us.id when objects were destroyed but not always incrementing when
created. The intent of modifying us.id is to make Python a "user" of the
data so it persists even when it is deleted from Blenders UI. The original
commit was unintentional but Ton thought the idea was OK.
- update for the old mathutils rewrite
- update for some other methods ive added
- added explaination of wrapped data
- added a .css file for epydoc gives nice blender/python colors :?
* add some protection against accessing MVerts/MEdges/MFaces which have
been deleted
* correct cut-and-paste bug in faces.delete() method
* correct some compiler warnings
-- calling mesh.getFromObject(obj) with mesh object now also
copies material properties
-- mesh.quadToTriangle() takes a parameter to duplicate both
Ctrl-TKEY and Shift-Ctrl-TKEY actions
-- assigning None to mesh.verts "clears" the memory allocated
to the mesh (equivalent of Mesh.New(), but on an existing
mesh)
-- exception handler message for mesh.faces[i].uv = [..] more
clear (uv attribute only accepts tuple, not list)
-- fixed bug for meshs with deformed verts when deleting verts
(deformed verts deleted and repacked correctly now, I think)
bug #3367 reported by Chris Want (thanks): ob.mat documentation was
wrong (was not updated); that attribute returns the world space matrix,
not the local one.
- don't get too excited
- allows you to get armatures from a scene
- makeEditable()/saveChanges() puts the armature into out of editmode (pythonically)
- Armature.bones is a dictionary that contains all the bones in the armature and can be iterated
- getters are available for:
name,
roll (dictionary) keys are BONESPACE, ARMATURESPACE
head (dictionary) keys are BONESPACE, ARMATURESPACE
tail (dictionary) keys are BONESPACE, ARMATURESPACE
matrix (dictionary) keys are BONESPACE, ARMATURESPACE
weight
deform_dist
subdivisions
options (list of constants)
parent
children
Setter work only in editmode. Some are not fully implemented.
Type class is embedded in the module. This means the construct is called as follows:
Blender.Armature.ArmatureType()
import Blender.Armature as Armature
arm = Armature.Get('myarm')
for name, bone in arm.bones.items():
...print name, bone, bone.matrix['ARMATURESPACE']
more documentation is forth coming. This is an alpha for this api.
remaining effect type, it didn't make much sense to leave things
implemented in two separate files. Changes include:
* two bug fixes (the getChild() and getMat() methods were using floats
instead of shorts)
* performing clamping on input values
* implementing attributes using tp_getset
* merging Effect and Particle functions: the Particle module exists in
name only, with the Particle.New() and Particle.Get() functions
remaining for backward compatibility (they are in fact identical to
Effect.New() and Effect.Get() functions)
* update of doc/Effect.py (including remove all old references to wave
and build effects)
so that it resides in the Effects module instead, with the intent of
soon merging Effect.c and Particle.c into a single file (and somedat
removing references to the Effects.Particle submodule)
its adrcode in addition to its string name (shape keys don't have fixed
or unique string names, and they are stored in the key, not the Ipo).
This will make it easier to later use constants from dictionaries to
access a curve.
(old code is just "#ifdef"-ed out for now.)
Please give this a try on all platforms and give us feedback on what you
see for your sys.path setting in Blender (run these two lines in text window)
import sys
print sys.path
returning Python exceptions. EXPP_ReturnPyObjError() always returns a
NULL because Python expects error conditions to return a NULL pointer
instead of an object. Since the pointer is cast to a PyObject *, it's
ugly to use for propagating the errors back in this case, so this fix just
uses PyErr_SetString() to set the error and return NULL (see the body
of EXPP_ReturnPyObjError() ).
It was calling PyModule_AddObject() with unassigned pointer. Crashed on
exit here (python 2.3.2). Weird thing was that python 2.3.5 didnt complain
CVSr ----------------------------------------------------------------------
NOTE: I had to fix NMesh.c, Mesh_fromNMesh(), that is a real bad
function... it was returning a Py object as a Mesh (on error).
This is still not really solved (NULL return is not handled).
- added faces.uvSel attribute: can get/set selection status of UV vertices in
UV Editor window
- make mesh.faceUV, mesh.vertexUV and mesh.vertexColor writable: users
can now enable/disable UV faces, vertex colors, "sticky" vertices
- fixed bug with mesh tool methods: before it would only work if an object
linked to the mesh was selected
- added mesh.quadToTriangle() and mesh.triangleToQuad() methods
- added selected() method to verts, edges, faces; returns list of indices
of selected items
- mesh.getFromObject() now gets derived mesh data
- ported vertex group methods from NMesh (required change to Object.c)
- ported module dictionaries from NMesh
- new methods from NMesh (transform, getFromObject, findEdges)
- new methods for deleting groups of verts, edges and faces
- new methods for accessing mesh editing tools: fill, flipNormals,
recalcNormals, remDoubles, smooth, subdivide, toSphere
- Added PVertType to Types module (not my favorite name; any suggestions?)
