<teepee>
I remember the cmake install discussion, I thought that was all merged already
ndnihil is now known as forest
forest is now known as nihil
nihil is now known as ndnihil
<peepsalot>
yeah, probably part of the confusion (on both our parts) is because I kept splitting off pieces of that PR into separate PRs and merging those... then rebasing the original and doing a force-push
<peepsalot>
I was planning to do a final test of CI build artifacts on my windows box before merging, and I guess my attention wandered elsewhere before I got to it
<teepee>
good thing github remembers :)
<teepee>
the multiple projects at work have left me with a quite reduced attention span in the evenings which is not helping for the things I'd like to do
KimK has quit [Remote host closed the connection]
KimK has joined #openscad
L29Ah has quit [Ping timeout: 264 seconds]
califax has quit [Remote host closed the connection]
califax has joined #openscad
L29Ah has joined #openscad
<Scopeuk>
linext isn't all3d basically just a content aggregator? Everything is shipped from elsewhere, they just got a good domain name
castaway has joined #openscad
milza has joined #openscad
GNUmoon has quit [Remote host closed the connection]
<J23644>
»Sipeed expects to sell it for less than $2 at launch.« seem affordable
<buZz>
so about same price as esp32?
<Scopeuk>
I think the $2 is the little dev board rather than the raw chip
<Scopeuk>
but yeh same ball park
<teepee>
specs are reading very similar to esp32-c3
<teepee>
maybe a bit more generous in terms of peripherals
<teepee>
I suppose the most important part will be software side support
milza has quit [Ping timeout: 246 seconds]
Lagopus has joined #openscad
<Scopeuk>
I think the bit for me was that it had wifi and zigbee so could act as a bridge, the ESP's appeared to be either or
<teepee>
yeah, I think there's no esp32 which has all
<teepee>
but then I'm not missing that much with the dresden electronic usb stick :)
<teepee>
they really packed everything... VPU – MJPEG video encoder
<teepee>
apache 2 license sdk on github, that's nice
<buZz>
esp32 can do rtmp without hw acceleration ;)
<buZz>
hehe even, esp32 has hw sha256 acceleration, which is -slower- than doing it in cpu :P
<teepee>
I suppose the idea is not just speed but maybe also parallelize + power if it's done right
rapha has joined #openscad
<rapha>
hi all!
<rapha>
if you import an .stl, is there a way to make that whole object solid? i.e. keep the outer shape but fill any and all holes it might have, possibly (but not necessarily) simplifying the outer shape in the process, too?
<teepee>
no
<teepee>
unless it's a completely convex object, but I assume that's really never the case in real world
<buZz>
rapha: yes, in meshmixer or blender or something
<buZz>
meshmixer (and meshlab iirc) can remesh to simplify shapes in .stl
<rapha>
oh, no openscad? sad. but thank you, buZz!
<teepee>
a 3d fill was actually discussed not long ago but the new feature is 2d only for now
<teepee>
I suppose a generic remesh would be nice, but nobody actually had a look at this yet
<buZz>
rapha: openscad is a poor stl editor :)
<teepee>
yeah, never going to be a replacement of low level editing like in blender
<teepee>
some more mesh processing would be possible and useful though
<InPhase>
buZz: It's a fine computational geometry modifier of stl files. It depends on one's mental model of what stl "editing" should be.
<buZz>
:)
<InPhase>
buZz: But with data=render() we might also get a lot more nuanced editing.
<buZz>
ah, yeah render() is something chatGPT keeps outputting aswell for openscad
<InPhase>
OpenSCAD is the worst supported language I've found so far of the ones chatgpt thinks that it knows.
<InPhase>
It hits at its achilles heel of having no visual awareness.
<InPhase>
Also, it really fouls up the syntax pretty bad sometimes too.
<InPhase>
rapha: And the core problem really is that the badly formed stl files don't even have a definition for "outer".
<InPhase>
rapha: There's no obvious dividing line between an stl file of a donut, and an stl file of a model that has some tiny triangles that spike a microscopic triangular tube through the middle that pops out the other side inverted.
<teepee>
I've read that not as fixing stl but removing enclosed holes?
<teepee>
like fill3d()
<teepee>
which would still have 2 interesting cases. the simple one being difference() { cube(10, center=true); cube(2, center=true); } = fully enclosed cavity
<teepee>
vs. a hole through the cube
<teepee>
the 2d fill would only cover the 1st version
califax has quit [Ping timeout: 255 seconds]
califax has joined #openscad
kintel has joined #openscad
<kintel>
peepsalot: Thx, your PR looks fine, albeit being a collection of unrelated fixes. Let me know if you plan to split out more PRs from it. Otherwise I can try to give the current PR a spin on macOS while I'm in the mood to fight build systems
<InPhase>
Removing fully enclosed properly manifold internal empty spaces would be a doable operation, and has an exact definition. I assumed rapha was looking more for a manifold-fixing remesh however. Which can be done in principle with a size scale. Just all the tools that attempt to do it struggle with it, so it's apparently a bit tricky.
<teepee>
yeah, I think that would be just a CGAL operation that follows the outer shell dropping anything else. but those operations are never really obvious
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<InPhase>
Good. I spent 15-20 minutes since my last comment attempting to design one. I convinced myself it's doable but involved. Expanding outward like a balloon involves some messy calculations for finding some edges, so shrink wrapping is probably the better choice.
kintel has joined #openscad
<teepee>
omg, this cheap-ish bluetooth headset says HI! when switching on :)
<InPhase>
Well that could freak a person out if they're not expecting it.
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kintel has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 255 seconds]
teepee_ is now known as teepee
Guest36 has joined #openscad
Guest36 has quit [Client Quit]
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<J23644>
•oO(when your headset becomes your friend)
kintel has joined #openscad
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<linext>
i don't like bluetooth speakers that talk when turned on
<kintel>
^ peepsalot: Suggestion for minor clang-tidy tweak.
<kintel>
I'm also wondering if we should run clang-tidy via the run-clang-tidy wrapper for more control over how it's executed. E.g that wrapper adds a -j argument which could free clang-tidy from being bound to a single core
<peepsalot>
kintel: tweak looks ok to me... as long as cmake doesn't care if HINTS keyword is given when nothing follows it. not sure if it matters, but maybe wrapping in quotes is more proper
<peepsalot>
... HINTS "${CLANG_TIDY_HINTS}")
<peepsalot>
kintel: i don't know anything about that wrapper. its currently being run through cmake, which is still capable of multiple threads
<peepsalot>
with the caveat that headers get processed multiple times and if you use "--fix", it will mangle things with multiple randomly interleaved "fixes" per offending line
<kintel>
peepsalot: find_program can handle zero or more HINTS arguments; it seems pretty common to just pass a naked variable. It expects a list so we shouldn't quote it. I've tested it to work with zero, one and two args.