[Freeciv-Dev] Re: BUG: Goto is executed too early

[zamar-fc@xxxxxxxxxxxxxxxxxxx (Joona Kiiski)]

>
>On Fri, Feb 23, 2001 at 02:08:48PM +0200, Joona Kiiski wrote:
>> 
>> I don't know if this a bug:
>> 
>> You attack on enemy city and destroy the last defending unit.
>> Your enemy buys new unit in the city to defend at next turn.
>> 
>> But if you now use goto-command (target empty city),
>> you occupy city before new unit is
>> created. I think that is a bug!
>
>No, it is a feature.  Goto is executed at the server side, before
>interaction from the clients is processed.  It is processed player
>by player, with the players in a random order.

Of course it's executed before interaction from the clients is
processed, but why can't we first execute city_build_stuff and
then execute goto commands and after that wait for client commands.

>>> The order should be:
>> First new stuff is built and then players (or goto functions) move units 
>
>So you think there should be a round of city movement for all players
>first, then a full round of unit management?

Absolutely, I was very suprised when this strategy was used against me
in multiplayer game.

>
>> This is easy to reproduce.
>
>Definitely, it is the same as the 'my opponents perform two moves at
>once' "bug".  It is a feature because it makes goto more powerful to
>help neutralize differences in hetwork latency.

IMO the current behaviour doesn't
neutralize differences in network latency.

And It's not a feature, it is a bug.
Of course goto must be executed before
commands given by players are executed, but _not_ before
new units are built in cities!

I'm quite sure that most people who have played a lot Civ2 (like me)
suppose that new units are built in the beginning of the turn and _after_
that units are moved (either by players or by goto).

--
Zamar