sort: --files0-from should not panic when invalid utf-8

Author: sylvestre

src/uu/sort/src/sort.rs

@@ -1291,26 +1291,34 @@ pub fn uumain(args: impl uucore::Args) -> UResult<()> {
         let reader = open_with_open_failed_error(&files0_from)?;
         let buf_reader = BufReader::new(reader);
         for (line_num, line) in buf_reader.split(b'\0').flatten().enumerate() {
-            let f = std::str::from_utf8(&line)
-                .expect("Could not parse string from zero terminated input.");
-            match f {
-                STDIN_FILE => {
-                    return Err(SortError::MinusInStdIn.into());
+            // Handle filenames as raw bytes to support non-UTF-8 paths
+            #[cfg(unix)]
+            let filename = {
+                use std::os::unix::ffi::OsStrExt;
+                OsStr::from_bytes(&line).to_owned()
+            };
+            #[cfg(not(unix))]
+            let filename = {
+                // On non-Unix systems, convert to UTF-8 with replacement chars
+                match std::str::from_utf8(&line) {
+                    Ok(s) => OsString::from(s),
+                    Err(_) => OsString::from(String::from_utf8_lossy(&line).into_owned()),
                 }
-                "" => {
-                    return Err(SortError::ZeroLengthFileName {
-                        file: files0_from,
-                        line_num: line_num + 1,
-                    }
-                    .into());
+            };
+
+            // Check for special cases using bytes comparison
+            if line == b"-" {
+                return Err(SortError::MinusInStdIn.into());
+            }
+            if line.is_empty() {
+                return Err(SortError::ZeroLengthFileName {
+                    file: files0_from,
+                    line_num: line_num + 1,
                 }
-                _ => {}
+                .into());
             }
 
-            files.push(OsString::from(
-                std::str::from_utf8(&line)
-                    .expect("Could not parse string from zero terminated input."),
-            ));
+            files.push(filename);
         }
         if files.is_empty() {
             return Err(SortError::EmptyInputFile { file: files0_from }.into());

tests/by-util/test_sort.rs

@@ -1544,6 +1544,37 @@ fn test_files0_from_zero_length() {
         .stderr_only("sort: -:2: invalid zero-length file name\n");
 }
 
+#[test]
+#[cfg(unix)]
+fn test_files0_from_non_utf8_filename() {
+    use std::ffi::OsStr;
+    use std::os::unix::ffi::OsStrExt;
+    use std::path::Path;
+    let (at, mut ucmd) = at_and_ucmd!();
+
+    at.touch("file1.txt");
+    at.append("file1.txt", "zebra\n");
+    at.touch("file2.txt");
+    at.append("file2.txt", "apple\n");
+
+    // Create a file with non-UTF-8 bytes in filename
+    let non_utf8_name = OsStr::from_bytes(b"\xff\xfefile3.txt");
+    let non_utf8_path = at.plus_as_string("").to_owned() + "/";
+    let full_path = Path::new(&non_utf8_path).join(non_utf8_name);
+    std::fs::write(&full_path, "banana\n").unwrap();
+
+    // Create files0-from input containing the non-UTF-8 filename bytes
+    let mut files0_input = Vec::new();
+    files0_input.extend_from_slice(b"file1.txt\0");
+    files0_input.extend_from_slice(b"\xff\xfefile3.txt\0");
+    files0_input.extend_from_slice(b"file2.txt\0");
+
+    ucmd.args(&["--files0-from", "-"])
+        .pipe_in(files0_input)
+        .succeeds()
+        .stdout_is("apple\nbanana\nzebra\n");
+}
+
 #[test]
 // Test for GNU tests/sort/sort-float.sh
 fn test_g_float() {