code design thought

[Simon Taeter]

I noticed the signature of List.Extra.splitWhen don't make completely sense:

-- current --
splitWhen : (a -> Bool) -> List a -> Maybe ( List a, List a )

splitWhen (\n -> n == 6) [ 1, 2, 3, 4, 5 ]
--> Nothing

-- problematic example --
case splitWhen someFunction list of
    Nothing ->
          -- do one thing

    Just (left, []) ->
         -- never happens

    Just (left, headR :: queueR) ->
         -- do something else

To me it makes more sense to remove the Maybe and if the given function never returns True then just give an empty list in Tuple.second.

-- imo --
splitWhen : (a -> Bool) -> List a -> ( List a, List a )

splitWhen (\n -> n == 6) [ 1, 2, 3, 4, 5 ]
--> ([ 1, 2, 3, 4, 5 ], [])

-- problematic example --
case splitWhen someFunction list of
    (left, []) ->
         -- do one thing

    (left, headR :: queueR) ->
         -- do something else

Right now the behavior is kind of oddly combining with List.any and there is a repeating of the information.

[Chadtech]

Thats a really good point. That is confusing.

If we go to a (List a, List a) is it confusing that..

   splitWhen fn []
   --> ([], [])

?

Maybe not right? I feel like thats alright.

Without the Maybe, the second list being empty would indicate that there was no point on which to split.

So far I like this idea of dropping the Maybe from splitWhen