From: Marco Costalba <mcostalba@gmail.com>
Date: Sat, 15 Nov 2008 23:47:55 +0000 (+0100)
Subject: MovePicker: fix a nasty bug in EvalInfo optimization
X-Git-Url: https://git.sesse.net/?p=stockfish;a=commitdiff_plain;h=75d001adddfaa7a2eb57af99cf6c57a12913d2d1;hp=7daaf03b397f6ac2b6b9436b1b80054a2fea7c81

MovePicker: fix a nasty bug in EvalInfo optimization

EvalInfo has missing attack info when a specialized
endgame function is used.

We missed this case and were using an empty attack bitboard
instead so that no captures were generated for endgames.

After testing the EvalInfo optimization gave worst results,
so after a (long) debug session this nasty bug was found.

Signed-off-by: Marco Costalba <mcostalba@gmail.com>
---

diff --git a/src/movepick.cpp b/src/movepick.cpp
index b1dcea01..43f30f77 100644
--- a/src/movepick.cpp
+++ b/src/movepick.cpp
@@ -78,11 +78,14 @@ MovePicker::MovePicker(const Position& p, bool pv, Move ttm,
   numOfBadCaptures = 0;
 
   // With EvalInfo we are able to know how many captures are possible before
-  // generating them. So avoid generating them in case we know are zero.
+  // generating them. So avoid generating in case we know are zero.
   Color us = pos.side_to_move();
   Color them = opposite_color(us);
-  bool noAttacks = ei && (ei->attackedBy[us][0] & pos.pieces_of_color(them)) == 0;
-  bool noCaptures = noAttacks && (pos.ep_square() == SQ_NONE) && !pos.has_pawn_on_7th(us);
+  bool noCaptures =    ei
+                   && (ei->attackedBy[us][0] & pos.pieces_of_color(them)) == 0
+                   && !ei->mi->specialized_eval_exists()
+                   && (pos.ep_square() == SQ_NONE)
+                   && !pos.has_pawn_on_7th(us);
 
   if (p.is_check())
       phaseIndex = EvasionsPhaseIndex;
@@ -93,7 +96,6 @@ MovePicker::MovePicker(const Position& p, bool pv, Move ttm,
   else
       phaseIndex = (noCaptures ? NoMovesPhaseIndex : QsearchWithoutChecksPhaseIndex);
 
-
   dc = p.discovered_check_candidates(us);
   pinned = p.pinned_pieces(p.side_to_move());