[Bf-committers] Boolean op holes error handled

Marc Freixas mfreixas at lsi.upc.edu
Mon Dec 5 11:24:16 CET 2005

Hi all,

After reading your messages we can draw some conclusions. I think that
the open meshes are allowed if are representing an infinite object, for
example a plane or a cylinder without cover. This kind of objects are
closed in the infinity point, like a parallel lines. But, their
nonmanifold edges do not have to intersect any face. Becouse the
intersection cretes a new faces and these must be classified inside or

By all this, I suggest to implement two level test. First, check if the
mesh is closed. If it is closed, go ahead. If not, check if some
nonmanifold edge intersect a face. Nevertheless, I recomend to use a
closed mesh.

On the other hand, one has said that the normal face could be used to
classify. I use the face vertex order to compute the normal face, using
the anti clockwise rule. But, I think that some times this don't works
fine. Knows any one an alternative way?


Arne Schmitz wrote:

>Am Donnerstag, 1. Dezember 2005 12:07 schrieb Tom M:
>>>Can you give an example on what kind of object you would like to do
>>>booleans with, and you would consider 'open'?
>>basically using any plane or series of connected planes to slice an
>>object. (ie aplane booleaned with a uv sphere leaving a partial UV
>>sphere with one side flat). It would be fairly easy to work around
>>(extrude the plane or series of planes thus creating a solid - but
>>with the proposed use the normal, it would just be an 'infinite
>>extude' to accomplish the same effect.
>In other words: With a plane you would define the direction in which the 
>normal is facing "outside", and the other side "inside". You would then treat 
>the plane es being infinite. This way the object is "closed", it has no 
>Bf-committers mailing list
>Bf-committers at projects.blender.org

More information about the Bf-committers mailing list