BooleanUnion Failing Because of Odd Intersection


#1

The BooeanUnion command is failing for a simple case because the intersection is not correct. The polysurfaces are both valid and manifold. There is one surface that is coplanar but I was told that the BooleanUnion command should work as long as the shared surface is planar. The file with the geometry is attached. Is there any way to fix this situation?

FAiled Boolean Intersect.3dm (90.2 KB)



(Chuck Welsh) #2

Hi Chris,

Although the non-box is not a “bad” object, it is not good. the What command shows that it has edges that are inaccurate by more than 0.3m. Explode, then RebuildEdges, and you will see how far off one of the surfaces is. Do you recall how it was made? That is where the bug is.

Thanks,
Chuck Welsh
Robert McNeel and Associates


(Pascal Golay) #3

Hi Chris- was the non-box made by moving edges around (sub-object selection) or by manipulating SolidPts?

thanks,
-Pascal


(Wim Dekeyser) #4

I know this is an old discussion but if you ask me, yes, there is a bug in a command somewhere but it is equally buggy that this is not reported as bad. I’m not sure what it will take to convince you guys but if this had triggered the CheckNewObjects mechanism, chances are that you would have fixed the actual bug by now. And the user wouldn’t have been frustrated by a command that doesn’t seem to work. A simple win-win situation, IMO.

Same here:

I hope you filed this one in the myjetbrains system…


(Chuck Welsh) #5

Yes, Check is a bad command to decide if your object is good or bad. It really just checks the validity of the data structure, not the goodness of the geometry. I have been working on geometry validation code for the past few months. High edge tolerances will definitely be identified.

Chuck


(Wim Dekeyser) #6

Looking forward to that, Chuck. Thanks! :smile: