Using querysets as an rvalue in filter() calls was causing an unnecessary

database query, due to a bool() call. This change stops that behaviour.

git-svn-id: http://code.djangoproject.com/svn/django/trunk@9715 bcc190cf-cafb-0310-a4f2-bffc1f526a37

1 files changed, 7 insertions, 0 deletions

diff --git a/tests/regressiontests/queries/models.py b/tests/regressiontests/queries/models.py
index a3247bb460..05842ebe7b 100644
--- a/tests/regressiontests/queries/models.py
+++ b/tests/regressiontests/queries/models.py
@@ -1015,6 +1015,13 @@ performance problems on backends like MySQL.
 >>> Annotation.objects.filter(notes__in=Note.objects.filter(note="n1"))
 [<Annotation: a1>]
 
+Nested queries should not evaluate the inner query as part of constructing the
+SQL. This test verifies this: if the inner query is evaluated, the outer "in"
+lookup will raise an EmptyResultSet exception (as the inner query returns
+nothing).
+>>> print Annotation.objects.filter(notes__in=Note.objects.filter(note="xyzzy")).query
+SELECT ...
+
 """}
 
 # In Python 2.3 and the Python 2.6 beta releases, exceptions raised in __len__