Re: [Bug-gnubg] Unexpected exit - Rollout cube decision

[Christopher D. Yep]

At 08:01 PM 8/17/2003 +0200, Nardy Pillards wrote:

> > >On Sat 16 Aug 2003 (12:20 -0400), Christopher D. Yep wrote:
> > > > Windows XP, GUI version
> > > > Version 0.14-devel 1.1260 030813 (build Aug 13 2003)
> > > >
> > > > Set up any position.  From the Analyse menu, select "Rollout cube
> > > > decision".  Rollout screen will now open.  Immediately click
> > > > "Stop" before gnubg even finishes the first trial.  Click on "Ok".
> > > >  Gnubg will now unexpectedly terminate.
>
> I propose this:
> copy .gnubgautorc to 'whatever' (so you have a backup)
> then run gnubg
> delete the file .gnubgautorc
> and now save the settings with gnubg
> Settings/Save Settings
> (this will create a new .gnubgautorc)
>
> Let know if it works/doesn't work?

I tried the following (reversing your steps 2 and 3, as I assume you meant 
to write):

1. Copy .gnubgautorc to backup
2. Delete .gnubgautorc
3. Run gnubg (since there is no .gnubgautorc it uses default values)
4. Save settings (it will now save the default values to a new .gnubgautorc)
5. Close gnubg (in case it matters, but I doubt it does)
6. Re-open gnubg (new .gnubgautorc is now used)

I then tried (Analyse - Rollout cube decision) on the following 
position.  This unusual position (X's checkers are all in O's inner board 
and vice-versa) was chosen so that I could stop the rollout quickly enough 
based on the default rollout settings (0-ply truncated at 11-ply)!

 GNU Backgammon  Position ID: AADc3TYAANzdNg
                 Match ID   : cAkAAAAAAAAA
 +13-14-15-16-17-18------19-20-21-22-23-24-+     O: gnubg
 |                  |   | X  X  X  X  X  X |     0 points
 |                  |   | X  X  X  X  X  X |
 |                  |   | X  X  X          |
 |                  |   |                  |
 |                  |   |                  |
v|                  |BAR|                  |     (Cube: 1)
 |                  |   |                  |
 |                  |   |                  |
 |                  |   | O  O  O          |
 |                  |   | O  O  O  O  O  O |     On roll
 |                  |   | O  O  O  O  O  O |     0 points
 +12-11-10--9--8--7-------6--5--4--3--2--1-+     X: user

The bug is still present.  Note: the bug only occurs with (Analyse - 
Rollout cube decision).  If I instead do (Analyse - Rollout) and stop 
before the first trial is completed, gnubg does not unexpectedly terminate.

If this bug isn't reproducible on other machines, I won't worry about it 
too much, since it will rarely affect me.  I can imagine myself 
occasionally stopping a rollout after 0 trials (if I set-up the position 
incorrectly, or accidentally chose a 6-ply rollout instead of a 2-ply 
rollout, etc.), but it shouldn't happen too frequently.

Chris