<ali1234>
snapd is finally fixed, i can use the snaps again :)
mmu_man has quit [Ping timeout: 272 seconds]
kintel has joined #openscad
TylerTork has joined #openscad
<kintel>
guso78 not annotation of edges, just wireframe overlay. Same method as before, just generalized a bit to support PolySet rendering
<kintel>
teepee Did you root-cause the print button issue? I see the same issue here, so I can take a look unless you're already on it
<teepee>
only so far that I could only find 2 lines making it invisible and none making it appear again
<kintel>
ok, I can try to bisect it,
<kintel>
J25k In terms of variable name warnings, I suggest responding to Jordan's announcement on the mailing list and express the pain point of lots of warnings. Perhaps he has some ideas for how to address it
<teepee>
I'm not going to have any spare time over the next 2 weeks
LordOfBikes has quit [Ping timeout: 245 seconds]
<kintel>
Time is never spare, it's meticulously carved out from a packed schedule :/
<teepee>
yep, that's the normal case :)
LordOfBikes has joined #openscad
<TylerTork>
I'm trying to understand what was meant by this, in the documentation of the text function:
<TylerTork>
Multi line text is not supported with `text()` but translating each line size × .72 will result in a leading (line spaceing) of 1 em (em= fonts body height or points). 20% (× 1.2) may be added.
<TylerTork>
So if the size is 10, this is saying translating subsequent lines by 10*.72 will give a reasonable line spacing?
<TylerTork>
https://prnt.sc/M2x2lZFMm6uC That's obviously not right. Can someone suggest how to interpret what was meant here?
<J25k96>
with errors it is much more a problem as they get slower with each error so when calculating points it often is necessary to kill and restart oSCAD. And the "stop at first warning" should be first 5 or 10 .. so repair with manifold doesn't cause a stop (in most cases)
<guso78k>
teepee, would you expect/support/wish that there is also a cheat-sheet for python in openscad ? So what shall go there ad python really aims to be identical to SCAD in function names and parametert names ?
dingodoppelt_ has joined #openscad
dingodoppelt has quit [Read error: Connection reset by peer]
dingodoppelt_ has quit [Remote host closed the connection]
dingodoppelt has joined #openscad
J25k96 has quit [Quit: Client closed]
J25k96 has joined #openscad
guso78k has quit [Ping timeout: 240 seconds]
ali12341 has joined #openscad
ali1234 has quit [Remote host closed the connection]
aiyion4 has quit [Remote host closed the connection]
aiyion4 has joined #openscad
J25k96 has quit [Quit: Client closed]
J25k96 has joined #openscad
inkmoth has joined #openscad
<inkmoth>
I've been testing one of my files in different versions, and I've noticed that I'm getting different rendering results in the latest 2025 nightly versus the older build I'd been running. Specifically, things seem to render one way in the 2021 stable [and the 2023 dev build I'd been previously using for a while] and another way in the 2025 snapshot
<inkmoth>
I just downloaded. Is there any place I can go to in order to see if the import() function has been modified lately? I'm expecting that to be related to the problem as it's a translation issue
<J25k96>
inkmoth it is a different view in the images
<J25k96>
the 2025 version uses manifold render (can be set in preferences)
<J25k96>
import has a "center" parameter
<inkmoth>
I'm already using center = true in both of those renders. I'll check for that renderer setting, but the 2025 version can't even perform a F6 render as the STL is in an invalid place in 2025 -
<inkmoth>
Rendering Polygon Mesh using CGAL...
<inkmoth>
ERROR: CGAL error in CGAL_Nef_polyhedron3(): CGAL ERROR: assertion violation!
<inkmoth>
I'm eventually planning on using the new feature for import() to take json, so I need to sort out why things are broken here :D
aiyion4 is now known as aiyion
<inkmoth>
I can probably fix the problem by adjusting the translation coordinates until it is solved, but I guess my question boils down to "Is the current import code in the nightly correct"? If it is going to remain this way, I have no problem just changing my files to match the new method
kintel has joined #openscad
<kintel>
I don't think center=true was implemented for STL imports in 2021.01, so that's a difference
<kintel>
2021.01 and the latest snapshot is more than 4 years apart. Lots and lots of changes has happened since then, but generally if you use the same parameters, you should get the same results, as long as all parameters are supported.
<kintel>
If you share your STL, I can take a look
<inkmoth>
Yeah, that's why I was hesitant to post at all. I can bisect it a little better than 2021, since I know that the original file was designed (and renders the same as the 2021 picture I posted above) in 2023.09.30 - but that's still pretty dang ancient
howiemnt has joined #openscad
<inkmoth>
The imported STL is from https://www.thingiverse.com/thing:5917616/files. I can confirm that the 2023.09.30 generated output file "was correct" as I actually 3D printed it successfully. I think the best step here is for me to quickly patch the corrected coordinates in so I can toggle between them and then upload my scad file - it's destined to be
<inkmoth>
open source anyway
<othx>
inkmoth linked to "Simple Game Boy Printer Stand by JPRanger18" on thingiverse => 1 IRC mentions
TheAssassin has quit [Remote host closed the connection]
teepee has quit [Remote host closed the connection]
kintel has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
L29Ah has joined #openscad
bertronika has quit [Quit: Konversation terminated!]
bertronika has joined #openscad
TheCoffeMaker has joined #openscad
ali12341 is now known as ali1234
<howiemnt>
Is there a way to import an openscad library INTO python and use it in blender?
<teepee>
well, mostly no, but maybe in theory ;-)
<howiemnt>
damn
<howiemnt>
so what exactly does pyopencad do?
<teepee>
guso78 had something that might go in that direction
<teepee>
no idea, there are hundreds of those projects, most generate openscad source code only
<teepee>
we do have the beginning of real python integration
<teepee>
and guso78 mentioned something installable via pip
<teepee>
so *maybe* that could cover what you asked
<teepee>
what specifically?
<howiemnt>
i have seen things that look to take python and use it inside of openscad, but not the reverse
<howiemnt>
im doing some wild stuff in blender and I showed it to some heavy hitters who said "blender is too complex for this, openscad is too basic, but running open scad inside python in blender would work"
<teepee>
one too complex, one too basic but the magic combination is just right?
<howiemnt>
im not really a programmer and i only messed around with opeenscad but had no need for it since i can do cad the normal way
<teepee>
a bit like a fairy tale :)
<howiemnt>
a bit
<howiemnt>
there was nothing openscad could do in 2013 i couldnt do with inventor
<howiemnt>
so i just played with it and forgot about it until yesterday
<teepee>
uh, is that what was discussed yesterday or so
<howiemnt>
YESTERDAY!
<teepee>
ping ali1234?
<howiemnt>
i have been talking with him
<ali1234>
hello, yes
<howiemnt>
i just figured i would come to the openscad pros and ask around
<howiemnt>
ali1234, lol
<teepee>
yep, same link
<howiemnt>
ali1234, wtf ?
<ali1234>
yes, "import openscad" is coming. its not ready yet though
<howiemnt>
ahhh
<howiemnt>
ali1234, nice to know
<howiemnt>
sorry for barging in here like a dipshit
<teepee>
all good, no worries
<ali1234>
teepee: you know i've been requesting this for years :)
<teepee>
people connecting is good :)
<howiemnt>
ali1234, see now that I have asked about it, maybe the universe will make it so
<teepee>
requesting what?
<ali1234>
openscad integrated into blender
<teepee>
what did they say?
<ali1234>
who?
<teepee>
blender developers
<ali1234>
they said "you can't do that"
<teepee>
I mean it's probably easier via python, but the C++ route was there always
<ali1234>
but if openscad is a python module, i can install it in blender that way
<ali1234>
C++ route means i have to maintain a fork of blender
<ali1234>
nobody got time for that
<teepee>
true, if you can't convince anyone else
<howiemnt>
if i ask chatgpt to make a module for openscad to work in python, what will happen?
<ali1234>
all i need from blender side is the ability to define objects/nodes/modifiers from python code
<ali1234>
howiemnt: it will generate a bunch of nonsense that doesn't work
<howiemnt>
lol
<ali1234>
AI can't do such complex things
<teepee>
also no do *new* stuff
<ali1234>
it can sometimes do novel things, if they are really simple and obvious
<teepee>
clever combining things yes, so with lots of python code around that can work
<ali1234>
but anything that takes more than about 100 lines of code it has basically zero chance
<ali1234>
blender devs are somewhat against implementing nodes in python, because it is "slow" and may freeze the GUI (much like it tends to do in openscad really)
<ali1234>
but i think they might come around eventually, i mean you can freeze blender already quite easily
<ali1234>
but even without that you can still do it with a plugin
<ali1234>
you just don't get a nice GUI for it
<ali1234>
teepee: did you ever play with geometry nodes?
<teepee>
comparing the funding they should be able to do that easily ;-)
<teepee>
in blender? a tiny bit
<teepee>
doing a tutorial
<ali1234>
what i'm imagining is basically a node for that: geometry input and output, and "what it does" is defined by a user-supplied python script (from a text resource)
<ali1234>
given that, and the ability to import openscad, you now have boolean that actually works
<ali1234>
but you could also define geometry completely from scratch, and leave the input unconnected
<ali1234>
for example you could define some shapes with openscad code, then output it into a blender bevel op
<ali1234>
one downside is you'd always have to render to get a preview
<ali1234>
but on the up side, your "preview" could have cool material shaders very very easily
inkmoth has quit [Ping timeout: 240 seconds]
J25k96 has quit [Quit: Client closed]
J25k96 has joined #openscad
inkmoth has joined #openscad
<howiemnt>
I have a question, why do programmers tend to prefer making objects using openscad versus traditional cad like inventor or solidworks?
inkmoth has quit [Ping timeout: 240 seconds]
TheAssassin has quit [Remote host closed the connection]
TheAssassin has joined #openscad
inkmoth has joined #openscad
bertronika has quit [Quit: Konversation terminated!]