build_helper: add recursive_remove helper

`recursive_remove` is intended to be a wrapper around
`std::fs::remove_dir_all`, but which also allows the removal target to
be a non-directory entry, i.e. a file or a symlink. It also tries to
remove read-only attributes from filesystem entities on Windows.

Author: jieyouxu

src/build_helper/src/fs/mod.rs

@@ -0,0 +1,107 @@
+//! Misc filesystem related helpers for use by bootstrap and tools.
+
+use std::path::Path;
+use std::{fs, io};
+
+#[cfg(test)]
+mod tests;
+
+/// Helper to ignore [`std::io::ErrorKind::NotFound`], but still propagate other
+/// [`std::io::ErrorKind`]s.
+pub fn ignore_not_found<Op>(mut op: Op) -> io::Result<()>
+where
+    Op: FnMut() -> io::Result<()>,
+{
+    match op() {
+        Ok(()) => Ok(()),
+        Err(e) if e.kind() == io::ErrorKind::NotFound => Ok(()),
+        Err(e) => Err(e),
+    }
+}
+
+/// A [`std::fs::remove_dir_all`]-like helper recursively remotes filesystem entities starting at
+/// the provided root path. This helper supports being used on a single non-directory entity, e.g. a
+/// symlink or file.
+///
+/// - This will produce an error if the target path is not found.
+/// - Like [`std::fs::remove_dir_all`], this helper does not traverse symbolic links, will remove
+///   symbolic link itself.
+/// - This helper will try to remove even read-only files and directories on Windows.
+/// - This helper is **not** robust against races on the underlying filesystem, behavior is
+///   unspecified if this helper is called concurrently.
+/// - This helper is not robust against TOCTOU problems.
+///
+/// FIXME: this implementation is insufficiently robust to replace bootstrap's clean `rm_rf`
+/// implementation as this implementation currently does not perform retries.
+#[track_caller]
+pub fn recursive_remove<P: AsRef<Path>>(path: P) -> io::Result<()> {
+    let path = path.as_ref();
+    let metadata = fs::symlink_metadata(path)?;
+    if metadata.is_dir() {
+        for e in fs::read_dir(path)? {
+            let entry = e?;
+            let entry_metadata = fs::symlink_metadata(&entry.path())?;
+            if entry_metadata.is_dir() {
+                recursive_remove(entry.path())?;
+            } else {
+                recursive_remove_single(&entry.path())?;
+            }
+        }
+        fs::remove_dir(&path)?;
+    } else {
+        recursive_remove_single(path)?;
+    }
+
+    Ok(())
+}
+
+fn recursive_remove_single(path: &Path) -> io::Result<()> {
+    let metadata = fs::symlink_metadata(path)?;
+
+    if metadata.is_file() {
+        try_remove_op_override_perm(fs::remove_file, path, metadata)
+    } else if metadata.is_symlink() {
+        #[cfg(windows)]
+        {
+            use std::os::windows::fs::FileTypeExt;
+            if metadata.file_type().is_symlink_dir() {
+                try_remove_op_override_perm(fs::remove_dir, path, metadata)
+            } else {
+                try_remove_op_override_perm(fs::remove_file, path, metadata)
+            }
+        }
+        #[cfg(not(windows))]
+        {
+            try_remove_op_override_perm(fs::remove_file, path, metadata)
+        }
+    } else if metadata.is_dir() {
+        unreachable!("`recursive_remove_single` does not expect dir: {}", path.display())
+    } else {
+        // Unrecognized file type, let it go.
+        Ok(())
+    }
+}
+
+fn try_remove_op_override_perm<'p, Op>(
+    mut remove_op: Op,
+    path: &'p Path,
+    metadata: fs::Metadata,
+) -> io::Result<()>
+where
+    Op: FnMut(&'p Path) -> io::Result<()>,
+{
+    match remove_op(path) {
+        Ok(()) => Ok(()),
+        // On Windows, we need to remove read-only attribute for a read-only file.
+        #[cfg(windows)]
+        Err(e) if e.kind() == io::ErrorKind::PermissionDenied => {
+            // Try to remove the read-only attribute.
+            let mut perms = metadata.permissions();
+            perms.set_readonly(false);
+            fs::set_permissions(path, perms)?;
+            // Then try again.
+            remove_op(path)
+        }
+        Err(e) => Err(e),
+    }
+}

src/build_helper/src/fs/tests.rs

@@ -0,0 +1,206 @@
+#![deny(unused_must_use)]
+
+use std::{env, fs, io};
+
+use super::recursive_remove;
+
+mod recursive_remove_tests {
+    use super::*;
+
+    // Basic cases
+
+    #[test]
+    fn nonexistent_path() {
+        let tmpdir = env::temp_dir();
+        let path = tmpdir.join("__INTERNAL_BOOTSTRAP_nonexistent_path");
+        assert!(fs::symlink_metadata(&path).is_err_and(|e| e.kind() == io::ErrorKind::NotFound));
+        assert!(recursive_remove(&path).is_err_and(|e| e.kind() == io::ErrorKind::NotFound));
+    }
+
+    #[test]
+    fn file() {
+        let tmpdir = env::temp_dir();
+        let path = tmpdir.join("__INTERNAL_BOOTSTRAP_file");
+        fs::write(&path, b"").unwrap();
+        assert!(fs::symlink_metadata(&path).is_ok());
+        assert!(recursive_remove(&path).is_ok());
+        assert!(fs::symlink_metadata(&path).is_err_and(|e| e.kind() == io::ErrorKind::NotFound));
+    }
+
+    mod dir_tests {
+        use super::*;
+
+        #[test]
+        fn dir_empty() {
+            let tmpdir = env::temp_dir();
+            let path = tmpdir.join("__INTERNAL_BOOTSTRAP_dir_tests_dir_empty");
+            fs::create_dir_all(&path).unwrap();
+            assert!(fs::symlink_metadata(&path).is_ok());
+            assert!(recursive_remove(&path).is_ok());
+            assert!(
+                fs::symlink_metadata(&path).is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+        }
+
+        #[test]
+        fn dir_recursive() {
+            let tmpdir = env::temp_dir();
+            let path = tmpdir.join("__INTERNAL_BOOTSTRAP_dir_tests_dir_recursive");
+            fs::create_dir_all(&path).unwrap();
+            assert!(fs::symlink_metadata(&path).is_ok());
+
+            let file_a = path.join("a.txt");
+            fs::write(&file_a, b"").unwrap();
+            assert!(fs::symlink_metadata(&file_a).is_ok());
+
+            let dir_b = path.join("b");
+            fs::create_dir_all(&dir_b).unwrap();
+            assert!(fs::symlink_metadata(&dir_b).is_ok());
+
+            let file_c = dir_b.join("c.rs");
+            fs::write(&file_c, b"").unwrap();
+            assert!(fs::symlink_metadata(&file_c).is_ok());
+
+            assert!(recursive_remove(&path).is_ok());
+
+            assert!(
+                fs::symlink_metadata(&file_a).is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+            assert!(
+                fs::symlink_metadata(&dir_b).is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+            assert!(
+                fs::symlink_metadata(&file_c).is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+        }
+    }
+
+    /// Check that [`recursive_remove`] does not traverse symlinks and only removes symlinks
+    /// themselves.
+    ///
+    /// Symlink-to-file versus symlink-to-dir is a distinction that's important on Windows, but not
+    /// on Unix.
+    mod symlink_tests {
+        use super::*;
+
+        #[cfg(unix)]
+        #[test]
+        fn unix_symlink() {
+            let tmpdir = env::temp_dir();
+            let path = tmpdir.join("__INTERNAL_BOOTSTRAP_symlink_tests_unix_symlink");
+            let symlink_path =
+                tmpdir.join("__INTERNAL_BOOTSTRAP__symlink_tests_unix_symlink_symlink");
+            fs::write(&path, b"").unwrap();
+
+            assert!(fs::symlink_metadata(&path).is_ok());
+            assert!(
+                fs::symlink_metadata(&symlink_path)
+                    .is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+
+            std::os::unix::fs::symlink(&path, &symlink_path).unwrap();
+
+            assert!(recursive_remove(&symlink_path).is_ok());
+
+            // Check that the symlink got removed...
+            assert!(
+                fs::symlink_metadata(&symlink_path)
+                    .is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+            // ... but pointed-to file still exists.
+            assert!(fs::symlink_metadata(&path).is_ok());
+
+            fs::remove_file(&path).unwrap();
+        }
+
+        #[cfg(windows)]
+        #[test]
+        fn windows_symlink_to_file() {
+            let tmpdir = env::temp_dir();
+            let path = tmpdir.join("__INTERNAL_BOOTSTRAP_symlink_tests_windows_symlink_to_file");
+            let symlink_path = tmpdir
+                .join("__INTERNAL_BOOTSTRAP_SYMLINK_symlink_tests_windows_symlink_to_file_symlink");
+            fs::write(&path, b"").unwrap();
+
+            assert!(fs::symlink_metadata(&path).is_ok());
+            assert!(
+                fs::symlink_metadata(&symlink_path)
+                    .is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+
+            std::os::windows::fs::symlink_file(&path, &symlink_path).unwrap();
+
+            assert!(recursive_remove(&symlink_path).is_ok());
+
+            // Check that the symlink-to-file got removed...
+            assert!(
+                fs::symlink_metadata(&symlink_path)
+                    .is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+            // ... but pointed-to file still exists.
+            assert!(fs::symlink_metadata(&path).is_ok());
+
+            fs::remove_file(&path).unwrap();
+        }
+
+        #[cfg(windows)]
+        #[test]
+        fn windows_symlink_to_dir() {
+            let tmpdir = env::temp_dir();
+            let path = tmpdir.join("__INTERNAL_BOOTSTRAP_symlink_tests_windows_symlink_to_dir");
+            let symlink_path =
+                tmpdir.join("__INTERNAL_BOOTSTRAP_symlink_tests_windows_symlink_to_dir_symlink");
+            fs::create_dir_all(&path).unwrap();
+
+            assert!(fs::symlink_metadata(&path).is_ok());
+            assert!(
+                fs::symlink_metadata(&symlink_path)
+                    .is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+
+            std::os::windows::fs::symlink_dir(&path, &symlink_path).unwrap();
+
+            assert!(recursive_remove(&symlink_path).is_ok());
+
+            // Check that the symlink-to-dir got removed...
+            assert!(
+                fs::symlink_metadata(&symlink_path)
+                    .is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+            // ... but pointed-to dir still exists.
+            assert!(fs::symlink_metadata(&path).is_ok());
+
+            fs::remove_dir_all(&path).unwrap();
+        }
+    }
+
+    /// Read-only file and directories only need special handling on Windows.
+    #[cfg(windows)]
+    mod readonly_tests {
+        use super::*;
+
+        #[test]
+        fn overrides_readonly() {
+            let tmpdir = env::temp_dir();
+            let path = tmpdir.join("__INTERNAL_BOOTSTRAP_readonly_tests_overrides_readonly");
+            fs::write(&path, b"").unwrap();
+
+            let mut perms = fs::symlink_metadata(&path).unwrap().permissions();
+            perms.set_readonly(true);
+            fs::set_permissions(&path, perms).unwrap();
+
+            // Check that file exists but is read-only, and that normal `std::fs::remove_file` fails
+            // to delete the file.
+            assert!(fs::symlink_metadata(&path).is_ok_and(|m| m.permissions().readonly()));
+            assert!(
+                fs::remove_file(&path).is_err_and(|e| e.kind() == io::ErrorKind::PermissionDenied)
+            );
+
+            assert!(recursive_remove(&path).is_ok());
+
+            assert!(
+                fs::symlink_metadata(&path).is_err_and(|e| e.kind() == io::ErrorKind::NotFound)
+            );
+        }
+    }
+}

src/build_helper/src/lib.rs

@@ -2,6 +2,7 @@
 
 pub mod ci;
 pub mod drop_bomb;
+pub mod fs;
 pub mod git;
 pub mod metrics;
 pub mod stage0_parser;