Re: [glob2-devel] I'm confused by the code in NewNicowar::control_attacks

[Joe Wells]

"Bradley Arsenault" <address@hidden> writes:

> On 8/10/07, Joe Wells <address@hidden> wrote:
>
>     "Bradley Arsenault" <address@hidden> writes:
>
>     gdb says b is NULL after this command, so get_building is clearly
>     returning NULL.
>
>
> Yes, its being given an invalid input somehow.
>
>     First, am I correct in understanding that b is actually supposed to be
>     an attack flag (not a real building)?
>
> b is supposed to be an attack flag, because the I that is used to 
> get_buildings
> comes from the vector attack_flags.

The i that is used is an _index_ of the vector attack_flags, not
something in one of the vector's slots.  The code looks like this:

  for(int i=0; i<attack_flags.size(); ++i)
  {
    Building* b = echo.get_building_register().get_building(i);
    ...
  }

I am guessing from what you say that it should look like this instead:

  for(int i=0; i<attack_flags.size(); ++i)
  {
    Building* b = echo.get_building_register().get_building(attack_flags[i]);
    ...
  }

Am I right?

>     Is it always the case that the building register contains only attack
>     flags?  Otherwise why should entries 0, 1, 2, etc., be guaranteed to
>     be only attack flags and not also other buildings that are being
>     managed?
>
> The difference between Echo and the actual game is that every building is 
> given
> a unique ID for the entire game. Rather than the GID's, which can be re-used,
> the echo ID's are only used once for a building. I store those ID's in the
> vector attack_flags and retrieve them in this function. Somehow, the vector
> contains an ID for a flag that has already been destroyed. 0, 1, 2, and
> probably up to 20 are most likely normal buildings cause thats what nicowar
> constructs first, somewhere in there would be a couple exploration flags. A
> flag is a building the game engines perspective, except in a few special
> circumstances.

The reason why I asked about 0, 1, 2, etc., is that these are the
values this loop actually uses.  See comments above.

The crash is with i == 0.

Yes, my analysis must be right.  What is happening is that it crashes
when I start killing the buildings with small echo numbers.  That's
because it's doing get_building(i) instead of
get_building(attack_flags[i]).

>     So b->locked[true] indicates whether the attack flag is accessible by
>     land *or* water, while b->locked[false] indicates whether it is
>     accessible by land only?  Right?
>
> Yes.
>  
>     For what it is worth, this happened when I had wiped out about 1/3 of
>     Nicowar's base area.  Would that cause Nicowar to delete one of its
>     attack flags in a strange way?

This is the cause.  I killed a building with a small echo number like
0 or 1.

> Hmm, it might have something to do with it, although doubtful. It might have
> removed the flag with that is_locked override right there since i've never
> actually seen nicowar do it in action (despite much hard testing)

-- 
Joe