1.4.0

[alexobviously]

• Custom move generation and processing (beta):
  • MoveGenerator: optionally supply a list of these in Variant.moveGenerators. These can be used to generate custom moves based on the state. You can generate both normal move types (e.g. add an extra move to a piece on a certain square or something), or totally different types of move that can be managed by MoveProcessor.
  • MoveProcessor: supply a list of these in Variant.moveProcessors to process custom move types.
• StateTransformer: a way to transform game states based on player perspectives.
  • Variant.stateTransformer takes a StateTransformer that defines how the state is transformed.
  • BishopState.transform([int? player]) or BuiltVariant.transformState() will return a transformed state (or the same state if there is no transformer). It's possible to define a transformer that takes player: null and transforms the state regardless of player. This option exists for cases where you always want to transform
  • VisionAreaStateTransformer applies a mask to the board based on a vision radius around pieces. MaskedState is a subclass of BishopState which contains the [mask] used to create it (for cases where e.g. you want to draw the mask on a board).
  • HideFlagsStateTransformer hides the flags for all of the player's or opponent's pieces, depending on configuration.
• King attacks are calculated in advance for mainline moves and can be accessed with state.meta!.checks!. They come in the form of a list of squares containing pieces that are attacking the king of each player.
• Game.evaluate() now takes pieces in hands into account.
• Gating move logic is no longer part of the standard move logic. GatingMove is its own type which contains a child move.
• SubtractRegion, XorRegion, SetRegion and DirectionSetRegion.
• Regions in Variant.regions are now built into a more efficient form when the variant is built.
• BoardRegions can now be combined using the +, -, & and ^ operators.
• Variants:
  • TenCubed (LargeVariants.tenCubed()): 10x10 board with fairy pieces.
  • Opulent (LargeVariants.opulent()): a variant of Grand.
  • Troitzky (ShapeVariants.troitzky()): vaguely circular shaped board.
  • Omega (ShapeVariants.omega()): experimental version - mostly working except for en-passant on long pawn moves.
  • Five Field Kono (OtherGames.kono()): win by moving all of your pieces to your opponent's start position.
  • Joust (OtherGames.joust()): the square a piece moves from is removed from the board each time it moves.
  • Wolf (FairyVariants.wolf()): an 8x10 variant with several fairy pieces.
  • Knightmate (MiscVariants.knightmate()): royal knight instead of king, common kings instead of knights.
  • Pocket Knight (MiscVariants.pocketKnight()): chess but each player has an extra knight in hand.
  • Breakthrough (OtherGames.breakthrough()): only (modified) pawns, move one to the end to win.
  • New Zealand Chess (FairyVariants.newZealand()): rooks capture like knights, knights capture like rooks.
  • Nightrider Chess (FairyVariants.nightrider()): knights are replaced with nightriders.
  • Shatranj (CommonVariants.shatranj()): an ancient form of chess.
• ActionFillRegionEnding: the end condition for Kono.
• ActionBlockOrigin: implements the behaviour for Joust.

---

This discussion was created from the release 1.4.0.