stylish-haskell
                                
                                 stylish-haskell copied to clipboard
                                
                                    stylish-haskell copied to clipboard
                            
                            
                            
                        Normalize constraints parantheses
Both of the following are valid:
equals1 :: (Eq a) => a -> a
equals2 :: Eq a => a -> a
Some people like the first because it keeps things consistent when you start introducing more constraints. Some people prefer the second because it's fewe characters. But I doubt anyone prefers both in the same codebase.
It would be nice to support normalizing to either of these based on the settings.