kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
UltimateCodeWarr has joined #openscad
<UltimateCodeWarr>
So is there anyway to setup some kind of 'pantomime' that will trace a 2D shape on a 3D Curved surface like a Cylinder or Sphere?
<peeps[zen]>
UltimateCodeWarr: do you mean like embossing a pattern onto a curved surface?
<UltimateCodeWarr>
Yes
<peeps[zen]>
not really any direct way to do so. i would imagine some libraries probably support an action for sphere and cylinder specifically, but don't know off top of my head
<peeps[zen]>
and if you wanted to support arbitrary 3d curved surfaces i think that would be quite difficult
<peeps[zen]>
without direct access to geometry data in userspace i think any solution would require some compromises, like breaking a 2d shape into small disjoint segments and projecting onto the surface
<pca006133>
I feel like even with direct access to geometry data, this is a difficult problem in general
<pca006133>
or not well-defined
<UltimateCodeWarr>
It's like you would want to load a skeleton file where black is the trace path, and then you would specify either a cut depth or raised amount and a tool shape (like a triangle or sphere), and then where to center it on a particular surface. The mechansim would just use some sort of polar cords to translate the 2D shape on the surface of the
<UltimateCodeWarr>
3D shape
<UltimateCodeWarr>
I've seen CNC laser/plasma torches take a 2d drawing and cut it out on pipe
<peeps[zen]>
usually that's with a rotary axis, unless you mean on a fancy robot arm
<UltimateCodeWarr>
Not the stuff in my price range, it's rotary 4-jaw chucks
<peeps[zen]>
in which case the rotary axis sorta does the transformation for you, you are basically sending 2d data to the CNC
<pca006133>
if the user can specify a transformation function, it is well-defined...
<pca006133>
because otherwise you have cases like mapping a square onto a sphere for example, or maybe into things with holes, in general when things have different topology you have to make arbitrary decisions
<peeps[zen]>
just remininded me of some fun trivia: a triangle on the surface of a sphere can have 3 right angles
<peeps[zen]>
*reminded
aiyion1 has quit [Remote host closed the connection]
aiyion1 has joined #openscad
extor has quit [Ping timeout: 245 seconds]
extor has joined #openscad
extor has quit [Quit: ZNC 1.8.2+deb2ubuntu0.1 - https://znc.in]
<J24k58>
f0lkert would it be possible to put all "tweakable" items at the top of the pov file? .. I have files with 500mb - so scrolling up down is something
<guso78>
howeverm when I try it with povray and looks very bad and i basically cannot recognize my model again. is it an error in the file, in my povray or simple wrong usage ?
<guso78>
i expext to see something like this and i plan to make it look like green transparent glass
<f0lkert>
i'm puzzled!
<f0lkert>
as if stuff is missin
<f0lkert>
g
<f0lkert>
do you have the code for me? the .scad-file?
<guso78>
sorry, no scad, i created with pythonscad. maybe STL is fine, too ?
<f0lkert>
no
<f0lkert>
*pondering*
<f0lkert>
guso78: does the stl look ok?
<guso78>
i have just seen, that my model is not completely closed in the top, could that be an issue ?
<guso78>
i will check stl
<f0lkert>
guso78: I don't think so no. yes please check the stl. I'll see if i can install pythonscad in the mean time.
<guso78>
yes, stl looks fine apart top closing
<guso78>
will share stl with you, compiling pythonscad could take some time
<guso78>
guenther-sohler.net/wrap.stl
<guso78>
I really like to add povray into my flow, I'd love to see it working in my place :)
<f0lkert>
yeah the results are better than I expected myself
<guso78>
.. aparently i need to improve my wrap function to handle more situations ;)
<f0lkert>
another one for the todo! :)
teepee_ has joined #openscad
teepee has quit [Ping timeout: 260 seconds]
teepee_ is now known as teepee
<f0lkert>
guso78: are you the pythonscad developer? because I get: /home/folkert/t/pythonscad/src/python/pydata.cc:5:10: fatal error: Python.h: No such file or directory
<f0lkert>
even though i installed libpython3.13-dev
<guso78>
f0lkert, this is interesting, which distro are you working with ?
<f0lkert>
ubuntu 24.10
<f0lkert>
I ran cmake, then make, then install libpythonbla, then ran cmake again and make
<guso78>
do you have experience with cmake ? maybe you can find out which include directories are and where your Python.h sits ?
<guso78>
that should be fine. later today i can test, if there is an issue with python3.13
<f0lkert>
my knowledge of cmake is minimal unfortunately, but I'll see what I can do
<guso78>
(i still use python 3.11 , because its much easier to handle, in respect to threads even though 3.13 is recent)
<f0lkert>
I specially upgraded to 24.10 because of 3.13 threading improvements :-)
<guso78>
:) but with python threads you have to lock and unlock the GIL and python will crash if you don' t do correctly '=( '=(
<UltimateCodeWarr>
Did you install anything with SNAP?
<f0lkert>
guso78: indeed
<f0lkert>
UltimateCodeWarr: is that a question for me?
<UltimateCodeWarr>
Yes, I've had quite a few issues with SNAP sandbox folders and things not working because paths are not what they have been.
<othx>
InPhase linked to YouTube video "4D Quaternion Hypercomplex Mandelbrot Set" => 1 IRC mentions
<InPhase>
That's the sort of thing you see before you get shredded into pulp by aliens from a higher dimension unconcerned with your lower order existence.
<f0lkert>
:)
<guso78>
the 4D thingy looks quite disturbing to me
<InPhase>
guso78: That probably means your mind is functioning.
<themind>
I've been using openscad-nightly with --enable=manifold, however with recent builds I'm getting "Ignoring request to enable unknown feature 'manifold'.". Wondered if maybe nightly builds had manifold feature turned off for builds, but did a build myself with it enabled and the feature still seems to be missing. What am I missing here?