Chris Hennes

@KeithSloan are you saying that this code is generating a valid shape in FreeCAD for you? ``` // 'inner' linear_extrude(height=25.002, twist=0.0, scale=[0.3, 2.85], $fn=128) { polygon(points=[[0.0, 10.0], [-0.4906767432741801, 9.987954562051724], [-0.9801714032956065,...

OK, good -- the OCC error is because the geometry becomes self-intersecting, the real problem is the incorrect twist being applied. I think I can track this one down...

If I replace the polygon with the circle it is approximating, everything works fine: ``` linear_extrude(height=25.002, twist=0.0, scale=[0.3, 2.85], $fn=128) { circle(r=10); } ```

The OpenSCAD also has a polygon -- it's just recognizably a circle, so to debug I tried switching. So OCC doesn't like that high-segment-count circle approximation.

It works fine without the scale.

> Am I grasping at straws? Well, probably, but that's kind of what we're reduced to here 🤷 I'll try to poke at it next week.

Could you please amend the commit message to include a brief explanation of why we are doing this? Otherwise some enterprising person is going to come along in a couple...

Yes, I followed that discussion, the explanation is not for me, it's for some future developer working on the clang-tidy file who does a `git blame` on that line of...

Alas, I "approved" it before I was well-versed in FreeCAD's source, that approval was based purely on testing, and should probably not be understood as an approval in my current...

I will re-test, but overall the code is solid and the feature a welcome improvement. As @realthunder notes above, the changeset is actually not all that large, there are probably...