<kintel>
Something is seriously wrong with the Windows CI
<kintel>
It went from mostly working to always failing around 4 days ago
<kintel>
Not even -j1 makes it go green again. Are we really using >7GB memory compiling a single file?
<InPhase>
Is that the thing running on Linux but cross-compiling?
<kintel>
No, msys2 on Windows Server 2022
<kintel>
Always the same file: manifold-applyops-minkowski.cc
KimK has quit [Ping timeout: 246 seconds]
KimK has joined #openscad
<InPhase>
kintel: Try something for me... In that msys2 environment, run this and tell me if it works or errors: /usr/bin/time -v /usr/bin/echo 2>&1 | grep "Maximum resident set size"
<InPhase>
kintel: I expect an output like, "Maximum resident set size (kbytes): 1924"
<kintel>
I tried running time locally yesterday, but it didn't give reasonable output
<kintel>
It gave resident size, but it didn't change if I compiler something bigger
<InPhase>
Excellent.
<InPhase>
That was just the preliminary test. Now, this script I just whipped up...
<InPhase>
Let me assembly that for you.
<kintel>
echo gives: Maximum resident set size (kbytes): 8784
<InPhase>
kintel: Okay, so we have this script, that can go into ./scripts/g++_mem_monitor.sh https://bpa.st/BLVPQ
<InPhase>
kintel: Then, the thing that runs cmake, needs to instead do something like: CXX="$(pwd)"/../scripts/g++_mem_monitor.sh cmake -DEXPERIMENTAL=ON ..
<InPhase>
So either hack that into the script launching cmake, or temporarily alter the CMakeLists.txt itself to call that script as the "compiler".
<InPhase>
This script is assuming g++ in the path is the right compiler. If that's false... I guess hack it worse and edit that script. But probably that's right.
<InPhase>
But yeah, don't forget the +x when you do. I should also sleep for now.
<InPhase>
Crap.
<InPhase>
I bet it's expecting g++.exe ?
<InPhase>
Maybe?
<kintel>
ninja probably doesn't want to run shell commands
<InPhase>
From a prior build: D:\a\_temp\msys64\ucrt64\bin\c++.exe
<InPhase>
An attempt to run that needs to go into where the g++ is, probably with the slashes converted.
<kintel>
I could try using makefiles instead, but not tonight, as that's a whole other can of worms. I assume ninja is default on Windows for a reason...
<InPhase>
Yep. Goodnight. Let me know if it actually works later. I'll be curious to see.
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
guso78 has joined #openscad
guso78 has quit [Write error: Broken pipe]
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
guso78 has joined #openscad
qeed__ has joined #openscad
qeed_ has quit [Ping timeout: 246 seconds]
xloem[m] has quit [Quit: Bridge terminating on SIGTERM]
usop[m] has quit [Quit: Bridge terminating on SIGTERM]
LinuxHackerman has quit [Quit: Bridge terminating on SIGTERM]
mikolajw has quit [Quit: Bridge terminating on SIGTERM]
Cadair has quit [Quit: Bridge terminating on SIGTERM]
mfg[m] has quit [Quit: Bridge terminating on SIGTERM]
evils[m] has quit [Quit: Bridge terminating on SIGTERM]
guso78[m] has quit [Quit: Bridge terminating on SIGTERM]
Notkea has quit [Quit: Bridge terminating on SIGTERM]
JordanBrown[m] has quit [Quit: Bridge terminating on SIGTERM]
isitaboat[m] has quit [Quit: Bridge terminating on SIGTERM]
anotherlucas[m] has quit [Quit: Bridge terminating on SIGTERM]
BZK[m] has quit [Quit: Bridge terminating on SIGTERM]
agnes_de_lion[m] has quit [Quit: Bridge terminating on SIGTERM]
<mmu_man>
oh, I found the later like yesterday, didn't know about the others
<InPhase>
mmu_man: Last I noticed it was stuck on the 2015 release of OpenSCAD, which is missing major shifts in the language. The chance of it working on current code is low.
<mmu_man>
eh
<mmu_man>
oh wow, does the 2nd one render on a server or with js?
<mmu_man>
not sure it'll like my complex parts though
<J2345>
using WASM so it renders in your browser not using server CPU
<mmu_man>
my old thinkpad probably won't survive that with minkowski & sphere differences
<J2345>
and as https://ochafik.com/openscad2/ also uses manifold it is quite fast as the WASM renders always as it is a stl display not a preview
<J2345>
makewithtech renders on the server
<mmu_man>
talking about manifold, I get a warning on this one, but couldn't figure out why:
<teepee>
good idea poking the newer sites, I never used thangs3d so far
<mmu_man>
thx :)
<guso78>
my personal successor to thinigverse is cults3d. me and thingverse relate like teeepee and birdie ...
<teepee>
yeah, thingiverse still has amazing stuff, but the website itself is just horrible
<teepee>
especially considering it's now owed by big pockets corp making that even more sad
<teepee>
I still don't get what they payed $300mil for
<teepee>
$403 even
<mmu_man>
guso78: yeah but they don't have the customizer, do they?
<guso78>
right, i am not aware of a customer in cults3d, but that is not a requirement for me
<guso78>
But i tried the thingiverse customer and i did not manage to customize a model without publishing the customized model at the same time(but maybe i just cannot operate it)
mmu_man has quit [Ping timeout: 240 seconds]
mmu_man has joined #openscad
mmu_man has quit [Ping timeout: 240 seconds]
tachoknight has joined #openscad
guso78 has quit [Ping timeout: 245 seconds]
mmu_man has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
mmu_man has quit [Ping timeout: 246 seconds]
mmu_man has joined #openscad
mmu_man has quit [Ping timeout: 265 seconds]
J2345 has quit [Quit: Client closed]
J2345 has joined #openscad
kow__ has quit [Ping timeout: 240 seconds]
Guest73 has joined #openscad
<Guest73>
$fn = 100; // Set the number of facets for curved surfaces
<Guest73>
/ Bottle dimensions
<Guest73>
$opening_diameter = 8; // Diameter of the bottle opening
<Guest73>
$neck_diameter = 11; // Diameter of the bottle neck
<Guest73>
$base_diameter = 40; // Diameter of the bottle base
<Guest73>
$height = 180; // Overall height of the bottle
<lf94>
you know how we've gotten really good at telling CG from reality
<lf94>
I feel that's the type of thing AIs gonna cause for a bit
<InPhase>
Yeah. We sort of have an uncanny valley of code generation going on.
<lf94>
I also wonder if that was with GPT4
<InPhase>
teepee: I think when the history of 3D printing is written, Stratasys's strategies are probably going to look like IBM's blunders during the emergence of the consumer PC market.
<InPhase>
They seem to think the future of 3D printing is in large manufacturing.
<lf94>
future of 3d printing is good f'ing quality on the current beds
<lf94>
also: color
<InPhase>
And there is a role in this, for now! It should however shift rapidly to more favoring of decentralized systems.
<InPhase>
On-site availability and customization is just too potent. And thingiverse was a far more important asset for this than a bunch of printer lines.
<J2345>
i am waiting for the 24nozzle printer
<teepee>
I would like decentralized, but I'm not seeing too much progress into that direction, just in general
<InPhase>
J2345: 10 should suffice.
<J2345>
i am not sure we see much decentralization in the consumer sector as there is postprocessing and assembly required
<InPhase>
J2345: As long as we can mix a bit.
<InPhase>
J2345: CMYKW with rigid and flexible and a mixing process, and then we have all colors and flexibilities by ratio of these.
<J2345>
InPhase i have seen a 6nozzle printer that sequentially uses the nozzles
<InPhase>
But a mixer probably requires a single nozzle and a small mixing chamber.
<J2345>
my approach was not color but more a printing line that build a print like a matrix
<InPhase>
And then with 10 lines coming into it.
<InPhase>
Realistically I could work with 5, because I probably don't need all colors concurrently with multi-flexibility, so I could change it out. But 10 would be best.
pie_ has quit []
<InPhase>
One might also want transparent, so maybe 6 or 12.
<J2345>
all the same color but a printhead 2cm wide
<J2345>
pushing 200mm³ per minute
<J2345>
second .. and more like 500 (20 per nozzle)
guso78 has quit [Quit: Client closed]
pie_ has joined #openscad
pie_ has quit [Client Quit]
pie_ has joined #openscad
pie_ has quit [Client Quit]
pie_ has joined #openscad
mmu_man has joined #openscad
<InPhase>
J2345: Ah, yeah. For speed one could get some nice results with a fat sweeping line.
<InPhase>
Also, a fat sweeping line would probably make stronger prints.
<InPhase>
But you'd need to be able to adjust how much of the line is active, like with some sort of sliding gate.
<InPhase>
Or perhaps simply some injectors that merge internally into a line right before coming out, and then you get the width by which ones are injecting. And that method could potentially even print lines with gaps in the middle if the software were clever about it.
<InPhase>
If you could do different extrusion results for one side of the line vs the other, then you could sweep it in a circle and get circular infill patterns / layer lines.
<J2345>
every nozzle would have a coil that moves a needle valve and then like a matrix printer