01:01 <joseph_> Is the forum down? I am getting an Unable to connect error

01:31 <teepee> joseph_: what link? https://lists.openscad.org/empathy/list/discuss.lists.openscad.org works

02:03 <joseph_> teepee: It seems to happen only with links to https://forum.openscad.org but using http instead will successfully redirect me to the correct site. I encountered this after clicking a link in Google results, which uses https

02:04 <teepee> ah, indeed. might be an error when the shortcut link was switched not long ago

02:10 <gbruno> [github] kintel pushed 7 modifications (Use gettext functionality from glib instead of directly from gettext) https://github.com/openscad/openscad/commit/8346b391975e97f2475250731551037ffe0af64c

02:18 <kintel> Another chapter in the world of python: Does anyone have a convenient, multi-platform way of correctly callinh python from a shell script?

02:19 <kintel> e.g. depending on system, the python interpreter is called "python", "python2", or "python3", but any of those may also be missing.

02:20 <kintel> I noticed this because scripts/translation-update.sh stopped working on macOS, since "python" is no longer shipped; it's only "python3" now..

02:21 <kintel> I can rename to "python3", just not sure if that would break other platforms

02:24 <InPhase> kintel: On everything but Windows, /usr/bin/env python3

02:25 <InPhase> "python" sometimes exists. "python3" always exists, except on python2 only systems which no one is supporting anymore.

02:25 <InPhase> So there's no need to call anything but python3.

02:26 <InPhase> The /usr/bin/env will work for automatically finding the path of python3 in a #!

02:27 <InPhase> And in any place on Windows that can actually handle on a #! line, /usr/bin/env is probably present there too.

02:28 <InPhase> If you just need to call it directly on a regular line, and it's in the path, then just plain old "python3" should work reliably on all systems worth caring about.

02:50 <kintel> InPhase Sounds good - I'll get the remaining scripts updated to python3 :)

03:17 <gbruno> [github] kintel pushed 3 modifications (Use python3 instead of python) https://github.com/openscad/openscad/commit/b71ee22e97a503eb0656403f7383ed92be1d2189

03:17 <gbruno> [github] kintel opened pull request #4657 (Use python3 instead of python) https://github.com/openscad/openscad/pull/4657

12:44 <gbruno> [github] kintel closed pull request #4657 (Use python3 instead of python) https://github.com/openscad/openscad/pull/4657

12:44 <gbruno> [github] kintel pushed 3 modifications (Use python3 instead of python (#4657)) https://github.com/openscad/openscad/commit/c8c246c0c0fbc762a06ea102ce2f377386a47707

12:45 <gbruno> [github] kintel pushed 13 modifications (Use gettext functionality from glib instead of directly from gettext) https://github.com/openscad/openscad/commit/b116b85c65d0e50c86bda0e8aaf0dc078fd2f5c6

12:45 <gbruno> [github] kintel synchronize pull request #4656 (Use gettext functionality from glib instead of directly from gettext) https://github.com/openscad/openscad/pull/4656

17:58 <pa> hi folks, [OT] does anyone happen to have an STL for a handle for a file?

17:59 <pa> nvm found one

19:21 <guso78[m]> Special question. Does an object stay manifold, when at an Edge only one of the halfedges IS Split into two colinear halfedges. That means two halfedges abut to one Long halfedge.

19:54 <J23k59> isn't manifold if none points are shared in two volumes?

20:09 <guso78> my question is just one volulme

20:16 <joseph_> kintel: Here's a summary of what I've found so far about the artifacts. For some context, I try to build all of my projects in Docker. Last year I had issues matching the libc version with my host OS, so I was actually also running OpenSCAD inside Docker the whole time (with X forwarding). That meant the flickering/z-fighting artifact didn't appear at all.

20:17 <guso78> https://imgpile.com/i/9f650N i refer to this. is this an manifold object ?

20:18 <joseph_> kintel: But I just fixed the dynamic library issues and I now can build OpenSCAD binaries that run directly on my Linux host. What I've found is that my code from last year on #4330 has the flickering on F5 show edges (no different from master). BUT, I haven't seen any artifacts in F6 mode when using GLSL shaders.

20:21 <teepee> guso78: I think that is manifold if that those 2 middle vertex are attaching to the new edge and to 2 half length edges each

20:24 <guso78> ok great. this is relevant when when creating data for my polyhedron. i believe such a  solid as shown can *ALWAYS* be reduced to have always mirrored halfedges, but generation of shown object is easier

20:27 <J23k59> guso78 you may get a problem if the other faces go over the additional center vertex.

20:28 <J23k59> imagine you move those vertex up a little .. then there would be an open gap to the side walls and not being watertight anymore due to floating point imperfection

20:34 <J23k59> https://imgpile.com/i/9f6zwk  is  what the faces need to be

20:47 <guso78> yes, but everything is calculated mathematically. no chance for imperfection.

20:47 <guso78> but i got the idea

21:05 <J23k59> you can not make a face with a vertex on the edge that is not used for the face.  And you always have floating point error which is why you need to overlap objects in scad

21:12 <J23k59> guso78 if you use a computer there is no perfect math https://0.30000000000000004.com/

21:13 <guso78> great site! i did not know this one

21:23 <J23k59> only hard to remember the url Ü

22:08 <guso78> you simply put the number of 0's into your brain. its  15 exactly ...

22:17 <teepee> I just starred the github repos so I can find it via "0.3" :)

22:19 <teepee> sadly searching for just 000 does not work

22:24 <guso78> haha, all contemporary stuff must be on github nowadays ...

22:25 <guso78> ... is printing his 1st qr code now ...

22:59 <kintel> joseph_ Makes sense, F6 mode doesn't use OpenCSG, so the artifacts would be gone. However, I would expect we'd want custom shaders to be applied during F5 as well, as we try to make the two modes as similar as possible.