I'm trying to build fonts with rounded, rather than square edges. I've
tried several methods, and finally found one that did NOT give
facetize errors, or so I thought! Due to some unknown peculiarity in
the facetization code, I ended up with edges that went around the
other way, i.e., CCW vs CW. This should make no difference, as all of
the points are the same (at least as far an the l(ist) command is
I have two test cases: fontTst-A.g, which fails and tstAof.g, which works.
these files match:
fontTst-A1.0.s <==> tstarsLC1.0.s
fontTst-A1.o.s <==> tstarsLC1.o.s
fontTst-A1.n.s <==> tstarsLC1.n.s
these files contain the same points, but in different orders
fontTst-A0.0.s <~~> tstarsLC0.0.s - strictly reversed order
fontTst-A0.o.s <~~> tstarsLC0.o.s
fontTst-A0.n.s <~~> tstarsLC0.n.s
fontTst-A0.o.s and fontTst-A0.n.s were initially generated as a single
ars, with the topology of a torus, where the two ends of the ars
wrapped around and touched each other. This gave facetize errors, so,
I removed the first segment of the ars, and made it a separate
ars. ".o." designates segment One, and ".n." designates the other n
these are identical in z, where it is important, and vary in x,y, which is not relevant.
fontBox13x13.s <~~> tstABox.s
these only vary to swap the files names above.
fontTst-A.c <~~> tstAof.c
facetize tstAof.f tstAof.c - completes with no errors
facetize fontTst-A.f fontTst-A.c - fails - Dangling faces error
fontTst-A1.c is a reduced version of fontTst-A.c that still
fails. Removing any more lines causes it to pass.
Am I doing something fundamentally wrong, or it this another facetize bug?
files can be downloaded from:
This still fails in brlcad version 7.16.10. I found it if ver 7.12, and was hoping it was corrected.
I created a bug report for this. The test case is slightly simplified in the bug report. The bug report has the same name as this post.