Fwd: Typing of Pattern Synonyms: Required vs Provided constraints

[Anthony Clayden]

Thanks Gergö, I do find Richard's tendency to use ViewPattern examples
distracts me from understanding the point. That arrow-to-nowhere or
arrow-to-the-wrong-value syntax is nausea-inducing.

Your "instructive" example is really nothing to do with PatSyns, just
ordinary Haskell numeric patterns: the Language Report (section 3.17.2
point 7., same wording in 2010 and '98) says a numeric literal in a pattern
gets desugared to an `(==)` test against `fromInteger  <something>`. So
that explains where the constraints come from. (Annoyingly, the <something>
you can't express in legit Haskell source.)

Avoiding distractions is why my o.p. used an `Ord` constraint, also to
follow SPJ's 1999 example. And the canonical example in those old
discussions (like Hughes' 1999 paper) is an `Ord` constraint for elements
of a `Set` structured as a BST.

With your "polymorphic literal" how do I follow the Language Report to
desugar? I'd rather be able to write Richard's example using an as-Pattern
like one of these:

>    pattern x at Is3 | x == fromInteger (#3#) = x   -- where #3# is the
inexpressible <something>
>    pattern x at 3@Is3 = x

(It's not that I'm "new to patsyn typing"; it's more that patsyn typing and
design keeps falling short of what I hope from such a promising feature. So
I always want to check I'm not mis-understanding. Specifically, the design
for Pattern Synonyms was a lost opportunity to expunge ViewPatterns
entirely.)

AntC

On Fri, 7 Jan 2022 at 09:43, Gergő Érdi <gergo at erdi.hu> wrote:

> Fwiw, a less contrived, and much more relatable, version of Richard's
> example would be
>
>
> pattern Is3 :: (Num a, Eq a) => a   -- only a Required constraint
> pattern Is3 = 3 -- a polymorphic literal!
>
> I think it can be quite instructive for people new to patsyn typing to
> work out why this is exactly the same as the one in Richard's email!
>
>
> On Thu, Jan 6, 2022, 21:11 Richard Eisenberg <lists at richarde.dev> wrote:
>
>>
>>
>> ...
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/glasgow-haskell-users/attachments/20220107/bf7d4f19/attachment.html>