califax has quit [Remote host closed the connection]
califax has joined #openscad
misterfish has quit [Ping timeout: 252 seconds]
mmu_man has quit [Ping timeout: 260 seconds]
mmu_man has joined #openscad
peepsalot has joined #openscad
epony has joined #openscad
ferdna has joined #openscad
SharpHammer has joined #openscad
<SharpHammer>
Hmm, why are the forums dead?
<teepee>
what is dead?
<teepee>
the old nabble forum provider closed down
<InPhase>
teepee: A correction to my previous note, because accuracy matters. :) Turns out Tuxedo did not ghost me on support. They just have support responses which get flagged as spam. They in fact responded in a timely manner, but I never saw the response until I complained, then the response to my complaint came straight through. (And I was able to dig out the original from amidst the deep pile of junk.)
<InPhase>
teepee: So I guess they have improved their support responsivity to about 24-48 hours, which is acceptable.
<teepee>
oh, interesting, yep, that's usually fine
<teepee>
I still think they focus on too many stuff we are not target group,
<teepee>
like "we have our own linux distro"
<teepee>
well, no, thanks, honestly wrong choice
<InPhase>
Yeah, I'm not interested in niche distros either. :) Likewise for System76.
<teepee>
it just about works fine for raspberry pi, looks like they can actually afford that now
<InPhase>
I was looking hard at the Framework stuff over the past week while fuming, incorrectly, about Tuxedo not responding. And the Framework stuff IS pretty interesting, but I think the coming 16 would probably be the one for me if I went with one of those.
<teepee>
I saw they have some qmk (or so?) stuff for keyboard which sounds like they hook into an existing keyboard design which is nice
<InPhase>
The only thing I really don't like about the Frameworks is the gaudy ethernet jack module. They really need to come up with one that's flush with the unit. But maybe they're having a sizing issue.
epony has quit [Remote host closed the connection]
<InPhase>
Having ethernet embedded is pretty important to me, but having a laptop with a constant thing jutting out the side seems wrong.
<teepee>
yes, same problem with all thin notebooks now
To_Aru_Shiroi_Ne has quit [Ping timeout: 256 seconds]
<InPhase>
I can keep dongles at places I sit, but I want to always be ready to plug into ethernet.
<teepee>
I think one of the thinner tuxedos has like a hinged door so it just about fits with the connector half open or something like that
epony has joined #openscad
To_Aru_Shiroi_Ne has joined #openscad
<teepee>
SharpHammer: I see last message from December 22, did you send a more recent message that did not show up?
<SharpHammer>
teepee nope just didn't see your reply till you mentioned me just now
<SharpHammer>
also dec 29th? not 22nd?
<SharpHammer>
https://nabble.com/ teepee looks like it sill exists? Maybe it re-branded?
<teepee>
redirect is possible in theory but who would spend a week or two collecting that data?
<teepee>
right now the name just points the mailing list archive which is hosted, we don't own or control that server
* teepee
crosses fingers
<teepee>
hoping at least that company stays with us, the founder died a while ago but it seems they at least did not go topsyturvy directly
<teepee>
well, that's what I heard, so that's rated rumors ;-)
LordOfBikes has quit [Ping timeout: 255 seconds]
LordOfBikes has joined #openscad
J24k14 has joined #openscad
<InPhase>
teepee: Yeah, my Tuxedo has precisely one of those hinged ethernet ports. I was pretty concerned at purchase time about that thing snapping off, but, well, it hasn't yet.
<InPhase>
teepee: Although most of the time I've had that laptop I've used a dongle or hub for ethernet, as that's my solution at my home desk and at work. But I am right now climbed into bed and using the side port with the hinge.
J24k71 has quit [Ping timeout: 250 seconds]
<teepee>
I don't remember when I last used ethernet on my notebook, maybe 4 years ago when the docking station was still working :)
<teepee>
but I do have the ethernet card for the framework sitting in a box
<teepee>
need to design an overflow case for the cards not in use
<InPhase>
Did they not come with one? I saw one for 6 of them in a video, and thought it must have come with it or something, since it was so perfectly sized for it.
<InPhase>
Althought that one was cardboard with inserts, and they laid flat in it. A better printed one would have them insert such that the ports are visible, with a customizable length, and it would maybe have a slide-in-place cover to keep the dust out.
<InPhase>
That way you could keep 1-4 of them in a pack and they stay in good shape.
<InPhase>
I'd trust a slide in place cover more than a snap on lid in a backpack.
<InPhase>
Like, with a groove.
califax has quit [Ping timeout: 240 seconds]
califax_ has joined #openscad
califax_ is now known as califax
<teepee>
on the plus side, they have a repo with scad designs for the card cases :)
<InPhase>
Hmm, nice.
guerdy has quit [Ping timeout: 256 seconds]
SharpHammer has quit [Quit: Client closed]
Timelord83 has joined #openscad
<Timelord83>
hello all i ham getting a render error while trying to import and STL and difference some rectangles out and output a new stl
<kintel>
We should probably set up some qt6 builds soon..
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mmu_man has quit [Ping timeout: 256 seconds]
mmu_man has joined #openscad
<teepee>
kintel: good idea, probably easiest via Ubuntu 22.04
ferdna has quit [Quit: Leaving]
LordOfBikes has quit [Ping timeout: 256 seconds]
LordOfBikes has joined #openscad
Guest97 has joined #openscad
<guso78k>
https://www.youtube.com/watch?v=XjoPUjRG9nk OpenSCAD is becoming more interactively slowly but steadily. Create handles, find them and use them again to combine objects.
<othx>
guso78k linked to YouTube video "OpenSCAD is becoming more interactively" => 1 IRC mentions
<guso78k>
resuing QT Tooltip an annotate the names
<teepee>
I think those should be vectors not points
<guso78k>
actually these handles are eigen matrices where only [0][3] . [1][3] and [2][3] are used as coordinates
<guso78k>
the rest is the orientation of the coordinate system, which could also be visulized
<teepee>
right, probably even better in the sense that it has a clear direction for attaching things
<teepee>
so it's not a point but something that does not leave many degrees of freedom
<guso78k>
you can also store scale, mirroring and even skew there . literally everything ...
<guso78k>
everything which multmatrix can do.
<teepee>
hmm, not sure for what that would be needed, but having the option for later might be useful
<guso78k>
next task is just to do the inverse matrix, so I am not fixed anymore to attach the 2nd object at its origin :')
<teepee>
I guess the interesting part is how to do this in a way that will allow it for scad too
<guso78k>
haha, i think scad is missing some language constructs yet to support these features
<teepee>
not really
<guso78k>
i believe that users which come from click&play world which dont have so much idea about coordinates and rotations will have it easier to combine objects
<guso78k>
i think the annotation thing can generally be used to display the coordinate of a selected vertex.
<teepee>
I'm pretty sure it's possible to have this with scad too, it probably would only limit to the built-in objects at first
<teepee>
yes, maybe even having a debug mode for polyhedrons
<guso78k>
I am happy to read, that "those" handles are possible with scad too, apparently i know to less about it
<teepee>
well, we need an implementation that works for both anyway
<guso78k>
i am just wondering how mapping additional information with a solid works out for SCAD
<teepee>
right now, internal only, it needs to attach to the Geometry object
<teepee>
so in first step, you would not be able to define the places via script, but the c++ code could just define the "obvious" ones
<guso78k>
yes, i know about relativity.scad, but apart from that i was missing such functionality for a long time in openscad. if its internal only, we should probably establish an API ?
<teepee>
so what I'm saying is, that should go into the general core code
<teepee>
yes, 2nd step it would make sense to allow via scripting like you show for python
<teepee>
which likely would be coming with the Object PR stuff
<teepee>
but just like we currently can already create objects internally (e.g. the JSON import) but not via script, same we could start out with the attachment/handle stuff
<guso78k>
i;d like that handles could also have parameters finally (like an angle e.g.)
<guso78k>
with that we could easily attach e.g. gears and their correct abutting in an animation.
<guso78k>
openscad would do the physics correctly just out of the box.
<guso78k>
looking forward to make use of the object PR soon (and explore new abilities then)
<teepee>
the attachment stuff might be possible to get in even much sooner I believe
<teepee>
if we can get the core behavior into the openscad core code that could be merged
<guso78k>
I believe so, too. I think more people will get aware about the idea and there is also a template implementation already. However, I fear i am not very good about parsing SCAD language.
<teepee>
why would you need that?
mmu_man has quit [Ping timeout: 246 seconds]
<teepee>
I don't think there's any need for that if we postpone the creation of user-level handles from scad script
<teepee>
that would be only possible from python for now
<guso78k>
well. suppose we will need a new "attach" function. In case I am wrong, i probably did not catch your idea ;)
<teepee>
so the minimal scad side change would adding a fixed set of handles in the cube builtin
<teepee>
yes, that is probably needed, but a new module does not need anything from the parser either
<guso78k>
for a cube, i could think of at least 30 different handles ,which could be useful ... will be hard to give names to them such users can immeditely translate the intent to handle name ...
<teepee>
I suppose we could start with what BOSL2 provides
<guso78k>
admit, that i never ever deeply looked into BOSL2 '=(
<guso78k>
maybe now could be a good time ...
<teepee>
hmm, I have not looked closely at the attachment stuff either, it has different styles it seems, like edges and points
<teepee>
yeah, it's something that has proven some real-world use, so it probably makes sense to see what happens there :)
<teepee>
same with the relativity and maybe that construct library too
<InPhase>
kintel: Qt6 is a minefield for exception handling, but in an attempt to grep wildly for cases where we are not handling exceptions adequately for Qt6, it appears we've made the necessary edits. And it looks like the OpenSCADApp::notify handler hasn't led to a submitted issue since the 2019 release.
<guso78k>
no, it has not. my handle stuff just defines eigen matrices , give them a name and map them to certain points the user can focus./select
<InPhase>
kintel: Obviously it's not possible in a brief audit to spot all possible flaws for this, but I gave it a good hunt effort.
mmu_man has joined #openscad
<InPhase>
kintel: So I think we're in fairly good shape. If there's one or more rare case missing, I guess we'll find it the hard way.
<guso78k>
ahh you refer to relativity.scad
<teepee>
yes
<guso78k>
sorry '=D
<teepee>
well, all of the libs basically
<teepee>
InPhase: yeah, arch and gentoo users :)
misterfish has quit [Ping timeout: 276 seconds]
<guso78k>
relativity.scad does quite a good trick in mapping the bounding box of an object to cube between [-1,-1,-1] to [1,1,1] with the abilities of openSCAD. it can achieve really a lot with it.
<guso78k>
teepee, so every primitive in openscad will get a array mapping a name to a vector3d/eigen matrix ?
<guso78k>
how could attaching two primitives look like in SCAD language ?
<teepee>
guso78k: yep, that sounds like a good start which should work fine with the logic you have too
<teepee>
not sure yet, initially maybe something like attach(["top", "bottom"]) { sphere(); cylinder(); }
<teepee>
we have to see in what oder data is available, it's sometimes easy to get things in the wrong order
<guso78k>
yeahhh! X-P X-P
<teepee>
hmm, actually it may need the target object as first and special one like difference()
<guso78k>
issue now is: the result of the attach must also derive the handles from the sources, else you can attach at most 2 primtives
<teepee>
well, 3 would be attach(["top", "bottom", "left"]) {...}
<teepee>
but how to pass the information on for later is probably also an interesting topic
<guso78k>
yes, if the handle names of 2 cubes are the same, you will get a name collision unless you prefix, but there is no object name ...
<teepee>
for a later "full" version, it might be needed to pass a normal accessor like object.subobject.left or so?
misterfish has joined #openscad
<guso78k>
yes: you refer the object with how you "reach" it.
<guso78k>
in python union, right now i only reuse the handles of the 1st object always ...
<guso78k>
these attach statement always use union to combine the 2 objects, but difference instead has at least equal value to make holes precisely.
kintel has joined #openscad
<kintel>
InPhase A good start after your initial audit would be to build qt6 snapshots alongside qt5 snapshots, and gradually promote qt6 as the default and see where that leads us. Too bad we don't have automated UI tests..
<guso78k>
effect on me would just be to make sure, that i have qt6-devel(or similar installed) ?
little_blossom has quit [Ping timeout: 264 seconds]
<teepee>
depends on how distros select things, if that de-installs qt5-devel that might be enough
<teepee>
for the qt4->qt5 switch, debian for example supported both and offered either a separate "make-default" package and environment variables for selection
<kintel>
Not sure about MXE
<kintel>
homebrew on macOS supports having both
<kintel>
Anyway just asked cmayo if he has any further plans, then we can open a tracking issue to deal with all the details
<teepee>
MXE has qt6
<teepee>
but we probably need a separate docker image, or build support for both and have it select at openscad compile time
<teepee>
separate image is clear I think
noob_001 has joined #openscad
<noob_001>
Hi guys, I have a nooby question regarding OpenSCAD syntax.
<noob_001>
How can I modify a vector element? Like this:
<noob_001>
vector1 = [10,20,30,40];
<noob_001>
vector1[0] = vector1[0] * 2
noob_001 has quit [Client Quit]
<teepee>
you can't in-line, you have to create a 2nd modified vector
<teepee>
uff
noob_001 has joined #openscad
<teepee>
and welcome back :)
<teepee>
you can't in-line, you have to create a 2nd modified vector
<teepee>
general rule is you can only assign variables once
<noob_001>
I see
<noob_001>
Thank you for the quick help, in my simple situation I found a way to create a new vector and copy over the values, but most likely I will return once I will need 'for loops'
<teepee>
easy to do
<teepee>
vector2 = [ for (v = vector1) 2 * v ];
<teepee>
would just multiply any entry by 2
<teepee>
or for the first case: vector3 = [ for (i = [0:len(vector1)-1]) if (i == 1) 2 * vector1[i] else vector1[i] ];
Guest97 has quit [Quit: Client closed]
<noob_001>
Jesus Christ
<noob_001>
Coming from C/C++, I'm having a hard time
<teepee>
yes, forget c++ :)
<noob_001>
Btw is this syntax or system similar to some other, older language?
<teepee>
that said, there's a syntax that *looks* very much like c
<teepee>
it's list comprehension, scripting languages like python have something similar
<noob_001>
JS too?
<noob_001>
Note: I haven't learned JS or Phyton yet, because of these