To make it clear up front: I don’t want to bother you by any means. 
I’m super happy with your tool. 
Today I milled this ring. And similar to the roughing operation I reported earlier, it changed milling direction on the last cycle.
The part turned out pretty good and I don’t need a fix.
Cylinder_Inner.kmz (347.4 KB)
verified bug. working up a fix. it was also adding dogbones on open polys around tabs, which I just fixed.
spent some time on this and it turns out that when tabs are present, cut direction can (and likely will be) invalidated by the code path that does that clever little bit of path clipping. and it’s hard to fix because nesting, direction, and other pertinent information is lost as it passes through several layers of abstraction.
this is fixable, but requires a little bit of either a) risky surgery or b) refactoring. I’m in favor of the latter because I am already planning a large refactoring of the CNC code as I get ready to implement 4th axis control, and other important features.
but it means it’ll take a little longer to get a fix.
Here’s a vote for carrying such info further into the tool path planning process.
From the outside looking in, it looks like that would enable some desirable improvements.
I have a ton of CNC work and refactoring scheduled for the next release cycle. This will be in there.
