Mathutils n-dimension patch by Andrew Hale

[ideasman42, 2011-12-05 03:49:29]

comments inline, mostly api usage.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
File source/blender/python/mathutils/mathutils.c (right):

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils.c:163: {
array_min is not getting checked in this case

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
File source/blender/python/mathutils/mathutils_Vector.c (right):

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:112: if
(!PyArg_ParseTuple(args, "i|f", &size, &fill)) {
to avoid hand writing exceptions for PyArg_ParseTuple you can replace "i|f" with
"i|f":Vector.Fill", then just return NULL.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:345: static PyObject
*Vector_resize(VectorObject *self, PyObject *args)
(picky) for functions which recieve a single argument you can use METH_O, in the
methoddef and then take the PyObject and get its an int (not deal with args
tuples). enough examples of this in the code.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:381: self->vec[i]= 0.0f;
(picky) why not use range_vn_fl fill function here? - just need to offset the
array.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:694: static PyObject
*Vector_dot(VectorObject *self, PyObject *value)
(picky), this is an example of where you can use cleanup...

697 » PyObject *ret= NULL;
....
 if (some error)
   goto cleanup;

cleanup:
    PyMem_Free(tvec);
    return ret;
----

since its simple func its not going to confuse anyway, but would do it this way
all the same.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:885: if
(mathutils_array_parse_alloc(&tvec, size, value, "Vector.lerp(other), invalid
'other' arg") == -1) {
again, suggest using a `cleanup` goto at the end.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:1128: vec=
PyMem_Malloc(vec1->size * sizeof(float));
no check on fail here.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:1192: vec=
PyMem_Malloc(vec1->size * sizeof(float));
also no check on fail.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:1279: tvec=
PyMem_Malloc(vec->size * sizeof(float));
no check on fail.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:1480: vec=
PyMem_Malloc(vec1->size * sizeof(float));
no check on fail.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:2422: /*{"func",
(PyCFunction) Vector_func, METH_VARARGS, NULL},*/
think the to_2/3/4 d funcs can stay for now, at least not to break the api, see
how resized is more flexible tho.

http://codereview.appspot.com/5449088/diff/1/source/blender/python/mathutils/...
source/blender/python/mathutils/mathutils_Vector.c:2598: static PyObject
*Vector_CreatePyObject_alloc(float *vec, const int size, const int type,
PyTypeObject *base_type)
no need for type argument