Haskell-style state monads.
State monads are a way to thread a modifiable, readable state record through a computation without resorting to mutability. At any point in a state-monad computation, we can peek at the present value of the state, or modify it, or do both.
We provide two signatures for state monads: one corresponding to the situation where the state type is fixed at the module level (S), and one leaving the state type as part of the monad type (S2). The former has a corresponding make functor; the latter has a direct implementation.
We also provide a functor To_S for fixing the state type in an arity-2 monad after the fact.
Like in Haskell, state monads are a special case of state transformers, corresponding to the application of those transformers to the identity monad.
If you need to run stateful computations that can fail, or are partial, or non-deterministic, and so on, use state transformers instead.
State_intf contains the signatures for State.
include module type of State_intfmodule type Generic = sig ... endGeneric contains the signature bits common to all state monad signatures.
module type S = sig ... endS is the signature of state monads parametrised over their value, but with a fixed state type.
module type S2 = sig ... endS2 is the signature of state monads parametrised over both value and state types.
Make makes an S (state monad with fixed state type) from a single state type.