Fixed #29249 -- Made JSON and YAML serializers use Unicode by default.

6 files changed, 25 insertions, 9 deletions

diff --git a/django/core/serializers/json.py b/django/core/serializers/json.py
index 5d35440a6c..886e8f894c 100644
--- a/django/core/serializers/json.py
+++ b/django/core/serializers/json.py
@@ -29,6 +29,7 @@ class Serializer(PythonSerializer):
             # Prevent trailing spaces
             self.json_kwargs['separators'] = (',', ': ')
         self.json_kwargs.setdefault('cls', DjangoJSONEncoder)
+        self.json_kwargs.setdefault('ensure_ascii', False)
 
     def start_serialization(self):
         self._init_options()
diff --git a/django/core/serializers/pyyaml.py b/django/core/serializers/pyyaml.py
index 778c933584..8f89a633c8 100644
--- a/django/core/serializers/pyyaml.py
+++ b/django/core/serializers/pyyaml.py
@@ -57,6 +57,7 @@ class Serializer(PythonSerializer):
             super().handle_field(obj, field)
 
     def end_serialization(self):
+        self.options.setdefault('allow_unicode', True)
         yaml.dump(self.objects, self.stream, Dumper=DjangoSafeDumper, **self.options)
 
     def getvalue(self):
diff --git a/docs/releases/3.1.txt b/docs/releases/3.1.txt
index fbefef589f..16cb9c4e6e 100644
--- a/docs/releases/3.1.txt
+++ b/docs/releases/3.1.txt
@@ -668,6 +668,11 @@ Miscellaneous
 * The undocumented ``version`` parameter to the
   :class:`~django.contrib.gis.db.models.functions.AsKML` function is removed.
 
+* :ref:`JSON and YAML serializers <serialization-formats>`, used by
+  :djadmin:`dumpdata`, now dump all data with Unicode by default. If you need
+  the previous behavior, pass ``ensure_ascii=True`` to JSON serializer, or
+  ``allow_unicode=False`` to YAML serializer.
+
 .. _deprecated-features-3.1:
 
 Features deprecated in 3.1
diff --git a/docs/topics/serialization.txt b/docs/topics/serialization.txt
index 93af47ef28..00322ac64b 100644
--- a/docs/topics/serialization.txt
+++ b/docs/topics/serialization.txt
@@ -274,6 +274,11 @@ function::
 Also note that GeoDjango provides a :doc:`customized GeoJSON serializer
 </ref/contrib/gis/serializers>`.
 
+.. versionchanged:: 3.1
+
+    All data is now dumped with Unicode. If you need the previous behavior,
+    pass ``ensure_ascii=True`` to the ``serializers.serialize()`` function.
+
 ``DjangoJSONEncoder``
 ~~~~~~~~~~~~~~~~~~~~~
 
@@ -315,6 +320,11 @@ again a mapping with the key being name of the field and the value the value::
 
 Referential fields are again represented by the PK or sequence of PKs.
 
+.. versionchanged:: 3.1
+
+    All data is now dumped with Unicode. If you need the previous behavior,
+    pass ``allow_unicode=False`` to the ``serializers.serialize()`` function.
+
 .. _topics-serialization-natural-keys:
 
 Natural keys
diff --git a/tests/serializers/test_xml.py b/tests/serializers/test_xml.py
index 4bbfcf18b2..b11cfdd864 100644
--- a/tests/serializers/test_xml.py
+++ b/tests/serializers/test_xml.py
@@ -4,7 +4,6 @@ from django.core import serializers
 from django.core.serializers.xml_serializer import DTDForbidden
 from django.test import TestCase, TransactionTestCase
 
-from .models import Author
 from .tests import SerializersTestBase, SerializersTransactionTestBase
 
 
@@ -87,13 +86,6 @@ class XmlSerializerTestCase(SerializersTestBase, TestCase):
         with self.assertRaises(DTDForbidden):
             next(serializers.deserialize('xml', xml))
 
-    def test_unicode_serialization(self):
-        unicode_name = 'יוניקוד'
-        data = serializers.serialize('xml', [Author(name=unicode_name)])
-        self.assertIn(unicode_name, data)
-        objs = list(serializers.deserialize('xml', data))
-        self.assertEqual(objs[0].object.name, unicode_name)
-
 
 class XmlSerializerTransactionTestCase(SerializersTransactionTestBase, TransactionTestCase):
     serializer_name = "xml"
diff --git a/tests/serializers/tests.py b/tests/serializers/tests.py
index 863ee00e89..9d1be81146 100644
--- a/tests/serializers/tests.py
+++ b/tests/serializers/tests.py
@@ -202,7 +202,7 @@ class SerializersTestBase:
         for field_name in valid_fields:
             self.assertTrue(self._get_field_values(serial_str, field_name))
 
-    def test_serialize_unicode(self):
+    def test_serialize_unicode_roundtrip(self):
         """Unicode makes the roundtrip intact"""
         actor_name = "Za\u017c\u00f3\u0142\u0107"
         movie_title = 'G\u0119\u015bl\u0105 ja\u017a\u0144'
@@ -219,6 +219,13 @@ class SerializersTestBase:
         mv_obj = obj_list[0].object
         self.assertEqual(mv_obj.title, movie_title)
 
+    def test_unicode_serialization(self):
+        unicode_name = 'יוניקוד'
+        data = serializers.serialize(self.serializer_name, [Author(name=unicode_name)])
+        self.assertIn(unicode_name, data)
+        objs = list(serializers.deserialize(self.serializer_name, data))
+        self.assertEqual(objs[0].object.name, unicode_name)
+
     def test_serialize_progressbar(self):
         fake_stdout = StringIO()
         serializers.serialize(