Inlining expressions

[nielstron]

Inlining expressions should be implemented. The main reason is that it makes more readable code more efficient. For example the following

def validator(x: int):
  y = x
  z = y * 2
  w = z + 4
  return w

will now be computationally as cheap as the spaghetti-ized version

def validator(x: int):
   return (x*2)+4

The optimization should by default only applied to variables that are read once. In the case of reading and writing exactly once, this optimization will alway lead to less CPU/Mem steps and smaller script size. When applying the optimization to values read more than once, it may lead to a tradeoff of less steps but larger size (depending on the size of the expression that needs to be re-computed it might also increase the steps)

A draft implementation is in feat/inline_single_defs