05:50 <J23k19> why does i get this error "Fontconfig error: Cannot load default config file: No such file: (null)" when using command line scad?  (v 08.2023 )

05:51 <guso78k> hmm , must be an QT issue, but it should not be deactivated in command line mode.

05:55 <teepee> J23k19: which OS is that?

05:56 <J23k19> windows

05:56 <teepee> can you try with full path for the application, like c:\...\openscad.com

05:57 <teepee> but on windows that should not be an issue

05:57 <J23k19> that is what i do

05:58 <teepee> right, that path thing should only happen on Linux

05:59 <teepee> I don't remember where it does lookup the fontconfig files

06:00 <teepee> well, work calls

06:01 <gbruno> [github] gsohler synchronize pull request #4703 (Store Points in PolySet as Indexed Points) https://github.com/openscad/openscad/pull/4703

06:02 <J23k> funny thing the file contain text only as "%"  but that is still evaluated for render as it seems

06:02 <guso78k> J23k reports, that this problem only happens in command line mode ... and i dont believe this path thing does different in command line mode than in gui mode

06:04 <guso78k> just managed to compile openscad successfully again with MSYS64 but it does not launch. anybody successfully run it lately ?

06:13 <J23k> so i tested a file with 3D text only and beside the error (on command line) i get a file with the 3D text ..    but i also get this error when just starting open scad via command line

06:16 <guso78k> maybe openscad tries to find fonts in several places, fails in 1st place and succeeds in 2nd ?(warning issued in 1st place) ?

06:16 <J23k> open a file to be precise as starting calls the launcher. But the fontconfig error shows first when closing the file

06:20 <J23k> hm ok opening a file now doesn't show this anymore - but rendering from command line does

06:29 <guso78k> J23k can you check if you see the error in GUI in the console window instead ?

06:30 <J23k> i have but nothing there

06:41 <J23k> just tried some versions and it first show with 2023.01  while 2022.11 doesn't  ..

06:42 <J23k> as nobody noticed it before maybe some windows update causing it?

06:44 <guso78k> as you did not update your windows installation  in between the difference must be with the different openSCAD version

06:44 <guso78k> suppose both openSCAD relases have a tag in the repository. maybe use "git difftool" on the src/gui folder ?

06:51 <J23k> there sure is a difference in scad but i thought that the error might be triggered because something in windows changed and wasn't there before - on the other hand i haven't run oSCAD on the command line with 3Dtext

06:55 <guso78k> Yes, environments do change. lately MacOS compiler became more picky and did not accept OpenSCAD source code, which did compile before. it had to be expressed more exactly.

13:36 <gbruno> [github] gsohler synchronize pull request #4702 (Python pr3) https://github.com/openscad/openscad/pull/4702

13:38 <guso78k> ctests for all geometry functions added.  see https://github.com/gsohler/openscad/blob/python-pr3/tests/regression/echotest_python/abuse-expected.echo

13:40 <guso78k> ideas  for further "DAU"  function invocations welcome

14:29 <gbruno> [github] gsohler synchronize pull request #4702 (Python pr3) https://github.com/openscad/openscad/pull/4702

18:25 <dTal> I wish we could assign anonymous functions to local variables within function definitions

18:33 <dTal> or can we?

18:34 <teepee> via let, yes, that should work

18:38 <teepee> like so: https://pastebin.com/kVi5busa ?

19:06 <teepee> joseph_: do you have any updates? it's a bit quiet

19:57 <joseph_> teepee: Thank you for checking in. I was going to message today because it's been a couple weeks. I got slowed down for a while on a mystery problem where shaders couldn't be loaded. I knew it was specific to my system because it didn't happen back when debugging entirely in Docker (which I'm no longer doing this year, so my tests are on the actual graphics hardware).

19:57 <joseph_> Last night I found workaround by moving the location of all the GLSL files to a hardcoded directory instead of finding them with PlatformUtils::resourcePath. This isn't a permanent solution, but it will unblock my work for now. I'll share details if I find a real fix.

19:59 <joseph_> In addition to that issue, a few hours of work time each week were displaced by various other commitments. I am considering an extension request, which seems to still be possible. I think it can happen any time before the final submission period begins (which is still 10 days away).

20:18 <guso78k> Analyzing the recent paste with the let statement: i was not used to it before , but i like the let statement as you can assign functions. Documentation in https://en.wikibooks.org/wiki/OpenSCAD_User_Manual/Conditional_and_Iterator_Functions#Let_Statement tells that let can redefine a variable for a subscope(also great) but in the shown pastbin the

20:18 <guso78k> sq variable was used outside of the let statement. what' s the actual validity scope of variables which are defined inside let ?

20:45 <teepee> joseph_: ok, please drop some progress report, even if small a bit more often so we can keep in touch

20:46 <teepee> not sure if you have seen that yet, the Matrix bridge is currently down

20:49 <teepee> guso78k: it's not outside, functions have no braces, imagine in would have: let(a = 2, b = a + 2) { return a * b }

20:49 <teepee> so that's only: let(a = 2, b = a + 2) a * b

20:58 <teepee> joseph_: incidently someone seems to have filed a bug report recently which sounds related to your folder problem https://github.com/openscad/openscad/issues/4715 - it has a pointer to some solutions

20:59 <guso78> Teepee thanks, Now i understand. I was confused with lisp/Skill language. Where let variable scope IS inside let( braces.

21:01 <guso78> From todays discussion do we know that this only applies for Windows?

21:01 <teepee> what is?

21:01 <teepee> oh, the fontconfig thing?

21:02 <guso78> 2nd Statement was Fontconfig

21:02 <teepee> I'm not sure that's the same issue. Windows has a dedicated API call to get the application folder

21:06 <guso78> Teepee do you know If anyone was recently able to successfully use Jordans instructions to build openscad in msys64 ?

21:08 <teepee> I don't know. didn't you say you built successfully?

21:08 <guso78> My binary Launches and immedialy finishes without an Error and i got No idea how to Debug that

21:08 <guso78> Build yes, Run no

21:09 <teepee> oh :(

21:09 <guso78> Not even Help displays

21:10 <guso78> Hmm maybe i try Checkout a 1 year old Version 😅

21:10 <teepee> you could try running from gdb

21:11 <guso78> Yes good hint. ITS available in msys env

21:53 <Guest61> Hey guys ! I am a newbie trying to see if openscad can be used for some monte-carlo style rendering and modelling of particle interactions with some integrated circuits, I see in the documentation that openscad is not meant for "rendering" , anybody tried material interaction with some sort of gaussian distribution added to this interaction ?

21:54 <teepee> that sounds pretty far off from what openscad can do

21:55 <InPhase> Guest61: Yep, way off on the target program for that one. It's more for making 3D printing models and similar type tasks.

21:56 <Guest61> Ahh ok, yea sounded like it, any opensource tools which can do the same ?

21:56 <Guest61> Mind you I am not trying to do the actual interaction, that can only be done in TCAD, but as a way to visualize these interactions is all I want, as an abstraction

21:57 <InPhase> Probably! But the last I did serious particle simulations was 22 years ago before most of the open source tools for that sort of thing existed.

21:58 <InPhase> ##physics is probably the better place to ask. I'm also in there, and won't know the answer any better in that channel, but maybe someone else will. :)

22:00 <Guest61> got it, there is a wrapper some people have developed to visualize integrated surface cross-sections, I though an extension beyond that would be to rotate a particle track on a variable sphere around that  and do a random walk on the same

22:00 <Guest61> https://codeberg.org/tok/gds3xtrude/src/branch/master

22:00 <Guest61> Thanks for the pointer though!!

22:02 <InPhase> I see how you found your way here from that repository.

22:02 <InPhase> OpenSCAD can in fact display things like shown in that image if fed the right input text.

22:04 <InPhase> Your ability to do display of other things would be limited to what you can feed in similarly from external tools.

22:05 <Guest61> (y) Ok if anyone wants to collaborate or has ideas on something similar do get in touch with me chand.leo101@gmail.com

22:06 <Guest61> Yea like I said I understand this is not a simulation tool, but lets say I had 100 input text files, each as an output of such a sim (whichever tool we do that in), we can then render this one openscad file at a time

22:22 <J23k> blender will give you  better render .. or povray.org

22:33 <Guest61> (y)

22:50 <dTal> J23k: openscad is much much faster than povray for rendering shapes, though it can't do fancy effects

22:51 <joseph_> teepee: I just read that issue and will investigate. I'm not sure yet what it means by when there is "no Qt" because Qt is a dependency of OpenSCAD. Also, do you have any opinions about the extension? As I recall from last year, they need to get requested through the org admin

22:52 <dTal> blender is kind of a pain in the butt, you can't write .blend files with external tools and there's no text format blender can load that lets you program shapes

22:53 <dTal> I suppose you could kinda get close by dumping python code that builds a scene, but that's not very declaritive and the api will be a pain

23:09 <teepee> joseph_: yes only org admins can setup changes via the gsoc dashboard, at least that was how it worked last year

23:10 <teepee> I assume "no Qt" means building with -DHEADLESS=ON which will disable the GUI. the WASM build does that automatically but it's possible to do "normal" builds without GUI/Qt too

23:11 <teepee> the issue might still appear when just using command line mode, even if Qt is compiled in though