Re: [Bug-gnubg] Resignation wrongly accepted by gnubg

[Jon Kinsey]

Petr Kadlec wrote:
> > It seems that my particular settings (I've attached the settings file), 
> > seem to work.  If I remove them then I get the problem too.  This is run 
> > on one of my work machines so the settings are pretty random...
>
>
> Well, if I try to use those settings, I get a crash after "resign 1" 
> (instruction at 004059C7 trying to read via NULL). A peek at the .gnubgautorc reveals an 
> interesting command:
> set variation 1-chequer-hypergammon
> (Note: I do not have HG database here.)
>
> After removing the line, everything works fine, and...
> gnubg accepts and wins a single game.
>
> Many questions come to my mind:
> 1. Gnubg should not crash, even without HG files.
> 2. Why do you have that command there?
> 3. How come it works for you?

Ah - I was looking at a hypergammon problem on this machine...  I did 
say the settings were quite random!
If I remove that line then gnubg wrongly accepts the game.

As to the other questions, I don't think hypergammon is used much so 
lots of the code is probably doing things wrong when that variation is 
set.  I guess the following should happen/be fixed:

1. It shouldn't crash if you haven't got the correct databases, although 
the gui doesn't let you pick that option.
2. It shouldn't let you enter positions with the wrong number of chequers.

Jon

signature.asc
Description: OpenPGP digital signature