Anyway, I uploaded it to

http://hgm.nubati.net/transfer/xboard.gz .

Looks like I struck out on both of these ideas.

After running ./configure, it said I needed to install makeinfo.
apt-get install makeinfo failed
Googled that, answer is apt-get install texinfo
now ./configure fails with error messages about three packages: cairo, pangocairo and librsvg-2.0
Based on synaptic package manager it appears I have newer versions of both pangocairo and librsvg installed

Next I tried running your xboard binary, but that didn't seem to work. It couldn't locate the chess engines, or the images for the chess pieces.

[hgm]

Well, the binary appears to run, and the error it reports can be cured. It just means its piece images are not found in the location where it expects them. So you couldjust move them there. I believe that for this binary that would be /usr/local/share/games/xboard/themes/default/. Just copy all files from the svg directory in the distribution rarr ball there. That should get you passed this error.

[rgrosz789]

Well, the binary appears to run, and the error it reports can be cured. It just means its piece images are not found in the location where it expects them. So you couldjust move them there. I believe that for this binary that would be /usr/local/share/games/xboard/themes/default/. Just copy all files from the svg directory in the distribution rarr ball there. That should get you passed this error.

I did copy the piece images. To save the settings, I copied xboard.conf from xboard-4.9.1 to /usr/local/etc

I have played duck chess against Fairy Stockfish and MayhemDuck. The engines play duck chess against each other - but only for one or two moves:

[rgrosz789]

Great to see more engines !

Thanks for the game. I added 1 tactic to MayhemDuck. I uploaded MayhemDuck-1.4 to the latest Mayhem release.
Some work has been done. Added "play" command so anybody can watch game in shell. Tons of bugs found and fixed.
My thinking is that v1.4 is much stronger. As it now solves all tactics ( + symmetry ) immediately.
Signature: 38185312
I recommend avx2 version as it's a bit faster than 64bit.

I finally got MayhemDuck 2.0 working under linux today. I used the file "mayhemduck-2.0-x86-linux-64bit", since I have a very old PC.

When I played this game, mayhem rejected my move 10 as illegal.
When I load the PGN file with mayhem as the engine, it does NOT reject move 10.
When I load the PGN file with FairyStockfish as the engine, it does reject move 10.

And I just noticed that the PGN file does not include the duck moves for Black - it shows the same square as the piece move for Black.

[Event "Computer Chess Game"]
[Site "rick-XPS-8300"]
[Date "2023.02.22"]
[Round "-"]
[White "rick"]
[Black "MayhemDuck 2.0 (UCI2WB)"]
[Result "*"]
[TimeControl "180+2"]
[Variant "duck"]
[Annotator "1... +0.25"]

1. e4,e6 Nf6,f6 {+0.25/6 9} 2. e5,d5 Ng4,g4 {-0.06/6 9} 3. f4,d6 h5,h5 {+0.17/5 8} 4. h3,h6 d6,d6 {-2.71/4 8} 5. hxg4,d7 dxe5,e5 {-2.71/4 8} 6.
Rxh5,d7 Rxh5,h5 {-2.13/7 8} 7. gxh5,d7 Nc6,c6 {-2.47/5 7} 8. d3,d7 e6,e6 {-2.22/5 7} 9. fxe5,e7 Qd4,d4 {-2.52/5 7} 10. Nf3,d5

[hgm]

OK, so XBoard seems to finally work. And now there is a problem with the move format of the engine. To know exactly what the engines do wrong, I would have to know what they do. I.e. see the log. Unfortunately XBoard was not sized large enough to see the full error mesage in the second case. And if the move in the first case really was d4,d4, that would seem illegal to me, and I am surprised XBoard accepts it.

So try to run the games again with the option -debug, and post the resulting xboard.debug file here.

[Edit] That was my answer to your first message, and now I see there is a second one. And the same applies: I must see the logs to see what Mayhem actually sends. Like you say, it seems to put the Duck on the same square as the piece it moves. XBoards legality testing is obviously not strict enough here; it should have rejected all Mayhem's moves, also when loading the PGN. Probably the bug is that it tests whether the Duck destination is empty in the position before the move.

That still leaves the matter of why it thinks Mayhem puts the Duck there, while Mayhem itself obviously thinks the Duck is elsewhere. Like on f3 in the position where it rejects your move 10.