Unlike some other bugs I have reported, brlcad does not issue any error or warning messages for this one, it just makes a bad part. The error is pretty obvious when raytraced, and can be seen in the wire frame, if you know what to look for, i.e., several extra edges near the ends of the part.
This test case makes a lower case "l" This problem happens on most, if not all, of the letters I have made. There is a good case (workaround) and a bad case.
good: fontET-l.c making fontET-l.f, bad: fontET-l0r.c making fontET-l0r.f
The top and bottom faces (constant z) are pointing the wrong way, i.e. towards the inside of the part, on the *0r.f part. Additionally, the inside faces of the fontET-l0.0.s part are NOT removed, as they should be, and are, in the good part.
The workaround and the only difference between these two parts, is the direction the points are listed in the fontET-l0.L.s ars and the fontET-l0r.L.s ars. This is the same workaround as bug ID: 3087757. This may indicate some common issue.
One perhaps peculiar thing, is that the *.L.s ars are topological tori, i.e., the first and last curves are identical. This has nothing to do with the problem, though. A test case in the Open Discussion forum post: "facetize error: Dangling faces", was done without topological tori,which did not fix the problem. Though I have not actually tried THIS EXACT case without the topological tori, I am convinced that it is not relevant to the problem.
Gilligan