Re: [pygame] Clone/game request: Chain of Command

[Jon Peirce <Jonathan.Peirce@nottingham.ac.uk>]

> Yes, I've done that, as at that time it was the only way to get textures
> loaded. Talk about an inefficient of doing it...
>
> And you still have to fight with cameras and viewports to get the 2D
> graphics shown, and getting something look correct (1:1 pixel content wrt
> the original data) can be tricky.
>  
As with most programming it's only hard to *find out* how to do it. Not 
hard actually to do.
e.g. to draw with 1:1 pixel content (using a numeric array as input 
rather than the inefficient texture step) its very simple:

glRasterPos2i(xPos,yPos)#set position to start
glDrawPixelsub(GL_RGB, numArray)

where numArray is an RGB Numeric array of shape (X,Y,3). glDrawPixels 
does no scaling so you always get 1:1. Using the version 'ub' above 
means that you give your data as unsigned bytes.

If you're mixing and matching 2D and 3D then you also want to precede 
the above 2 lines with the following. (NB This is the "fighting with 
cameras and viewports" that you mention):
#go to 2D mode
glMatrixMode(GL_PROJECTION)                      
glPushMatrix()                                 
glLoadIdentity()                      
glOrtho( 0, winSize[0], winSize[1], 0, 0, 1 )    #sets the 0,0 to be 
top-left
glMatrixMode(GL_MODELVIEW)                          
glLoadIdentity()

...and follow the 2 lines with:
#return to previous drawing mode
glMatrixMode( GL_PROJECTION )                  
glPopMatrix()
glMatrixMode( GL_MODELVIEW )

but none of this counts as hard to do - just maybe hard to discover!

I agree with stuart - once you learn how to use openGL you end up with a 
lot more power.

--
Jon Peirce
Nottingham University
http://www.psychology.nottingham.ac.uk/staff/jwp/