Greetings and a few questions

General Discussion
1

Hello Panda community! Decided to jump head first into coding a project with Panda along with some other helpers (no one is a native python programmer). Also, I am not a programmer by trade but I built the working model of what our backend processes will be in our game so now we’re on to the rendering side of things.

Have a few questions (not looking for code snippets just want to know where to look and if it is possible). Enjoy learning the code but I do so from basically taking apart working demos and/or trial and error. The following is what I couldn’t find answers for directly.

[1] Is it possible to stop the clients from rendering actors outside of say range 30 (modified by scale)? If so where do I look for that? I want to keep the scene intact just anything that is an actor to not render it until it’s close enough. I want to assume that is shaders but from previous game playing (not developing) experience I took that to mean something else entirely.

[2] Where’s the stability issues with python thread classes and Panda if my backend processes never touch a panda class and those threads never call in any panda modules? Is there a reason to change to panda’s threading classes?

[3] I know distributed nodes were mainly built for servers but is it a better alternative to pickling data for actors (x y z, hpr and animations)? Our game is intended to be networked and hosted on one of the player’s machines. We managed it well pre-rendering phases (not a big accomplishment to some but we were happy with the results). Seems like if we can get a handle on the nodes the builtin sync functions would be perfect for us.

[4] Are using sprites or putting up with wxPython imported into Panda the only ways to create pixel perfect GUIs? Have to say this is the only frustration that I ran into so far. Everything else has been downright enjoyable. My XML/HTML brain just doesn’t cope well with L,R,B,T and then X,Y,Z for positioning then trying to fill something based on a scale for the image.

[5] Haven’t played with PandaAI but the demos that I have seen show flat planes being used. This does work for terrains where Z changes right? Hate to waste time building a new map in 3ds max with the ai points and then find out it was all for naught.

That’s it for now.

Wanted to thank everyone who posted up code snippets! Really helped me out in a lot of ways to reverse engineer things. Will contribute when I can but still in the learning process and my code is chock full of inefficiencies. Gotta start somewhere though, right?

2

Hi, welcome!

I don’t have an answer to all of your questions, but I’ll answer the ones I can.

Perhaps you mean you want to stop rendering an actor when it gets farther than 30 units away from the camera? An LODNode or FadeLODNode is ideally suited for this purpose.

If your Python threads never touch Panda, they’re perfectly safe. You do have to promise that they never even store a Panda object, since that would be manipulating Panda’s reference count.

This is indeed one of the problems that DistributedNodes were specifically built to solve; but the whole DistributedObject system is quite complex, and if you are using it only to solve this problem it is probably not worth the complexity, especially if you already have another working solution in place.

Several people have taken a hand at contributing a pixel-perfect GUI system. You can try searching the forums for some of them. But one easy answer is, if you attach your geometry to pixel2d instead of render2d, it will inherit a more-traditional pixel-based coordinate system. This feature is new as of 1.7.0, I believe.

David

3

My solution to networking is more of a hack (in my opinion). If the distributed object system is designed to be efficient, fast and low bandwith I might be able to change from UDP to TCP (which is what I want).

Thanks for the quick response will look into all of those suggestions!