<InPhase>
The syntax is also potentially intuitive. I'm only a little confused about the structure of that blend result with the cube vertices deforming outward near the ball.
<InPhase>
Oh. Is that the 0.5 in the blend, meaning a square root blend function, which then extends things beyond their original sizing into the outer zones of their sdf?
<Freysgodi>
Hello. Is there way to extrude shape parametric by extrusion dimension? Say when I do 'linear_extrude() circle(r)' I get cylinder of radius r, but what if I want it to be 'variable width cylinder', so radius is defined as function of z, say r(z)=2+sin(z). Is it possible?
<teepee>
Freysgodi: there's no simple way like passing the function to linear_extrude, but there's various libraries allowing such things in different way
<Freysgodi>
can you name a few?
<teepee>
I assume the ClosePoints would be an option, and BOSL2 provides extended extrude options
<Freysgodi>
thanks, will take a look
guso78k has joined #openscad
dustinm` has quit [Quit: Leaving]
<peeps[win]>
what's with new users always wanting complete rewrites of the language
<InPhase>
I guess everyone seeks the familiar.
dustinm` has joined #openscad
<buZz>
Freysgodi: usually i then just do a forloop with very low linear_extrudes in steps over the full height i want, and then hull() two of them ;)
<buZz>
then increase steps until output is 'that'll do'
<buZz>
and maybe lower for quicker renders
<buZz>
Freysgodi: you catch my drift?
<Freysgodi>
yes, contemplating
<buZz>
:)
<buZz>
its more of a cheat, but still could be totally suitable
<Freysgodi>
kinda works, thanks
<buZz>
yw :)
guso78k31 has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 252 seconds]
teepee_ is now known as teepee
cart_ has quit [Ping timeout: 255 seconds]
lastrodamo has quit [Quit: Leaving]
guso78k has quit [Quit: Client closed]
guso78k has joined #openscad
L29Ah has quit [Ping timeout: 258 seconds]
<lf94>
InPhase: just saw your comment: I think you're right, but I'd have to check
<lf94>
peeps[win]: who just said that, me?
<lf94>
"rewrites of the language"
<guso78k>
lf94 you fixed the segfault by reordering your code ?
<lf94>
I feel like I missed something lol
<lf94>
guso78k: no
<lf94>
guso78k: I PM'd you
<guso78k>
did not yet notice ;)
<lf94>
InPhase: I'm starting to feel, deep down, that this how OpenSCAD started
<lf94>
as in
<lf94>
*forgets username* wanted to one day use code to make a few models
<lf94>
there were libraries around to do geometry booleans
<lf94>
he collected the bare minimum together
<guso78k>
lf94 which username did you pm to ?
<lf94>
and boom: openscad
<lf94>
that one
<lf94>
hehe
<lf94>
is the bouncer not working for you? ;p
<lf94>
ahhhh
<lf94>
78k31
mmu_man has joined #openscad
<lf94>
ill pm this one
<peeps[win]>
lf94, it was mostly in response to the overall trend of github issues and mailing list topics etc. the bytecode interpreter issue that was just opened is the latest
<teepee>
that would not change the language though
<teepee>
but yeah, it seems a bit odd lately
<lf94>
peeps[win]: ah lol.
<lf94>
have you guys been following KittyCAD at all?
<guso78k>
lf94 i better write here because bigger keyboard
<guso78k>
i have proven in my openscad fork that the program crashes with this "double free error" when using calling libfive and not even using its result.
<guso78k>
when not invoking libfive, program does not crash
<guso78k>
lf94 do you own a mac ?
<lf94>
Haha, I dont x)
<lf94>
Linux4ever
<lf94>
(Jokes aside - I own a thinkpad and a regular desktop)
<guso78k>
there are people who would never use something else than a mac(my brother included)
<guso78k>
probably also OpenSCAD personalities ...
<teepee>
guso78k: try disabling the internal multithreading of libfive
<guso78k>
teepee, this is a receipe against the double free error ?
<teepee>
it's possible
<guso78k>
teepee, good hint, thank you!
<teepee>
the example PR will crash too with threads enabled as the openscad part is not guarding against that
<guso78k>
teepee how is it going with todays' OLED case ?
<teepee>
which case?
<guso78k>
today you have starred a fancy case and i believe i have seen photos of your make already ...
<lf94>
teepee: ouuu, neat idea.
<lf94>
that may be the reason for crashing in summonscript too
<teepee>
guso78k: oh, on github? nah, I'm just uploading a bunch of my designs to github for web page deployment and to printables - https://www.printables.com/@Torsten_199456
<teepee>
the self-starring is for my backup script which will clone and sync all starred github / gitlab repos to the raspi-nas :)
<guso78k>
ahh, no recent event, but still interesting
<peeps[win]>
I guess i'm just annoyed that its seemingly the people with the least invested that come in demanding the most wide-sweeping and arduous changes
<lf94>
could you link the issue? sounds entertaining
<guso78k>
peepsalot: i understand; people which did least , request most ?
<teepee>
the byte-code interpreter issue is from one of the most active manifold developers, that seems an honest suggestion with even example code
<teepee>
in general some sort of add/diff feature would be neat, but so far I've not seen any good and readable idea
<guso78k>
i believe most time in openscad is spent on 3d evaluation time, not on script parsing time
<teepee>
yeah, except when using BOSL2 :)
<teepee>
also Manifold makes a huge difference in that regard
<guso78k>
SCAD libraries are very powerful, even if they try to cope/process on a (crippled) environent
<teepee>
my position is still that's due to limited geometry engine possibilities and not the language itself
<lf94>
how many x faster is manifold? can we do surface-level details now with reasonable resources?
<teepee>
unfortunately it seems everyone just looks at language to solve this - which is not going to work as it's not the main problem
<lf94>
this pca person
<lf94>
has been very active across code CAD projects recently.
<teepee>
sometimes multiple orders of magnitude
<lf94>
they commented on a few threads Ive been in
<lf94>
yeah openscad's problems are engine based
<lf94>
not the language, like you said
<lf94>
(although I prefer JS over openscad, openscad is at least better than curv...)
<lf94>
lol, that second example... terrible
<lf94>
(add/del/etc.)
<guso78k>
openscad has other advantages than curv. my fork has both ...
<guso78k>
teepee, what the effort to establish libnettle in the CI images ?
<teepee>
for everything except macOS minimal
<teepee>
at least in theory as the lib is available according to my web search
<guso78k>
teepee, i was able to compile and install libnettle on macos. do you have close relationship to an MAC ?
<teepee>
yes, but not a good one ;-)
<teepee>
I'd prefer if all macs would be tossed in the trash
<guso78k>
me neither : i have 10.15 Catalina ..
<teepee>
or as alternative apple hires someone friendly to development and open source
<guso78k>
Haha, there is far too much audience about MAC yet, and we would not want to loose Marius ..
<teepee>
unfortunately nothing of those two will happen before hell freezes over
<teepee>
next best thing, someone else takes care of it :)
<guso78k>
BTW I managed to deploy python-openscad for MAC, but with every python: It will only run if there are these python environmental python files are in "place"
<teepee>
maybe that can be solved with a special venv?
<guso78k>
teepee, I still dont understand "venv" . i just dont understand ...
<guso78k>
virtual environment, but no more
<teepee>
well I never looked at the technical details, but it's like a VM just specialized for python and probably solved by only pointing some files and environment variables instead of having an actual container or VM
<teepee>
maybe more comparable to a chroot environment
<guso78k>
in any case: python needs a set of python files in place to properly start up
<lf94>
f the apple machines
<lf94>
SummonScript is *nix only
<lf94>
I think only Apple users should port FOSSware to Apple machines
<guso78k>
lf94 its effort for me too, but still my intent is to reach a broad audience
<lf94>
> guso78k │ openscad has other advantages than curv. my fork has both ...
<lf94>
hehe
<guso78k>
lf94 in my life i even try to convert BREP world to FREP world. not 100% successful yet
califax has quit [Remote host closed the connection]
califax has joined #openscad
<lf94>
yeah that's nuts
<lf94>
truly a research level project x)
<guso78k>
at least foundation is done
guso78k has quit [Quit: Client closed]
guso78k has joined #openscad
<InPhase>
guso78k: A venv is just a path Python looks in for the libraries it needs to load. When you launch a python executable, it looks relative to whatever path it was launched from. Therefore, a link to a python executable (or a copy of one) within a subdirectory tree will look within that subdirectory tree according to a standard pattern for packages and library files.
<InPhase>
guso78k: "Activating" a venv is literally just the really simple step of putting that venv's directory with python in it first in the path.
<InPhase>
guso78k: There's no more magic to it than that.
<InPhase>
guso78k: If you run pip out of that venv directory, it installs packages into the venv directory as well, after recognizing where it was launched from.
<InPhase>
The point of having them is just so you can setup specific versions of packages a particular project needs without conflicting with the needs of other projects (a very common Python problem).