Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Dutchling]

Well, I did find an installer which gives me errors because network sources are unavailable.

edit: I'm off to doing uni stuff now. I'll see if IDLE has the same problem or not when I get home.

[Max White]

Just get somebody to punch you in the face or something so you can go home and program in a dark room, alone, talking only to text on the internet LIKE US NORMAL PEOPLE.

[Mego]

Just get somebody to punch you in the face or something so you can go home and program in a dark room, alone, talking only to text on the internet LIKE US NORMAL PEOPLE.

The way programming should be done.

I'm getting good work done on my latest non-uni project, but making Java and C++ communicate via sockets is a pain. Just using sockets in C/C++ is bad enough.

[Dutchling]

Sooo, another module related problem.
I basically have this:

Code: [Select]

# families.py

from simple import *

def main():
    global WORLD
    WORLD = <yadda yadda>
    loop()

main()
   

Code: [Select]

# simple.py

def loop():
    <yadda yadda>
    map_mod()

def map_mod():
    for X in WORLD:  #  //NameError: global name 'WORLD' is not defined//
        <yadda yadda>
How do I prevent the error? using one of the following:

Code: [Select]

import families # at the beginning of simply.py
from families import WORLD # at the beginning of map_modresults in restarting the whole thing, as it runs main() in families.py again.

Is there no way to have global variables? I'm thinking of just coding everything in one large file, as this is really annoying.

[Levi]

I'm not familiar enough with python to know how to do it properly, but one way I have to avoid using globals is to create an object (I call it Game) that contains everything important that you want everyone to have access too.  Then whenever creating an object, I pass in the game object and assign it to self.game.  Eventually every object has access to the game object, which has access to pretty much every thing else. 

So I've got something like:

Code: [Select]

class Game:
  def __init__(self):
    self.dungeon = dungeon.Dungeon(self)
    self.item_manager = item_manager.ItemManager(self)
    self.monster_list = []
    self.whatever = {}

class Monster:
  def __init__(self, game)
    self.game = game
    //blah blah blah

So basically I make one instance of game, and just pass it around to everything.

[Dutchling]

Thanks! Will be very useful.

edit: still one question, is there a way to import a module without running it? I don't need to do that now any more, so I'm not interested in workarounds, just wondering if it is possible.

[Levi]

edit: still one question, is there a way to import a module without running it? I don't need to do that now any more, so I'm not interested in workarounds, just wondering if it is possible.

I only deal in workarounds.     I have no idea other than wrapping the code into a function, and then not calling the function.

[Darvi]

Wrong thread.

[fergus]

Sooo, another module related problem.
I basically have this:

Code: [Select]

# families.py

from simple import *

def main():
    global WORLD
    WORLD = <yadda yadda>
    loop()

main()
   

Replace

Code: [Select]

main()
with

Code: [Select]

if __name__ == '__main__':
    main()
and import families.py

[Skyrunner]

So, I want to make a simple, inefficient n-body simulation (By inefficient, I mean O(n²) inefficient). The problem is I have no idea how to display it. What could be a way to display what's going on in the simulation? o_O Possible ways I've thought up, in order of increasing time probably needed and complexity. Most likely readability, too.

1. Export as lots of text with X, Y coordinates every once in a while. Only 2D. Can use Excel to show all the particles.
2. Use a low resolution, easy-to-use display library (like libtcod). It has 144 times less resolution than you would expect, since each tile is 12 x 12 pixels.
3. Export images with white pixels as bodies and black as nothing, again every once in a while. Doesn't work with 3D simulations.
4. Learn a graphics library and implement it.

...Dx Are there any suggestions as to which is best? And maybe even tutorials. c: Number four seems pretty darn imposing.

[alway]

Supposedly libtcod can work with any size of tileset; a 1x1 or 2x2 tile size should work for your needs. It would probably be slow as all hell, but you could use that to render. It would work for 3D as well; you just use an orthographic view. Which basically just means 'toss out the z axis and render it on the x-y axes.' You could do a full perspective view, but that would require transformation and projection matrices, which would make it both take longer to render and longer to code.

As for learning a graphics library, it's best to do that when you are just focusing on learning a graphics library. Particularly for a project like n-body simulation, where the things being rendered are dots, it would be difficult to figure out where things went wrong; as they most assuredly will when one does graphics programming for the first time.

[Urist McScoopbeard]

Is anyone here familiar with Slick2D?

[Virex]

Alway: SDL is simple enough to learn and ported to nearly every language.

[alway]

Ah, true, I forgot about SDL. I was thinking more along the lines of shader-based OpenGL or DirectX.

[Skyrunner]

So the verdict is 'use libtcod with a 1x1 tileset' for now?
Hrm. If there's lots of dots, it might be slow enough that it interferes with the simulation...