[ghc-steering-committee] Proposal #402 (changes to GADT syntax); rec: accept

[Joachim Breitner]

Hi,

TL;DR: Not in favor of (1), but of (2).

I agree about (1), and would lean towards not changing the behavior,
and still allow parentheses.

Am Freitag, den 19.03.2021, 09:18 -0700 schrieb Iavor Diatchki:
> I don't see any benefits to (2), it seems just as easy to write the
> quantifiers at the beginning of the signature, which is what they'd
> presumably mean anyway. 

I don't remember the occasion, but ran into a related restriction very
recently, where I wanted to list all the constraints a

  C1 =>
  C2 =>
  T1 ->
  T2 ->
  …

but was told by the compiler I better use 

  (C1, C2) =>
  T1 ->
  T2 ->
  …


so I am in favor of the change (2), simply because it may reduce
friction and gives me more leeway in writing it the way it suits me.

I could imagine that in some cases it might be nice to mix forall’s and
constraints, if it somehow serves the narrative better.

Cheers,
Joachim


-- 
Joachim Breitner
  mail at joachim-breitner.de
  http://www.joachim-breitner.de/