wc: fix word undercount with invalid byte sequences (#10348)

* wc: fix word undercount with invalid byte sequences

* wc: update utf8 test counts to account invalid byte sequences

* wc: remove unnecessary borrow in test

Author: dhr412

src/uu/wc/src/wc.rs

@@ -624,10 +624,18 @@ fn process_chunk<
     total.max_line_length = max(*current_len, total.max_line_length);
 }
 
-fn handle_error(error: BufReadDecoderError<'_>, total: &mut WordCount) -> Option<io::Error> {
+fn handle_error(
+    error: BufReadDecoderError<'_>,
+    total: &mut WordCount,
+    in_word: &mut bool,
+) -> Option<io::Error> {
     match error {
         BufReadDecoderError::InvalidByteSequence(bytes) => {
             total.bytes += bytes.len();
+            if !(*in_word) {
+                *in_word = true;
+                total.words += 1;
+            }
         }
         BufReadDecoderError::Io(e) => return Some(e),
     }
@@ -660,7 +668,7 @@ fn word_count_from_reader_specialized<
                 );
             }
             Err(e) => {
-                if let Some(e) = handle_error(e, &mut total) {
+                if let Some(e) = handle_error(e, &mut total, &mut in_word) {
                     return (total, Some(e));
                 }
             }

tests/by-util/test_wc.rs

@@ -65,7 +65,7 @@ fn test_utf8() {
         .args(&["-lwmcL"])
         .pipe_in_fixture("UTF_8_test.txt")
         .succeeds()
-        .stdout_is("    303    2119   22457   23025      79\n");
+        .stdout_is("    303    2178   22457   23025      79\n");
 }
 
 #[test]
@@ -826,6 +826,16 @@ fn wc_w_words_with_emoji_separator() {
         .stdout_contains("3");
 }
 
+#[test]
+fn test_invalid_byte_sequence_word_count() {
+    // wc should count invalid byte sequences as words
+    // Input: "a \xff b\n" should produce: 1 line, 3 words, 6 bytes
+    new_ucmd!()
+        .pipe_in([b'a', b' ', 0xff, b' ', b'b', b'\n'])
+        .succeeds()
+        .stdout_is("      1       3       6\n");
+}
+
 #[cfg(unix)]
 #[test]
 fn test_simd_respects_glibc_tunables() {