linter: formatter reports utf-8 index while eslint reports in utf-16

[hi-ogawa]

What version of Oxlint are you using?

0.15.10

What command did you run?

pnpm dlx oxlint -D no-debugger repro.js

What does your .oxlintrc.json config file look like?

None

What happened?

The code of repro.js

"🤨"; debugger;

The commands output

$ pnpm dlx eslint --no-config-lookup --rule 'no-debugger: error' repro.js

/home/hiroshi/code/others/oxc/repro.js
  1:7  error  Unexpected 'debugger' statement  no-debugger

✖ 1 problem (1 error, 0 warnings)


$ pnpm dlx oxlint -D no-debugger repro.js

  × eslint(no-debugger): `debugger` statement is not allowed
   ╭─[repro.js:1:9]
 1 │ "🤨"; debugger;
   ·       ─────────
   ╰────
  help: Delete this code.

Found 0 warnings and 1 error.
Finished in 5ms on 1 file with 97 rules using 16 threads.

I'm not sure if this has been discussed, but bringing this up just in case as I'm wondering whether terminal output is fine with column index in utf-8. For example, on Vsocde's terminal, clicking repro.js:1:9 jumps to the 3rd character b of debugger.

Interestingly, language server protocol seems to have a knowledge about encoding, so it's reported correctly on Vscode extension.