[Freeciv-Dev] Re: (PR#2223) cma: server knowledge of island != client knowledge

["Raimar Falke via RT" <rt@xxxxxxxxxxxxxx>]

On Thu, Jan 09, 2003 at 02:22:35AM -0800, Jason Short via RT wrote:
> On Thu, 2003-01-09 at 04:17, Raimar Falke via RT wrote:
> > On Wed, Jan 08, 2003 at 09:47:15PM -0800, Anthony J. Stuckey via RT wrote:
> > > On Wed, Jan 08, 2003 at 10:44:47AM -0800, Raimar Falke via RT wrote:
> > > > 
> > > > [rfalke - Wed Oct 23 07:43:28 2002]:
> > > > 
> > > > The recently committed solution doesn't fix this issue. The reason
> > > > is that trade_between_cities uses the map knowledge of two players.
> > > > This is bad since the client only has this knowledge for itself.
> > > > The attached patch fixes this and also adds a comment.
> > > > With the patch this issue is resolved and you get an output of:
> > > >   2: Harstad (19) and Tromsø (9) -> 8
> > > >   2: Tromsø (9) and Harstad (19) -> 4
> > > > indicating that Harstad gets 8 trade from this route and Tromsø 
> > > > only 4.
> > > 
> > >   Yikes!
> > > 
> > >   How can this be a good thing?!?!?!?
> > >   Trade should be the same value at both ends.
> 
> Definitely!
> 
> > Mike and the others decided over xmas that effects and traderoutes
> > should be based on player knowledge. While the committed code didn't
> > show the above different results for the two cities it also didn't fix
> > the issue which lead to this discussion: the above 2223. I fixed this
> > under the condition of player based knowledge. If you don't like this
> > you have to be on Davide side who preferred a solution where not the
> > player knowledge but the server knowledge (which is constant for all
> > players) decides.
> 
> Mike's explanation of the islands patch is that it prevents most
> cheating from extra knowledge by the player.  This is reasonable, and
> perhaps worth the extra overhead it takes.  But to change the rules to
> hack our way around a client-side-ai bug is IMO very bad.  It is better
> (and almost as easy) to have the server send the client all the
> information it needs to correctly do its calculations.  In this case
> this means the server should tell the client whether the trade route is
> same-island or not.

This would get us back to square one: the server tells the client if a
traderoute is on the same island or not. And if the server calculates
this information based on the knowledge of the player or just take it
from the "true" data is only a very small detail then.

Please note that we have to same situation with the effects. There may
be an effect which affects all players of an island. Or doesn't such
effect exists.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 checking for the vaidity of the Maxwell laws on this machine... ok
 checking if e=mc^2... ok
 checking if we can safely swap on /dev/fd0... yes
    -- kvirc 2.0.0's configure