Save results of Iterative Deepening in IterativeDeepening-Instance instead of statically

[jlux98]

Saving the results of iterative deepening in a single static variable might lead to a race condition where the result of an old calculation overwrites the result of the new calculation. My suggestion would be to save the results in the IterativeDeepening-instance and either

• provide a static link to the instance, that can be swapped out to a link to another instance
  • might be easier to program
  • could potentially lead to the same race condition as earlier
• provide MiniMax-variants with a reference to the IterativeDeepening-instance, that created them and let them write the result directly in the instance. The conductor will still then only query the relevant instance and throw it away when creating a new one
  • probably needs more programming effort
  • should fix the racecondition more reliably than the other option

[M-oritz]

Depending on the implementation of IterativeDeepening, I dont think the MiniMax-Instance even has to hold a reference to the IterativeDeepening-Instance. If IterativeDeepening holds a reference to the Minimax-Variant and invokes it (as it currently does), this variant returns it's result (best move) and the IterativeDeepening instance can store it.

public void evaluateTree(
                    Tree<? extends Node<ContentType>> tree, TreeEvaluator<ContentType> evaluator, 
                     boolean whitesTurn, int secondsToCompute, int maxDepth) {

        .
        .
        .
        while (depth <= maxDepth && !Conductor.stopCalculating) {
            bestMove = evaluator.evaluateTree(tree, depth, whitesTurn);
            saveMove(bestMove, depth); // <- change implementation of saveMove() to save inside this instance 
                                                         // rather than in a static variable
            depth += 1;
        }
        Conductor.stopCalculating = true; //can this kill a new search?
    }