[Haskell-beginners] opengl and application state

[Sean Charles]

How!?!?!  LOL

I am getting back into OpenGL again with Haskell now and I have a 
program that listens to all the callbacks and prints stuff out, 
terminates on Ctrl+X and draws a grey sphere. Earth shattering.

If I was using C/C++ I'd be using a (shudder) global pointer to a record 
containing state that the render callback would use to know what to do.

I smell state monads or something but I cannot see how to pass them into 
"the loop".

I want to organise my code so that all input events cause a record to be 
updated and then the render part just uses the current values to draw 
the right things, with the obvious caveat of "what if a key event 
happens half way through the render phase." MVars? Threads? What??? I'm 
a real newb on this right now!

How do I,or what is a "best practice" for passing a continuously 
changing record across calls given that the "displayCallback" doesnt 
have any obvious place to plug one on? In Erlang or even Haskell, there 
is the idion of making a tail call back to yourself with the new 
modified record and so on but I cannot figure out where to actually pass 
it around!

Thanks in advance.
Sean.