astral-sh/ruff · error

path `{}` is not valid UTF-8

Error message

path `{}` is not valid UTF-8

What it means

While copying a fixture's files into the eval temp directory, the walk converts every visited entry to a `SystemPath`; any file or directory inside the fixture whose name is not valid UTF-8 aborts with this message and the path echoed. Dot-prefixed entries are filtered out before this, so hidden dirs like `.venv` are not the cause.

Source

Thrown at crates/ty_completion_eval/src/main.rs:561

fn copy_project(src_dir: &SystemPath, dst_dir: &SystemPath) -> anyhow::Result<Vec<Cursor>> {
    std::fs::create_dir_all(dst_dir).with_context(|| dst_dir.to_string())?;

    let mut cursors = vec![];
    let it = walkdir::WalkDir::new(src_dir.as_std_path())
        .sort_by_file_name()
        .into_iter()
        .filter_entry(|dent| {
            !dent
                .file_name()
                .to_str()
                .is_some_and(|name| name.starts_with('.'))
        });
    for result in it {
        let dent =
            result.with_context(|| format!("failed to get directory entry from {src_dir}"))?;

        let src = SystemPath::from_std_path(dent.path()).ok_or_else(|| {
            anyhow::anyhow!("path `{}` is not valid UTF-8", dent.path().display())
        })?;
        let name = src
            .strip_prefix(src_dir)
            .expect("descendent of `src_dir` must start with `src`");
        // let name = src
        // .file_name()
        // .ok_or_else(|| anyhow::anyhow!("path `{src}` is missing a basename"))?;
        let dst = dst_dir.join(name);
        if dent.file_type().is_dir() {
            std::fs::create_dir_all(dst.as_std_path())
                .with_context(|| format!("failed to create directory `{dst}`"))?;
        } else {
            cursors.extend(copy_file(src, &dst)?);
        }
    }
    anyhow::ensure!(
        !cursors.is_empty(),
        "could not find any `<CURSOR>` directives in any of the files in `{src_dir}`",

View on GitHub (pinned to 672bb4edf0)

Solutions

1. Detect offenders: `find <fixture-dir> -print0 | while IFS= read -r -d '' f; do printf '%s' "$f" | iconv -f utf-8 -t utf-8 >/dev/null 2>&1 || echo "non-UTF-8: $f"; done`
2. Delete or rename the reported paths, and keep fixtures to committed files only (`git status` / `git clean` on the fixture)
3. Review what created the file and disable it for the truth tree (e.g. run builds outside the fixtures)

Example fix

# before: run aborts with `path ... is not valid UTF-8`
# after: clean the fixture, then re-run
git clean -fdx crates/ty_completion_eval/truth/<fixture>
cargo run -p ty_completion_eval -- all

find crates/ty_completion_eval/truth -print0 \
| while IFS= read -r -d '' f; do
    printf '%s' "$f" | iconv -f utf-8 -t utf-8 >/dev/null 2>&1 \
      || echo "non-UTF-8 path: $f" >&2
  done

When it happens

Trigger: A non-UTF-8-named file or directory inside a truth fixture project — build droppings, cache files, or extraction residue — encountered during the recursive copy from `truth/` into the temp eval dir.

Common situations: Fixtures that once held real project artifacts (`.tox`, caches with byte names); files synced from Windows with mangled encodings; fuzzing residue committed accidentally.

Related errors

• truth source directory `{path}` contains invalid UTF-8
• The current working directory `{}` contains non-Unicode char
• The current working directory `{}` contains non-Unicode char
• Detected project directory `{path}` contains non-Unicode cha
• {truth} does not exist: ty's completion evaluation must be r

AI-assisted analysis of astral-sh/ruff@672bb4edf0 (2026-08-16). Data as JSON: /api/errors/02b142eef365a479. Report an issue: GitHub.