🐛 Fix too small JSON payload triggering md with high ratio (#59)

Ousret · Jul 14, 2021 · 595514e · 595514e
1 parent 53b2dab
commit 595514e
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 1 deletion.
diff --git a/charset_normalizer/md.py b/charset_normalizer/md.py
@@ -56,7 +56,7 @@ def eligible(self, character: str) -> bool:
     def feed(self, character: str) -> None:
         self._character_count += 1
 
-        if character != self._last_printable_char and character not in ["<", ">", "=", ":", "/", "&", ";"]:
+        if character != self._last_printable_char and character not in ["<", ">", "=", ":", "/", "&", ";", "{", "}", "[", "]"]:
             if is_punctuation(character):
                 self._punctuation_count += 1
             elif character.isdigit() is False and is_symbol(character):

diff --git a/tests/test_on_byte.py b/tests/test_on_byte.py
@@ -27,6 +27,18 @@ def test_empty_bytes(self):
             len(r.alphabets)
         )
 
+    def test_on_empty_json(self):
+
+        with self.subTest("Detecting empty JSON as ASCII"):
+            results = from_bytes(b"{}").best()
+            self.assertIsNotNone(
+                results.best()
+            )
+            self.assertEqual(
+                results.best().encoding,
+                "ascii"
+            )
+
     def test_bom_detection(self):
         with self.subTest('GB18030 UNAVAILABLE SIG'):
             self.assertFalse(