Suppose I have a function func :: a -> m (Maybe b), where m is a monad. Then, in a do for m, I can write
example :: a -> m (Maybe b)
example a = do
Just x <- func a
.....
If func a = Nothing, this will trigger the fail method of m. Is there a way to still use this notation while specify a catch all pattern matching case? I'm trying to avoid having to add a separate case of expression. I vaguely remember something like this being possible, but I cannot seem to find where I originally saw it. I remember something along the lines of
example :: a -> m (Maybe b)
example a = do
Just x <- func a | Nothing -> myfail
.....
I'm worried that I may be recollecting something from a different language, since I can't find anything about this after searching for a while.
You can write:
do
x <- func >>= \case
Just x -> pure x
Nothing -> myfail
...
You do still need to write a case, but it doesn't increase the indentation level like the "obvious" case does:
do
mx <- func
case mx of
Just x -> do
...
Nothing -> myfail
Notice how much shallower the ... is in the first example than the second.