Re: Faster `elem` "Hack" for Data.Set?

I haven't found a downside to requiring that the elements of sets be `Ord` to be `MonoFoldable`. Yes, the dictionary is ignored by all the generic methods, but you still needed it to construct the `Set`. `MonoFoldable` includes some questionable design choices, but the `Ord` requirement for `Set` is not one. Sent from my phone with K-9 Mail. On 30 January 2022 05:19:46 UTC, David Feuer wrote:

MonoFoldable isn't the answer either. That also has the problem of requiring the same constraint for folding as for membership tests.

On Sun, Jan 30, 2022, 12:12 AM PHO wrote:

...

On 1/30/22 9:59 AM, David Feuer wrote: [snip]

...

. I think the real answer is to remove `elem` from `Foldable` and put it somewhere more appropriate.

In a separate class? That would only help `elem` but not any other methods of Foldable. I'd rather love to see MonoFoldable[1] to be moved to base then.

[1]:

https://hackage.haskell.org/package/mono-traversable-1.0.15.3/docs/Data-Mono... _______________________________________________ Libraries mailing list Libraries@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries