clipPolygon3dHP is not fully implemented

[oqilipo]

There are still two todos in clipPolygon3dHP

• What is the output if the plane is same as polygon plane?
• Rename this function to 'clipPolygon3d'

[oqilipo]

Two questions remain: What should be the output if the plane is same as polygon plane? Should we rename this function to: clipPolygonByPlane3d?

Maybe reopen till final decision?

[dlegland]

Yes, I agree.

my opinion about the two questions:

• if plane is the same as polygon plane, this should return the same polygon. If I had to start from scratch I would maybe rather throw an error. However, several functions in MatGeom return default values (like NaN or Inf) in similar cases. It seems preferable to be consistent with the existing.
• For renaming: yes, cleary more obvious!

best, David

[oqilipo]

Sounds good!

One more question. Before processing the input polygon it is checked if all points are in a plane. What tolerance would you use here? At the moment it is. 1e-8. But maybe you use some other (default) tolerance in other functions?

Kind regards

[oqilipo]

Sounds good!

One more question. Before processing the input polygon it is checked if all points are in a plane. What tolerance would you use here? At the moment it is. 1e-8. But maybe you use some other (default) tolerance in other functions?

Kind regards

I've changed it to 1e-14 now.

[dlegland]

Hi,

there is no absolute default for tolerance within the library. I often use 1e-8, but other functions use 1e-12. Value 1e-14 is fine as well for me.

The problem is that this often depends on the range of input data. I usually try to make the tests on normalized values, but there stille may be some uncoverd cases;