Introduce Data.Enum and Data.Bounded

[bgamari]

In #146 I noticed that the Enum and Bounded typeclasses currently have no proper home (currently exposed only via Prelude and GHC.Enum). I propose that we remedy this by introducing a pair of new modules, Data.Bounded and Data.Enum, to base:

module Data.Bounded where

class Bounded a where
  minBound :: a
  maxBound :: a

module Data.Enum where

class Enum a where
  succ :: a -> a
  pred :: a -> a
  toEnum :: GHC.Types.Int -> a
  fromEnum :: a -> GHC.Types.Int
  enumFrom :: a -> [a]
  enumFromThen :: a -> a -> [a]
  enumFromTo :: a -> a -> [a]
  enumFromThenTo :: a -> a -> a -> [a]

This has been implemented in !10422 and !11347.

[mixphix]

+1, and please, oh please, let's include enumerate = [minBound .. maxBound] as an export.

[int-index]

Why the name enumerate? I expected enumerate = zip [0..] because there is a similar function in Python.

[Kleidukos]

Ah indeed that would be not too great to have a conflict on this. I've seen universe in Relude. That would be a reasonable alternative?

[tomjaguarpaw]

I think discussion about a new function needs a new proposal. Can we please avoid scope creep in this discussion?

[bgamari]

Given there is some disagreement on the details here, I have dropped enumerate from the proposal. As I don't have a particular interest here, I will leave a proposal introducing it to someone else (but they should feel free to reference !11321).

[tomjaguarpaw]

I'm confused. ghc-!0422 looks like it has already been merged.

[bgamari]

@tomjaguarpaw , yes, this proposal arose out of a miscommunication in #146.

[tomjaguarpaw]

Aha, I see. Thanks!

[ocharles]

Does Bounded make sense to be in Data.Enum? An alternative is to house that in Data.Bounded.

[bgamari]

My initial thought was that Enum and Bounded tend to appear together (and indeed they are both defined in the GHC.Enum module) but yes, I agree that exposing them as separate modules makes some amount of sense. I have amended the proposal to do so.

[Bodigrim]

Dear CLC members, let's vote on the proposal to provide canonical home for class Enum and class Bounded in Data.* land (namely, Data.Enum and Data.Bounded) as opposed to being exported only from GHC.Enum and Prelude. Status quo is unfortunate: if you do not want to import entire Prelude, you have to import GHC.Enum which has a flavour of internal module, prone to rapid changes. We discussed the issue earlier in #64 and #146.

@tomjaguarpaw @hasufell @mixphix @angerman @parsonsmatt @velveteer

---

+1 from me.

[velveteer]

+1

[hasufell]

+1

[tomjaguarpaw]

+1

[parsonsmatt]

+1

[mixphix]

+1

[Bodigrim]

Thanks all, that's enough votes to approve.

@bgamari please make sure that both MRs land into the same GHC release.

[bgamari]

Implementation in !11347.

[Bodigrim]

@bgamari I'm not sure what's happened here. The proposal we approved meant to introduce Data.Bounded for class Bounded and Data.Enum for class Enum. Yet in base-4.20 I observe Data.Enum with both type classes, but no Data.Bounded at all. There is neither changelog nor @since pragma. Even more confusingly, the MR which was meant to implement the proposal remains unmerged: https://gitlab.haskell.org/ghc/ghc/-/merge_requests/11347

I'm not sure what to do, the genie is out of the bottle. We cannot reasonably remove class Bounded from Data.Enum now. Does it still make sense to introduce Data.Bounded?..