[Bf-python] Got independent copy of mesh, now how to free it later?

[lynx.abraxas at freenet.de]

If someone is interested: I got it to work with:

   ellm= ell.getData(mesh=1)
   ellmesh= ellm.__copy__() #that's the one!!!

I must have overlooked this function because it starts with __. Does that mean
it's an inherited copy constructor or  "just  use  if  you  know  what  you're
doing"?

That  way  I  can  use  the real copy to apply any transform for pointInside()
while keeping the original as it is.

I'm now wondering how to free a mesh in python? Just like this?

ellmesh= None #is this the correct way to free the mesh?

These don't work:
bpy.data.meshes.unlink(ellmesh.name)
ellmesh.unlink()

Regards,
Lynx

On 21/05/09 16:09:06, joe wrote:
> I'd suggest asking for help on blenderartists or
> irc://irc.freenode.net/blendercoders , I'm a little too tired to give
> good answers at the moment.  :-/
> 
> Anyway, there should be a way of extracting the scale and rotation
> part of a matrix (I believe the API has stuff for that) so I'd suggest
> doing that, then creating a function that puts vertices through a
> round-trip, first applying rot/scale based on the object matrix (don't
> forget to zero the rot/scale fields after applying), then deapplying
> it.  This will ensure that you have the order of transformations
> correct.
> 
> Joe
> 
> On Thu, May 21, 2009 at 3:59 PM,  <lynx.abraxas at freenet.de> wrote:
> > On 21/05/09 15:23:38, joe wrote:
> >> What do you want?  The global positions of each vertex?  Or in
> >> relation to the object center?
> >
> > I  would like the ellipsoid-obj to be translated where as their rot. and scale
> > should be applied to the mesh. I think I got that with:
> >
> >   mat= B.Mathutils.Matrix([dl[6],dl[7],dl[8], 0],  [dl[9],dl[10],dl[11],  0],
> > [dl[12],dl[13],dl[14], 0], [0,0,0,1])#octave: v' #why is this correct???
> >   #mat=  B.Mathutils.Matrix([dl[6],dl[9],dl[12],  0],   [dl[7],dl[10],dl[13],
> > 0], [dl[8],dl[11],dl[14], 0], [0,0,0,1])#octave: v
> >
> >   local_rot(ob, mat)
> >   ob.setSize(dl[0], dl[1], dl[2])
> >   ###ob.setLocation(dl[3], dl[4], dl[5])#transl obj and the mesh
> >   me= ob.getData(mesh=1) #blender shows the obj. according to its matrix
> >   me.transform(ob.matrix)#but does not apply it on the obj verts
> >   ob.setMatrix(ob.matrix.identity())#set identity to avoid double effect
> >   ob.setLocation(dl[3], dl[4], dl[5])#transl obj only not the mesh
> >
> >
> > In the next step I want a  copy  of  an  ellipsoid  mesh  which  I  apply  the
> > trnslation  of  the  ellipsoid  object on (just for pointInside). But the copy
> > should not be linked to the ellipsoid object so that the mesh  of  the  object
> > stays untouched by the translation of the (copied) mesh.
> >
> > All types of coping I tried some how kept a link to the original object:
> >
> >   ellmesh= ell.getData(mesh=1)
> >   ellmesh= bpy.data.meshes.get(ell.getData(name_only=1,mesh=1))
> >   ellmesh.get(ell.getData(name_only=1,mesh=1))
> >   ellmesh= copy.copy(ell.getData(mesh=1))
> >
> > Is it possible?
> > Do You know how?
> >
> > Thanks for Your help
> > Lynx
> >
> >  To only apply the rot/scale, I believe
> >> you can simply multiply (do this temporarily though, don't store the
> >> result back in the mesh) with the rotation/scale matrix, which I
> >> believe is something like:
> >>
> >> (matrix.rotationPart() * matrix.scalePart()) * vert_co
> >>
> >> . . .but I could be wrong, look it up in the docs.  It'd help if you
> >> could give a more complete description of what your trying to do.
> >>
> >> Joe
> >>
> >> On Thu, May 21, 2009 at 11:17 AM,  <lynx.abraxas at freenet.de> wrote:
> >> > On 20/05/09 17:33:20, joe wrote:
> >> >> If applying rot/scale works, then it means your script isn't properly
> >> >> taking the object matrix into account.  To calculate the final
> >> >> transformed position of a vertex, you have to multiply it with the
> >> >> object's matrix, like so:
> >> >
> >> > Is this OK too?
> >> >
> >> >   ellmesh= ell.getData(mesh=1)
> >> >   ellmesh.transform(ell.matrix)  #apply any transform, just in case it wasn't done already
> >> >   ell.setMatrix(ell.matrix.identity())
> >> >
> >> >
> >> > This sadly has the unwanted side effect that the ellipsoids have their blender
> >> > centres at the global origin.
> >> > Without  the matrix application they have their blender centers at the centers
> >> > of each ellipsoid.
> >> > I created the ellipsoids by adding a UVsphere at the origin, rotating, scaling
> >> > and displacing it:
> >> >
> >> >
> >> >   me= B.Mesh.Primitives.UVsphere(segments, rings, diameter)
> >> >
> >> >   sc= B.Scene.GetCurrent()          # get current scene
> >> >   #http://www.blender.org/documentation/246PythonDoc/Object-module.html
> >> >   obn= "Ellipsoid_%0.4d" % dl[15]
> >> >   ob= sc.objects.new(me, obn) # add a new mesh-type object to the scene
> >> >
> >> >   #http://www.blender.org/documentation/246PythonDoc/Object.Object-class.html
> >> >   mat= B.Mathutils.Matrix([dl[6],dl[7],dl[8], 0],  [dl[9],dl[10],dl[11], 0], [dl[12],dl[13],dl[14], 0], [0,0,0,1])
> >> >
> >> >   local_rot(ob, mat)
> >> >   ob.setSize(dl[0], dl[1], dl[2])
> >> >   ob.setLocation(dl[3], dl[4], dl[5])
> >> >   me= ob.getData(mesh=1) #blender shows the obj. according to its matrix
> >> >   me.transform(ob.matrix)#but does not apply it on the obj verts
> >> >   ob.setMatrix(ob.matrix.identity())#set identity to avoid double effect
> >> >
> >> >
> >> > Why is blender showing the ellipsoids correctly where as their verts are still
> >> > the unit sphere until I apply the matrix to the mesh?
> >> > Can You point me to some docs explaining this behavior?
> >> > Is there a way to apply only rot and scale to the mesh but not the translation
> >> > at creation time?
> >> > How  can  I  later  some  how translate the mesh (for pointInside to work) but
> >> > leave the object as it is?
> >> >
> >> > Well, at least I'm slowly getting what I want;)
> >> >
> >> > Thanks for any help
> >> > Lynx
> >> > _______________________________________________
> >> > Bf-python mailing list
> >> > Bf-python at blender.org
> >> > http://lists.blender.org/mailman/listinfo/bf-python
> >> >
> >> _______________________________________________
> >> Bf-python mailing list
> >> Bf-python at blender.org
> >> http://lists.blender.org/mailman/listinfo/bf-python
> > _______________________________________________
> > Bf-python mailing list
> > Bf-python at blender.org
> > http://lists.blender.org/mailman/listinfo/bf-python
> >
> _______________________________________________
> Bf-python mailing list
> Bf-python at blender.org
> http://lists.blender.org/mailman/listinfo/bf-python