diff --git a/src/uu/sort/src/sort.rs b/src/uu/sort/src/sort.rs
index 8c27910dce8..2a887249ee5 100644
--- a/src/uu/sort/src/sort.rs
+++ b/src/uu/sort/src/sort.rs
@@ -1846,11 +1846,12 @@ pub fn uumain(args: impl uucore::Args) -> UResult<()> {
         uucore::clap_localization::handle_clap_result_with_exit_code(uu_app(), processed_args, 2)?;
 
     // Prevent -o/--output to be specified multiple times
-    if matches
-        .get_occurrences::<OsString>(options::OUTPUT)
-        .is_some_and(|out| out.len() > 1)
-    {
-        return Err(SortError::MultipleOutputFiles.into());
+    if let Some(mut outputs) = matches.get_many::<OsString>(options::OUTPUT) {
+        if let Some(first) = outputs.next() {
+            if outputs.any(|out| out != first) {
+                return Err(SortError::MultipleOutputFiles.into());
+            }
+        }
     }
 
     settings.debug = matches.get_flag(options::DEBUG);
@@ -2616,17 +2617,17 @@ fn compare_by<'a>(
 }
 
 /// Compare two byte slices in ASCII case-insensitive order without allocating.
-/// We lower each byte on the fly so that binary input (including `NUL`) stays
+/// We upper each byte on the fly so that binary input (including `NUL`) stays
 /// untouched and we avoid locale-sensitive routines such as `strcasecmp`.
 fn ascii_case_insensitive_cmp(a: &[u8], b: &[u8]) -> Ordering {
     #[inline]
-    fn lower(byte: u8) -> u8 {
-        byte.to_ascii_lowercase()
+    fn fold(byte: u8) -> u8 {
+        byte.to_ascii_uppercase()
     }
 
     for (lhs, rhs) in a.iter().copied().zip(b.iter().copied()) {
-        let l = lower(lhs);
-        let r = lower(rhs);
+        let l = fold(lhs);
+        let r = fold(rhs);
         if l != r {
             return l.cmp(&r);
         }
diff --git a/tests/by-util/test_sort.rs b/tests/by-util/test_sort.rs
index 0106d719fad..0972cae0392 100644
--- a/tests/by-util/test_sort.rs
+++ b/tests/by-util/test_sort.rs
@@ -208,6 +208,24 @@ fn test_version_sort_stable() {
         .stdout_is("0.1\n0.02\n0.2\n0.002\n0.3\n");
 }
 
+#[test]
+fn test_ignore_case_orders_punctuation_after_letters() {
+    new_ucmd!()
+        .arg("-f")
+        .pipe_in("A\na\n_\n")
+        .succeeds()
+        .stdout_is("A\na\n_\n");
+}
+
+#[test]
+fn test_ignore_case_unique_orders_punctuation_after_letters() {
+    new_ucmd!()
+        .arg("-fu")
+        .pipe_in("a\n_\n")
+        .succeeds()
+        .stdout_is("a\n_\n");
+}
+
 #[test]
 fn test_human_numeric_whitespace() {
     test_helper(
@@ -1450,6 +1468,16 @@ fn test_multiple_output_files() {
         .stderr_is("sort: multiple output files specified\n");
 }
 
+#[test]
+// Test for GNU tests/sort/sort.pl "o3"
+fn test_duplicate_output_files_allowed() {
+    new_ucmd!()
+        .args(&["-o", "foo", "-o", "foo"])
+        .pipe_in("")
+        .succeeds()
+        .no_stderr();
+}
+
 #[test]
 fn test_output_file_with_leading_dash() {
     let test_cases = [