Hi Michael,

I have some notes...

When I cut a sphere by line, which lies on the x or y axis exactly, using Boolean-Merge, -Diff and also -Isect command, the sphere is not sliced perpendicular to the xy-plane. I don't think it is bug... but maybe it is a trace of a small code mistake that might lead to real bug elsewhere...


From time to time I encounter that filleting fails on closed curve in case the small part of curve are completely eaten by fillet. On the other hand, all works fine when I separate the curve (and re-join it into two open curves), then run fillet command again.

To see an example, download a 3dm file...

Petr

Hi Petr, thanks for reporting these!

Re: sphere/line merge - I'm sure you guessed what is happening here, there is some detection if the line is on one of the world planes and if it is it uses that plane's normal as the projection direction. But in that case you are on 2 world planes, and when that happened it would just pick the first one of x, y, or z axes.

I've made a slight adjustment to this so that the z axis will be preferred in this case instead.


Re: curve fillet - thanks for narrowing down the cause of this failure! It will be difficult to allow segments to be completely consumed in general when there were fillet vertices at the ends of the segment (like a small line out of a polyline being consumed by a large radius), but I should be able to tweak things to handle this case to allow a G1 segment to be consumed, where it was not crossing a fillet vertex. I'll dig into that tomorrow.

Do you think that I should put something into Join to automatically glue arcs of the same circle into larger single arcs?

- Michael

>>Do you think that I should put something into Join to
>>automatically glue arcs of the same circle into larger single arcs?

I usually encountered that fillet issue after making a curve boolean operation, so I am not sure if it helps directly.
I'd rather vote on including that mechanism into the new "Merge" command so nobody would be surprised that Join can produce "smooth" curve...

Hi Petr,

> I usually encountered that fillet issue after making a curve boolean
> operation, so I am not sure if it helps directly.

Aha - this seems to be the root problem, I am glad you mentioned it!

It looks like I should be able to tune this up so that curve booleans won't cause an extra segment split at the "seam" point of closed curves.

- Michael

Hi Petr, I was able to get the curve boolean fix in for this next beta, but not the fillet fix yet. I should be able to get the fillet fix into a patch or something after the documentation is fully wrapped up.

- Michael

>>It looks like I should be able to tune this up so that curve booleans won't cause
>>an extra segment split at the "seam" point of closed curves.

hehe... it really works!!! No tiny curveeeeeee.... :)

Same result with yours Petr....