00:00 <peepsalot> a quick commit to improve the messages would be great though :)

00:02 <InPhase> https://bpa.st/PTOQ The comments from CGAL source about the exception it's throwing.

00:02 <InPhase> Which I was only able to obtain with this edit. So yeah, worth committing.

00:03 <linext> i tried studying the https://ochafik.com/openscad/ example today

00:03 <linext> i removed unused code and made all the libaries/dependencies load locally

00:05 <gbruno> [github] rcolyer pushed 3 modifications (Catch and log std::exception for compilation) https://github.com/openscad/openscad/commit/902b23715867702442c6f2dbbbd77f65f7073170

00:06 <gbruno> [github] rcolyer opened pull request #4354 (Catch and log std::exception for compilation) https://github.com/openscad/openscad/pull/4354

00:10 <InPhase> There, PR opened and issue comment added to document.

00:11 <InPhase> peepsalot: Hopefully you find "compilation" adequately generic to suffice. The other place this can be triggered is during a tab open. TabManager has some function that will trigger the same.

00:12 <InPhase> But at the very least, the .what() messages should be the real information going forward.

00:12 <peepsalot> yeah, definitely better than "parsing"

00:13 <InPhase> Now I suppose none of that says what's actually the problem. But it's at least in the ballpark. :)

00:14 <InPhase> Maybe some sort of zero-volume piece issue?

00:14 <InPhase> Like with two vertices the same?

00:35 <JordanBrown> So the simple summary is "Something incomprehensible failed. You lose.".

00:38 <linext> i wish there were some more examples like this: http://3dcustomizer.net/openscad-gear-configurator/configurator/index.html

00:38 <linext> ochafik.com/openscad/ is pretty complex

00:38 <linext> probably about 50 files, some code is generated automatically and is not designed to be edited or debugged

00:39 <linext> minified

00:42 <peepsalot> JordanBrown: welcome to CGAL :P

00:43 <JordanBrown> Yeah, it conveyed just about exactly as much information as any other CGAL message.

00:43 <JordanBrown> "It Failed. Have a Nice Day."

00:44 <JordanBrown> linext examples like that in what way?

00:44 <linext> importing libraries/fonts/other.scad files, reading stdout and stderr

00:45 <JordanBrown> Ah, you mean the web page that runs that OpenSCAD customizer, as opposed to the particular model that it offers for customization.

00:45 <linext> you can get a feel for it: view-source:https://ochafik.com/openscad/

00:46 <linext> yea, for my 3dcustomizer.net idea, it needs more features

00:49 <linext> the ochafik.com/openscad demo actually unzips compressed libraries and puts them into the web assembly file system

01:22 <linext> is it possible to output a copy of the input.scad file but with the variable substitutions made?

01:23 <linext> openscad input.scad -p ... -P ... -o output.scad

01:26 <peepsalot> can you even throw something that doesn't inherit from std::exception?

01:27 <InPhase> Try /msg geordi { throw 1; }

01:28 <InPhase> It's not really a great idea, but you can. :)

01:32 <peepsalot> huh, interesting

01:49 <JordanBrown> What is geordi?

01:50 <peepsalot> bot from #C++

01:52 <Friithian> geordi is a very useful bot

08:12 <J1A8433188237> so it printed more or less https://www.printables.com/model/276612-honycomb-projection-lampshade

16:12 <J1A84> so i integrated the support into the design .. now it is a bowl https://www.printables.com/model/276787-hexagon-bowl

21:16 <gbruno> [github] PerthGoat opened issue #4355 (Put an msys2 build on the website in the downloads section to fix reliance on Microsoft Media Feature Pack binaries) https://github.com/openscad/openscad/issues/4355

21:16 <gbruno> [github] PerthGoat edited issue #4355 (Put an msys2 build on the website in the downloads section to fix dependency on Microsoft Media Feature Pack binaries) https://github.com/openscad/openscad/issues/4355

22:33 <dalias> is there any trick for using openscad to assemble imported STLs that are entirely non-overlapping without gigantic useless CGAL computations to determine that they're non-overlapping?

22:37 <gbruno> [github] rcolyer synchronize pull request #4354 (Catch and log std::exception for compilation) https://github.com/openscad/openscad/pull/4354

22:37 <gbruno> [github] rcolyer pushed 1 modifications (const reference exceptions Co-authored-by: Torsten Paul <Torsten.Paul@gmx.de>) https://github.com/openscad/openscad/commit/962c8fab57fb01e3a6dd696160ec48df1cdd116a

22:37 <gbruno> [github] rcolyer pushed 1 modifications (const reference exceptions Co-authored-by: Torsten Paul <Torsten.Paul@gmx.de>) https://github.com/openscad/openscad/commit/05d11d06605eaba851ab64dc359f23ef34f075fa

22:37 <gbruno> [github] rcolyer synchronize pull request #4354 (Catch and log std::exception for compilation) https://github.com/openscad/openscad/pull/4354

22:38 <joseph_> teepee: I forgot to mention earlier that I fixed the segfault issues occurring with 2D test cases. Now there are only 4 failing tests with my PR. One of them is just a visual issue related to wireframe not being shown when toggled (hopefully a trivial fix). The remaining 3 failures are in lazy-union-root-for test cases. The output render makes it through 5 columns of cubes, but fails before the 6th can be rendered. What seems to

22:38 <joseph_> be happening is that evaluateGeometry produces a polyset, which would normally be a nef polyhedron when the lazy union feature is turned off. Something about this polyset is incompatible with my changes to the attribute construction in VBO/CGALRenderer.

22:39 <joseph_> GeometryEvaluator.cc line 500 has a couple notes in the comments, and from the blame it looks like peepsalot was the last one to work on it. I wonder if the issues I'm having could be related to the temporary approach for root nodes that is mentioned in that comment. Because the only real error across the whole test suite is for the lazy union root modifier.

22:40 <InPhase> joseph_: peepsalot gets blamed for a lot of stuff from having fixed up some whitespace. ;)

22:40 <InPhase> Although he was actually working in that area not too long ago.

22:43 <JordanBrown> I need to print a big copy of https://xkcd.com/386/ and post it on my wall.

22:43 <JordanBrown> Nothing to do with OpenSCAD today... I just had to restrain myself from jumping into a discussion on another mailing list.

22:43 <InPhase> Does he not sell xkcd posters? Because he should really be selling them.

22:47 <dalias> the idea is to assemble print-in-place articulated models from constituent parts modelled outside openscad

22:47 <dalias> but it's unusably slow because openscad insists on doing CGAL unioning on imports that don't intersect each other

22:50 <peepsalot> joseph_: define "fails before the 6th can be rendered". are you saying the 6th column is just not displayed or are you getting an exception of some sort?

22:55 <joseph_> peepsalot: In the CI openscad exits with 0 and doesnt render the column. When debugging, it's an in valid operation error from OGL and exception in VertexState destructor. I'm about to be driving for 45min but then I'll be back online

22:56 <peepsalot> can you upload your test report html somewhere?

22:56 <peepsalot> or, if its using CI, then i guess just link to the specific CI build

23:03 <InPhase> dalias: Did you try enabling lazy union?

23:06 <dalias> hm?

23:08 <InPhase> dalias: Well with fast-csg and lazy-union both enabled on the master branch, my two separated $fn=300 spheres render in 0.05 seconds.

23:08 <dalias> oh i guess it's something new i don't have yet

23:08 <dalias> i guess i need to build from soource. distro doesn't have a new ver because there hasn't been a tag in forever :/

23:08 <InPhase> There's an AppImage.

23:09 <InPhase> https://openscad.org/downloads.html#snapshots

23:09 <dalias> yeah that's not gonna work. either it would just not run at all because it assumes a glibc host or it would pull in its own glibc ecosystem (including opengl drivers?) that would be gigantic :-p

23:10 <InPhase> Well it's a 45MB file. How bad could it be?

23:10 <dalias> i'm guessing it just wouldn't run

23:10 <InPhase> What sort of computer is this?

23:11 <InPhase> Until you start rendering giant stuff, it's pretty lightweight. :)