<peeps[zen]>
are the CI builds doing anything in particular for TBB to build? i am not seeing ENABLE_TBB=OFF or anything like that
LordOfBikes has joined #openscad
JordanBrown has joined #openscad
Gene has joined #openscad
<Gene>
Thanks for the incredibly quick responses
<JordanBrown>
Is this w1ebr from github?
<Gene>
Yes
<Gene>
That's my Amateur radio call sign
<JordanBrown>
OK, so you have 34 STL files. Are you trying to combine them into one file, or what?
<Gene>
each one is a unique, fractal snowflake. They are mirrored in the Z axis and to 3d print them, I need to split them in half.
<Gene>
I do this every year (since 2016) as a Christmas present for my family
<JordanBrown>
OK, so you want to process each of the 34 files and split them in half, producing two STL files each. Right?
<JordanBrown>
Actually, since they're mirrored in Z I guess you only need one file each.
<Gene>
I am going through the process to chose 1 for this year
<Gene>
Correct. Right now, I use difference to split them
<JordanBrown>
What platform? Windows, Linux, MacOS?
<Gene>
flake minus a cube
<Gene>
MacOS
<JordanBrown>
Know how to write simple shell scripts?
<Gene>
Windows is also ok
<Gene>
yes, I can figure it out ( I am new to MacOS )
<JordanBrown>
And are they all the same thickness, or centered on Z=0?
<Gene>
not the same thickness, but all centered on Z=0
<Gene>
and they are manifold
<JordanBrown>
You have already confirmed that you can do the difference and render with F6 and it works? That's always a challenge with imported STLs, because OpenSCAD (or actually the underlying CGAL engine) is very picky about correctness and many STLs are not acceptable.
<Gene>
Yes, it works great !
<JordanBrown>
So all you need is hints on wrapping some automation around it.
<Gene>
right, and it looks like the file input is inside the OpenSCAD program while the output file is via the command line
<Gene>
I guess I could try to create the OpenSCAD program on the fly with the needed import in it, but that seems overly complicated
<JordanBrown>
You can do it with -D.
<JordanBrown>
Hold a moment and I will whip up an example.
<JordanBrown>
Call them split.scad and split.sh respectively.
<JordanBrown>
Oh, sorry, delete the "b/" from the shell script.
<Gene>
OK, thank you so much !
<JordanBrown>
Run the shell script as, e.g., "./split.sh *.stl".
<Gene>
What is your email address? I want to send you a picture of last year's snowflake that OpenSCAD has made possible for me to give
<JordanBrown>
Do make backup copies of your STL files first, in case you or I got something wrong.
<JordanBrown>
openscad@jordan.maileater.net is the one I use for OpenSCAD traffic.
mmu_man has quit [Ping timeout: 252 seconds]
<JordanBrown>
Most of the complexity in the script is pulling apart the file name to make the result file name; it takes whatever.stl and produces whatever.out.stl.
<JordanBrown>
Also note that this is cheap and dirty and (because it uses "basename" to strip off the .stl) won't work on files that aren't in the current directory.
<JordanBrown>
But maybe it will give you ideas on how other kinds of automation can be done.
<JordanBrown>
Picture is beautiful.
<Gene>
Thanks, your software helped make it possible
<JordanBrown>
I am only a minor contributor.
<Gene>
Not for me right now
<JordanBrown>
But the others do hang out here in this IRC chat and may see the comment in the logs.
<Gene>
Thanks, Jordan
<JordanBrown>
I try to help out, but have only done a small amount to improve the program proper.
<Gene>
Is there a "manifold" option ?
<Gene>
I read something about it being faster
<JordanBrown>
Manifold is an alternative geometry engine replacing CGAL. It is indeed much faster.
<JordanBrown>
It is available only in the development snapshot, not in the 2021.01 production release.
<Gene>
I am running the dev snapshot
<JordanBrown>
(And I see suggestions that maybe it's broken at the moment, so you might want to see if you can pick up a development snapshot from a couple of weeks ago.)
<Gene>
oops, ok
<Gene>
I have 2023.11.06
<JordanBrown>
From the application, you can enable it using Edit / Preferences / Features / manifold.
<Gene>
it is only a time saver, so if it doesn't work, that's ok
<Gene>
ok !
<Gene>
I will try it out
<JordanBrown>
For the CLI, you can --enable=manifold .
<Gene>
ok, I will test it out first
<JordanBrown>
By all means try it. But don't be shocked if it crashes.
<JordanBrown>
Oh, wait, you're on MacOS. The problems have been on Windows, so you might be OK.
<JordanBrown>
The guy to thank is Marius Kintel. I'm not clear on whether he's the original author or just a major long-time contributor, but I believe he's the single biggest current contributor. He does hang out here on IRC, but he's in Toronto so maybe it's getting to be his bedtime. The guy who is doing
<JordanBrown>
Torsten Paul, teepee, who also hangs out here - but he's in Germany and so it's definitely past his bedtime.
<JordanBrown>
most of the current release management type stuff is
<Gene>
Wow, that is fast !
<JordanBrown>
I would expect the simple difference to be pretty fast no matter what.
<Gene>
Manifold worked on one example
<Gene>
No, it isnt
<Gene>
A single snowflake has about 175,000 triangles
<JordanBrown>
Huh. OK. Can't say that I've tried to do booleans on a complex polyhedron.
<Gene>
So, on my 2018 i7 MacBook Pro it can take about 15 seconds
<Gene>
vs about 1 second
<JordanBrown>
Oh, and in case you didn't notice: another quick-and-dirty thing about split.scad is that it differences with a fixed-size 200x200x200 cube, so if your object extends > 100 in X or Y, or < -200 in Z, it won't get all of it.
<JordanBrown>
Or < -100 in X or Y.
<Gene>
I do the same thing in my quick program, it works just fine !
<InPhase>
Gene: Any little kids in the family, like age 2-5?
<JordanBrown>
AFK for a few mins
<Gene>
My niece's daughter is 5 and we spend a lot of time with her
<othx>
InPhase linked to "Modular Train Track Set, Standard Compatible, and Customizable by rcolyer" on thingiverse => 6 IRC mentions
<Gene>
So cool, I will check it out. Thanks!
<InPhase>
Gene: I invented a snap-fit mechanism for it that is designed to be more printer friendly, and it worked even better than I expected. It turns out it's easier to use for kids than all the commercial stuff, if you have your printing calibrated right.
<Gene>
That is very cool !
<Gene>
Thanks for sharing it
<InPhase>
Tis the season to start printing such things, but I ran out of little kids in my family for now. :)
<Gene>
I am 67. Wait a while and thew next generation comes along !
<InPhase>
Heh, yeah, we'll see. My kids are 15 and 7, so it'll be a bit yet.
<JordanBrown>
For me, one generation is 19-25 and the next generation is 5mo :-)
<othx>
InPhase linked to "Baby Hair Comb with Personalized Text by rcolyer" on thingiverse => 2 IRC mentions
<Gene>
Hehe. Mine are 37, 35 and 29
<JordanBrown>
InPhase WARNING: maximum_head_curvature_radius was assigned on line 18 but was overwritten in file Baby_HairComb.scad, line 19
<InPhase>
JordanBrown: :) Well, wasn't a warning in 2018. But file that under WONTFIX I guess. I won't maintain thingiverse code unless they fix the customizer and invest in making the site less annoying.
<Gene>
I have been sitting in front of this laptop form most of the day... time to take a break. Thank you, gentlemen, for such a kind welcome !
<InPhase>
Although I see the layout has randomly changed yet again.
<InPhase>
The layout kind of looks more cartoonish now.
<JordanBrown>
Sure thing. As I said, if you have questions then this IRC channel and the mailing list are good places to ask.
<JordanBrown>
IRC may be a bit faster (and gets a few top-notch people who aren't on the mailing list) but can be hit-or-miss; if nobody is around then there's not really any good way to get a reply later.
<JordanBrown>
InPhase two-level Minkowski. No wonder it took 2m40s to preview.
<InPhase>
JordanBrown: Yeah... That's 2016 era code from me.
<InPhase>
JordanBrown: It used to be a 25 minute render for me.
<JordanBrown>
Random WONTFIX suggestion: move the label, the thing most likely to be customized, to the top of the parameters.
<InPhase>
JordanBrown: Might still be 25+ minutes on your 386 or whatever you're running there.
<JordanBrown>
Feh. It's not a 386.
<InPhase>
Although the master branch might do a lot better with it.
<JordanBrown>
Preview was OK; that's what I got with F6.
<JordanBrown>
Switching to CGAL.
<JordanBrown>
The comb proper rendered OK, but the name was missing.
<InPhase>
Hmm, still not looking to fix it, but I'm curious enough to try it.
<JordanBrown>
I may get a new system for Christmas.
<JordanBrown>
Probably not super hot, but a lot hotter than what I have now.
<InPhase>
I love the design of that comb, but I actually kind of hate the code I used. It was the first model I made in 2016 that wasn't a printer part or a library for printer parts, and thus I did a lot of things that I would never think of doing now.
<InPhase>
Rendered fine for me in the master branch actually.
<InPhase>
And surprisingly quickly.
<InPhase>
66 second preview, followed by a 5 second render.
<InPhase>
That's a slight improvement over 25 minutes.
<JordanBrown>
Might depend on the details of the font and the name.
<InPhase>
The only difference in my local copy is apparently I already commented out that redundant define.
<InPhase>
Might.
<JordanBrown>
Mine was a 2m40s preview, changed the name, 53s preview, 3s failed render.
<JordanBrown>
Flushed caches and rendering now with CGAL.
<JordanBrown>
Should probably have tried flushing and Manifold first.
Gene has quit [Ping timeout: 250 seconds]
<JordanBrown>
But I think the CGAL render may take long enough for a nap.
<JordanBrown>
InPhase have you tried printing that comb in TPU? That might be a good child-proof variation.
<InPhase>
I haven't, but I'm pretty sure it would be too floppy at those scales to function properly as a comb.
<InPhase>
You'd need something slightly in the middle.
<JordanBrown>
I don't have much experience with TPU.
<InPhase>
My kids were in fact prone to snapping off teeth from them during a certain age range. Almost certainly from messing with trying to break it. But they're cheap to print more.
<JordanBrown>
But PLA is pretty brittle and I would worry about a little kid snapping off a tooth and eating it.
<InPhase>
PLA is a bad choice because it's too uncomfortable.
<InPhase>
PETG is much more comfortable.
<InPhase>
Slightly more flex than PETG would be okay and maybe even better, but not TPU levels of flex.
<JordanBrown>
I have an idea that I'm sure will sell tens of units: a 3D printer that prints ice.
<JordanBrown>
It would push water through a refrigerated head into a refrigerated build volume. The head would supercool the water but it would stay liquid (or liquid-ish) because it's under pressure, and then would freeze instantly as soon as it cleared the head.
<InPhase>
I think I saw a video of one of those...
<JordanBrown>
Indeed. There's also a more recent project that does micro-scale printing. I'm not sure exactly what scale they are talking about, but it sounds like they are trying to be at a blood-vessel sort of size. They're doing it for lost-ice casting; I was only thinking of ice sculptures.
<JordanBrown>
Well, darn. There goes the very very small fortune I was going to make.
<JordanBrown>
So far two failed comb prints... the teeth keep curling up.
<InPhase>
Glue stick.
<InPhase>
It's a particularly sensitive design for initial adhesion.
<t4nk_fn>
a hair-comb comb?
<InPhase>
Oh, apparently I don't have that suggestion in the baby one. But I wrote it out in the original version under Print Settings: https://www.thingiverse.com/thing:1689741
<othx>
InPhase linked to "Ergonomic Hair Comb With Personalized Text by rcolyer" on thingiverse => 6 IRC mentions
ccox_ has quit [Read error: Connection reset by peer]
ccox has joined #openscad
<t4nk_fn>
maybe you could print the text, and then glue it onto a comb btw?
<InPhase>
JordanBrown: I got distracted and forgot to rerun it after pulling the master branch and building again, but just did it. The baby comb still renders fine for me.
<InPhase>
JordanBrown: I did have problems with the handle disappearing in an old version with manifold, and I think I even put that as an example failing design in an issue once. But I thought that was fixed, and it seems to be fixed here.
stefanct has quit [Server closed connection]
stefanct has joined #openscad
<JordanBrown>
InPhase shrug.
<JordanBrown>
Yes, glue stick. Actually, it might be good if I were to clean off the accumulated glue stick and put down a new clean layer.
<JordanBrown>
Also might be good if I disabled my usual 0.2mm elephant foot compensation, so that the first layer would be just a tad bigger.
<JordanBrown>
Though it seems to be doing OK on initial adhesion; it's after N layers that it starts to curl.
<JordanBrown>
Maybe less cooling, or heating the print volume.
<J23k46>
simple way to make a vase .. crashes openSCAD with manifold (and also clear the clipboard) » union()for(i=[0:10:360])translate([0,0,i/100]) cylinder(0.1,d1=sin(i)+2,d2=sin(i+10)+2,$fn=60);
<teepee>
looks like that tbb update mentioned earlier is the fix
<J23k46>
ah cool - wasn't aware we already had several "windows crash" tickets
guso78k has joined #openscad
bcn has quit [Server closed connection]
bcn has joined #openscad
teepee_ has joined #openscad
guso78k3 has joined #openscad
<J23k46>
As we already safe backup files .. wouldn't it be nice if you get a dialog to open last " backup files" or "remove" - or is there a reason you build up these backup files over time?
teepee has quit [Ping timeout: 264 seconds]
teepee_ is now known as teepee
<J23k46>
or just include the backup in the "last opened"
<J23k46>
if everything went fine it the backup is removed and so the entry from recent files
<J23k46>
And if not you have a nice easy way to access it
<J23k46>
I would assume lots of people are not aware that these files exist and can recover last sessions.
<J23k46>
(should i open an improve issue for that?)
guso78k86 has joined #openscad
guso78k3 has quit [Ping timeout: 250 seconds]
guso78k has quit [Ping timeout: 250 seconds]
<guso78k86>
yes, very good idea. But once you choose to use the backup file, you need to make sure that you dont continue to edit the backup file but move back to the original file name.
<Scopeuk>
I guess the Lazy way is to create a "dirty file" which on open we mark as dirty and when closing in a controlled manner set at clean. at next application load check if that is clean or dirty, if dirty offer/restore backup
guso78k has quit [Client Quit]
rawgreaze has quit [Ping timeout: 260 seconds]
rawgreaze has joined #openscad
arebil has quit [Quit: arebil]
guso78k86 has quit [Quit: Client closed]
rawgreaze has quit [Ping timeout: 246 seconds]
rawgreaze has joined #openscad
bozo16a has joined #openscad
bozo16 has quit [Ping timeout: 256 seconds]
castawayc has quit [Server closed connection]
castawayc has joined #openscad
J23k46 has quit [Quit: Client closed]
J23k has joined #openscad
<J23k>
guso78 I guess you could preventing saving to the backup folder and offer saveAS
<J23k>
Or open files from the backup folder as readonly
califax has quit [Ping timeout: 264 seconds]
califax_ has joined #openscad
califax_ is now known as califax
<InPhase>
JordanBrown: Is that all with an ABS attempt?
<InPhase>
JordanBrown: I hear that filament is more challenging in general for these issues, so a print of this shape might require special compensations to handle it.
guerd87 has quit [Read error: Connection reset by peer]
juri__ has joined #openscad
juri_ has quit [Ping timeout: 258 seconds]
juri__ has quit [Read error: Connection reset by peer]
juri_ has joined #openscad
snaked has quit [Quit: Leaving]
arebil has joined #openscad
mmu_man has joined #openscad
juri_ has quit [Ping timeout: 246 seconds]
juri_ has joined #openscad
bozo16a has quit [Remote host closed the connection]
juri_ has quit [Read error: Connection reset by peer]
bozo16 has joined #openscad
juri_ has joined #openscad
bozo16 has quit [Remote host closed the connection]
<JordanBrown>
But it isn't like there's paid staff monitoring it.
<Guest10>
Of course. LOL.
<JordanBrown>
InPhase no, PLA.
<JordanBrown>
Guest10: The Thingiverse customizer is a Thingiverse feature built atop OpenSCAD. If it's not working for you, you'd have to talk to them.
<JordanBrown>
OpenSCAD has its own Customizer UI inside the application, based on the Thingiverse UI, but it's a completely independent implementation.
<Guest10>
Right, I'm talking about the OpenSCAD customizer
<Guest10>
Ah, I missed that those instructions are for thingverse Customizer scripts. So, I assume this wasn't implemented in OpenSCAD since it already has the View dropdown?
<JordanBrown>
OpenSCAD has a built-in customizer that's based on the Thingiverse UI.
<JordanBrown>
Some designs look best when viewed from a certain angle. For instance, perhaps your script adds text to the bottom of your object and you want people to see the bottom when the page loads. You can control the default view by adding a comment formatted like this:
<JordanBrown>
// preview[view:south, tilt:top]
<JordanBrown>
(IRC didn't like the line starting with '/')
mmu_man has quit [Ping timeout: 268 seconds]
<InPhase>
And yeah, you can set the $vpt and $vpr variables like that, but it requires manually processing the designated labels inside the script.
<InPhase>
Guest10: Are you an educator or student?
<Guest10>
Hobbyist
<InPhase>
Ok.
<InPhase>
Just wanted to appropriately structure the guidance. :)
<Guest10>
"but it requires manually processing the designated labels inside the script." So, there's no way to "bind" a line using a viewpoint variable to a button or similar in the OpenSCAD Customizer?
<JordanBrown>
I don't understand the question. But yes, you'd have to implement it in your script.
<Guest10>
I want to bring custom views forward to the UI (Customizer).
<Guest10>
I'd like to establish a view (using these viewport variables or whatever), then somehow make them selectable in the Customizer
<Guest10>
*it
<JordanBrown>
Check out the paste above.
<JordanBrown>
That's what it does.
<JordanBrown>
It puts a top/bottom/left/right/front/back pulldown in the Customizer; selecting one selects that view.
<Guest10>
Nevermind. I was putting it in the wrong section of my existing script. That'll work. Thanks!
<JordanBrown>
Happy to be of assistance.
<Guest10>
So, who normally hangs out in here? What do both of you do for a living?
Guest10 is now known as GizmoGoody
<JordanBrown>
I'm a software engineer for Oracle living in Los Angeles.
<guso78k>
FROM LosAngles SELECT * ;
<JordanBrown>
At a guess there's maybe ten "regulars" who show up for some part of the day. I believe we're scattered across about ten time zones, so times vary a lot.
<GizmoGoody>
Gotcha. I used to be stationed at Edwards Air Force Base in the Mojave Desert
<JordanBrown>
Sure, I know Edwards.
<JordanBrown>
guso78k no, I'm in the organization that used to be Sun Microsystems.
<JordanBrown>
Want to buy a really big file server?
<JordanBrown>
if you've got a couple of million dollars, we'll bring one over on a fork lift right away.
<JordanBrown>
There are smaller configurations available, but if you really want to go for it...
<JordanBrown>
oh, sorry, misread the spec sheet a little. 64 processor cores per system, usually shipped in a two-system cluster pair, each with 2TB RAM.
<InPhase>
GizmoGoody: I tried to do some other solutions, but this was the cleanest I came up with for cardinal directions. https://bpa.st/2FYA
<InPhase>
Everything else I came up with either fails or is "too clever" for the problem scope.
<JordanBrown>
You can do it cooler if you have my experimental build.
<GizmoGoody>
InPhase Thanks! It's good to have options. I'll play around with it a bit. Thanks to both of you. Heading to lunch.
<JordanBrown>
Yeah, I need to pretend to do real work.
<GizmoGoody>
LOL
mmu_man has joined #openscad
<InPhase>
JordanBrown: Yes, that's nicer. :)
<InPhase>
JordanBrown: Although I still need to put some attention on that excellent work you did and resurface those syntax thoughts and discussions we were having.
GizmoGoody has quit [Quit: Client closed]
<JordanBrown>
it's a little tidier if you spell the 45° directions as one word (northeast) so that they don't need the quotes.
J23k has quit [Quit: Client closed]
J23k has joined #openscad
<JordanBrown>
I need to update the OEP so that it describes what I've currently implemented.
<JordanBrown>
(My work meeting starts in 5m, but I can hang out here for another couple.)
Guest79 has joined #openscad
Guest79 has quit [Client Quit]
Lagopus has joined #openscad
cart_ has joined #openscad
rawgreaze has quit [Ping timeout: 268 seconds]
rawgreaze has joined #openscad
mmu_man has quit [Ping timeout: 260 seconds]
J23k has quit [Quit: Client closed]
J23k has joined #openscad
cart_ has quit [Ping timeout: 268 seconds]
J23k has quit [Quit: Client closed]
J23k has joined #openscad
Guest52 has joined #openscad
<Guest52>
Hello. Anyone online to answer (hopefully) an easy question?
<peepsalot>
don't ask to ask
lastrodamo has quit [Quit: Leaving]
<Guest52>
Is there a function to check if an integer is odd? Also, is there one that will take an even integer and round it down to the next closest odd integer?
<InPhase>
Guest52: function is_odd(x) = x%2==1;
<InPhase>
Down toward negative infinity, or down toward 0?
<Guest52>
toward 0. can assume integer is positive.
<Guest52>
function is_odd(x)=x%2==1;
<Guest52>
echo(function is_odd(5));
<Guest52>
Should that return true?
<Guest52>
nm...didn't need the function is the second line
<InPhase>
function oddfloor(x) = 2*floor((x+1)/2)-1;
<InPhase>
That does toward negative infinity, which is equivalent for positive values.
<InPhase>
0 is otherwise hard to define for round toward the nearest odd in the direction of 0. :)