[General] Finding World Object Base & Size Offsets

[Menlaben]

Hello all,

My question is about how to find (in memory) the list of pointers to the objects surrounding your character in an MMO. I'm not sure exactly what this is called, so my searches turned up nothing.

I'm assuming that MMOs store a list of pointers to these world objects in 4-byte offsets one after another (or in some other organization).

In designing a Radar, I would need to cycle through these pointers to access the object structures.

How do I find (general approach) where such a pointer list starts and where the indication of how large the list is at a given point? In studying other open source radars, they only include the hardcoded offset for these pointers, which is no use to me if I'm trying to do something from scratch.

Use as technical a language as you please. I've very well versed with Olly/IDA and Cheat Engine/T-Search, but I keep getting taken around in circles.

One thing I did notice is that each object structure (NPCs/HarvestNodes) all start with the same value in the first 4-byte offset, but this leads me nowhere.

Thanks all!

[miceiken]

I'm not sure exactly what you are looking for, but some objects are static and is a part of the collision/terrain files, but some objects are accessible from the objectmanager, they are named game and dynamic objects.

[Menlaben]

I'm looking for whatever list of pointers that a typical Radar program would cycle through to know where all the world objects are. By World Objects, I mean NPCs, Harvest Nodes, Ground Spawns, or any interactable object. I don't know what else to call it.

It seems to me that most MMOs would keep a list in the Client's memory of pointers to each and all objects within a certain distance of the player. It is this list that I'm trying to locate, so that my Radar can cycle through and locate the structure of each object near the player.

I believe your description of the "ObjectManager" fits best.

Thank you.

[!@^^@!]

Search the forum for ObjectManager and you'll find all the info you need

[Menlaben]

Wow, that's perfect! Thank you!

For anyone else looking at this thread, I've deteremined this to answer my question:
[Guide-kind of] How I handle objects.

[eLaps]

http://www.mmowned.com/forums/world-...ml#post1650322

int32_t (*EnumVisibleObjects)(uint32_t (*)(uint64_t, uint32_t), uint32_t) = (int32_t (*)(uint32_t (*)(uint64_t, uint32_t), uint32_t))0x004D4B30;
uint32_t (*ClntObjMgrObjectPtr)(uint64_t, uint32_t) = (uint32_t (*)(uint64_t, uint32_t))0x004D4DB0;
uint32_t (*ClntObjMgrGetActivePlayerObj)() = (uint32_t (*)())0x4038F0;

bool ObjectMgr::update() {
	m_objects.clear();
	m_me = ClntObjMgrGetActivePlayerObj();
	if (m_me)
		EnumVisibleObjects(populate_callback, 0);
	
	return m_me;
}

uint32_t ObjectMgr::populate_callback(uint64_t _guid, uint32_t _filter) {
	m_objects.push_back(ClntObjMgrObjectPtr(_guid, -1));
	return 1;
}

Yes pretty simple.

[Menlaben]

Thanks for all the replies. My weakness was in understanding this particular structure of Online 3D games, and you've helped me understand it.

I've completed the object dumping section of my program and can begin training the Neural Net which I've designed for Runes of Magic. Perhaps I'll post a youtube video of how well a Neural Network can play an MMO It won't be near the capacity of a rule-based AI, but this is just for fun

Take Care.

[Apoc]

Thanks for all the replies. My weakness was in understanding this particular structure of Online 3D games, and you've helped me understand it.

I've completed the object dumping section of my program and can begin training the Neural Net which I've designed for Runes of Magic. Perhaps I'll post a youtube video of how well a Neural Network can play an MMO It won't be near the capacity of a rule-based AI, but this is just for fun

Take Care.

Neural nets are way overboard for a botting platform. Especially when you effectively have only 1 agent.

Neural nets are more for 'giant' collections of agents (say, RTSs, or massive opposing forces). You're going way overboard. A simple GOAP engine would do you far better than a NN.

[Menlaben]

Right! This is just for fun and personal knowledge. I'm interested in NN-based AI to mimic human behaviour and not in whatever it is a traditional bot gains you in a game.

And so, I'm not looking at what's already been done. The training would take place as I play the game, attempting to play as normally as possible and going through as many important facets during this period. An ART-based NN and a Genetic Algorithm will take care of decisions. Rules will be used in things like skill/talent/attribute upgrades and a few other things (humans have automatic responses where no cognitive decision is made).

I have a few goals I'd like for the bot to accomplish, but I'll keep those secret.

Anyways, we'll see how it goes I'll FRAPS it when I'm happy with it.

Take care.