[Bf-committers] Carve vs. Bmesh booleans

Kai Kostack kaikostack at gmx.net
Thu Apr 19 12:50:27 CEST 2018


Hi Mikhail, 
 
I wonder what makes you so certain that the Bmesh way is the way how a boolean 
tool is supposed to work? 
 
> Incorrect input should result in incorrect output. 
 
Sure, you can declare everything being incorrect input that your tool can't 
handle properly. Or you just can make it work better. Carve has proven that 
this can work better. 
 
 
> If FM result geometry must not contain loose edges, than it should do a 
> post process 
> (check for loose edges) after boolean operation finished. 
 
Of course, we can and will do this if necessary, but what's about the other 
users? Don't they deserve clean outputs? As I said you cannot remove these 
edges in a non-destructive way using modifiers. In Python I would need to do 
expensive mesh editing then, too. I'm using the cutter plane approach to 
discretize elements for scientific physics simulations... a lot of elements. 
 
These edges are wrong, they don't belong there. And if the plane is subdivided 
then there are even faces generated not belonging there.
 
 
> Please do not try to make boolean tool more than it supposed to be. 
 
I question that a boolean tool is supposed to work exactly that way. It's 
inconsistent to ignore the normal of a plane and to make the decision of which 
half of the mesh is to be removed depending on the target object matrix rather 
than on the normals of the operator object. If you rotate the plane around 180 
degrees you get the same output again, this is not useful and makes results 
unnecessarily hard to predict. 
 
Of course, I can accept your opinion on the matter. I don't want to enforce 
anything, I'm just providing my feedback and explaining our use cases in the 
hope that it might be helpful to improve Blender. 
 
Best regards, 
Kai Kostack 


More information about the Bf-committers mailing list