#645 Breeders are explosive
The limits you are placing on breeders is not working.
They filled
the lawful quest, before I managed to work my way over
to the left
side of the screen. My character had ESP, and was
skilled at detect
monsters, so most monsters were detected all the time.
I assumed that this was a feature of the lawful quest,
until I found
that they had overflowed in large numbers to the branch
level of the
main dungeon, and then proceeded to grow from there.
I suspect that part of the problem is that for all
other monster actions it
doesn't really matter if the think they are close, when
they are not.
Attempting to hunt down the breeders is complicated by
the fact that
if there is not room for the offspring adjacent to the
parent, then it
is placed as close as possible, often behind the
player, even if the
player is in a doorway. Eventually they will fill the
entire level as
they are placed further away from the parents.
In looking at the code, I can find no place where
(struct monst).mux
and (struct monst).muy is initialized, if the player's
position is not
known to the monster. This means that they are using
the values of
the last allocation of the struct. (i.e. garbage from
before the call to
newmonst()). The other place where this runs into
trouble is when the
character moves out of range. This causes the breeder
to believe that
the character is nearby, and continue to breed. As the
breeder can't
move, due to its neighbors, it just continues to breed.
I would suggest that you change your breeder goals to
allow breeding
when not near the player, but restrict it to be less
explosive in
other ways.
- All children must be placed next to their parent(s).
- Only those animals which are not too crowded may breed.
- For example, only a pair with no other neighbors
breeds.
- Children occupy all open squares which are
adjacent to both
parents.
- This would allow bugs to loosely fill a room, but
they would not
be able to breed in corridors, except at
intersections and
turns.
- Breeding only happens once animals reach a certain
age, and with a
limited frequency.
Logged In: YES
user_id=64456
If you could specify which version of Slash'EM was giving
you these problems and the type of monster(s) involved that
would allow some investigation to be done.
By the way, all monsters created with newmonst() are
initialised (all fields, including mux and muy) within a
couple of lines of the call to newmonst(). In the case of
clone_mon() this means the parent's values of mux and muy
are inherited.
Logged In: NO
Sorry about the missing version. It is 0.0.7E6F2. I'm on
Windows in case it matters.
Both giant fleas and giant louses were present initially.
The fleas grew to encompass the entire quest level, then
later the branch level.
I think I have the sources which go with this version, but I
can't be sure, as I don't have a compiler. However the
patchlevel file indicates that I have the correct version.
The place I suspect, the call to newmonst() is at line 1237
of makemon.c. Neither mux nor muy appears textually in the
file.
The only occurances I can find appear to either be for pets
(in dog.c), or used to inform the monster it has found the
player. I can't see that they ever lose the player, they
just have their last best guess.
Logged In: YES
user_id=64456
Thanks for the extra details. I'll do some simulations in
due course.
If you look at how newmonst() is called from clone_mon() you
will see the following:
m2 = newmonst(0);
*m2 = *mon; /* copy condition of
old monster */
The second line sets all the fields of the new monster
(including mux and muy) to the values that the first monster
has. The following lines then sets specific fields for which
the old monster's values would not be appropriate.
Logged In: YES
user_id=1048205
I guess I'm too old. When I learned C, you couldn't count
on struct assignment if you wanted portability, so I didn't
look for it.
So much for that guess, maybe my other guess is relevent,
who knows. Both times I did the Lawful quest, I ran into
the same problem. I suspect it relates to the proximity
created by the maze, without an ability to actually reach to
kill the bugs.
Logged In: NO
A friend of mine also ran across this bug. In his case, it
happened after polymorphing a vampire into a giant flea in
the crypt version of Mine's End. The vampire turned into an
invisible giant flea, and soon the entire level was full of
invisible giant fleas (incidentally, should they really all have
been invisible since it was probably a cloak of invisibility
which made the vampire invisible in the first place?), which
quickly wiped out his character.
We later stumbled across it as a bones level while playing
together with his laptop connected to a TV so we could both
see. Even when his square was Elberethed with a wand of
fire and the fleas couldn't attack him they continued to
multiply as fast as they were killed. Eventually we waded our
way to the stairs, but (possibly because they were originally
a vampire) the adjacent ones actually followed him up the
stairs and filled the level above it, finally ripping his poor
character to shreds.
Basically, there are two points to this. First, could the fact
that the giant flea was something else polymorphed have an
effect? And second, should the giant flea and its offspring
that result really keep the invisibility and ability to follow
characters upstairs of the original creature? I can
understand that the first one might, but unless my friend
accidentally fathered an unstoppable race of intelligent giant
fleas to spread across the world it seems odd that the
offspring would inherit the intrinsics of their parents rather
than being generated as the default.
Eric "Critic of the Dawn"
Logged In: YES
user_id=64456
Finally found the problem. Although the comments said that
breeders could only multiply if they were next to you they
could actually multiply if they were within 6 squares of
you. If you are in anything like an open space that can be a
lot of breeders each cloning themselves every turn (or
faster if you're slower than them).