Replace std::path with camino for UTF-8 path handling

Cargo.toml

@@ -49,6 +49,7 @@ async-compression = { version = "0.4.18", features = ["gzip", "xz", "tokio"] }
 async_zip = { git = "https://github.com/astral-sh/rs-async-zip", rev = "c909fda63fcafe4af496a07bfda28a5aae97e58d", features = ["deflate", "tokio"] }
 axoupdater = { version = "0.9.0", default-features = false, features = ["github_releases"], optional = true }
 bstr = { version = "1.11.0" }
+camino = { version = "1.1.9", features = ["serde1"] }
 clap = { version = "4.5.16", features = ["derive", "env", "string", "wrap_help"] }
 clap_complete = { version = "4.5.37", features = ["unstable-dynamic"] }
 constants = { workspace = true }
@@ -71,7 +72,6 @@ levenshtein = { version = "1.0.5" }
 liblzma = { version = "*", features = ["static"] }
 memchr = { version = "2.7.5" }
 owo-colors = { version = "4.1.0" }
-path-clean = { version = "1.0.1" }
 quick-xml = { version = "0.38" }
 rand = { version = "0.9.0" }
 rayon = { version = "1.10.0" }

src/builtin/meta_hooks.rs

@@ -1,7 +1,7 @@
 use std::io::Write;
-use std::path::Path;
 
 use anyhow::Result;
+use camino::Utf8Path;
 use fancy_regex::Regex;
 use itertools::Itertools;
 
@@ -22,7 +22,7 @@ use crate::workspace::Project;
 pub(crate) async fn check_hooks_apply(
     store: &Store,
     hook: &Hook,
-    filenames: &[&Path],
+    filenames: &[&Utf8Path],
 ) -> Result<(i32, Vec<u8>)> {
     let relative_path = hook.project().relative_path();
     // Collect all files in the project
@@ -64,7 +64,7 @@ pub(crate) async fn check_hooks_apply(
 
 // Returns true if the exclude pattern matches any files matching the include pattern.
 fn excludes_any(
-    files: &[impl AsRef<Path>],
+    files: &[impl AsRef<Utf8Path>],
     include: Option<&Regex>,
     exclude: Option<&Regex>,
 ) -> bool {
@@ -73,9 +73,7 @@ fn excludes_any(
     }
 
     files.iter().any(|f| {
-        let Some(f) = f.as_ref().to_str() else {
-            return false; // Skip files that cannot be converted to a string
-        };
+        let f = f.as_ref().as_str();
 
         if let Some(re) = &include {
             if !re.is_match(f).unwrap_or(false) {
@@ -94,7 +92,7 @@ fn excludes_any(
 /// Ensures that exclude directives apply to any file in the repository.
 pub(crate) async fn check_useless_excludes(
     hook: &Hook,
-    filenames: &[&Path],
+    filenames: &[&Utf8Path],
 ) -> Result<(i32, Vec<u8>)> {
     let relative_path = hook.project().relative_path();
     let input = collect_files(hook.work_dir(), CollectOptions::all_files()).await?;
@@ -154,12 +152,12 @@ pub(crate) async fn check_useless_excludes(
 }
 
 /// Prints all arguments passed to the hook. Useful for debugging.
-pub fn identity(_hook: &Hook, filenames: &[&Path]) -> (i32, Vec<u8>) {
+pub fn identity(_hook: &Hook, filenames: &[&Utf8Path]) -> (i32, Vec<u8>) {
     (
         0,
         filenames
             .iter()
-            .map(|f| f.to_string_lossy())
+            .map(ToString::to_string)
             .join("\n")
             .into_bytes(),
     )
@@ -172,9 +170,9 @@ mod tests {
     #[test]
     fn test_excludes_any() {
         let files = vec![
-            Path::new("file1.txt"),
-            Path::new("file2.txt"),
-            Path::new("file3.txt"),
+            Utf8Path::new("file1.txt"),
+            Utf8Path::new("file2.txt"),
+            Utf8Path::new("file3.txt"),
         ];
         assert!(excludes_any(
             &files,
@@ -188,7 +186,10 @@ mod tests {
         ));
         assert!(excludes_any(&files, None, None));
 
-        let files = vec![Path::new("html/file1.html"), Path::new("html/file2.html")];
+        let files = vec![
+            Utf8Path::new("html/file1.html"),
+            Utf8Path::new("html/file2.html"),
+        ];
         assert!(excludes_any(
             &files,
             None,

src/builtin/mod.rs

@@ -1,4 +1,4 @@
-use std::path::Path;
+use camino::Utf8Path;
 use std::str::FromStr;
 use std::sync::LazyLock;
 
@@ -33,7 +33,7 @@ pub fn check_fast_path(hook: &Hook) -> bool {
 pub async fn run_fast_path(
     store: &Store,
     hook: &Hook,
-    filenames: &[&Path],
+    filenames: &[&Utf8Path],
 ) -> anyhow::Result<(i32, Vec<u8>)> {
     match hook.repo() {
         Repo::Meta { .. } => run_meta_hook(store, hook, filenames).await,
@@ -50,7 +50,7 @@ pub async fn run_fast_path(
 async fn run_meta_hook(
     store: &Store,
     hook: &Hook,
-    filenames: &[&Path],
+    filenames: &[&Utf8Path],
 ) -> anyhow::Result<(i32, Vec<u8>)> {
     match hook.id.as_str() {
         "check-hooks-apply" => meta_hooks::check_hooks_apply(store, hook, filenames).await,

src/builtin/pre_commit_hooks/check_added_large_files.rs

@@ -1,5 +1,4 @@
-use std::path::{Path, PathBuf};
-
+use camino::{Utf8Path, Utf8PathBuf};
 use clap::Parser;
 use futures::StreamExt;
 use rustc_hash::FxHashSet;
@@ -10,11 +9,11 @@ use crate::run::CONCURRENCY;
 
 enum FileFilter {
     NoFilter,
-    Files(FxHashSet<PathBuf>),
+    Files(FxHashSet<Utf8PathBuf>),
 }
 
 impl FileFilter {
-    fn contains(&self, path: &Path) -> bool {
+    fn contains(&self, path: &Utf8Path) -> bool {
         match self {
             FileFilter::NoFilter => true,
             FileFilter::Files(files) => files.contains(path),
@@ -35,7 +34,7 @@ struct Args {
 
 pub(crate) async fn check_added_large_files(
     hook: &Hook,
-    filenames: &[&Path],
+    filenames: &[&Utf8Path],
 ) -> anyhow::Result<(i32, Vec<u8>)> {
     let args = Args::try_parse_from(hook.entry.resolve(None)?.iter().chain(&hook.args))?;
 
@@ -64,8 +63,7 @@ pub(crate) async fn check_added_large_files(
         if size > args.max_kb {
             anyhow::Ok(Some(format!(
                 "{} ({size} KB) exceeds {} KB\n",
-                filename.display(),
-                args.max_kb
+                filename, args.max_kb
             )))
         } else {
             anyhow::Ok(None)

src/builtin/pre_commit_hooks/check_json.rs

@@ -1,4 +1,4 @@
-use std::path::Path;
+use camino::Utf8Path;
 
 use anyhow::Result;
 use futures::StreamExt;
@@ -18,7 +18,7 @@ enum JsonValue {
     Null,
 }
 
-pub(crate) async fn check_json(hook: &Hook, filenames: &[&Path]) -> Result<(i32, Vec<u8>)> {
+pub(crate) async fn check_json(hook: &Hook, filenames: &[&Utf8Path]) -> Result<(i32, Vec<u8>)> {
     let mut tasks = futures::stream::iter(filenames)
         .map(async |filename| check_file(hook.project().relative_path(), filename).await)
         .buffered(*CONCURRENCY);
@@ -35,7 +35,7 @@ pub(crate) async fn check_json(hook: &Hook, filenames: &[&Path]) -> Result<(i32,
     Ok((code, output))
 }
 
-async fn check_file(file_base: &Path, filename: &Path) -> Result<(i32, Vec<u8>)> {
+async fn check_file(file_base: &Utf8Path, filename: &Utf8Path) -> Result<(i32, Vec<u8>)> {
     let file_path = file_base.join(filename);
     let content = fs_err::tokio::read(file_path).await?;
     if content.is_empty() {
@@ -53,7 +53,7 @@ async fn check_file(file_base: &Path, filename: &Path) -> Result<(i32, Vec<u8>)>
             Ok((0, Vec::new()))
         }
         Err(e) => {
-            let error_message = format!("{}: Failed to json decode ({e})\n", filename.display());
+            let error_message = format!("{filename}: Failed to json decode ({e})\n");
             Ok((1, error_message.into_bytes()))
         }
     }
@@ -151,25 +151,26 @@ impl<'de> Deserialize<'de> for JsonValue {
 #[cfg(test)]
 mod tests {
     use super::*;
-    use std::path::{Path, PathBuf};
+    use crate::path::IntoUtf8PathBuf;
+    use camino::{Utf8Path, Utf8PathBuf};
     use tempfile::tempdir;
 
     async fn create_test_file(
         dir: &tempfile::TempDir,
         name: &str,
         content: &[u8],
-    ) -> Result<PathBuf> {
+    ) -> Result<Utf8PathBuf> {
         let file_path = dir.path().join(name);
         fs_err::tokio::write(&file_path, content).await?;
-        Ok(file_path)
+        Ok(file_path.into_utf8_path_buf())
     }
 
     #[tokio::test]
     async fn test_valid_json() -> Result<()> {
         let dir = tempdir()?;
         let content = br#"{"key1": "value1", "key2": "value2"}"#;
         let file_path = create_test_file(&dir, "valid.json", content).await?;
-        let (code, output) = check_file(Path::new(""), &file_path).await?;
+        let (code, output) = check_file(Utf8Path::new(""), &file_path).await?;
         assert_eq!(code, 0);
         assert!(output.is_empty());
 
@@ -181,7 +182,7 @@ mod tests {
         let dir = tempdir()?;
         let content = br#"{"key1": "value1", "key2": "value2""#;
         let file_path = create_test_file(&dir, "invalid.json", content).await?;
-        let (code, output) = check_file(Path::new(""), &file_path).await?;
+        let (code, output) = check_file(Utf8Path::new(""), &file_path).await?;
         assert_eq!(code, 1);
         assert!(!output.is_empty());
 
@@ -193,7 +194,7 @@ mod tests {
         let dir = tempdir()?;
         let content = br#"{"key1": "value1", "key1": "value2"}"#;
         let file_path = create_test_file(&dir, "duplicate.json", content).await?;
-        let (code, output) = check_file(Path::new(""), &file_path).await?;
+        let (code, output) = check_file(Utf8Path::new(""), &file_path).await?;
         assert_eq!(code, 1);
         assert!(!output.is_empty());
 
@@ -205,7 +206,7 @@ mod tests {
         let dir = tempdir()?;
         let content = b"";
         let file_path = create_test_file(&dir, "empty.json", content).await?;
-        let (code, output) = check_file(Path::new(""), &file_path).await?;
+        let (code, output) = check_file(Utf8Path::new(""), &file_path).await?;
         assert_eq!(code, 0);
         assert!(output.is_empty());
 
@@ -217,7 +218,7 @@ mod tests {
         let dir = tempdir()?;
         let content = br#"[{"key1": "value1"}, {"key2": "value2"}]"#;
         let file_path = create_test_file(&dir, "valid_array.json", content).await?;
-        let (code, output) = check_file(Path::new(""), &file_path).await?;
+        let (code, output) = check_file(Utf8Path::new(""), &file_path).await?;
         assert_eq!(code, 0);
         assert!(output.is_empty());
 
@@ -229,7 +230,7 @@ mod tests {
         let dir = tempdir()?;
         let content = br#"{"key1": "value1", "key2": {"nested_key": 1, "nested_key": 2}}"#;
         let file_path = create_test_file(&dir, "nested_duplicate.json", content).await?;
-        let (code, output) = check_file(Path::new(""), &file_path).await?;
+        let (code, output) = check_file(Utf8Path::new(""), &file_path).await?;
         assert_eq!(code, 1);
         assert!(!output.is_empty());
 
@@ -246,7 +247,7 @@ mod tests {
         }
 
         let file_path = create_test_file(&dir, "deeply_nested.json", json.as_bytes()).await?;
-        let (code, output) = check_file(Path::new(""), &file_path).await?;
+        let (code, output) = check_file(Utf8Path::new(""), &file_path).await?;
         assert_eq!(code, 0);
         assert!(output.is_empty());