dantb/chess-with-tdd

A chess game and engine built using test-driven development

chess-with-tdd's Issues

En Passant should not be allowed when marked by pawn of same colour

Issue was brought to light by chess engine - should be simple enough to fix, add a validator into MoveValidator. Perhaps EnPassantManager can do the job

Performance issues

There are performance issues when we get to end-game. This is due to the fact that making a new board that's the same as the current board applies every move that has been applied to the current board. It's a long-winded form of cloning.

This also blocks stalemate being implemented since that requires a lot of move validation (to test that one team cannot move at all), meaning it gets it heavily by these issues.

Look into other ways of cloning the board - has to be a deep clone - cloning the squares, pieces, board cache as well as every other property of the board.

Hypothetical move logic doesn't always consider pawn special case

For move logic that is hypothetical, pawn pieces are special cases since they may be able to move currently but not if the hypothetical move was executed, since they can only take diagonally.

For example currently the a king can't move in front of a pawn because the pawn can move there. Should implement a fake move executor perhaps, to play out the hypothetical situation. This could be an IDisposable that undoes all the fake moves after during disposal.

This specific case is from MoveIntoCheckValidator when checking that a king can't move in front of a pawn - it should be able to.
Another potential case is from CheckMateEscapeManager when checking whether the king can escape itself.

Cloning failing reference equality between Board and BoardCache

Noticed this when changing the engine to use the cloning. It appears the board cache pieces are not the same pieces as those on the board - I find it odd at this hasn't manifested itself in a different way.

Try to drill down the specific scenario and raise to the creator of the deep copying at: https://www.codeproject.com/Articles/1111658/Fast-Deep-Copy-by-Expression-Trees-C-Sharp

Check mate when not in check

I've found that checkmate without being in check does not currently get flagged as checkmate because the king isn't directly under threat at this moment. This should be simple enough to fix, in CheckManager/CheckMateManager. Already saved the board position in which this was found - I'll use it for another integration test.