CAD->bisect node. Some glitches appear on skewed quads. Fixed.

[satabol]

Windows 11, Blender 3.3.1, Sverchok-master

Glitches on skewed quads:

Expected:

and now it works. https://docs.blender.org/api/current/bmesh.types.html#bmesh.types.BMesh.normal_update

I think this patch has to be tested in other functions where bmesh.ops.bisect_plane used.

Test:

Bisect_2022_12_13_23_11.zip

v002.0001-0600.mp4

[satabol]

In the library .update_normals is used:

sverchok/utils/sv_bmesh_utils.py

Lines 855 to 861 in b7c5278

	def bmesh_bisect(bm, point, normal, fill):
	bm.normal_update()
	geom_in = bm.verts[:] + bm.edges[:] + bm.faces[:]
	res = bmesh.ops.bisect_plane(
	bm, geom=geom_in, dist=0.00001,
	plane_co=point, plane_no=normal, use_snap_center=False,
	clear_outer=True, clear_inner=False)

bit fill used an old method of filling:

sverchok/utils/sv_bmesh_utils.py

Lines 862 to 866 in b7c5278

	if fill:
	fres = bmesh.ops.edgenet_prepare(
	bm, edges=[e for e in res['geom_cut'] if isinstance(e, bmesh.types.BMEdge)]
	)
	bmesh.ops.edgeloop_fill(bm, edges=fres['edges'])

[mifth]

It's hard to get what's going on. I guess the problem is in NGons? Maybe if you triangulate NGons the issue will be solved?

[satabol]

@mifth Good remark. Triangulation of mesh can be used as additional method to fix some errors that bisect can not calculate (self intersections, single points and other). It is good to denote it in docs (I do it later, after some tests) . I try to maximize usefulness for this node. In some tests manual bisect and bisect_plane in Sverchok has different results on same geometry. I could not understand why? And now I think this is it! (with #4817 )

[satabol]

More info:

For tests with current Sverchok version: https://gist.github.com/b9d4a75befd5b82f3466317bbbb8c4c1

I spend not much time to find problem with current bisect node:

And try to repeat it with Blender bisect:

No problemo. )

If set .normal_update:

then Sverchok result and Blender result are equal!