Documentation

Run an action forever, using termination handler to notify of failure events.

This is generally most useful to guard a thread against un-noticed termination.

A reasonable usage would be to add monitoring / notifications in termination handler and just let this loop keep your code alive. It is recommened that you use the termination handler to control number of retried.

An equally reasonable alternative is to call exitImmediately (or some equivalent - perhaps you have an MVar controlling program termination?), and use this to ensure that the entire process dies if any of the supervised threads die.

The action passed to guard should run forever, if you want to run a short-lived action repeatedly in a supervised fashion see repeatedly. This function will still work, but the onGraceful handler will be called (a lot).

Common usage (where expectation is the do block never return): void . forkIO . guard (TerminationHandler ...) . forever $ do doThis andThis

Run an action repeatedly with a fixed delay between runs, using termination handler to notify of failure events.

See guard for more description. This offers the same behaviour for actions that don't run forever.

Common usage (where expectation is the do block does returns): void . forkIO . repeatedly (seconds 5) (TerminationHandler ...) $ do doThis andThis