Addition to Mathutils

 vec.normalized()
 mat.inverted()
 mat.transposed()
 made vec/float possible

 normalize/invert/transpose now return None because they modify the data in place.
 use the ...(ed) versions to return a modified copy.


Fixed Memory leaks from not decreffing PyFloat_AS_DOUBLE from these python functions...
(found when testing above functions)
  ob.rbMass
  ob.rbRadius
  matrix.determinant()
  quat*float
  vec*float
  matrix.transpose()
  EXPP_setModuleConstant

Checked all instances of PyFloat_AS_DOUBLE so I dont think there are any mroe leaks there.

1 files changed, 2 insertions, 0 deletions

diff --git a/source/blender/python/api2_2x/quat.c b/source/blender/python/api2_2x/quat.c
index 1cd07b0f588..af778d27f07 100644
--- a/source/blender/python/api2_2x/quat.c
+++ b/source/blender/python/api2_2x/quat.c
@@ -462,6 +462,7 @@ static PyObject *Quaternion_mul(PyObject * q1, PyObject * q2)
 			}
 
 			scalar = (float)PyFloat_AS_DOUBLE(f);
+			Py_DECREF(f);
 			for(x = 0; x < 4; x++) {
 				quat[x] = quat2->quat[x] * scalar;
 			}
@@ -478,6 +479,7 @@ static PyObject *Quaternion_mul(PyObject * q1, PyObject * q2)
 				}
 
 				scalar = (float)PyFloat_AS_DOUBLE(f);
+				Py_DECREF(f);
 				for(x = 0; x < 4; x++) {
 					quat[x] = quat1->quat[x] * scalar;
 				}