<kintel>
There is no such test. Did you mean "examples_roof" you think?
<kintel>
(trying to figure out some off behaviors)
marcus has quit [Remote host closed the connection]
marcus has joined #openscad
<peeps[zen]>
uh oh, a pop quiz on a 1yr old commit?
<peeps[zen]>
the test naming is something I don't know well off the top of my head, but iirc they are composed from "type" openscsg/cgal etc + some general scad naming
<peeps[zen]>
i mean, the *roof pngs from that PR should have been generated using normal TEST_GENERATE run afaik, so maybe it was something later moved into examples if its not valid anymore?
snaked has joined #openscad
snaked has quit [Remote host closed the connection]
snaked has joined #openscad
snaked has quit [Remote host closed the connection]
<kintel>
peeps: OK, will dig more, was just hoping for a clear memory :)
<peeps[zen]>
kintel: sorry. i am looking at the repo from that commit state, and trying to understand, but i guess the whole naming of tests and generated images has always been fuzzy/confusing to me
<kintel>
the filenames are correct, it's just the test name which doesn't match
<kintel>
Might just have been an oversight, due to this:
<kintel>
I'll take a look together with the test issues I'm working on
<peeps[zen]>
i have a (bad?) habit of running "All" tests somewhat often, (usually as a sort of lazy performance benchmark) where strange things sometimes come up ;)
LordOfBikes has joined #openscad
snaked has joined #openscad
<kintel>
peeps It's a good habit, we should do it more often :)
<teepee>
I denounce a place which cannot decide to at least use 2 ee :)
ccox_ has joined #openscad
ccox has quit [Ping timeout: 255 seconds]
J23k5 has quit [Quit: Client closed]
J23k5 has joined #openscad
snaked has quit [Quit: Leaving]
mmu_man has joined #openscad
J23k5 has quit [Quit: Client closed]
J23k5 has joined #openscad
J23k5 has quit [Quit: Client closed]
J23k5 has joined #openscad
arebil has quit [Ping timeout: 264 seconds]
qeed has joined #openscad
qeed_ has quit [Ping timeout: 272 seconds]
qeed has quit [Quit: qeed]
J23k5 has quit [Quit: Client closed]
J23k5 has joined #openscad
qeed has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
misterfish has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
kintel has joined #openscad
<teepee>
just FYI, I've mailed matrix if they have some sort of room policy, lets see if we can get openscad:matrix.org, I guess otherwise we can always try running an official one ourselves :)
<teepee>
I might have a look at a self-hosted instance anyway at some point
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
J23k5 has quit [Quit: Client closed]
J23k5 has joined #openscad
qeed_ has joined #openscad
qeed has quit [Ping timeout: 248 seconds]
L29Ah has joined #openscad
kintel has joined #openscad
<JordanBrown>
You can just create one, though I think the obvious name is already taken as a gateway to freenode.
<JordanBrown>
That is, you can just create a room.
ferdna_ has joined #openscad
<JordanBrown>
Perversely, there's no straightforward way to *delete* a room once you've created it.
<teepee>
that's the same with IRC but there's a clear rule on assignments at libera.chat
<teepee>
otherwise that "freenode" one seems to be owned by some actual person not some bridge account, so trying to contact this person could be next step
Guest77 has joined #openscad
Guest77 has quit [Client Quit]
<JordanBrown>
The question would be whether Matrix would agree that we are the "real" OpenSCAD.
<teepee>
yes, and what their policy regarding rooms is, I did not find anything on the website yet. libera.chat has some clear rules on rooms and accounts
<teepee>
even if they agree we are "real" they still could have a policy of who comes first gets the room
qeed_ has quit [Ping timeout: 255 seconds]
qeed has joined #openscad
<JordanBrown>
right
Guest53 has joined #openscad
<Guest53>
Any ideas on a good way to model a warped torus? (ie. A donut which sweeps from high, to low, to high when looking at it across the X axis). My current strategy is using a python script to generate a list of points and saving into a .scad file with a list. And then in SCAD loop over the list with a hull of spheres between each point. That is kind
<Guest53>
of an annoying workflow, and is quite slow to render.
<JordanBrown>
Do you mean that your torus is skewed?
<Guest53>
That might be a better way to describe it
<JordanBrown>
are the "sweeps" straight lines?
<Guest53>
Yeah, although I'm just approximating a continuous curve
<JordanBrown>
Skewing introduces a linear relationship between the axes - for instance, for each step in +X there would be a step in +Z.
<JordanBrown>
But if you want a curve that's something else entirely.
<Guest53>
Yes skewing would be the right word then. Generally something along the lines of Z = X^2
<Guest53>
Although not linear
<JordanBrown>
skew, at least as I'm using it, is always linear.
<Guest53>
Not a skew then
<JordanBrown>
I don't think any standard transform can do anything non-linear.
<JordanBrown>
If you want to play with skew to see what it does, https://bpa.st/NENQ is an example.
<JordanBrown>
But I suspect that it will not do what you want.
<Guest53>
I don't think I would be able to model it with a linear transformation
<JordanBrown>
Bending something, a change that would turn a straight line into a not-straight-line, is different and harder.
<Guest53>
Right. Is just generating a list of points and doing a hull between each point the way to go?
<JordanBrown>
Much much faster, but maybe harder, is to generate a polyhedron.
<JordanBrown>
Depending on how you're generating your list of points it might or might not be hard to generate a polyhedron.
<JordanBrown>
(At least I think it's much faster than the hull scheme. I can't say that I use hulls very much.)
<JordanBrown>
The polyhedron scheme is very fast, but you have to figure out how to represent your shape as a whole bunch of faces.
<Guest53>
Might be a solved problem to come up with a triangulation for a torus already. That'
<Guest53>
That's a good idea, ill look into that.
<Guest53>
Thanks, appreciate the help.
<JordanBrown>
If you're not too picky about your triangulation, I expect that it's not very hard. But simple answers probably produce larger triangles on the outside than on the inside.
<JordanBrown>
Not very hard if you don't mind doing the trig, that is.
<Guest53>
Probably not that big of a deal, anyways, I'm sure i can push more vertices by generating a polyhedra in the same time it takes to build the shape out of hulls
<JordanBrown>
And of course while you can continue to generate your points from Python, it's tidier if you generate them from OpenSCAD and probably not very different.
mmu_man has quit [Ping timeout: 245 seconds]
cart_ has quit [Ping timeout: 240 seconds]
qeed has quit [Ping timeout: 255 seconds]
tritcher has joined #openscad
mmu_man has joined #openscad
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]