Thread: reborn and jv-bots ?

I'm also interested how current bots (SP) work and why they can't be used for MP. Or what are their limits. There's a lot of code in the engine we could try to use from the scripts if I knew something more. I never looked into that.

Yes, with traces. One would do it beforehand since this'll only work for navigational nodes, the rest is up to the modder.

A node is just an entity of which the most important property is their location. Thus nodes can be dynamically added during runtime by the pathfinding algorithm,
if necessary. Here we can choose whether they stay or are disposed of after use. In the latter case, we don't even need the entity, just a vector.
In the other case I think it makes more sense to introduce a cooldown. If they aren't used after X amount of time, they are disposed of. But here they may be
node entities with their own set of data.

I didn't know quake3/MoH had such a system. Damn.. why do I even bother thinking up optimization for pathnodes
I wonder, if this was possible, why singleplayer uses pathnodes instead...

Originally Posted by Razo[R]apiD

I'm also interested how current bots (SP) work and why they can't be used for MP. Or what are their limits. There's a lot of code in the engine we could try to use from the scripts if I knew something more. I never looked into that.

I can't really say for sure, I glanced over the q3 source code before but I think this has been significantly modified in MoHAA. I mean, the original Q3 bots were designed for multiplayer.
MoHAA works with this "leash" business (used to keep bots from venturing too far from the 'origin') and the scripting commands available for actors also suggest some dramatic modifications.

If you have it reversed (but I assume not) I could take a look at it. Doing the logic in C definitely has its benefits ...too bad I'm not C-worthy yet.

But can we load bots into MP and make them move, animate and so on? Or do you need to emulate this? If we can, then the only problem is to make them roam properly in the environement and make "intelligent" decisions.

How does the lean portion of the script work / what's the easiest way to make bots lean? I'm trying to understand it, but I don't get it. I've made simple changes in JV Bot (using cvars instead of hard coding task priority), and I'm liking what I'm seeing. I'm trying to grab what's in the proof of concept to make bots lean and jump. Assuming I won't be using the variables from the proof of concept, how would I make bots lean and jump using just self?

For leaning, I tried to add this, but again, I'm missing variables such as level.bots. and self.cueued_upperanim --- I can get the values for level.bots.x but I don't get this cueued_upperanim thing... I can't find anywhere in the script that actually does the leaning in the POC.

Code:

// randomly sets lean
lean:

  // Took out these conditions
  // && self.position && self.position != "prone" && self.position != "crouch"
  if(self.task && self.task != "defendbomb"){
  	self.leandir = 0
  	
  	while(isAlive self)
  	{
  		self.leandir = randomint 3 - 1
      
      		self.key_lean = self.leandir
      		local.lean = self.key_lean
		local.target_lean_angle = local.lean * level.bots.lean_angle
		local.delta = local.target_lean_angle - self.lean_angle
		self.lean_angle += local.delta * level.bots.lean_speed * 0.05
      
      		if(abs self.pitch < 5.0 && abs self.lean_angle < 5.0)
	        {
			 // this anim does play
			 self setupperanim self.cueued_upperanim
			 self waittill upperanimdone
 	        }
      
  		wait 0.2
  		wait (randomfloat 1.0 * (1.0 - level.bots.skill))
  	}
  }

end

Is there an easy way to make a bot jump randomly (normal height)?

Great to see and read so much interest and effort into the bots - If you need test / Dev servers give me a shout never thought would see an advancement on this side of MOHAA

I'm trying, but I need help making them lean / jump... hoping someone understands how to make this work.