I am currently trialing Rhino 8 to upgrade from Rhino 7 and I have noticed the results of CurveBoolean are sometimes not accurately representing the input. As far as I have observed this happens only with curved segments and seems to be more prevalent when the amount of selected input-curves is higher. Interestingly the result deviates always in the same way but only with exactly the same input curves, regardless if all the curves touch the area or not. If the input changes the deviation for the same area does too.
This is a problem. It leads to drawings and models that are missaligned and potentially a lot of trouble later.
I would appreciate if someone would have a look if they can replicate this, and if not point me at something that might be causing this.
For anyone interested: This occured because of the tolerance setting being too high. I added a couple of zeros and now it works just fine. I knew of the setting, just somehow didnt consider it.
@thilo I could reproduce the error, but also in Rhino 7. I think however this is not a tolerance issue, but rather caused by the fact that that shape is closed and overlapping with the larger arc.
I am not sure I understand what you mean, are you reffering to the file i attached? So you are saying there is something wrong?
I also dont really know how CurveBoolean works, I would have thought it would test the input curves for closed areas and then trim and join the adjacent curves, however this seems like an oversimplyfication on my part.
I see, I did produce that line with curveboolean aswell. But it is also happening at other spots, like the railing of the stairs, and there is only one curve.
