00:28 <rapha> InPhase: i'm back at that thing i'm trying to make where i asked about using minkowski on an imported stl. so, like you said, this does work, and is much faster with manifold enabled. unfortunately i'd misunderstood minkowski completely.

00:29 <rapha> sooo, new question ... if i have a random 2d shape, say, a circle for simplicity's sake. is there a way of "blowing" that "up" like a balloon, so that, in the case of the circle, it'd become a half-sphere?

00:33 <rapha> my current thing looks like this https://gist.github.com/sixtyfive/5e65d329f110e3431a7f17d4efaf5740 ... but i'd like for it to be flat on the bottom like it is, but rounded, as if it were a balloon, on the top. not sure if this is at all possible.

03:24 <InPhase> rapha: Starting with an stl for that is not particularly ideal.

03:26 <InPhase> rapha: I suppose you could do it somewhat clumsily with an epsilon-overlapped series of roof-based layers, each stretched by the trig functions that would produce a circular curvature, and then truncated by a big cube intersection?

03:29 <InPhase> It would take a combination like: for (...) { translate(...) intersection() { translate(...) cube(...); scale([1,1,...]) roof() offset(-...) shape(); }

03:29 <InPhase> }

03:31 <InPhase> rapha: It's late and I'm on the cusp of asleep, so I'm not going to attempt the trig math. :) Only the big picture part of my brain is still operating adequately.

03:33 <gbruno> [github] rcolyer review_requested pull request #5165 (Reseed rng at compile for unseeded initial vals) https://github.com/openscad/openscad/pull/5165

03:34 <InPhase> teepee: I dropped a review request for you on #5165. It's like 7 active lines of code, but it's a nuanced intentional change of behavior, wedged into an unusual place, so it should get another pair of eyes to sanity check. :)

03:39 <InPhase> teepee: Specifically, I wasn't 100% certain the location in MainWindow::compile where initialize_rng was placed was the right place. It is, however, the place that seemed to work right on testing. There might have been a smarter spot I didn't see. The behavior is, as far as I can reason, unneeded for any command line usage.

03:42 <InPhase> teepee: The treatment of the static seed_val is also a wee bit weird, but I think it's a valid approach, and was the simplest I could think of for the objective of not over-tapping the slow clock.

04:10 <gbruno> [github] rcolyer pushed 2 additions (Add daylight gem and nocturnal gem color schemes) https://github.com/openscad/openscad/commit/d9193ef740a030c92c62d924e891266446c499c1

04:12 <gbruno> [github] rcolyer opened pull request #5171 (Add daylight gem and nocturnal gem color schemes) https://github.com/openscad/openscad/pull/5171

05:39 <InPhase> kintel: I hope you are still magically noticing whenever someone mentions you. Did something break in tests? I was aiming for a second PR for the evening, but it was not possible because cgalstlcgalpngtest and cgalbinstlcgalpngtest have both been running for over 30 minutes with manifold, which seems weird.

08:35 <kintel> InPhase Out traveling so I don't have time to look into it, but I'm unaware of any breakage

08:37 <rapha> Good morning InPhase, I wasn't expecting an answer in the middle of the night, no worries :) STLs are not particularly ideal for anything I guess. Unfortunately the author seems to not be monitoring his account anymore, so, no source files. What you described is a little advanced for me, but I wonder if the intersection technique could be used to get close "enough". Will play with that some more,

08:37 <rapha> thank you!

12:02 <InPhase> kintel: Oh, I think I see the issue... I fell asleep while trying to rerun it, and so it actually finished before I woke up. I think the two tests it was trying to run were probably utilizing CGAL rather than Manifold in some way that interacted poorly with my attempt to change the default $fa and $fs. I will look into the details of those two tests later, now that I have the actual completion of the

12:02 <InPhase> report.

12:04 <InPhase> kintel: I'm trying to prepare a pretty-by-default PR to improve OpenSCAD first impressions.

12:05 <InPhase> kintel: I think now is our opportunity to actually do this for a release.

14:28 <rapha> Hmm. Why does cube() have center=, but sphere() doesn't. That seems inconsistent, no?

14:31 <J24k81> why would you put a sphere not to center?

14:32 <J24k81> cube has a vec3  sphere only r

14:32 <J24k81> cylinder has center that is centering only h

14:33 <J24k81> sphere don't have h so this is consistent with cylinder, also circle has no center

14:34 <J24k81> rapha but you can always define your own sphere module that has a center parameter if you need that

14:50 <rapha> Hmm.

14:51 <rapha> Okay, I guess that makes sense.

14:51 <Scopeuk> it is not necessarily intuitive but you can usually rationalise it

15:31 <rapha> yeah, it'll forever be a learning curve :) ... but using intersection() and a lot of difference() i managed a kind of shape that should work and doesn't look too bad, either. https://i.imgur.com/xVvcUQ4.png

15:39 <J24k81> rapha looks quite cool - some drone frame?

15:40 <J24k81> seems so just duckgo TKS GTR

15:44 <rapha> yes, TarkusX used to be a designer of really cool micro frames

15:45 <rapha> He built this one for 2S, though, and I'm flying it on 4S and with a GPS module up top. So no space for battery there. So I put it underneath, but that means I land on it. Not good for the lipo, better to land on some TPU.

21:34 <J24k81> why is the customizer not supporting arrays of strings?

21:35 <J24k81> label=["yes","no"];

21:35 <J24k81> up to 4 this should work, i think