<gbruno>
[github] kintel closed pull request #4822 (Rename --render=cgal to --render=force to force-convert to the current backend-specific geometry) https://github.com/openscad/openscad/pull/4822
TheCoffeMaker has quit [Read error: Connection reset by peer]
TheCoffeMaker has joined #openscad
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
fling has quit [Remote host closed the connection]
fling has joined #openscad
fling has quit [Remote host closed the connection]
fling has joined #openscad
fling has quit [Remote host closed the connection]
fling has joined #openscad
lockywolf has quit [Killed (ozone (No Spam))]
lockywolf has joined #openscad
hyperair has joined #openscad
J24k13 has quit [Quit: Client closed]
J24k13 has joined #openscad
zauberfisch has joined #openscad
<J24k13>
teepee doesn't the snap cheatsheet automatically copy changes from release?
J24k13 has quit [Quit: Client closed]
J24k13 has joined #openscad
J24k13 has quit [Client Quit]
J24k13 has joined #openscad
erectus has quit [Ping timeout: 260 seconds]
erectus has joined #openscad
erectus has quit [Ping timeout: 260 seconds]
hyperair has quit [Read error: Connection reset by peer]
erectus has joined #openscad
JakeSays_ has joined #openscad
JakeSays has quit [Ping timeout: 256 seconds]
alycidyne has joined #openscad
<alycidyne>
Hello, don't know if anyone else here uses OpenSCAD on Gentoo, but I'm getting this build error here and I didn't get an answer on #gentoo: https://paste.gentoo.zip/z1qgDpmG
hyperair has joined #openscad
<teepee>
alycidyne: not sure, but it sounds like the current code base does not support compiling without the experimental features turned on
<alycidyne>
teepee: that's what I'm thinking too
<alycidyne>
FWIW I think this involves Manifold
<alycidyne>
Note this is on the live ebuild so it pulls before building
<teepee>
the largely tested and supported case is "enable-all-experimental-features" which includes manifold
<teepee>
I don't know where this combination is coming from, that it tries a release build but still wants to enable manifold separately
<alycidyne>
More specifically in Geometry.h, it seems obvious that ENABLE_MANIFOLD is not defined so the declaration `virtual void visit(const ManifoldGeometry& node) = 0;` never is inserted, but VISITABLE_GEOMETRY inserts a call to it nevertheless; typically the ebuild is supposed to -DENABLE_MANIFOLD
<alycidyne>
Also, the ebuild has USE="experimental", but enabling that makes it fail on tbb - because the ebuild doesn't even pull it in: https://paste.gentoo.zip/blwtZpQy
<alycidyne>
This _does_ seem like a Portage bug
cart_ has joined #openscad
JakeSays has joined #openscad
JakeSays_ has quit [Ping timeout: 264 seconds]
cart_ has quit [Ping timeout: 272 seconds]
To_Aru_Shiroi_Ne has quit [Ping timeout: 264 seconds]
R2robot has quit [Quit: Pull the lever, Kronk. Wrong leverrrrrrr!]
R2robot has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 260 seconds]
teepee_ is now known as teepee
guso78k has joined #openscad
cart_ has quit [Ping timeout: 255 seconds]
GNUmoon has quit [Ping timeout: 260 seconds]
GNUmoon has joined #openscad
sellorm has quit [Ping timeout: 255 seconds]
R2robot has quit [Quit: Pull the lever, Kronk. Wrong leverrrrrrr!]
R2robot has joined #openscad
J24k13 has quit [Quit: Client closed]
J24k13 has joined #openscad
snaked has quit [Remote host closed the connection]
guso78k has quit [Quit: Client closed]
J24k13 has quit [Quit: Client closed]
J24k13 has joined #openscad
cart_ has joined #openscad
sellorm has joined #openscad
sellorm has quit [Ping timeout: 256 seconds]
kintel has joined #openscad
<kintel>
alycidyne t4nk_fn Not sure if building an experimental OpenSCAD but disabling Manifold is a good idea. I would probably just disable experimental until a full build can be made.
<kintel>
Also, ENABLE_MANIFOLD is a temporary, internal, macro which will be killed at some point as we integrate Manifold more tightly
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mmu_man has quit [Ping timeout: 255 seconds]
<t4nk_fn>
hey.. I just tried to build without portage, it installed pillow and numpy with pip install.. is that just a venv?
<t4nk_fn>
probably, right? I think it might be using pyhton3.12, my active version is 3.11
kintel has joined #openscad
<kintel>
t4nk_fn Yes, it's a venv, and those tools are only needed by the test framework
<t4nk_fn>
-- venv setup for *snip*/openscad/build/tests/venv/bin//python .. that looks a bit dodgy ;)
<t4nk_fn>
oh btw, I never had the experimental USE flag enabled on my openscad-9999
<kintel>
you mean the double // ?
<t4nk_fn>
yes kintel
<t4nk_fn>
anyhow, I got a bit interested in the experimental options after I saw mention of the issues with gentoo here
<t4nk_fn>
(wondering if any of that multicore stuff is any good of course)
<t4nk_fn>
couldn't find any docs on the experimental options
<kintel>
t4nk_fn Not sure if the // matters, but it can probably be fixed with more clever CMake usage; not sure if we can rely on CMake's find_path() to always yield a trailing slash
<kintel>
The experimental options are not really documented; they're meant to be short-ish lived and documented as they make it out of experimental mode
<t4nk_fn>
(I must remark that I did not run the set-env script thingie.. I just checked the dependencies (which were ok of course since I'm building from portage normally))
<kintel>
That's why I was wondering why package experimental in Gentoo at all. ..unless there are two parallel versions; openscad and openscad-nightly or smth.
<kintel>
Basically if you're using the experimental build, you're kind of expected to keep updating as it's in constant flux.
<kintel>
"flux" as we may change API and stuff, but all tests are going to pass
<t4nk_fn>
mmm well, there's openscad 'stable' and openscad-9999 which is the 'live build' from git
<kintel>
ok, cool, makes sense
<t4nk_fn>
openscad-2021.01-r4.ebuild seems to be the stable one at this moment
<kintel>
yeah, that's the latest stable release
<t4nk_fn>
and I'm on Mon Feb 12 18:26:57 2024 >>> media-gfx/openscad-9999
<kintel>
Anyway, we're expecting some friction with Manifold and distro packaging, and this is likely one of the early attempts at doing that.
mmu_man has joined #openscad
<t4nk_fn>
I'm not completely versed enough to be very helpful I guess, but I'm quite interested in this experimental rubbish ;) so I guess I can just override the commit to perhaps feb 12 and build that one
<kintel>
The later the better, as master is pretty active, and we also update Manifold from upstream regularly
<kintel>
( and Manifold is perhaps the most exciting part of experimental ;) )
<t4nk_fn>
hey, does this default to trying to use ccache btw?
<kintel>
Yes
<kintel>
..but you can -DUSE_CCACHE=OFF
<kintel>
We also support CMAKE_UNITY_BUILD
<t4nk_fn>
funny.. that's not in the ebuild, wonder how much of a difference it would make
<t4nk_fn>
ks
<t4nk_fn>
that was supposed to be ls.. and meant for my console ;)
<t4nk_fn>
OpenSCAD version 2024.02.28 .. well, it built
<t4nk_fn>
what the?!
<t4nk_fn>
I just tested some 'leadscrew holder' I made; normal rendering took 12+ secs, with manifold it's friggin' instant!?
<pca006132>
yes
<pca006132>
try something that requires multiple hours
<pca006132>
usually they can finish within a few minutes, if you don't require minkowski
TheCoffeMaker has quit [Quit: So long and thanks for all the fish]
ToAruShiroiNeko has quit [Ping timeout: 256 seconds]
ToAruShiroiNeko has joined #openscad
<t4nk_fn>
not sure what's going on.. I took the listing of smoothed-antennas.scad from that page, and pasted it into a new file in openscad
<t4nk_fn>
and it's been at it for a decent while now already, from pressing F5 alone
<t4nk_fn>
openscad is using 6% cpu so it seems to be doing things on one core
<t4nk_fn>
not sure if I have that manifold disabled or not, I thought not
<t4nk_fn>
too bad I didn't time it, now I won't know
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kintel has joined #openscad
<kintel>
F5 generally doesn't trigger manifold, unless you have a `render()` statement in your code
<pca006132>
yeah, there can be cases in which preview is slower than render...
<t4nk_fn>
not quite sure, but it might be going on close to half an hour now
<t4nk_fn>
with the stuff I normally do, F5 is always instant
<t4nk_fn>
and I guess with manifold, pretty much everything will be instantaneous.. I'm a simple guy ;)
TheCoffeMaker has joined #openscad
<kintel>
It's very possible that you found a bug. There has been a bunch of changes to the rendering infrastructure lately.
<kintel>
If you have time to try to slim down your design to something smallish that still fails, we'd appreciate if you opened a GitHub ticket
<t4nk_fn>
I was just looking at the fact that they mention a 'mac M2 Max' .. I'm doing this on a ryzen 7 2700x
<t4nk_fn>
well, memory usage has just shot up by ten-fold in the last minute or so
<t4nk_fn>
for 1 to now 11G
<t4nk_fn>
yeah, I'll kill this
<kintel>
It uses minkowski, which is the exception mentioned by pca006132
<kintel>
..but it should perform better
<t4nk_fn>
though the document states "Note that minkowski operations get a specific boost thanks to the introduction of parallelism in the algorithm itself (and then the union of parts it generates benefits from Manifold's own parallelism)"
<t4nk_fn>
heh
<kintel>
I didn't realize until now where you pulled that example from. We'll take a look at it
cart_ has quit [Ping timeout: 260 seconds]
<t4nk_fn>
I tried it again, and put it in a render()
<kintel>
..should probably re-run those performance tests to look for regressions
<t4nk_fn>
it's there now
<t4nk_fn>
and it was using multiple cores
L29Ah has quit [Ping timeout: 252 seconds]
To_Aru_Shiroi_Ne has joined #openscad
ToAruShiroiNeko has quit [Ping timeout: 264 seconds]
To_Aru_Shiroi_Ne is now known as ToAruShiroiNeko
<JordanBrown>
j24k the {{anchor|circleres}} was already there. I didn't know that you could pass multiple anchors in a single invocation.
<JordanBrown>
I agree that individual anchors for $fa, $fa, et cetera would be good; I was only repairing the broken links and leaving improvement for another day.
<J24k13>
well i think that circleres makes no sense - so we should remove that one
<J24k13>
ah it probably means circle resolution .. still weird
<JordanBrown>
Yeah, the word made sense, and in fact if I was going to have a single word for that section I would consider it to be an okay one. But having multiple anchors using the variables names seems better.
<J24k13>
I already changed the links in the cheatsheet (not snap)
<J24k13>
else something like circleRes would be less confusion for me ..
<JordanBrown>
I figure that once you move away from using the actual text of the heading, you're creating symbols akin to variable names, and there's any number of good and bad patterns for doing that.
<kintel>
t4nk_fn fwiw, the design in question rendered fine with F5 on my build
<JordanBrown>
But alas I should now pretend to do real work...
L29Ah has joined #openscad
qeed has quit [Quit: qeed]
<t4nk_fn>
kintel: with my original openscad build from feb 12 it started eating mem way sooner, and I hit: 'CGAL error: assertion violation! Expression : e_below != SHalfedge_handle() File : /usr/include/CGAL/Nef_3/SNC_FM_decorator.h Line : 418'
<t4nk_fn>
also I killed it when ' 23.1 GiB + 19.2 MiB = 23.1 GiB openscad"
qeed has joined #openscad
J24k13 has quit [Quit: Client closed]
J24k13 has joined #openscad
<t4nk_fn>
with the manifold version, put into a render() ... took probably a minute or two
<t4nk_fn>
and it's now using 6.8G
aiyion2 has quit [Ping timeout: 260 seconds]
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
aiyion2 has joined #openscad
<t4nk_fn>
really need to get some sleep..
<t4nk_fn>
anyhow, for the things I normally do this will be a massive improvement, that's very nice indeed
<t4nk_fn>
thnx for your time, later.
aiyion2 has quit [Remote host closed the connection]
aiyion2 has joined #openscad
SamantazFox has quit [Remote host closed the connection]