00:40 <doctorpepstein> I want to generalize that idea of the hexagons on a rectangle that I saw the other day. I want to specify a rectangle and then have an array of hexagons maximally placed, and I think I have the basic idea in place, but need to pick my constraints. I think I need a "space between gons" and "gon radius" parameter in addition

00:49 <doctorpepstein> Mm, i found a couple ones on here

00:50 <MP71> If I wanted to make a patch, would I need to upload it to a repository somehow or attach it to a comment in the corresponding bug thread?

00:54 <teepee> MP71: normal process is: 1) clone repo into your own account 2) create a branch for the change 3) make changes and check-in into this branch 4) push to github 5) create pull request

02:48 <kintel> guso78k I think the sorting really hinges on performance. Automated re-sorting every 60-degree rotation sounds weird, but acceptable if the user experience is OK. My main concern is that VBO building can be _very_ expensive for larger designs, so we may need to optimize that a bit, e.g. use existing VBOs but only re-sort triangles inline for the necessary subset of triangles. ..plus making it all work across rendering modes may take some effort

02:48 <kintel> (haven't thought about that much), but visuals are easily testable, performance is more an art form at the moment.

02:54 <kintel> guso78k In terms of throwntogether renderer: F6 will disable that; it's only available as an alternative to preview (F5) rendering.

02:55 <kintel> It looks like you (still) have some oddness going on for axis rendering. Do you think that's just broken OpenGL drivers or a symptom of an OpenSCAD rendering bug?

06:37 <guso78k> kintel, my only intent  soi far is adding ansort algorithm ONCE per F6 key. this would make display of transparent models way better. User can hit F6 again to "corect"  the view, if they roated too much.

06:38 <guso78k> and for thorown-together i was actually wrong, i tought its the partner of CGALRenderer.

06:40 <guso78k> yes, my distorted axes are still there, did not perform any counter-measures as its very likely my  very own  local problem . This distortion is a combination of my negative dotted axis lines, my VirtualBox7.1 version and the Linux SVGA driver

06:41 <guso78k> Often i get an "SVGA Context mismatch" in the console after quit and even segfaults (gdb on the core tells me)

11:24 <h1> Hi how do I create the the end of a trumpet shape in openscad?

11:26 <h1> https://pasteboard.co/gT7NlBkEwiyw.png

11:26 <h1> the bell part

11:51 <ali1234> with difficulty

11:51 <ali1234> you could make something close with rotate_extrude on a 1/4 ellipse

11:52 <ali1234> to get it perfect you'd need to use BOSL2 to get a spline

11:52 <ali1234> or something like that

14:22 <gbruno> [github] kintel closed issue #3719 (Windows (at least) Snapshot badly broken - Main Window not displaying content) https://github.com/openscad/openscad/issues/3719

14:25 <kintel> guso78k Yeah, having the possibility of rendering a single image with correct transparency can be a good feature, to at least allow people to make great screenshots.

14:26 <kintel> We do have a number of issues related to transparency, perhaps I need to go through them and structure it a bit + add tests, so we understand the scope of the issue.

14:28 <kintel> Not sure how to best attack it - perhaps make an new, overall ticket, then link all individual issues and start collecting tests

14:30 <kintel> My main topic right now is to move Manifold to become the default renderer, but I can probably start chipping away at the transparency issue as a side task, as it would be very good to improve that before the next release, and it may help expose weaknesses in our current renderers

14:55 <guso78k> apart from the need to pass one parameter along the hole virtual derived class tree, the code change is in a very compacted, I would  collect the change in a PR if you agree.

14:55 <guso78k> prepare a PR if you agree

16:06 <h1> @ali1234 I thought some gurus on here would solve it easily :)

16:31 <juri_> curves are hard. :)

16:32 <teepee> yeah, unless you have the math :)

16:33 <teepee> h1: on the mailing list there was someone working on saxophone shapes, maybe some of the technique could be repurposed

16:37 <teepee> uff, the web frontend is really messing up some of the posts :( - https://lists.openscad.org/empathy/thread/64EXVJQHLQRBMRZKK7JDGYFYTV7COCHL

16:40 <teepee> not seeing a conclusive solution maybe except the last post from David Eccles who is a bit of a math magician

17:05 <h1> teepee: how do I see the mailing list ?

17:06 <teepee> link I posted ^

17:06 <h1> teepee: wha math do you need the equation for the suface ?

17:06 <h1> surface

17:06 <h1> thank will check it out

17:07 <teepee> yes, if you have a formula for the surface, a library like the one from InPhase may be able to turn that into an object quite easily

17:10 <teepee> hmm, I think I need to dial in that PETG printing quite a bit

17:11 <InPhase> Let's see. What is that functional shape for a trumpet.

17:12 <teepee> is there something like a trumpet formula?

17:13 <teepee> looks like those black hole images :-)

17:13 <h1> teepee: some like https://en.wikipedia.org/wiki/Gabriel%27s_horn

17:14 <teepee> h1: is that part completely straight, or in other words rotational symmetric

17:14 <h1> y = 1/x with the domain x ≥ 1 {\displaystyle x\geq 1} and rotating it in three dimensions about the x axis.

17:15 <h1> teepee: not sure what you mean

17:15 <teepee> like you can define the 2D shape of the outline and then rotate that 360 degrees to make it 3D

17:15 <teepee> that makes things so much easier

17:16 <h1> yes so y = 1/x rotated by 360 degrees

17:19 <h1> where x is positive

17:23 <teepee> $fa = 1; $fs = 0.2; s = 0.6; e = 8; rotate_extrude() rotate(90) polygon([[s, 0], for (x = [s:0.1:e]) [ x, 1/x ], [e, 0]]);

17:38 <InPhase> h1: https://bpa.st/BWXQ

17:41 <h1> InPhase: I am getting a syntax error

17:41 <InPhase> Oops.

17:41 <InPhase> Stray character got in there.

17:42 <InPhase> h1: https://bpa.st/3TDQ

17:42 <h1> teepee: it should be hollow ish

17:43 <h1> InPhase: great thanks

17:44 <InPhase> h1: Some values should be parameterized there to generalize it for different scalings, but hopefully you can sort that out from the example. :)

17:49 <teepee> nice! I needed a convexity=3 for the rotate_extrudes() to prevent preview glitches

18:17 <gbruno> [github] t-paul closed issue #4886 (Ctrl-C in the editor deletes selection) https://github.com/openscad/openscad/issues/4886

18:35 <InPhase> teepee: Always unpredictable with those... My preview was fine this time.

18:35 <teepee> yeah, it depends on the algorithm and gpu driver it seems

18:35 <InPhase> I look forward to the purge of preview and its quirks. :)

18:36 <InPhase> At least then we will fail only one way!

18:36 <teepee> I'm not fully convinced, but we'll see

18:36 <InPhase> Not for this release I think, as that's far too ambitious, but after it, I look forward to a "prove we shouldn't" aggregation of issues. I bet it won't be too big.

18:42 <h1> InPhase: thank you

18:46 <InPhase> h1: Sure. Hope it works out for what you're working on. :)

19:21 <h1> thanks

20:17 <ali1234> what ever happened with the mailing list and forum? i remember people were talking about it like 4 years ago

20:17 <ali1234> what is empathy and why does it look like vaguely discourse?

20:19 <teepee> the old "forum" called nabble died

20:20 <teepee> the mailing list is mostly unchanged (I think), they just added those more or less working web frontend