[freeciv-ai] Re: Patch Army

[Jordi Negrevernis i Font <jorneg@xxxxxxxxxxx>]

   All the objections are done (i think) and it's against cvs Jan 25.

   Still it does stupid thinks, but ...

   Mainly, the coordination with the naval units is very poor, and i 
can't make army advisor function to produce a valid choice struct even i 
tryed it hard. I don't know what happens inside 'process_attacker_want'.
En/na Per I. Mathisen ha escrit:

> On Thu, 9 Jan 2003, Jordi Negrevernis i Font wrote:
>  
>
> >  Well, i've been developing a patch for giving to the ai the capacity
> > to make coordinate mass attacks against a city. It does handle a total
> > of 8 attacking units, 4 defensive units and 4 naval units.
> >    
> Looks interesting. I need to give it some testing, but the patch doesn't
> apply anymore, so could you fix Raahul objections and repost the patch?
>
> Before you do, please
>  - fix style issues (lots of these)
>  - use ai_unit_goto() and ai_unit_gothere() instead of do_unit_goto()
>  - use handle_unit_activity_request() instead of set_unit_activity()
>  - use ai_unit_new_role() instead of punit->ai.ai_role = ...
>  - fix your comments, some of them are totally wrong and look like
>    they've been cut&pasted from somewhere else
>
>  
>
> >  But, i have 3 questions:
> >
> >  - How do i know if there is a city in the attack direction?
> >    
> I don't understand the question.
>
>  
>
> >  - Which are the wants for the units? While debugging i get want >
> > 101 many times, before i evaluate army_advisor_choice!! I knew they were
> > between 0 and 100!!
> >    
> You knew wrong, as did Raahul. A lot of wants are over 101, the want for
> attacking units are among them.
>
>  
>
> >  If that thing is likely to go to CVS i can continue to develop it.
> >    
> I can't say at this point, I'd have to look more closely at it while
> testing it.
>
>  - Per
>
>
>
>

army4.diff.gz
Description: application/compressed