Stop the deprecation of Data.Monoid.{First,Last}

[Simon Jakobi]

Hi!

I have just rebased my MR
https://gitlab.haskell.org/ghc/ghc/-/merge_requests/3154.

Carter, Daniel: Both of you should have the necessary permissions to
approve the MR now. Please do so or let me know why you do not
approve.

Carter: To expand on my somewhat brusque answer (apologies!) to your questions:

> are we merely slowing down the deprecation path? are there technical reasons aside from paucity of effort for eventual removal?

As I tried to explain in the original email of this thread
(https://mail.haskell.org/pipermail/libraries/2020-April/030357.html),
it took me quite a lot of effort to discover the issues with the old
migration path.
I'd expect that vetting any other migration plan would take a similar
amount of effort. As long as no one volunteers to explore a new
migration path, I believe we shouldn't speculate that the deprecation
will eventually happen.

In any case, I currently believe that any removal of
Data.Monoid.{First,Last} would cause substantial inconvenience for
their current users.

If you would still like to resolve the name conflict, I believe that
the easier path would be to rename Data.Semigroup.{First,Last} which
AFAICT are still used much less than the Monoid versions.

I'll leave it to others to propose any such change.

Cheers!


Simon