printf: support for extract chars

sylvestre · jtracey · commit 8f4c286c75b5 · 2025-01-24T17:40:10.000-05:00
Should fix tests/printf/printf-mb.sh
diff --git a/src/uucore/src/lib/features/format/argument.rs b/src/uucore/src/lib/features/format/argument.rs
@@ -49,6 +49,7 @@ impl<'a, T: Iterator<Item = &'a FormatArgument>> ArgumentIter<'a> for T {
             _ => b'\0',
         }
     }
+
     fn get_u64(&mut self) -> u64 {
         let Some(next) = self.next() else {
             return 0;
@@ -57,13 +58,21 @@ impl<'a, T: Iterator<Item = &'a FormatArgument>> ArgumentIter<'a> for T {
             FormatArgument::UnsignedInt(n) => *n,
             FormatArgument::Unparsed(s) => {
                 // Check if the string is a character literal enclosed in quotes
-                if s.starts_with(['"', '\'']) && s.len() > 2 {
-                    // Extract the content between the quotes safely
-                    let chars: Vec<char> =
-                        s.trim_matches(|c| c == '"' || c == '\'').chars().collect();
-                    if chars.len() == 1 {
-                        return chars[0] as u64; // Return the Unicode code point
+                if s.starts_with(['"', '\'']) {
+                    // Extract the content between the quotes safely using chars
+                    let mut chars = s.trim_matches(|c| c == '"' || c == '\'').chars();
+                    if let Some(first_char) = chars.next() {
+                        if chars.clone().count() > 0 {
+                            // Emit a warning if there are additional characters
+                            let remaining: String = chars.collect();
+                            show_warning!(
+                                "{}: character(s) following character constant have been ignored",
+                                remaining
+                            );
+                        }
+                        return first_char as u64; // Use only the first character
                     }
+                    return 0; // Empty quotes
                 }
                 extract_value(ParsedNumber::parse_u64(s), s)
             }
diff --git a/tests/by-util/test_printf.rs b/tests/by-util/test_printf.rs
@@ -869,12 +869,27 @@ fn float_switch_switch_decimal_scientific() {
 
 #[test]
 fn mb_input() {
-    for format in ["\"á", "\'á"] {
+    for format in ["\"á", "\'á", "'\u{e1}"] {
         new_ucmd!()
             .args(&["%04x\n", format])
             .succeeds()
             .stdout_only("00e1\n");
     }
+
+    let cases = vec![
+        ("\"á=", "="),
+        ("\'á-", "-"),
+        ("\'á=-==", "=-=="),
+        ("'\u{e1}++", "++"),
+    ];
+
+    for (format, expected) in cases {
+        new_ucmd!()
+            .args(&["%04x\n", format])
+            .succeeds()
+            .stdout_is("00e1\n")
+            .stderr_is(format!("printf: warning: {expected}: character(s) following character constant have been ignored\n"));
+    }
 }
 
 #[test]
