Re: [Haskell-cafe] Announce: EnumMap-0.0.1

Allow me to answer my own question:

EnumMap.insertWithKey f k v m = IntMap.insertWithKey (f . Key . toEnum) (unKey k) v (unEnumMap m)

On Mon, Aug 10, 2009 at 2:32 PM, John Van Enk wrote:

I've been trying to implement EnumMap as a wrapper for IntMap. Here's the first problem I ran into:

...

IntMap.insertWithKey :: (Key -> a -> a -> a) -> Key -> a -> IntMap a -> IntMap a

I'd like to translate this to something like:

...

EnumMap.insertWithKey :: Enum k => (Key k -> a -> a -> a) -> Key k -> a -> EnumMap k a -> EnumMap k a

My initial thought was just to make it a normal wrapper:

...

EnumMap.insertWithKey f k v m = IntMap.insertWithKey f (unKey k) v (unEnumMap m)

The obvious problem here is the type of `f' expected by the wrapper function and the internal function. Either we force the wrapper to take a function that takes an Int as the first parameter, or we rewrite the logic of insertWithKey to allow us to use the proper (Key k) type.

I don't see an obvious way around this--am I missing something?

On Sat, Aug 8, 2009 at 4:41 PM, Henning Thielemann wrote:

...

On Sat, 8 Aug 2009, John Van Enk wrote:

...

Hi List,

I've uploaded a first version of EnumMap to hackage.

EnumMap is a generalization of IntMap that constrains the key to Enum rather than forcing it to be Int. I have no idea what impact this has on performance, but it still passes all the tests that ship with IntMap. (My guess is that performance will be similar/identical unless I've missed something.)

Could that be implemented as wrapper around IntMap?