Author: Ed Schröder
Date: 12:45:09 01/17/01
Go up one level in this thread
On January 17, 2001 at 15:38:37, Peter McKenzie wrote: >On January 17, 2001 at 15:16:31, Ed Schröder wrote: > >>On January 17, 2001 at 14:47:52, Peter McKenzie wrote: >> >>>On January 17, 2001 at 14:05:21, Ed Schröder wrote: >>> >>>>On January 17, 2001 at 12:49:50, James T. Walker wrote: >>><snip> >>>>Your log-file is certainly extreme. The number in brackets is the total >>>>number of positions searched. This should be 3707485. If not the "engine >>>>damaged" message is displayed. >> >>>Ed, >> >>Hi Peter, > >Hi Ed > >> >>>this situation obviously isn't helping your cause at all. I think it would be >>>in everybody's best interests if the bug (where ever it may be) is found and >>>fixed. >> >>I agree. >> >>But after 2-3 years of searching I still have no clue. So I build the >>engine check and when I saw its results I dropped the whole case as >>un-solved. Hi Peter, >What I was wanting to know is have you tried putting an engine check after EVERY >MOVE (of Rebel and Rebel's opponent)? No. This would disturb the autoplayer keyboard transmission (missing moves) so I do it only after a game is over (won or lost). >Doing that, you might detect some sort of pattern to the problem which would >make it easy for the author of autoplayer to track down. > >By the way, who wrote the autoplayer software? Do they still maintain it? It is old software from the Nimzo author and isn't maintained any longer as Rebel is almost the only commercial program which uses it. The Windows autoplayer code is freely available on the Internet. Ed >>Since many of my most loyal customers want to autoplay Rebel after all >>(buggy or not) I left the autoplayer software. Also the autoplayer software >>is useful for a new feature of Century 3 (CAT). >> >> >>>If the bug is in the autoplayer, then it would seem likely that whoever makes >>>the autoplayer is likely to fix the bug IF you can isolate its exact cause. >>>Have you tried pinning down exactly when the corruption of Rebel occurs? Eg. is >>>it always late in the game, or early, or in a particular type of position etc >>>etc. Perhaps it is timing related? >>> >>>Do you know what part of Rebel is being corrupted? If so, is it always the same >>>part? >> >>The problem is that the autoplayer is an external driver driven by >>interrups (an old fashioned DOS TSR program) and I don't have access >>to the source code. The suspect is the TSR program writes outside its >>available memory. When it writes in Rebel itself the program is most >>likely going to crash (which does not happen frequently). If the TSR >>program writes in the hash table of Rebel (more likely) anything can >>happen, from zero damage (likely) to much damage (unlikely). Rubbish >>information in the hash table is not going to crash Rebel. >> >>I have no idea how to trace a TSR program writing outside its own >>memory. Do you? > >No, I have no experience with TSR programs. > >> >>Ed
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.