<russnelson>
When they're both there, it's not manifold. If I remove either one, it becomes manifold. Damifino why.
<russnelson>
If it matters, this is version 2019.05 .
<InPhase>
Well did you check for purple?
<InPhase>
Hmm. This is polygon, not polyhedron.
<InPhase>
And also, I reproduce your problem with the last release, but the master branch is accepting it just fine through the render.
<InPhase>
russnelson: When I change: [-133.856, 154.96] to [-133.866, 154.96], the problem disappears.
<InPhase>
russnelson: I was suspicious about the structure of those two very close points at the start and end of the A.
<peeps[zen]>
russnelson: you know there is a text() module, right?
<InPhase>
lol. Well yeah, there is also that.
<peeps[zen]>
also 2019 release is really ancient at this point, i'm surprised people are still running anything before at least the latest official release
<russnelson>
apt install openscad, for better or worse.
<teepee>
apt full-upgrade ;-)
<russnelson>
Yes, well maybe that too.
<russnelson>
HA! Thanks, InPhase, that fixed it in the real (not just the gist) model. Yes, I know about text() but this is actually an SVG that I imported.
<russnelson>
Somebody else's logo that I'm trying to reproduce in a 3D print.
rawgreaze has quit [Ping timeout: 240 seconds]
<InPhase>
russnelson: You know also about the svg import feature?
rawgreaze has joined #openscad
<InPhase>
You might still have to make the same adjustment, but it's a nice feature.
<InPhase>
... Or at least it is. I think it was there in 2019. Although that won't have our latest support for importing svg parts separately.
<kintel>
right, well, ideally our transparency rendering shouldn't be that sensitive, but I think we'd have to depth-sort triangles to make it consistent
<teepee>
that would solve the ordering issue too, so that would be nice
<peeps[zen]>
teepee: sorting would be depedndent on camera angle which would mean redoing every frame. i think the compute costs could get out of hand rather quickly
<teepee>
true, probably only viable in a shader
<kintel>
yeah, or convex decomposition, but we all know where that leads :)
<gbruno>
[github] kintel pushed 4 additions 22 modifications (Update test configs (#4806) * Set Heavy tests to > 5 seconds on a decent workstation * Updated example tests with green surfaces * Fix z fighting in hull example * Change font to Liberation Sans, to be certain it's available on all systems) https://github.com/openscad/openscad/commit/bfe4770324754110a238fc4f0c283c29e2e5e1ae
<quiliro>
the picture does not have catenaries...but it should
<quiliro>
the loft is a cylinder which becomes a square as it goes down
<quiliro>
the picture is not correct because the vertical supports for the roof should be catenaries and not straight lines
<kintel>
That might end up involving doing everything using math. Sounds like a fair bit of math : /
<quiliro>
or else the roof will not be made of catenaries
Guest82 has joined #openscad
<gbruno>
[github] kintel pushed 9 modifications (Remove -direct and -prealloc features. These are now always enabled. The main reason for enabling them is that glBufferSubData() is quite slow on macOS, so we'll standardize on one common, fast, code path.) https://github.com/openscad/openscad/commit/e2f3aa3f88ac4b5db1c9ffd4d5404c5c2da9395f