but this seems like a bit of a hack. Waiting on an otherwise-unused MVar or doing 'atomically retry' don't work because GHC kills them off straight away.
Not that this will help you any but this might actually be a bug. From: https://hackage.haskell.org/package/base-4.8.1.0/docs/Control-Concurrent.htm... Holding a normal ThreadId reference will prevent the delivery of BlockedIndefinitely exceptions because the reference could be used as the target of throwTo at any time, which would unblock the thread. So this will in fact run forever because threadId is still in scope. main = do threadId <-forkIO $ do var <- newEmptyMVar takeMVar var forever $ threadDelay maxBound so wouldn't it make sense to keep this alive main = do var <- newEmptyMVar takeMVar var because a user interrupt may ocure. Silvio