califax has quit [Remote host closed the connection]
<peepsalot>
retrosenator: then i'd guess you are using some operation which forces render of a subtree, like render() or minkowski() (there's a handle of other operations that also force mesh calculation) because 2min preview is pretty rare
<peepsalot>
and it should cache results between previews, unless the one thing you are changing happens to also be the thing which takes all the time... or your cache size settings are too small for the geometry
<peepsalot>
retrosenator: if you pastebin a sample which takes long for you, we might be able to suggest some optimizations
ur5us_ has quit [Ping timeout: 248 seconds]
<peepsalot>
there isn't any multithread capability in the main codebase yet. but if you run a dev snapshot, there are some optimizations like fast-csg and vbo which should improve performance substantially
<peepsalot>
they have to be manually enabled under Preferences -> Features
AndrewStone[m] has quit [Ping timeout: 240 seconds]
one-star-chef has quit [Ping timeout: 252 seconds]
mikolajw has quit [Ping timeout: 240 seconds]
dcg[m] has quit [Ping timeout: 256 seconds]
Cadair has quit [Ping timeout: 248 seconds]
gru3hunt3r[m] has quit [Ping timeout: 240 seconds]
M6piz7wk[m] has quit [Ping timeout: 252 seconds]
Notkea has quit [Ping timeout: 248 seconds]
aiyion has quit [Ping timeout: 240 seconds]
aiyion has joined #openscad
joseph_ has joined #openscad
<retrosenator>
i have minkowski yes
J225 has joined #openscad
J22 has quit [Ping timeout: 252 seconds]
ur5us_ has joined #openscad
one-star-chef has joined #openscad
dcg[m] has joined #openscad
KimK has joined #openscad
Cadair has joined #openscad
qeed has quit [Quit: qeed]
qeed has joined #openscad
mikolajw has joined #openscad
fling_ has joined #openscad
fling has quit [Remote host closed the connection]
dcg[m] has quit [Ping timeout: 240 seconds]
one-star-chef has quit [Ping timeout: 250 seconds]
fling_ is now known as fling
LordOfBikes has quit [Ping timeout: 252 seconds]
gru3hunt3r[m] has joined #openscad
M6piz7wk[m] has joined #openscad
AndrewStone[m] has joined #openscad
Notkea has joined #openscad
LordOfBikes has joined #openscad
one-star-chef has joined #openscad
dcg[m] has joined #openscad
qeed has quit [Read error: Connection reset by peer]
qeed has joined #openscad
snaked has joined #openscad
retrosenator has quit [Read error: Connection reset by peer]
neur0 has quit [Remote host closed the connection]
<retrosenator>
any languages that compile to openscad?
<retrosenator>
i'm wondering about something with different syntax
<teepee>
pretty much any
<retrosenator>
I mean an object description language that a program converts into scad format
<retrosenator>
so that I can use + and - for union and difference etc
<retrosenator>
and many many more syntactic differences to allow a more condensed representation
<teepee>
pick your language and check if it exists, everything else seems a bit waste of time
<teepee>
I know at least python, ruby, go, javascript, elixir and closure
<retrosenator>
do you use any of those for 3d modeling?
<linext>
i got a response from the printables contact form
<linext>
"I'm still postponing the customizer in the development plan because I think it will be appreciated by fewer users than the features being developed."
<dalias>
:/
<dalias>
like isn't this the main missing thingiverse feature on printables?
<dalias>
otherwise they're on feature par -- even to the point of the "feature" of needing to make an account with a particular printer vendor to use the site :-P
<linext>
i suggested they outsource the customizer development and use a project manager from their organization
<teepee>
yeah, that's pretty much what I said on twitter too
<teepee>
they basically just would need to discuss interfaces for file storage and integration
<dalias>
(like, the rebranding was great, but i wish they'd go another step and make it so you can have accounts detached from prusa accounts. the /image/ it'd convey would be so much better)
<teepee>
that said, openscad.cloud just posted addition of a customizer, so they may get some extra pressure, well, or they are right and it's just 100 people and those could use openscad.cloud
<linext>
i tried to get a count of how many customizers and customized objects there are on thingiverse
<teepee>
retrosenator: nope, my point of view is that the language is not the interesting part, but the geometry engine, that's where we need to get things moving
<linext>
20 items per page, 500+ pages
<linext>
so at least 10,000
<teepee>
that might count the auto generated stuff?
<teepee>
well, the customizer saved I mean
<retrosenator>
well i am trying to keep all of that in the box
<feep>
J225: fwiw your design (plus tweaking) is working very well :)
retrosenator has quit [Read error: Connection reset by peer]
<teepee>
wait, what?
<teepee>
Ultimaker merges with Makerbot?
<J225>
so do we now get shitty printer with shitty slicer and shitty platform for higher price?
<mlaga97>
Ah great, so Stratasys will capture even more talent and lock it behind patents for 20 years
pah has joined #openscad
pa has quit [Ping timeout: 246 seconds]
<teepee>
yep, that's probably the idea
<dalias>
threat to Cura is the only thing relevant there imo
<dalias>
ultimaker and makerbot have both been rather irrelevant for a long time
<dalias>
at least in consumer space and state of the art space
<dalias>
they're probably relevant in some industrial spaces the same way oracle is relevant in software domain :-p
<dalias>
(legacy enterprise dinosaurs)
<mlaga97>
Stratasys does make a lot of "big iron" printers for commercial/industrial use.
<mlaga97>
MakerBot is used a lot in education and commercial.
<mlaga97>
Ultimaker has a similar market to MakerBot, so it makes sense they would combine those brands.
<dalias>
yeah
<dalias>
i just hope cura development doesn't get more closed/inaccessible
<dalias>
they already have a separate hidden ultimaker bugtracker where the real bug triage and fixing work gets done
<dalias>
and commits refer to bug numbers on it without much detail on what they're actually doing/fixing
<mlaga97>
Oh wow I had no idea about that
<mlaga97>
I'm not too much of a fan of Cura as it is
<mlaga97>
That's probably unfounded at this point though
<mlaga97>
Last time I tried it was many years ago and it just kinda spit out garbage and didn't have any of the knobs I wanted
<mlaga97>
So I just kinda old man yells at sky'd away from it and went back to Slic3r and more recently PrusaSlicer.
<dalias>
Cura has by far the best support generation algorithms and now with Arachne engine (Cura 5) the best wall generation too (including support for features narrower than nominal/base line width, and not leaving nasty gaps that need zigzag fill)
<mlaga97>
Just design your parts better so they don't need supports in the first place, duh /s
<mlaga97>
But yeah, I know it does supports and a few other things better
<dalias>
that's good for mechanical parts, usually, but not always possible. and it's not possible for many things where appearance is the whole purpose (or a dominant purpose) of the print
<mlaga97>
Yep, I use my printers pretty much exclusively for mechanical parts that I design myself.
<dalias>
anyway speaking of "design your parts so that...", that's one of the huge advances of Arachne
<mlaga97>
Pretty much all of the open-source slicers cross-polinate functionality in the long term anyways.
<dalias>
that you don't have to design around "this needs to be an exact multiple of 0.4 or it's going to print slow or have structural problems"
<dalias>
you just design things the thickness they should actually be, and get non-idiotic toolhead paths out of the slicer
<mlaga97>
I am unfamiliar with Arachne outside of arachne-pnr the FPGA toolchain component
<dalias>
completely different thing :)
<mlaga97>
Googling it just brings up...spider girls...
<mlaga97>
Oh, I just saw it in the previous message, the new slicing engine in Cura
<dalias>
there's not much material out there on it yet except the "cura 5 beta" youtube videos where some folks have shown their results
<mlaga97>
Got it
<dalias>
but most of them don't understand all the ways it's awesome yet
<dalias>
i need to check... i think the cura team has a set of test models that really showcase it
<mlaga97>
I expect that PrusaSlicer will get similar functionality at some point so that I don't have to think about it lol
<dalias>
:)
<dalias>
hopefully
<mlaga97>
But yeah, it seems like one or the other will have a major innovation and then the other will get that functionality a few months later.
<mlaga97>
It's a good model, because different slicers have different potential for experimenting with the process
<dalias>
this has been in the works for over a year, but few ppl outside cura devs and ppl hacking on cura from source know about it
<mlaga97>
I'm sure it has, it seems like a fairly major set of changes.
pie_ has quit []
pie_ has joined #openscad
<dalias>
anyway it's really cool to have difference() { cylinder(d=40,h=20); cylinder(d=38,h=50,center=true); } slice properly as circular paths rather than a bunch of zigzag nonsense between single outer walls
<teepee>
I hope they fix the AppImage of the new Cura, the one I tried maybe 2 weeks ago just crashes for me :(
<teepee>
there's always a fallback though
<teepee>
we all learn Haskell and help juri_ :D
<teepee>
now with Nvidia doing open source drivers, literally anything can happen ;-)
<ecraven>
well, they still contain blobs, afaiu
<teepee>
I know, and no power management. At this point it's not a solution but a signal
<teepee>
also no user space libs as open source, so there's no way to mainline that anytime soon, which would be the more interesting achievement
<ecraven>
yea, it is definitely better than before!
<teepee>
in my totally irrelevant opinion, the best approach would be to keep open sourcing the stuff mainly for information and reference, but work together improving the nouveau driver which already works and has user space mesa support
<peepsalot>
J225: non-uniform scaling can cause curved surfaces as well: linear_extrude(height=10, scale=[0.1,2]) rotate(45) square(10, center=true);
<peepsalot>
J225: a global $fs setting of 0.2 is kind of excessive. setting it higher and scaling your shape in the other direction would also help reduce complexity
<J225>
peepsalot oh that looks nice - didn't thought about the different scaling factors
<J225>
yeah i think he done that for the offset to be roundish
<peepsalot>
yeah, i guess the curved surface is not a thing when the edges are axis aligned though (the longest edges of the half spike)
<peepsalot>
but the rounded edges should still encounter that effect, though to a much less noticeable degree. so I don't know if there's any better way of checking/handling non-uniform scale but that's what we've got
<J225>
the render will reduce the many faces .. probably difficult to sort planar faces before
<J225>
oh funny render (with show edges) show edges on a 45 rotated version .. not on a 44° rotated
<J225>
(added a cube for cgal)
<peepsalot>
hmm?
<J225>
v2022.05.04
<peepsalot>
45 rotated version of what? what is funny?
<J225>
your example with the 45 rotated square
<J225>
when adding a cube and show edges will show blue lines in render .. but not when rotated anything other than 45°
<peepsalot>
dunno, can't reproduce
<peepsalot>
maybe specific to the features you have enabled?
<J225>
ah was cache related i switched fast CSG off and on and so the blue lines were only when the cache had that geometry without fastCSG
<J225>
fastCSG can't generate a wireframe
J225 has quit [Quit: Client closed]
J225 has joined #openscad
<mlaga97>
fastCSG? 2022.05.04? Those certainly piqued my interest. I poked around on the GitHub but couldn't find: Is there an in-progress changelog for OpenSCAD to see what major features are in development or have been added since the last full release?