Re: [Freeciv-Dev] Problems loading saved games with aifill > current number of players.

[David Pfitzner <dwp@xxxxxxxxxxxxxx>]

Anders Wegge Jakobsen wrote:
> 
>  I recently had a game of freeciv, that lasted long enough that I was
> unable to finish it in one session. So I save the game, and wanders
> merrily down the road ...
> 
>  The next day, when I expected to finish of the last two ai's, I
> expect to be able to continue by loading the saved game,

A very reasonable expectation!  (I do this all the time.)

>  but! A new
> AI-player were created at the start of the game, probably because
> there was fewer active players than the aifill value. 

This would seem to be the problem -- aifill should not create new
players on reload, whether or not players have died.  (Though you 
don't mention which version of freeciv you are using.)  One way I 
could see this happening is if you had used the server command
"remove [player]".  In that case I can reproduce the problem.

>  I was able to cure the symptoms, ie. the runaway recursion, by
> providing a way out of find_prerequisites() [See enclosed patch], but
> I have a feeling that somebody should take a look at the poor AI that
> seemingly gets created without means.

Yes, or rather why the new player was created in the first place.
I think putting "if(is_new_game)" in front of the call to 
generate_ai_players() in civserver.c may suffice.

Regards,
-- David