[Bf-committers] Boolean op holes error handled

[Alexander Ewering]

On Wed, 30 Nov 2005, Marc Freixas wrote:

> Hi all,
>
> The boolean op is defined only when both meshes are a closed mesh
> (without holes). This happen because I use a BSP tree to classify the
> mesh faces (IN, ON & OUT). If a no closed mesh is used the BSP don't
> works. In other words, how can I classify a face IN/OUT using a mesh
> with holes?
>
> For this reason I add a test before do any thing to chech the meshes. If
> one mesh is no closed a error message is showed:
>
> "Both meshes must be a closed mesh"
>
> I think this is better than have a freeze or crash. What do you think
> about this?

Hi Marc,

yes, of course this is the right solution. The decimator does it similarly
I think - it tests if the mesh it is supposed to decimate is manifold before
applying the operation. Maybe this is a similar case.

Another interesting question would be - especially for using booleans
as a modifier (realtime) - how much time does this "sanity check" take?
Maybe it should be made optional if it takes more than 20 or 30% of the
total boolop time - then there could be a switch "Faster booleans -
but they may crash!" for users who (for a change) know what they
are doing! :)

| alexander ewering              instinctive mediaworks
| ae[@]instinctive[.]de   http://www[.]instinctive[.]de