Remove unused `is_error` import in Tokenize module

[ericphanson]

I was testing ExplicitImports.jl and noticed this unnecessary import:

julia> using JuliaSyntax, ExplicitImports

julia> print_explicit_imports(JuliaSyntax)
Module JuliaSyntax is not relying on any implicit imports.
Module JuliaSyntax.Tokenize is not relying on any implicit imports.

Additionally JuliaSyntax.Tokenize has stale explicit imports for these unused names:
JuliaSyntax
is_error

I left the JuliaSyntax one since I figured it could be annoying if you are working in that module and the name is unexpectedly unavailable. I feel that removing the is_error one improves clarity though, as it makes it easier to see that Tokenize does not add a method for is_error (or even use it).

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 96.17%. Comparing base (a6f2d15) to head (832dd41).

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #420 +/- ## ========================================== + Coverage 94.61% 96.17% +1.56% ========================================== Files 14 14 Lines 4178 4185 +7 ========================================== + Hits 3953 4025 +72 + Misses 225 160 -65 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[ericphanson]

I think the CI failures are unrelated (since I wouldn't expect the import is needed on some Julia versions and not others), though I don't really understand the failures so I'm not 100% sure.