[ghc-steering-committee] GHC 2020

Eric Seidel eric at seidel.io
Mon Aug 31 03:51:14 UTC 2020


On Sun, Aug 30, 2020, at 03:11, Joachim Breitner wrote:
> I wouldn’t worry about enabling the extension; you’d still have to
> actually write {..} to get the behavior, readers will at least know
> whether it is hard to tell where variables are abound.

I agree, but this is a concern that comes up often, so we should be prepared to address it.

> > at least consider adding OverloadedStrings and perhaps
> > OverloadedLists.
> 
> 
> I regularly write code where I use a library that only works nice if
> "foo" is monomorphic, because such strings are meant to be used in
> polymorphic contexts. With OverloadedStrings, you get ambiguity errors
> that you have to resolve with explicit type annotations. So
> unfortunately, that one probably can’t be on by default. Which is a
> shame, of course.

Interesting, I think I can count on one hand the number of times OverloadedStrings has caused an ambiguity error for me. What libraries do you use that are polymorphic in the string type?


More information about the ghc-steering-committee mailing list