#336 Using Nalimov TB gives wrong evaluation when black is to move

[Luca Daniel]

When using Nalimov Tablebase it shows wrong evaluation when black is to move.
It looks like it uses Syzygy-7 before switching to Nalimov but after the switch it shows wrong evaluation when black is to move.
For example:

[Event ""]
[Site ""]
[Date ""]
[Round ""]
[White ""]
[Black ""]
[Result ""]
[FEN "8/p5pp/2K5/P7/6k1/8/6P1/8 w - - 3 41"]
[Setup "1"]

{ Using Nalimov 6 pieces Tablebase } 41. Kb7 { Tablebase: 41...Kg3 - White wins
(reset in 1 move(s)) === 41...Kf4 41...Kh4 41...Kf5 41...Kg5 41...Kh5 41...a6
41...g6 41...h6 41...g5 41...h5} 41... h5 { Tablebase: 42.Kxa7 - White wins
(reset in 1 move(s)) === 42.a6 42.Ka6 42.Ka8 42.Kb8} 42. Kxa7 { Tablebase:
42...a6 - Black mates in 15 move(s) === 42...Kb6 (after entering 6 pieces
tablebase is shows wrong evaluation)} 42... Kg3 { Tablebase: 43.a6 - White
mates in 16 move(s) === 43.Kb6 (shows correct evaluation when white is to
move)} 43. a6 { Tablebase: 43...Kb6 - Black mates in 14 move(s) (shows wrong
evaluation when black is to move)} 43... Kxg2 { Tablebase: 44.Kb6 - White
mates in 15 move(s) (correct)} 44. Kb6 { Tablebase: 44...a7 - Black mates in
13 move(s) === 44...Kc5 (suggets 44...a7 as if black is to move but it is
white's turn)} 44... Kg3 { Tablebase: 45.a7 - White mates in 13 move(s) ===
45.Kc5 (again correct after black moves)} 45. a7 { Tablebase: 45...Kc5 -
Black mates in 12 move(s) === 45...a8=Q+}

Syzygy-7 works fine BTW.

[Jens Nissen]

Thanks for posting this issue. It appears indeed only on black moves in the Nalimov base due to an incompatibility between QRegExp (deprecated in Qt5 / missing in Qt6) and QRegularExpression.