[Freeciv-Dev] Re: (PR#13262) include pubserver jobs v5

["Mike Kaufman" <kaufman@xxxxxxxxxxxxxxxxxxxxxx>]

<URL: http://bugs.freeciv.org/Ticket/Display.html?id=13262 >


On Wed, Jul 20, 2005 at 12:26:16AM -0700, Jason Short wrote:
>
> Either you allow taking, in which case you have to do scoring based on
> the new user not the original user, or you don't allow taking in which
> case scoring is easy but continuing games is hard because you can't
> substitute.  Taking in a savegame isn't much different than taking in a
> running game - even in a running game your player can be "taken over" by
> another user if you happen to get disconnected for a moment (though
> obviously few users would do this).

disallowing taking on savegames is awful. Basically it means that you can't
offer prebuilt scenarios or "classic" games on pubserver, which I think is
one of the primary reasons to even offer game loading.

On Wed, Jul 20, 2005 at 12:39:15AM -0700, Per I. Mathisen wrote:
> 
> <URL: http://bugs.freeciv.org/Ticket/Display.html?id=13262 >
> 
> On Tue, 19 Jul 2005, Mike Kaufman wrote:
> > much better to allow scoring only one time per game instance
> 
> How would this solve scoring for long games, where the server would go
> down (planned or otherwise) during the game? Should we just decree that
> once the server goes down, the scoring ends there? I think we can do
> better than that.

Of course we can do better than that. The savegame stores the game number.
Once that game (via number) is scored, it cannot be scored again. I don't
think that's too hard do you?
 
> > but in fact this will not work, since the user who started the game will
> > be credited with the win not the user who is doing the taking.
> 
> I do not understand this.

For the "new" gamelog system, the user who plays the first 10 turns or so
is the user who is scored, not any other user who joins (or takes) later.
So your scenario of abuse is just not feasible (it would boost the original
user's score though, without a protection which I outlined above).

> > I'd say that this is a poor solution for the problem. The better solution
> > is on the scoring end.
> 
> I do not see a solution on the scoring end. Do you?

The one I mentioned above. The pubserver scripts will do the job, not the
server.

-mike