teepee has quit [Remote host closed the connection]
teepee has joined #openscad
guso48 has quit [Quit: Client closed]
snaked has joined #openscad
mmu_man has quit [Ping timeout: 250 seconds]
guso19 has quit [Quit: Client closed]
LordOfBikes has quit [Ping timeout: 260 seconds]
qeed_ has joined #openscad
qeed has quit [Ping timeout: 260 seconds]
LordOfBikes has joined #openscad
<J23k89>
teepee you probably need gold wire if that should work for longer or at least tinned wire
JakeSays_ is now known as JakeSays
TheAssassin has quit [Ping timeout: 240 seconds]
TheAssassin has joined #openscad
hyperair has quit [Remote host closed the connection]
hyperair has joined #openscad
J23k42 has joined #openscad
J23k89 has quit [Ping timeout: 246 seconds]
misterfish has joined #openscad
guso37 has joined #openscad
<guso37>
teepee, i also appreciate printed keypads, but even with this design its not clear to me: copper wire is not elastic. not sure how a key can be pressed and release reliable more than 100 times ...
qeed has joined #openscad
qeed_ has quit [Ping timeout: 264 seconds]
fling has quit [Ping timeout: 240 seconds]
fling has joined #openscad
misterfish has quit [Ping timeout: 260 seconds]
hyperair has quit [Quit: Segmentation Fault]
hyperair has joined #openscad
<J23k42>
guso37 there is a printed structure of crossing bridges that pull the wire apart after contact
J23k42 has quit [Quit: Client closed]
J23k42 has joined #openscad
guso37 has quit [Quit: Client closed]
misterfish has joined #openscad
Lagopus has quit [Read error: Connection reset by peer]
<teepee>
yeah, better wires are probably needed if it's supposed to work for more than a month. the bridges for contacts look interesting. not 100% clear from the pictures, so downloading the scad is on the to-do list :)
<J23k42>
checked on amazon "18k gold plated copper wire" .. says material: bras .. guess they are all fake
<teepee>
yeah, looks like it's crashing with some memory corruption: terminate called after throwing an instance of 'std::bad_alloc'
<teepee>
usually valgrind is a good option although it's a bit tricky to setup due to CGAL and valgrind having some compatibility issues
<teepee>
but kintel added some build options for various instrumentations, that might help already too
<teepee>
e.g. making it crash on linux
<teepee>
-DSANITIZE_MEMORY=ON when running cmake
<teepee>
there's also -DSANITIZE_ADDRESS=ON -DSANITIZE_THREAD=ON -DSANITIZE_UNDEFINED=ON
<guso78>
Yeah with These Options i can probably also See in Linux. Thank you!
J23k42 has quit [Quit: Client closed]
J23k42 has joined #openscad
<InPhase>
guso78: When I search for exit code 8 and windows, I see it described as "No driver was found that can be installed on the current device"
<InPhase>
guso78: Did something change the opengl stuff?
Lagopus has joined #openscad
<guso78>
Inphase in gl i Just Retrieve the coordinates in slightly Other way. No memory Handling changes there.
<guso78>
Did Not rebase after the PR merge of Josef. Could that cause an issue?
<InPhase>
guso78: I don't see anything about memory.
<InPhase>
Not sure why you were mentioning it.
<guso78>
Maybe a shared ptr content was Used after IT got freed. This IS the drawback of using get() method
<guso78>
I mentioned memory because Most of my issues we're memory. But i will find Out.
<Scopeuk>
The log suggests usemimalloc was enabled for that build
<Scopeuk>
I thought we disabled that on windows at the moment as it wasn't really working
<guso78>
Error Code 8 in the ci build was Always segfault on command line
juri__ has joined #openscad
juri_ has quit [Ping timeout: 245 seconds]
juri__ has quit [Ping timeout: 246 seconds]
juri_ has joined #openscad
mmu_man has quit [Ping timeout: 260 seconds]
misterfish has quit [Ping timeout: 250 seconds]
mmu_man has joined #openscad
misterfish has joined #openscad
J23k42 has quit [Quit: Client closed]
erectus has quit [Ping timeout: 240 seconds]
J23k42 has joined #openscad
erectus has joined #openscad
<joseph_>
teepee: I saw that you merged that my PR, thank you. You may have seen there was a CI failure but it was due to that flaky test 1240. As a procedural question, should I be squashing all my commits before future merges in order to make the history easier to read on master?
<joseph_>
Also I just created a second PR (with only a couple small changes)
<teepee>
joseph_: nice, thanks for the update, I'll have a look at the PR. it mentions removing GL_TRACE - is it replaced with some better way for debugging?
<teepee>
as for the test, yeah, it's again ubuntu-20.04. so I'm now supecting it's the old version of tbb. I've seen the 22.04 runs fail lately. maybe I can get the newer one to build on OBS
<teepee>
while in your branch, you can squash/fixup commits if it starts getting difficult to read. but there's no need to always squash everything. a couple of reasonable smaller commits are even better than a big one
<teepee>
basically: there's no fixed rule :)
guso57 has joined #openscad
<guso57>
hmm in my linux i could not find any error with any of -DSANITIZE_ADDRESS=ON -DSANITIZE_THREAD=ON -DSANITIZE_UNDEFINED=ON, -DSANITIZE_MEMORY=ON. Of course i had to compile and test 3 times as most of the defines will not work together.
<teepee>
ok, that's surprising then. maybe it's indeed the mimalloc thing Scopeuk pointed to. I don't remember what the last inteded state on Windows is / or what it should be right now
<guso57>
is it possible to definitely disable mimalloc for windows and rerun my windows ctest ?
<guso57>
on the other hand. i feel its very likely that I either created memory leak or used already-freed memory because all this shared-ptr thing is just too new for my knowledge.
<guso57>
The fact that you cant safely convert a shared_ptr back to a normal ptr without creating a complete clone make me crazy.
<teepee>
depends on what you mean by that. you can convert as long as you hold a copy and pass it down
<teepee>
but that raw pointer must never be stored anywhere
<guso57>
if i work on a function, which returns a raw ptr for historical reasons, but it gets a data from a subfunction which returns shared ptr. i would write:
<guso57>
std::shared_ptr tmp=subfunc();
<guso57>
return subfunc.get();
<guso57>
sorry:
<guso57>
return tmp.get();
<teepee>
that looks dangerous
<guso57>
this will definitily fail, because the memory could be freed before the caller of the func could access it.
<guso57>
in other words: i dont know of any way to convert a shared ptr to a aw ptr
<teepee>
I don't think there's one in that direction (= returning values)
<teepee>
you can pass the raw pointer down, e.g. to printf() while you are holding at least one reference so it's safe it will not go away
<guso57>
maybe a code review could help, but its very difficult , as the changed file set is over 40 ....
TheAssassin has quit [Ping timeout: 240 seconds]
TheAssassin has joined #openscad
<teepee>
worth a try I suppose
guso57 has quit [Ping timeout: 246 seconds]
J23k42 has quit [Quit: Client closed]
J23k42 has joined #openscad
J23k42 has quit [Client Quit]
J23k42 has joined #openscad
GNUmoon2 has quit [Ping timeout: 240 seconds]
GNUmoon2 has joined #openscad
pa has quit [Quit: quit.]
pah has joined #openscad
snaked has joined #openscad
<lf94>
when looking up "cad" on linkedin I get no actual CAD jobs lol
<lf94>
just a ton of software people jobs
<lf94>
Any suggestions?
<lf94>
Secondly: any suggestions on a robust gear design? I gotta build something to try and thresh the wheat I grew
<lf94>
Im open to ideas
<lf94>
So far the neatest thing I've seen is put it into a bucket and shake it with a couple of heavy rocks
snaked has quit [Ping timeout: 260 seconds]
teepee has quit [Remote host closed the connection]