The IsString1 solution seems okay, though it requires writing an IsString1 instance for all types which have an IsString instance now, and I'm not sure that the breakage is worth it. 

On 3 Sep 2017 4:33 p.m., "Henning Thielemann" <lemming@henning-thielemann.de> wrote:

On Sun, 3 Sep 2017, Tony Morris wrote:

Why not:

class isString1 a where fromString1 :: NonEmpty Char -> a

This would be a clean solution. GHC could reject empty string literals. (Though I do not know how difficult or sensible it is to extend GHC this way.)
_______________________________________________
Libraries mailing list
Libraries@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries