Zugg Software :: View topic

Apprentice Joined: 26 Sep 2001 Posts: 124

Recently on the mud I play, a few new areas have been introduced into the game, and ive mapped them.

Ive only ever used one main map for the entire mud. There are now 20 000 or so rooms on that map. (Actually a lot less, because chunks were deleted, and later chuncks were added)

But now, after ive mapped the new areas, im having trouble with some mapper commands. Like #walk for example.
It keeps throwing up stack overflow errors, even if im trying to walk from somewhere very near.

Is there a room limit to the zmud automapper? (im using zmud 6.16)
If there is, is there anyway i can get around this problem short of splitting the entire thing up into zones?

Thanks

That which does not kill us, makes us stronger.

SubAdmin Joined: 18 Nov 2001 Posts: 5182

As far as I am aware, there is no limit on the number of rooms. There should be a technical limit due to the programming structure of about 2 billion. You can try following the directions in the migration guide found in the support section to see if that will correct any corruption that may have occured.

Wanderer Joined: 25 Jan 2002 Posts: 77 Location: Ireland

not sure about this - my map isnt that big - how the #walk algorithm is implemented, but no matter how good it is, it is still the old 'travelling salesman' problem. Beyond a certain size, any computer will not be able to cope with it. May be that simple, may not - stack overflow sounds like it could be though.

"I've got an allergy to Perrier, daylight and responsibility" - Marillion

Check out Moral Decay on 131.247.113.1 3003

Apprentice Joined: 26 Sep 2001 Posts: 124

No, im sure it's not the travelling salesman problem, because the system seems to only crash if the #walk command, (or even just the %walk function), are used to travel between rooms very close together.

It's very peculiar and has me at a complete loss. It's always the same room that crashes. It, and it wont crash from every location.

For some reason it crashes from the room directly north of it. But not the room directly south west of it.

Even when i remove the link between these two rooms, it still crashes.

Yet if i try the exact same thing from one more room away it works. I have no idea what could be wrong, so ive copied the Stack values the error message brings up, apon crashing, in the hope that someone might be able to see some meaning in it.

---------------------------------------
ZMUD caused an invalid page fault in
module ZMUD.EXE at 0167:00517e74.
Registers:
EAX=02a3c524 CS=0167 EIP=00517e74 EFLGS=00010213
EBX=02a3cccc SS=016f ESP=00a20000 EBP=00a2001c
ECX=0000011b DS=016f ESI=00000008 FS=31ef
EDX=02a3cccc ES=016f EDI=00000001 GS=0000
Bytes at CS:EIP:
57 89 4d f8 89 55 fc 8b d8 33 c0 89 45 f0 8b 55
Stack dump:
00000008 02a3cccc 00000000 00000000 00000000 02a3f288 00000008 00a20048 00517fd3 00b1ea1c 00000001 00000005 02a3cd58 00000000 00000001 01000000
-------------------------------------------

Im running Windows 98 Se OS, if that helps.

Thanks

That which does not kill us, makes us stronger.

GURU Joined: 10 Oct 2000 Posts: 4379 Location: USA

What you describe keeps sounding more and more like map corruption. Did you try Vijilante's solution?

Kjata

SubAdmin Joined: 18 Nov 2001 Posts: 5182

If memory serves, the addresses indicated there are outside the actual code of zMud. This could be cause by corrupted settings, corrupted map (most likely), viral infection of zMud (least likely), misbehaving system drivers, bug that wasn't caught during beta testing.
Think about what you did shortly before first noticing the problem, if you haven't installed or updated any software or drivers then you can eliminate that one.
Try another 2 room set that is as similar as possible to the problem room and use #WALK to see if the same crash occurs. If it does then try a few more, if they are all crashing then corrupted settings is more likely then the corrupted map, and the bug possibility stays in. If no other set crashed then the bug possbility is completely eliminated, and corrupted map is pretty much the only one left worth considering.
Link for the migrtion guide http://www.zuggsoft.com/zmud/migrating.htm

Apprentice Joined: 26 Sep 2001 Posts: 124

I have indeed followed the Migration guide, it appears that both the settings and the map files are not in any way corrupt.

Keeping this in mind, I decided to test if other places of the map still worked with the speed walks. I found that they did not. Places where I used the speedwalks quite regularly no longer function.

So I made a back up and deleted all the new rooms which I have added (an entire city actually) and then I tested the old areas again. This time they work fine, just as they did before.

So, this leads me to belive one of two things. There is infact a room limit where commands like #walk or %walk will not work in the automapper, OR it may be, that my PC just cannot handle the #walk algorithm with that many rooms on the map.

Another thing which seems strange to me is that, the new city which was added was in no way connected to the old areas on the same map, where the #walk was tested. But still those rooms seem to have had an effect on the way that the #walk command executes.

Does anyone know the General implementation of the #walk command? That might help in figuring out what went wrong.

That which does not kill us, makes us stronger.

GURU Joined: 10 Oct 2000 Posts: 4379 Location: USA

You might want to try e-mailing Zugg with this problem. As far as I know, there is no limit to the #WALK command. I have heard of people with 40K+ and 60K+ rooms that use #WALK with no problems. I myself have around 10K rooms and don't have any problems.

Kjata

GURU Joined: 15 Jun 2001 Posts: 654 Location: USA

quote:

not sure about this - my map isnt that big - how the #walk algorithm is implemented, but no matter how good it is, it is still the old 'travelling salesman' problem. Beyond a certain size, any computer will not be able to cope with it. May be that simple, may not - stack overflow sounds like it could be though.

This is not the traveling salesman problem. The traveling salesman problem deals with finding a the shortest circuit among all rooms and is NP-complete. This is not. For more information about the traveling salesman problem, go here.

In fact, the shortest path problem can be solved in polynomial time. See here for information and implementations.

- Charbal

Wanderer Joined: 25 Jan 2002 Posts: 77 Location: Ireland

Very true. Remind me not to answer questions at almost 2 a.m. :)

"I've got an allergy to Perrier, daylight and responsibility" - Marillion

Check out Moral Decay on 131.247.113.1 3003