I've encountered a bizarre problem analyzing with Scid. When analyzing starting from this FEN position:
b2r4/3r1pkp/p2ppn2/1p2q1p1/3NP3/PBPR1P2/1P3QPP/4R1K1 w - - 1 29 which is the position after Black's 29th move in Game 16 of the 1993 Kasparov-Short championship match, available here: https://www.chessgames.com/perl/chessgame?gid=1070688 Scid crashes after analyzing two moves. I am analyzing with Stockfish 230620 64 BMI2 in Scid version 4.6.4 (using tcl/tk version 8.6.9) on a Debian Buster machine. The program exits with this error message: *** buffer overflow detected ***: /usr/games/tkscid terminated Aborted It happens whether Stockfish is configured to use half of the installed RAM on the machine, or one quarter of the installed RAM, or one eighth of the installed RAM. The crash happens after precisely two moves of analysis: if I analyze at 3600 seconds per move, Scid crashes after 7200 seconds; if I analyze at 30 seconds per move, Scid crashes after 60 seconds. It continued to happen after I installed a new developmental version of Stockfish. Scid analysis with Stockfish did not crash when analyzing other games, or the earlier positions of this game. The only online information I've found on this is a Scid-users message from 10 years ago, which was left unresolved: https://sourceforge.net/p/scid/mailman/message/24627767/ (As a test, I tried analyzing the Greco game in that thread starting from move 7; Scid did not crash.) My file ~/.scid4.6/log/engine1.log contains no useful information (it ends with "Log file size limit reached; closing log file"). The Stockfish log file also shows nothing out of the ordinary. This has happened multiple times, including after installing a new developmental version of Stockfish. Does anyone know how to diagnose the problem? Regards, Greg Marks
signature.asc
Description: PGP signature
_______________________________________________ Scid-users mailing list Scid-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scid-users
