[Freeciv-Dev] Re: buglet in normalization in init_new_game() (PR#1063)

[Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>]

On Mon, Nov 12, 2001 at 01:09:39PM -0800, jdorje@xxxxxxxxxxxxxxxxxxxxx wrote:
> This causes the assertion in CHECK_MAP_POS to fail on many games.
> 
> civserver: maphand.c:742: show_area: Assertion
> `is_normal_map_pos((x),(y))' failed.
> 
> The backtrace shows:
> 
> (gdb) bt
> #0  0x400888d1 in __kill () from /lib/libc.so.6
> #1  0x4008864d in raise (sig=6) at ../sysdeps/posix/raise.c:27
> #2  0x40089cb8 in abort () at ../sysdeps/generic/abort.c:88
> #3  0x40081d71 in __assert_fail (assertion=0x80cbd93
> "is_normal_map_pos((x),(y))", file=0x80cbd59 "maphand.c", line=742, 
>     function=0x80cbf53 "show_area") at assert.c:74
> #4  0x08072bd8 in show_area (pplayer=0x81313d0, x=-1, y=12, len=1) at
> maphand.c:742
> #5  0x080684ea in init_new_game () at gamehand.c:136
> #6  0x0804f5f0 in srv_main () at srv_main.c:1964
> #7  0x08049cdd in main (argc=1, argv=0xbffff954) at civserver.c:147
> #8  0x40076306 in __libc_start_main (main=0x8049794 <main>, argc=1,
> ubp_av=0xbffff954, init=0x804910c <_init>, fini=0x80b9e40 <_fini>, 
>     rtld_fini=0x4000d2cc <_dl_fini>, stack_end=0xbffff94c) at
> ../sysdeps/generic/libc-start.c:129
> (gdb) p game.randseed
> $2 = 1005598569
> (gdb) p map.seed
> $3 = 157894863
> 
> This seems straightforward enough.  Unfortunately, I'm not too familiar
> with this part of the code, so I don't know how best to fix it.  The
> easiest way may be to just put a normalize_map_pos into the beginning of
> show_area; however, this is wasteful because show_area is used in a lot
> more places other than this problem area.  It also may be possible for
> the center of the square_iterate to be unreal, which would mean the
> CHECK_MAP_POS check doesn't belong there at all.  However, I can't think
> of any reason why one would want this.
> 
> It's certainly not a dangerous bug, since square_iterate calls
> normalize_map_pos itself so it'll always give only normal positions.  To
> avoid it, compile without debugging.  There should be no side effects to
> the code in init_new_game that produces the non-normal coordinates.

IMHO it looks like a job for a new circle_iterate.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  Microsoft does have a year 2000 problem. I'm part of it. I'm running Linux.