F# computation expression for nested Boolean tests?

问题

I think I've got enough understanding of F# monads (workflows) that I see a few places in my code where implementing them makes sense.

For example, I've got a function with multiple nested if/thens, i.e. the function should continue only so long as the data pass certain "tests" along the way.

I'm familiar with the "maybe" monad, but in all the examples that I've seen, it's coded to operate on let! bindings, which I'm not doing. I'm hoping that someone can provide me with an example of the "maybe" workflow tailored for nested Boolean tests, not let binds.

回答1:

I offered up a conditional workflow in response to a similar problem. I'll copy it here, for reference.

module Condition =
  type ConditionBuilder() =
    member x.Bind(v, f) = if v then f() else false
    member x.Return(v) = v
  let condition = ConditionBuilder()

open Condition

let eval() =
  condition {
    // do some work
    do! conditionA
    // do some work
    do! conditionB
    // do some work
    do! conditionC
    return true
  }

As you can see in the comments to my previous answer, not everyone's a fan. But it's interesting nonetheless.

回答2:

You could pull this off without defining a new monad. Just define

let test b = if b then Some () else None

which you can now use with maybe:

maybe {
    do! test (1 > 0)
    printfn "1"
    do! test (2 > 3)
    printfn "2"

    return ()
}

回答3:

Both good answers, Daniel and eirik.

Daniels' answer led me to a solution that I had actually seen before, Tomas Petricek's implementation of an imperative workflow. I've decided to go with that, since it provides the best readability and is also useful in non-Boolean contexts.

Thanks, all.

来源：https://stackoverflow.com/questions/13710700/f-computation-expression-for-nested-boolean-tests

标签

monads

maybe

computation-expression