Add IORef-based State carrier. #422
Conversation
This was copied straight from `fused-effects-exceptions`.
There was a problem hiding this comment.
Looking good!
Can we add tests? I know it's a bit awkward given the whole IO thing, but I'd rather not have this untested if possible. It would also pave the way toward testing its interactions with nondeterministic carriers like EmptyC, ErrorC, and ChooseC, which would be good to pin down properly.
src/Control/Carrier/State/IORef.hs
Outdated
|
|
||
| {- | A carrier for the 'State' effect. It uses an 'IORef' internally to handle its state, and thus admits a 'MonadUnliftIO' instance. Because the state operations are performed impurely, this carrier will not lose state effects even with nefarious uses of 'liftWith'. | ||
|
|
||
| Unlike the other carriers for 'State', this carrier's effects will not backtrack when run in conjuction with 'NonDet' effects. |
|
(Note also that this PR has been correctly retargeted to merge into |
|
@robrix I think this is ready for another look. I agree we should test this, but propose to do this in a separate PR, as generalizing the existing machinery to either Identity or IO is somewhat fraught, and I agree that the existing tests are those that exercise paths like lifting |
This was copied straight from
fused-effects-exceptions.