[freeciv-ai] Re: ai bodyguard code cleanup part#1

["Per I. Mathisen" <per@xxxxxxxxxxx>]

On Sun, 11 Aug 2002, Gregory Berkolaiko wrote:
> > That is one problem. Before mech inf, maybe making smart use of our fast
> > movement (not stopping next to city) is better than bodyguard. But when we
>
> I am not 100% convinced.Depends on the potential danger, especially
> strength of fast units of the enemy.If enemy only has Horsemen, then
> Pikemen is a perfect guard.The final dash to the city will have to be
> made by diplo alone.

The question is how to make this happen with existing code. In
ai_unit_move, we could check if target is in a goto run, and if we have
enough movement to reach the target this turn, allow it to leave the
bodyguard behind?

> One way to model it is to run path-finding to the city for diplo alone and
> for the bodyguard.And see whose route comes out more dangerous.

If we have very fancy path-finding, yes. Right now we'll have to do with
warmaps. Also, you suppose the charge will go to the bodyguard, instead of
vica versa. In reality, I suppose they will meet en route and go a third
route.

> The current code assigns MechInf to be bodyguards for Riflemen.

The bug is that it builds Riflemen and consider them worthy attack units,
not that it assigns them good bodyguards.

Though I suppose we could check if the potential bodyguard has higher
attack power && the charge has no special flags.

Yours
Per