[Haskell-cafe] SqlPersistT action in MonadError issue

Matt parsonsmatt at gmail.com
Sat Jul 14 18:09:09 UTC 2018

The type of `runPool` is given here:

    runPool :: MonadUnliftIO
 m => c -> PersistConfigBackend
 c m a -> PersistConfigPool
 c -> m a

The return type of `runPool`, then, is `m a`, for some `m` satisfying

The type of `liftIO` is `liftIO :: (MonadIO m) => IO a -> m a`. This means
that the first argument to `liftIO` must be an `IO a`.

When we say `liftIO (runPool ...)`, GHC tries to unify the `m a` from
`runPool` and the `IO a` from `liftIO`. It is able to do this, as `IO` is
an instance of `MonadUnliftIO`. Then, the concrete type of `runPool pgConf
f1 pgPool` becomes `IO Bool`.

GHC now tries to unify the `m` in `f1 :: (MonadIO m, MonadError
AppException m) => SqlPersistT m Bool` with `IO`. It tries to satisfy the
constraints: `MonadIO` is easily satisfied, but `MonadError AppException m`
triggers the problem you mention.

Because `ExceptT` does not have an instance for `MonadUnliftIO`, we cannot
use it with the `m` in `SqlPersistT m Bool` directly. We can, instead, use
`mapReaderT` to push the `Either` into the return type, like this:

pushException :: SqlPersistT (ExceptT e m) a -> SqlPersistT m (Either e a)
pushException = mapReaderT runExceptT

Now, we can write:

liftIO $ runPool pgConf (pushException f1) pgPool

This gives us an `IO (Either AppException Bool)`, which, after lifting,
gives us `AppM IO (Either AppException Bool)`. You can then use `either
throwError pure` to pull the `AppException` into the `ExceptT` again.


I would suggest that you reconsider this approach, however. The
`persistent` library uses transactions and exceptions in a way that is
*almost always* what you want, but can be surprising, and using `ExceptT`
will break this system. Transactions are automatically rolled back on a
`MonadCatch`-style exception, but they are not automatically rolled back on
an `ExceptT`-style exception. Having a single `AppException` type that
represents errors that can occur in database transactions *and* the rest of
your application is also going to be a cause for unsafety and errors, as
the type cannot possibly be precise enough to provide any safety.

Matt Parsons

On Sat, Jul 14, 2018 at 10:22 AM, Vlatko Basic <vlatko.basic at gmail.com>

> Hello,
> I'm trying to use SqlPersistT funcs with MonadError, but am failing in
> writing the runDB correctly.
> I tried several things, but always getting:
>     • Couldn't match type ‘IOException’ with ‘AppException’
>         arising from a functional dependency between:
>           constraint ‘MonadError AppException IO’ arising from a use of
> ‘f1’
>           instance ‘MonadError IOException IO’ at <no location info>
>     • In the second argument of ‘runPool’, namely ‘f1’
>       In the second argument of ‘($)’, namely ‘runPool pgConf f1 pgPool’
>       In a stmt of a 'do' block: liftIO $ runPool pgConf f1 pgPool
> I understand there already is "instance MonadError IOException IO" and
> fundep says it can be only one for IO.
> How to make it compile?
> Best regards,
> vlatko
> Here is the minimal reproducible code:
> module Test where
> import Prelude
> import Control.Exception.Base
> import Control.Monad.Except
> import Control.Monad.Trans.Reader
> import Database.Persist.Postgresql
> data AppException =
>     ExcText String
>   | ExcIO IOException
>   deriving (Show)
> type AppM m = ExceptT AppException (ReaderT App m)
> data App    = App { pgConf :: PostgresConf, pgPool :: ConnectionPool}
> runDB :: (MonadIO m, MonadError AppException m) => AppM m Bool
> runDB = do
>   App{..} <- lift ask
>   liftIO $ runPool pgConf *f1* pgPool    -- error
>   -- liftIO $ runPool pgConf *f2* pgPool  -- OK
> f1 :: (MonadIO m, MonadError AppException m) => SqlPersistT m Bool
> f1 = throwError $ ExcText "exception"
> f2 :: MonadIO m => SqlPersistT m Bool
> f2 = return True
> _______________________________________________
> Haskell-Cafe mailing list
> To (un)subscribe, modify options or view archives go to:
> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> Only members subscribed via the mailman list are allowed to post.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20180714/ce116343/attachment-0001.html>

More information about the Haskell-Cafe mailing list