I just ran into a really bizarre bug, apparently in minkowski(). The following code renders incorrectly on my system, in both preview and rendering: $fn = 7; minkowski() { cube(0.1); union() { translate([5, 0, 0]) cylinder(h=10, r=5, center=true); cylinder(h=5, r=10, center=true); } } It should output a slightly warped wide cylinder with a longer, thin cylinder poking through it along one side. Instead it shows what looks like the longer, thin cylinder, with the wider cylinder differenced from it. It acts like that union() is a difference(). It'll render correctly if you remove the translate(). If you translate by a value other than 5, then it works. It works if you set $fn to any of 5, 6, 8, 9, or 12, but it fails when $fn is 7, 10, or 11. Very odd! Here's an image of the failed render, and an image of a successful render with $fn=8: - Revar

Yes very bizarre end result, but I think it is due to doing a union of very nearly coincident faces. Changing the translate to 5 - 1/1024 fixes it. On Mon, 4 Mar 2019 at 12:14, Revar Desmera <revarbat@gmail.com> wrote: > I just ran into a really bizarre bug, apparently in minkowski(). The > following code renders incorrectly on my system, in both preview and > rendering: > > $fn = 7; > minkowski() { > cube(0.1); > union() { > translate([5, 0, 0]) cylinder(h=10, r=5, center=true); > cylinder(h=5, r=10, center=true); > } > } > > > It should output a slightly warped wide cylinder with a longer, thin > cylinder poking through it along one side. > Instead it shows what looks like the longer, thin cylinder, with the wider > cylinder differenced from it. It acts like that union() is a difference(). > > It'll render correctly if you remove the translate(). If you translate by > a value other than 5, then it works. It works if you set $fn to any of 5, > 6, 8, 9, or 12, but it fails when $fn is 7, 10, or 11. Very odd! > > Here's an image of the failed render, and an image of a successful render > with $fn=8: > > > - Revar > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >