Make CharField prohibit surrogate characters (#7026) (#7067)

* CharField: Detect and prohibit surrogate characters

* CharField: Cover handling of surrogate characters

Author: hartwork

rest_framework/fields.py

@@ -36,6 +36,7 @@
 from rest_framework.settings import api_settings
 from rest_framework.utils import html, humanize_datetime, json, representation
 from rest_framework.utils.formatting import lazy_format
+from rest_framework.validators import ProhibitSurrogateCharactersValidator
 
 
 class empty:
@@ -818,6 +819,7 @@ def __init__(self, **kwargs):
         # ProhibitNullCharactersValidator is None on Django < 2.0
         if ProhibitNullCharactersValidator is not None:
             self.validators.append(ProhibitNullCharactersValidator())
+        self.validators.append(ProhibitSurrogateCharactersValidator())
 
     def run_validation(self, data=empty):
         # Test for the empty string here so that it does not get validated,

rest_framework/validators.py

@@ -167,6 +167,17 @@ def __repr__(self):
         )
 
 
+class ProhibitSurrogateCharactersValidator:
+    message = _('Surrogate characters are not allowed: U+{code_point:X}.')
+    code = 'surrogate_characters_not_allowed'
+
+    def __call__(self, value):
+        for surrogate_character in (ch for ch in str(value)
+                                    if 0xD800 <= ord(ch) <= 0xDFFF):
+            message = self.message.format(code_point=ord(surrogate_character))
+            raise ValidationError(message, code=self.code)
+
+
 class BaseUniqueForValidator:
     message = None
     missing_message = _('This field is required.')

tests/test_fields.py

@@ -758,6 +758,21 @@ def test_null_bytes(self):
                 'Null characters are not allowed.'
             ]
 
+    def test_surrogate_characters(self):
+        field = serializers.CharField()
+
+        for code_point, expected_message in (
+                (0xD800, 'Surrogate characters are not allowed: U+D800.'),
+                (0xDFFF, 'Surrogate characters are not allowed: U+DFFF.'),
+        ):
+            with pytest.raises(serializers.ValidationError) as exc_info:
+                field.run_validation(chr(code_point))
+            assert exc_info.value.detail[0].code == 'surrogate_characters_not_allowed'
+            assert str(exc_info.value.detail[0]) == expected_message
+
+        for code_point in (0xD800 - 1, 0xDFFF + 1):
+            field.run_validation(chr(code_point))
+
     def test_iterable_validators(self):
         """
         Ensure `validators` parameter is compatible with reasonable iterables.