Could it be the locale is not UTF8 but the comment contains such?

I'm seeing the same error with Deque-0.2 using stack hoogle --nix.

$ locale
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
LC_MESSAGES="en_US.UTF-8"
LC_PAPER="en_US.UTF-8"
LC_NAME="en_US.UTF-8"
LC_ADDRESS="en_US.UTF-8"
LC_TELEPHONE="en_US.UTF-8"
LC_MEASUREMENT="en_US.UTF-8"
LC_IDENTIFICATION="en_US.UTF-8"
LC_ALL=

@domenkozar stack hoogle --nix --no-nix-pure? perhaps the locale is only set outside of the nix shell

As far as I understand it, the issue is similar to others, e.g. sol/markdown-unlit#8 and the right way to resolve it is to set utf8 as encoding on all the file handles before reading from / writing to them (good thing there is already Documentation.haddock.Utf8, which is a perfect place for such a helper 😉). Does that sound right to you?

@domenkozar Which version of haddock did you use? I just tried building the documentation using haddock-2.18.1 and it worked with cd deque-0.2 && cabal new-haddock (with LANG=et_EE.iso88591).

What @Fuuzetsu suggest fixes the issue, haddock fails to build with LANG=C

What @kirelagin suggests is sensible for the future. Using shell env to determine filesystem encoding is one of the craziest ideas in CS history :)

Using shell env to determine filesystem encoding

Not like that. According to the Language Report, Haskell programs use Unicode so it is a reasonable assumption that the all the Haskell files are always in utf8. The only act in which env is involved is determining the encoding of the console if we are printing something to it (e.g. a Haskell identifier in an error message).

According to the Language Report, Haskell programs use Unicode so it is a reasonable assumption that the all the Haskell files are always in utf8

Didn't know this. I'm not even sure if GHC itself respects this. But if it's in there then we definitely just should use UTF8 everywhere and call it a day.

@Fuuzetsu Some of it was fixed in GHC a couple of years ago:
https://phabricator.haskell.org/D1151
https://phabricator.haskell.org/D1153

I believe most other things related to Unicode have been working fine for a while.