[Freeciv-Dev] Re: (PR#8791) crash with civ1 ruleset

["Gregory Berkolaiko" <Gregory.Berkolaiko@xxxxxxxxxxxxx>]

<URL: http://rt.freeciv.org/Ticket/Display.html?id=8791 >

On Fri, 21 May 2004, Raimar Falke wrote:

> <URL: http://rt.freeciv.org/Ticket/Display.html?id=8791 >
> 
> On Fri, May 21, 2004 at 06:12:33PM -0700, Jason Short wrote:
> > 
> > <URL: http://rt.freeciv.org/Ticket/Display.html?id=8791 >
> > 
> > Gregory Berkolaiko wrote:
> > 
> > > I am not too sure what shall we do about it.  multiplication by 10 is for
> > > veteran bonuses not to get lost.  Later division by 30 is to keep numbers
> > > sane for larger units in default/civ2 rulesets.
> > 
> > What about either:
> > 
> > - Making it use floating point values?
> > 
> > or
> > 
> > - Calculating the POWER_DIVIDER when we load the ruleset?
> 
> What about this patch? It works for both RCc you posted.

I am applying the attached patch, which is a modification of Raimars 
patch.  I added a comment and removed the assert.  In fact if only attack 
= 0, kill_desire returns sane values (negative desire to attack).  And if 
both attack and vuln are zero, you get SIGFPE, so why bother checking for 
it?

Index: ai/advmilitary.c
===================================================================
RCS file: /home/freeciv/CVS/freeciv/ai/advmilitary.c,v
retrieving revision 1.165
diff -u -u -r1.165 advmilitary.c
--- ai/advmilitary.c    2 May 2004 12:31:05 -0000       1.165
+++ ai/advmilitary.c    22 May 2004 06:15:58 -0000
@@ -798,6 +798,12 @@
       
       /* Take into account reinforcements strength */
       if (acity) attack += acity->ai.attack;
+
+      if (attack == 0) {
+       /* Yes, it can happen that a military unit has attack=1,
+        * for example militia with HP=1 (in civ1 ruleset). */ 
+       continue;
+      }
       
       attack *= attack;