The real fix will happen when Data.Text moves from UTF-16 to UTF-8: https://discourse.haskell.org/t/hf-tech-proposal-1-utf-8-encoded-text/2499 https://discours

The real fix will happen when Data.Text moves from UTF-16 to UTF-8: https://discourse.haskell.org/t/hf-tech-proposal-1-utf-8-encoded-text/2499 https://discourse.haskell.org/t/hf-tech-proposal-1-utf-8-enc... Fortunately this proposal has been accepted, but I don't know the timeline for its implementation in GHC. Until then, working with UTF-8 is kind of convoluted https://serokell.io/blog/haskell-with-utf8 https://serokell.io/blog/haskell-with-utf8 Also: libraries need to be fixed to accept Data.Text instead of String. IsString helps (it's a typeclass that contains all string types) but only if APIs take it instead of defaulting to String. Adding random string conversions to cope with legacy APIs is very annoying.