lens doesn't offer such a combinator in a ready-made form. If instead of runFoo
you had some g :: a -> a
function, you would be able to use (%=)
, which is to over
what use
is to view
, and write simply foo %= g
. Your desired functionality, though, involves interleaving the m
effects introduced by runFoo
in between the getting and setting, which precludes a solution this straightforward.
Searching for missing State Combinator for Lens
-
27-06-2022 - |
Question
I currently have code that looks like this:
do
x <- use foo
foo <~ runFoo x
where foo
is a Lens
to a Foo
field and runFoo :: MonadState m => Foo -> m Foo
I think there should be a way to do this operation in one line, but I cannot find it. I think it ought to have a comment like:
(????) :: MonadState s m => Lens' s a -> (a -> m a) -> m a
Questions:
- Does such a combinator exist? if so what is it?
- When I run into another question like this, what is the best way to search for it (i.e. Usually I would just type this into Hoogle but I've not had good luck doing that with the lens library)
- Is this actually a primitive combinator found in Control.Monad? (I will be slightly embarrassed if this is yet another job for the kleisli arrow)
Solution
OTHER TIPS
The type signature of the function you are looking for is suspiciously close to that of >>=
, but with the first argument different.
In particular, it appears as if use foo :: Monad m => m Foo
and foo <~ runFoo :: Foo -> m Foo
, so you could use bind directly as use foo >>= (\x -> foo <~ runFoo x)
. There's probably an easier/shorter way, but I don't have the full code to play around with. I'd say just stick with using two lines, and if it's a common enough operation, then define your own combinator for it.