test: add regression coverage for ordinal indicators (º, ª)

Issue #172 reported that slugify() doesn't transliterate the masculine/
feminine ordinal indicators (º, ª) common on Portuguese keyboards.
Verified against both supported transliteration backends
(text-unidecode>=1.3 and Unidecode>=1.1.1): both already map these
correctly (º->o, ª->a) via their Latin-1 Supplement tables, so the bug
no longer reproduces on the current codebase. Adding a regression test
to lock in the correct behavior and close out the issue.

Closes #172

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: mmaxjr

test.py

@@ -49,6 +49,13 @@ def test_accented_text(self):
         r = slugify(txt)
         self.assertEqual(r, "nin-hao-wo-shi-zhong-guo-ren")
 
+    def test_ordinal_indicators(self):
+        # Portuguese/Spanish masculine/feminine ordinal indicators (º, ª),
+        # common on Portuguese keyboards - see issue #172.
+        txt = '1º lugar, 1ª colocada'
+        r = slugify(txt)
+        self.assertEqual(r, "1o-lugar-1a-colocada")
+
     def test_accented_text_with_non_word_characters(self):
         txt = 'jaja---lol-méméméoo--a'
         r = slugify(txt)