Run a BasicGuest, and return when the Event returned by the BasicGuest fires.

Each host runs on a separate spider timeline, so you can launch multiple hosts via forkIO or forkOS and they will not mutex each other.

NOTE: If you want to capture values from a build before the network starts firing (e.g., to hand off event triggers to another thread), populate an MVar (if threading) or IORef as you build the network. If you receive errors about untouchable type variables while doing this, add type annotations to constrain the MVar or IORef contents before passing them to the function that returns your BasicGuest. See the Multithread.hs example for a demonstration of this pattern, and where to put the type annotations.

Run a BasicGuest without a quit Event.

basicHostForever guest = basicHostWithQuit $ never <$ guest 

The basic guest type. Try not to code against it directly; instead ask for the features you need MTL-style:

myFunction :: (Reflex t, MonadHold m) => ... 

Constraints provided by a BasicGuest, when run by basicHostWithQuit or basicHostForever.

Augment a BasicGuest with an action that is repeatedly run until the provided Event fires. Each time the action completes, the returned Event will fire.

Example - providing a 'tick' Event to a network

myNetwork :: (Reflex t, MonadHold t m, MonadFix m) => Event t () -> m (Dynamic t Int) myNetwork = count myGuest :: BasicGuestConstraints t m => BasicGuest t m (Event t ()) myGuest = mdo eTick <- repeatUntilQuit (void $ threadDelay 1000000) eQuit let eCountUpdated = updated dCount eQuit = () <$ ffilter (==5) eCountUpdated dCount <- myNetwork eTick performEvent_ $ liftIO . print <$> eCountUpdated pure eQuit main :: IO () main = basicHostWithQuit myGuest 

Like repeatUntilQuit, but it doesn't do anything with the result of the action. May be a little more efficient if you don't need it.