Fix warnings (that will be) introduced by ghc upgrade

[jfly]

We're upgrading to GHC 9.10.3, and it will start to emit warnings about risky use of partial functions.

AFAICT, the old code wasn't wrong (we were always checking for empty list before trying to call head on it, for exapmle), but it's nice when a compiler can check things for us!

$ cabal build
...
src/Nixfmt/Types.hs:375:68: warning: [GHC-63394] [-Wx-partial]
    In the use of ‘head’
    (imported from Prelude, but defined in GHC.Internal.List):
    "This is a partial function, it throws an error on empty lists. Use pattern matching, 'Data.List.uncons' or 'Data.Maybe.listToMaybe' instead. Consider refactoring to use "Data.List.NonEmpty"."
    |
375 |     (List _ items _) | Prelude.length (unItems items) == 1 -> case Prelude.head (unItems items) of
    |                                                                    ^^^^^^^^^^^^

src/Nixfmt/Types.hs:384:69: warning: [GHC-63394] [-Wx-partial]
    In the use of ‘head’
    (imported from Prelude, but defined in GHC.Internal.List):
    "This is a partial function, it throws an error on empty lists. Use pattern matching, 'Data.List.uncons' or 'Data.Maybe.listToMaybe' instead. Consider refactoring to use "Data.List.NonEmpty"."
    |
384 |     (Set _ _ items _) | Prelude.length (unItems items) == 1 -> case Prelude.head (unItems items) of
    |                                                                     ^^^^^^^^^^^^

[3 of 8] Compiling Nixfmt.Predoc    ( src/Nixfmt/Predoc.hs, dist-newstyle/build/x86_64-linux/ghc-9.10.3/nixfmt-1.2.0/build/Nixfmt/Predoc.o, dist-newstyle/build/x86_64-linux/ghc-9.10.3/nixfmt-1.2.0/build/Nixfmt/Predoc.dyn_o )
src/Nixfmt/Predoc.hs:169:35: warning: [GHC-63394] [-Wx-partial]
    In the use of ‘head’
    (imported from Prelude, but defined in GHC.Internal.List):
    "This is a partial function, it throws an error on empty lists. Use pattern matching, 'Data.List.uncons' or 'Data.Maybe.listToMaybe' instead. Consider refactoring to use "Data.List.NonEmpty"."
    |
169 |     if p /= [] && (isSoftSpacing (head p) || isSoftSpacing (last p))
    |                                   ^^^^

src/Nixfmt/Predoc.hs:634:27: warning: [GHC-63394] [-Wx-partial]
    In the use of ‘head’
    (imported from Prelude, but defined in GHC.Internal.List):
    "This is a partial function, it throws an error on empty lists. Use pattern matching, 'Data.List.uncons' or 'Data.Maybe.listToMaybe' instead. Consider refactoring to use "Data.List.NonEmpty"."
    |
634 |               grp' = case head grp of
    |                           ^^^^

src/Nixfmt/Predoc.hs:635:30: warning: [GHC-63394] [-Wx-partial]
    In the use of ‘tail’
    (imported from Prelude, but defined in GHC.Internal.List):
    "This is a partial function, it throws an error on empty lists. Replace it with 'drop' 1, or use pattern matching or 'GHC.Internal.Data.List.uncons' instead. Consider refactoring to use "Data.List.NonEmpty"."
    |
635 |                 Spacing _ -> tail grp
    |                              ^^^^

src/Nixfmt/Predoc.hs:636:70: warning: [GHC-63394] [-Wx-partial]
    In the use of ‘tail’
    (imported from Prelude, but defined in GHC.Internal.List):
    "This is a partial function, it throws an error on empty lists. Replace it with 'drop' 1, or use pattern matching or 'GHC.Internal.Data.List.uncons' instead. Consider refactoring to use "Data.List.NonEmpty"."
    |
636 |                 Group ann ((Spacing _) : inner) -> Group ann inner : tail grp
    |                                                                      ^^^^

[github-actions]

Nixpkgs diff

[MattSturgeon]

Seems reasonable to me. No objections.

I'll let someone more comfortable in Haskell approve, though. @dyegoaurelio are you able to review?