added some reasoning about deadlocks and removed unneeded code

app/Environment.hs

@@ -11,11 +11,10 @@ module Environment(Env (..), EnvM, runEnvM, myLogAction,defaultEnv) where
 import Prelude hiding (log)
 import Colog
 import Control.Monad.Reader (ReaderT (..), MonadReader)
-import Control.Monad.Trans
+import Control.Monad.Trans(lift)
 import Control.Monad.IO.Unlift
 import Control.Concurrent.MVar (MVar,modifyMVar_,newMVar)
 import System.IO.Unsafe(unsafePerformIO)
-import qualified Control.Concurrent.MVar as M
 import qualified Data.Text as T
 import qualified Data.Text.IO as T
 import Control.Parallel(pseq)
@@ -43,17 +42,17 @@ instance MonadUnliftIO EnvM where
   withRunInIO inner = EnvM $ withRunInIO (\fromReader -> inner (fromReader . unEnvM))
   {-# INLINE withRunInIO #-}
 
-modifyMVar :: M.MVar a -> (a -> EnvM (a,b)) -> EnvM b
-modifyMVar mvar action = withRunInIO $ \run -> M.modifyMVar mvar (run . action)
-
-
 runEnvM :: Env EnvM -> EnvM a -> IO a
 runEnvM env envm = runReaderT (unEnvM envm) env
 
+-- we do not export this MVar (which makes it easier to reason about deadlocks
 syncMVar :: MVar ()
 syncMVar = unsafePerformIO $ newMVar ()
 {-# NOINLINE syncMVar #-}
 
+-- Every time you do something like this you should ask yourself: Is this a potential deadlock? And the answer
+-- is in this case is NO. Even if m diverges it won't be able to lock up other threads as it will be forced to
+-- evaluate before we take the MVar. So as long printing to stdout terminates we will always put back the mvar.
 myLogAction :: LogAction EnvM Message
 myLogAction = LogAction $ \x -> let m = formatMessage x in EnvM . lift $ m `pseq` modifyMVar_ syncMVar (const (T.putStrLn m))