califax has quit [Remote host closed the connection]
califax has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 252 seconds]
teepee_ is now known as teepee
bozo16 has joined #openscad
mmu_man has joined #openscad
J23k has joined #openscad
misterfish has joined #openscad
misterfish has quit [Ping timeout: 264 seconds]
mmu_man has quit [Quit: reboot]
mmu_man has joined #openscad
misterfish has joined #openscad
pa has quit [Quit: quit.]
teepee_ has joined #openscad
teepee has quit [Ping timeout: 252 seconds]
teepee_ is now known as teepee
ferdna_ has quit [Quit: Leaving]
mmu_man has quit [Ping timeout: 255 seconds]
prev has joined #openscad
<prev>
Hi, I'm struggling to build a portable executable on Mac OS X. I've followed the build instructions on the Github page and have successfully built a working application, but the libraries and frameworks aren't included within the application itself and it stops working if I move the ../libraries/ directory that was created by
<prev>
./scripts/macosx-build-dependencies.sh. I feel like I'm missing something obvious but haven't had any luck with "make install" or "make package" yet. What should I be looking at?
mmu_man has joined #openscad
<teepee>
prev: macdeployqt or whatever it's called
<teepee>
it's called from release-common.sh
<teepee>
that script does the DMG packaging
<prev>
teepee: thank you! i'll look into that now.
kintel has joined #openscad
<kintel>
Humm, I wonder why "ctest -R echotest_issue4172-echo-vector-stack-exhaust" take 31 sec on Mac but 0.2 sec Linux..
<kintel>
Perhaps stack sizes on Mac are crazy huge?
<kintel>
One important detail: Run this script to validate that the app is actually portable, and built for the correct macOS version: ./scripts/macosx-sanity-check.py build/OpenSCAD.app/Contents/MacOS/OpenSCAD
<teepee>
oh, talking of macos versions, Manifold now allows 10.13
<kintel>
oh, nice, did they change TBB usage to avoid 10.14-specific functinos?
<teepee>
I did not have the time yet to revisit the Manifold-update PR yet
<teepee>
no it seems they only changed a couple of places in manifold directly, so maybe the 10.13 issues in TBB did not hit. I was wondering about that too re that comment in CMakeLists
<teepee>
at least someone said compiling on an actual 10.13. system worked
<teepee>
worst case, could be a difference between actual 10.13. system and later version compiling for -minver 10.13
<kintel>
ok, should we bump your Manifold PR to the current master?
<prev>
The person compiling for 10.13 was me, probably :) Beyond a few other issues (with the build system support on 10.13, not Manifold itself,) Manifold compiles nicely with 10.13 now.
<kintel>
prev Using the latest upstream manifold?
<prev>
Yep, the latest after the 10.13 patches, commit bec8164 it looks like.
ferdna_ has joined #openscad
<kintel>
humm, that's a different error than I was getting; I was hitting an issue with TBB
<kintel>
Anyway, I'll try again when I have some time
<prev>
I have a working build of the latest OpenSCAD commit (b23dd28) on OS X 10.13 as well, but it requires a workaround that's probably not easily resolved and risks issues at runtime. Related to the use of std::visit in some OpenSCAD code.
<teepee>
sorting may now be needed as manifold does not have a guarantied tesselation
<kintel>
Even then, if tessellation isn't guaranteed, no amount of sorting would help
<teepee>
true
<teepee>
hoping for our simple test cases it's enough
<kintel>
As long as there are no quads, or collinear faces, it should work I think
<teepee>
otherwise we may have to find a different way of comparing, like just looking at vertices or something
<kintel>
yeah, alternatively, just list all vertices. Otoh, since manifold does mesh optimization, it may move vertices around..
<kintel>
..but for simple tests, we should be able to make it work
<kintel>
prev What workaround, I believe OpenSCAD without manifold should still work on 10.13
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
guso78k has quit [Quit: Client closed]
<prev>
kintel I apologize for being unclear. Version 2021.01 works perfectly on OS X 10.13, the latest development snapshot (with Manifold) won't run on 10.13.
<prev>
The workaround to compile a current snapshot on 10.13 is to adjust CMakeLists.txt to unconditionally set(DEPLOYMENT_TARGET "10.13") and also add_compile_definitions(_LIBCPP_DISABLE_AVAILABILITY), as there is functionality technically present that was simply marked unavailable until 10.14.
<prev>
The problems I'm having are likely only because I have to compile on 10.13, since I don't have access to a newer machine. I imagine set(DEPLOYMENT_TARGET "10.13") would be enough if I was compiling on a newer version.
kintel has joined #openscad
<prev>
kintel: I'm happy to repost the explanations of why I'm using workarounds, in case those didn't come through to you.
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kintel has joined #openscad
<kintel>
prev no worries, I read the history
<prev>
sounds good, i'm not used to IRC in general.
mmu_man has quit [Ping timeout: 264 seconds]
<kintel>
prev I'm still getting this error:
<kintel>
include/oneapi/tbb/detail/_task.h:216:13: error: aligned deallocation function of type 'void (void *, std::align_val_t) noexcept' is only available on macOS 10.14 or newer
<kintel>
I haven't tried to upgrade tbb though. Did you?
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kintel has joined #openscad
kintel has quit [Client Quit]
<prev>
kintel: I haven't upgraded or changed TBB, looks like ./scripts/macosx-build-dependencies.sh downloaded version oneTBB-2021.8.0.
kintel has joined #openscad
mmu_man has joined #openscad
<kintel>
prev Oh, this error message comes from clang, and some versions clang seems to have a bug identifying aligned allocation as not available in macOS 10.13
<kintel>
(or rather llvm)
<prev>
Ah, that could do it. Glad there's a (relatively) obvious cause
<kintel>
But it's an old bug, don't understand why I still see it on my machine
<prev>
No idea, unfortunately, but which clang is being invoked? I have an updated version through MacPorts but CMake appears to find the version included in the XCode application.
<prev>
oh, nm, I don't have one installed through MacPorts
<kintel>
I've only used clang from Xcode, I don't trust homebrew enough :)
<prev>
Awesome, thank you folks, the built application is now standalone and portable! I really appreciate the help, it's good to have a snapshot even on my ancient devices.
<kintel>
Cool that you got it working! I might not be as easy for us, but hopefully the bug is fixed in Xcode 15, so once we can run the CI on Xcode 15, we should be able to fix it
ferdna_ has quit [Ping timeout: 272 seconds]
<prev>
Sounds good! Hopefully I'll learn enough C++ to be able to contribute by then ;)
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<teepee>
"You shall not share, sub-license, sell, rent, host, transfer, or distribute in any way the digital file or 3D printed versions of this object"
<teepee>
the file, ok, fine
<teepee>
the 3d printed object?
<teepee>
now I want a block button :(
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openscad
<J23k>
just don't follow print bunny
<J23k>
but just found some creator that have some hundred club member and getting several thousand bucks per month
<J23k>
seems that license model is new
<teepee>
well, it's BS, big time
<teepee>
restricting the all and any use of the printed object is not making sense
<teepee>
that's really a new invention also big corporations try which should never happen
<J23k>
you can buy the rights via the clubs and those pay a percentage to prusa
<teepee>
if I buy something in a shop for money, they also can't just come back to you and prevent reselling
<teepee>
with all that "it's not sold, but just licensed" crap they are trying to break that
<teepee>
and that's just a horrible idea
<J23k>
well not quite - if you buy a tesla and resell it .. the features deactivate
<teepee>
yes, and that is BS
<J23k>
but there are so many sheeple that just go along with that
<teepee>
yeah, it's awesome if Amazon can just pull out all the books from your book shelf 5 years after you bought them
<J23k>
and if your audience is using solidworks they are already used to pay money so it is quite normal for them
<J23k>
i think it was over 10yr ago when there where cases when MS mediacenter deleted files from your harddisk because of copyright
<teepee>
solidworks well, DS as company has no copyright on the stuff you make, I hope
<teepee>
yes, and that gets ever more widespread with those locked stores like on game consoles, for music and such stuff
<J23k>
copyright maybe not but in fusion360 you can't use it if your internet is down (even when your models are stored locally)
<J23k>
everything is just a service is going strong everywhere
<teepee>
which is why I like mastodon, no central service anymore
<J23k>
yes but if you don't own the instance .. i was on diaspora and then they switched of the server i was on
<teepee>
yes, but at least you should be able to move and for example Heise can run their own
<teepee>
so it's not an idiot billionaire who decides what they can publish
<teepee>
at least while we still have *some* competition in the low level internet transport
Guest29 has joined #openscad
<J23k>
i remember over 20yr ago when i found out that my mobile can be configured remotely which was normal for US markets as nobody was owning their devices.
<Guest29>
/ Defina as dimensões da placa
<Guest29>
largura = 0.04; // 40mm
<Guest29>
comprimento = 0.04; // 40mm
<Guest29>
espessura = 0.002; // 2mm
<teepee>
anyway, I'm not sure there's a license that can restrict that you can't give a printed object to someone else