Fixed transaction handling for a number of operations on related objects.

Thanks Anssi and Aymeric for the reviews. Refs #21174.

django/contrib/contenttypes/fields.py

@@ -394,9 +394,12 @@ class ReverseGenericRelatedObjectsDescriptor(object):
 
     def __set__(self, instance, value):
         manager = self.__get__(instance)
-        manager.clear()
+
-        for obj in value:
+        db = router.db_for_write(manager.model, instance=manager.instance)
-            manager.add(obj)
+        with transaction.atomic(using=db, savepoint=False):
+            manager.clear()
+            for obj in value:
+                manager.add(obj)
 
 
 def create_generic_related_manager(superclass):
@@ -474,12 +477,14 @@ def create_generic_related_manager(superclass):
                     self.prefetch_cache_name)
 
         def add(self, *objs):
-            for obj in objs:
+            db = router.db_for_write(self.model, instance=self.instance)
-                if not isinstance(obj, self.model):
+            with transaction.atomic(using=db, savepoint=False):
-                    raise TypeError("'%s' instance expected" % self.model._meta.object_name)
+                for obj in objs:
-                setattr(obj, self.content_type_field_name, self.content_type)
+                    if not isinstance(obj, self.model):
-                setattr(obj, self.object_id_field_name, self.pk_val)
+                        raise TypeError("'%s' instance expected" % self.model._meta.object_name)
-                obj.save()
+                    setattr(obj, self.content_type_field_name, self.content_type)
+                    setattr(obj, self.object_id_field_name, self.pk_val)
+                    obj.save()
         add.alters_data = True
 
         def remove(self, *objs, **kwargs):
@@ -498,6 +503,8 @@ def create_generic_related_manager(superclass):
             db = router.db_for_write(self.model, instance=self.instance)
             queryset = queryset.using(db)
             if bulk:
+                # `QuerySet.delete()` creates its own atomic block which
+                # contains the `pre_delete` and `post_delete` signal handlers.
                 queryset.delete()
             else:
                 with transaction.atomic(using=db, savepoint=False):

django/db/models/fields/related.py

@@ -735,6 +735,7 @@ def create_foreign_related_manager(superclass, rel_field, rel_model):
                 db = router.db_for_write(self.model, instance=self.instance)
                 queryset = queryset.using(db)
                 if bulk:
+                    # `QuerySet.update()` is intrinsically atomic.
                     queryset.update(**{rel_field.name: None})
                 else:
                     with transaction.atomic(using=db, savepoint=False):
@@ -763,11 +764,14 @@ class ForeignRelatedObjectsDescriptor(object):
 
     def __set__(self, instance, value):
         manager = self.__get__(instance)
-        # If the foreign key can support nulls, then completely clear the related set.
+
-        # Otherwise, just move the named objects into the set.
+        db = router.db_for_write(manager.model, instance=manager.instance)
-        if self.related.field.null:
+        with transaction.atomic(using=db, savepoint=False):
-            manager.clear()
+            # If the foreign key can support nulls, then completely clear the related set.
-        manager.add(*value)
+            # Otherwise, just move the named objects into the set.
+            if self.related.field.null:
+                manager.clear()
+            manager.add(*value)
 
     @cached_property
     def related_manager_cls(self):
@@ -901,11 +905,14 @@ def create_many_related_manager(superclass, rel):
                     "Cannot use add() on a ManyToManyField which specifies an intermediary model. Use %s.%s's Manager instead." %
                     (opts.app_label, opts.object_name)
                 )
-            self._add_items(self.source_field_name, self.target_field_name, *objs)
 
-            # If this is a symmetrical m2m relation to self, add the mirror entry in the m2m table
+            db = router.db_for_write(self.through, instance=self.instance)
-            if self.symmetrical:
+            with transaction.atomic(using=db, savepoint=False):
-                self._add_items(self.target_field_name, self.source_field_name, *objs)
+                self._add_items(self.source_field_name, self.target_field_name, *objs)
+
+                # If this is a symmetrical m2m relation to self, add the mirror entry in the m2m table
+                if self.symmetrical:
+                    self._add_items(self.target_field_name, self.source_field_name, *objs)
         add.alters_data = True
 
         def remove(self, *objs):
@@ -920,17 +927,17 @@ def create_many_related_manager(superclass, rel):
 
         def clear(self):
             db = router.db_for_write(self.through, instance=self.instance)
+            with transaction.atomic(using=db, savepoint=False):
+                signals.m2m_changed.send(sender=self.through, action="pre_clear",
+                    instance=self.instance, reverse=self.reverse,
+                    model=self.model, pk_set=None, using=db)
 
-            signals.m2m_changed.send(sender=self.through, action="pre_clear",
+                filters = self._build_remove_filters(super(ManyRelatedManager, self).get_queryset().using(db))
-                instance=self.instance, reverse=self.reverse,
+                self.through._default_manager.using(db).filter(filters).delete()
-                model=self.model, pk_set=None, using=db)
 
-            filters = self._build_remove_filters(super(ManyRelatedManager, self).get_queryset().using(db))
+                signals.m2m_changed.send(sender=self.through, action="post_clear",
-            self.through._default_manager.using(db).filter(filters).delete()
+                    instance=self.instance, reverse=self.reverse,
-
+                    model=self.model, pk_set=None, using=db)
-            signals.m2m_changed.send(sender=self.through, action="post_clear",
-                instance=self.instance, reverse=self.reverse,
-                model=self.model, pk_set=None, using=db)
         clear.alters_data = True
 
         def create(self, **kwargs):
@@ -990,35 +997,39 @@ def create_many_related_manager(superclass, rel):
                         )
                     else:
                         new_ids.add(obj)
+
                 db = router.db_for_write(self.through, instance=self.instance)
-                vals = self.through._default_manager.using(db).values_list(target_field_name, flat=True)
+                vals = (self.through._default_manager.using(db)
-                vals = vals.filter(**{
+                        .values_list(target_field_name, flat=True)
-                    source_field_name: self.related_val[0],
+                        .filter(**{
-                    '%s__in' % target_field_name: new_ids,
+                            source_field_name: self.related_val[0],
-                })
+                            '%s__in' % target_field_name: new_ids,
+                        }))
                 new_ids = new_ids - set(vals)
 
-                if self.reverse or source_field_name == self.source_field_name:
+                with transaction.atomic(using=db, savepoint=False):
-                    # Don't send the signal when we are inserting the
+                    if self.reverse or source_field_name == self.source_field_name:
-                    # duplicate data row for symmetrical reverse entries.
+                        # Don't send the signal when we are inserting the
-                    signals.m2m_changed.send(sender=self.through, action='pre_add',
+                        # duplicate data row for symmetrical reverse entries.
-                        instance=self.instance, reverse=self.reverse,
+                        signals.m2m_changed.send(sender=self.through, action='pre_add',
-                        model=self.model, pk_set=new_ids, using=db)
+                            instance=self.instance, reverse=self.reverse,
-                # Add the ones that aren't there already
+                            model=self.model, pk_set=new_ids, using=db)
-                self.through._default_manager.using(db).bulk_create([
-                    self.through(**{
-                        '%s_id' % source_field_name: self.related_val[0],
-                        '%s_id' % target_field_name: obj_id,
-                    })
-                    for obj_id in new_ids
-                ])
 
-                if self.reverse or source_field_name == self.source_field_name:
+                    # Add the ones that aren't there already
-                    # Don't send the signal when we are inserting the
+                    self.through._default_manager.using(db).bulk_create([
-                    # duplicate data row for symmetrical reverse entries.
+                        self.through(**{
-                    signals.m2m_changed.send(sender=self.through, action='post_add',
+                            '%s_id' % source_field_name: self.related_val[0],
-                        instance=self.instance, reverse=self.reverse,
+                            '%s_id' % target_field_name: obj_id,
-                        model=self.model, pk_set=new_ids, using=db)
+                        })
+                        for obj_id in new_ids
+                    ])
+
+                    if self.reverse or source_field_name == self.source_field_name:
+                        # Don't send the signal when we are inserting the
+                        # duplicate data row for symmetrical reverse entries.
+                        signals.m2m_changed.send(sender=self.through, action='post_add',
+                            instance=self.instance, reverse=self.reverse,
+                            model=self.model, pk_set=new_ids, using=db)
 
         def _remove_items(self, source_field_name, target_field_name, *objs):
             # source_field_name: the PK colname in join table for the source object
@@ -1037,23 +1048,23 @@ def create_many_related_manager(superclass, rel):
                     old_ids.add(obj)
 
             db = router.db_for_write(self.through, instance=self.instance)
+            with transaction.atomic(using=db, savepoint=False):
+                # Send a signal to the other end if need be.
+                signals.m2m_changed.send(sender=self.through, action="pre_remove",
+                    instance=self.instance, reverse=self.reverse,
+                    model=self.model, pk_set=old_ids, using=db)
+                target_model_qs = super(ManyRelatedManager, self).get_queryset()
+                if target_model_qs._has_filters():
+                    old_vals = target_model_qs.using(db).filter(**{
+                        '%s__in' % self.target_field.related_field.attname: old_ids})
+                else:
+                    old_vals = old_ids
+                filters = self._build_remove_filters(old_vals)
+                self.through._default_manager.using(db).filter(filters).delete()
 
-            # Send a signal to the other end if need be.
+                signals.m2m_changed.send(sender=self.through, action="post_remove",
-            signals.m2m_changed.send(sender=self.through, action="pre_remove",
+                    instance=self.instance, reverse=self.reverse,
-                instance=self.instance, reverse=self.reverse,
+                    model=self.model, pk_set=old_ids, using=db)
-                model=self.model, pk_set=old_ids, using=db)
-            target_model_qs = super(ManyRelatedManager, self).get_queryset()
-            if target_model_qs._has_filters():
-                old_vals = target_model_qs.using(db).filter(**{
-                    '%s__in' % self.target_field.related_field.attname: old_ids})
-            else:
-                old_vals = old_ids
-            filters = self._build_remove_filters(old_vals)
-            self.through._default_manager.using(db).filter(filters).delete()
-
-            signals.m2m_changed.send(sender=self.through, action="post_remove",
-                instance=self.instance, reverse=self.reverse,
-                model=self.model, pk_set=old_ids, using=db)
 
     return ManyRelatedManager
 
@@ -1103,8 +1114,11 @@ class ManyRelatedObjectsDescriptor(object):
             raise AttributeError("Cannot set values on a ManyToManyField which specifies an intermediary model. Use %s.%s's Manager instead." % (opts.app_label, opts.object_name))
 
         manager = self.__get__(instance)
-        manager.clear()
+
-        manager.add(*value)
+        db = router.db_for_write(manager.through, instance=manager.instance)
+        with transaction.atomic(using=db, savepoint=False):
+            manager.clear()
+            manager.add(*value)
 
 
 class ReverseManyRelatedObjectsDescriptor(object):
@@ -1157,11 +1171,15 @@ class ReverseManyRelatedObjectsDescriptor(object):
             raise AttributeError("Cannot set values on a ManyToManyField which specifies an intermediary model.  Use %s.%s's Manager instead." % (opts.app_label, opts.object_name))
 
         manager = self.__get__(instance)
+
         # clear() can change expected output of 'value' queryset, we force evaluation
         # of queryset before clear; ticket #19816
         value = tuple(value)
-        manager.clear()
+
-        manager.add(*value)
+        db = router.db_for_write(manager.through, instance=manager.instance)
+        with transaction.atomic(using=db, savepoint=False):
+            manager.clear()
+            manager.add(*value)
 
 
 class ForeignObjectRel(object):

docs/releases/1.8.txt

@@ -168,7 +168,19 @@ Backwards incompatible changes in 1.8
     deprecation timeline for a given feature, its removal may appear as a
     backwards incompatible change.
 
-...
+* Some operations on related objects such as
+  :meth:`~django.db.models.fields.related.RelatedManager.add()` or
+  :ref:`direct assignment<direct-assignment>` ran multiple data modifying
+  queries without wrapping them in transactions. To reduce the risk of data
+  corruption, all data modifying methods that affect multiple related objects
+  (i.e. ``add()``, ``remove()``, ``clear()``, and
+  :ref:`direct assignment<direct-assignment>`) now perform their data modifying
+  queries from within a transaction, provided your database supports
+  transactions.
+
+  This has one backwards incompatible side effect, signal handlers triggered
+  from these methods are now executed within the method's transaction and
+  any exception in a signal handler will prevent the whole operation.
 
 Miscellaneous
 ~~~~~~~~~~~~~

tests/many_to_many/tests.py

@@ -1,5 +1,6 @@
 from __future__ import unicode_literals
 
+from django.db import transaction
 from django.test import TestCase
 from django.utils import six
 
@@ -54,7 +55,9 @@ class ManyToManyTests(TestCase):
 
         # Adding an object of the wrong type raises TypeError
         with six.assertRaisesRegex(self, TypeError, "'Publication' instance expected, got <Article.*"):
-            a6.publications.add(a5)
+            with transaction.atomic():
+                a6.publications.add(a5)
+
         # Add a Publication directly via publications.add by using keyword arguments.
         a6.publications.create(title='Highlights for Adults')
         self.assertQuerysetEqual(a6.publications.all(),

tests/multiple_database/tests.py

@@ -295,19 +295,23 @@ class QueryTestCase(TestCase):
 
         # Add to an m2m with an object from a different database
         with self.assertRaises(ValueError):
-            marty.book_set.add(dive)
+            with transaction.atomic(using='default'):
+                marty.book_set.add(dive)
 
         # Set a m2m with an object from a different database
         with self.assertRaises(ValueError):
-            marty.book_set = [pro, dive]
+            with transaction.atomic(using='default'):
+                marty.book_set = [pro, dive]
 
         # Add to a reverse m2m with an object from a different database
         with self.assertRaises(ValueError):
-            dive.authors.add(marty)
+            with transaction.atomic(using='other'):
+                dive.authors.add(marty)
 
         # Set a reverse m2m with an object from a different database
         with self.assertRaises(ValueError):
-            dive.authors = [mark, marty]
+            with transaction.atomic(using='other'):
+                dive.authors = [mark, marty]
 
     def test_m2m_deletion(self):
         "Cascaded deletions of m2m relations issue queries on the right database"
@@ -762,7 +766,8 @@ class QueryTestCase(TestCase):
 
         # Add to a foreign key set with an object from a different database
         with self.assertRaises(ValueError):
-            dive.reviews.add(review1)
+            with transaction.atomic(using='other'):
+                dive.reviews.add(review1)
 
         # BUT! if you assign a FK object when the base object hasn't
         # been saved yet, you implicitly assign the database for the