🚀 go-pugleaf

Hi,

I've encountered a bug in GNU Chess 5.0 on a Sun Sparc Solaris platform and
attempted to send a report to bug-gnu-chess@ai.mit.edu (the address
mentioned in the README file) however, my mail bounces back from that
address.

Hopefully posting the problem to this newsgroup will be helpful.

The bug:
  I played a game where gnuchess captured a piece with a knight, by moving
the knight diagonally forward one square, i.e.  it captured the way a pawn
would capture.

I figured someone might  care about this type of behavior.
Unfortunately I can not reproduce the problem.

The moves of the game are as follows (Please no comments on my game playing
:-) I was playing white:
    1. e4 Nc6
    2. Bc4 e5
    3. Nc3 Nf6
    4. Nf3 b5
    5. Bxf7+ Kxf7
    6. O-O Kg8
    7. d4 exd4
    8. Nxd4 b4
    9. Ncb5 Nc6xb5
    10. Nxb5 d5
    11. Bf4 Bg4
    12. Qd4 b8c6

The problem occurred on move 9 ... Nc6xb5  where it explicitly moved a
knight diagonally to capture.  I did not notice the problem (was making
moves in between editing/compiling stuff) until move 12 ... b8c6 produced no
visible difference in the board (since b8 was an empty square).

By the way, I launched gnuchess with "xboard -size slim -depth 1" if that
matters, and I am running it without a book file since the install targets
in the Solaris Makefile are broken. Yes,  I can not say for sure if this is
a gnuchess bug or an xboard bug.


--
Stephen
1nagroms at retsehcor dot rr dot moc
(reverse the letters, replace the at and dots with symbols)