[Freeciv-Dev] Re: core file on civserver, http://civserver.freeciv.org/games/44166

[Mike Kaufman <kaufman@xxxxxxxxxxxxxxxxxxxxxx>]

On Fri, Oct 18, 2002 at 12:42:15PM -0500, Jason Short wrote:
> Per I. Mathisen wrote:
> > On Thu, 17 Oct 2002 freeciv@xxxxxxxxxxxxxxxxxxxxx wrote:
> > 
> >>#20x4202a8c2 in abort () from /lib/i686/libc.so.6
> >>#30x0806c76e in create_start_positions () at mapgen.c:1164
> >>#40x0804e765 in srv_main () at srv_main.c:1943
> >>#50x080499b1 in main (argc=10, argv=0xbffffc34) at civserver.c:153
> >>#60x42017589 in __libc_start_main () from /lib/i686/libc.so.6
> > 
> > 
> > I haven't seen this one before. Jason, is this the same bug you reported a
> > while ago?
> 
> This bug is because starting placement of the players failed:
> 
>      if (counter > MAXTRIES) {
>        freelog(LOG_FATAL,
>             "The server appears to have gotten into an infinite loop "
>             "in the allocation of starting positions, and will abort.\n"
>             "Please report this bug at " WEBSITE_URL);
>        abort();
>      }
> 
> I'm not sure that there is anything we can do about this (even with a 
> full rewrite of mapgen, it is always conceivable that the map will be 
> too small for the number of players).  But getting cores about it all 
> the time isn't so useful either.
> 
> Perhaps the mapgen rewrite will addreess it.

a suggestion:

if (counter > MAXTRIES) {
  map.seed++;
  generate_map_again();
  goto beginning of create_start_positions();
}

-mike