<kintel>
Exactly that - they spin up a container in a VM which likely is dedicated to that repo, so the owner of all files is always UID 1000 both in the container and on the host
<kintel>
..but I want to do that on ci.openscad.org under my own user account
<kintel>
workaround is of course to run containers as root and deal with the fallout later by doing a bunch of recursive sudo operations. What can possibly go wrong?
<teepee>
yeah, the user management is bad, I've been fighting with that quite a bit too
<teepee>
like restricting ssh access to sftp uses chroot internally with has it's own idea of what user settings need to happen on the directory tree, using that then as docker mount and things get a bit confusing
<kintel>
Not sure how they make that fly with Docker, but reading the VS Code source code may reveal some useful tricks
<teepee>
docker run has a --user parameter
<teepee>
time to get some sleep :)
<kintel>
Yeah, but if I run e.g. "docker run --user 2001:100", I also need to actually create a user with that UID in the Dockerfile, or I'm screwed
<kintel>
Right, sleep. I'm in Europe right now, so this applies to me too..
<teepee>
aha, good night then
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mmu_man has quit [Ping timeout: 246 seconds]
foul_owl_ has quit [Ping timeout: 240 seconds]
LordOfBikes has quit [Ping timeout: 240 seconds]
foul_owl_ has joined #openscad
LordOfBikes has joined #openscad
J23k25 has joined #openscad
J23k59 has quit [Ping timeout: 245 seconds]
qeed has joined #openscad
qeed_ has quit [Ping timeout: 260 seconds]
Alexer has quit [Ping timeout: 245 seconds]
J23k25 has quit [Quit: Client closed]
J23k25 has joined #openscad
<gbruno>
[github] paenian opened issue #4749 (MacOS build failure: github/openscad/build/objects/lexer.cxx:955:2: error: ISO C++17 does not allow 'register' storage class specifier) https://github.com/openscad/openscad/issues/4749
fling has quit [Remote host closed the connection]
erectus has quit [Remote host closed the connection]
GNUmoon has quit [Remote host closed the connection]
califax has quit [Remote host closed the connection]
califax has joined #openscad
fling has joined #openscad
GNUmoon has joined #openscad
erectus has joined #openscad
erectus has quit [*.net *.split]
GNUmoon has quit [*.net *.split]
califax has quit [*.net *.split]
fling has quit [*.net *.split]
teepee has quit [*.net *.split]
aiyion has quit [*.net *.split]
TheAssassin has quit [*.net *.split]
yogadude has quit [*.net *.split]
erectus has joined #openscad
califax has joined #openscad
TheAssassin has joined #openscad
yogadude has joined #openscad
fling has joined #openscad
teepee has joined #openscad
Alexer has joined #openscad
aiyion has joined #openscad
lostapathy has quit [Ping timeout: 244 seconds]
peeps[zen] has quit [Read error: Connection reset by peer]
peeps[zen] has joined #openscad
peeps has joined #openscad
peeps[work] has quit [Remote host closed the connection]
RichardPotthoff has quit [Ping timeout: 258 seconds]
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
erectus has quit [Remote host closed the connection]
erectus has joined #openscad
misterfish has joined #openscad
J23k25 has quit [Quit: Client closed]
J23k has joined #openscad
cart_ has joined #openscad
GNUmoon has joined #openscad
cart_ has quit [Ping timeout: 245 seconds]
cart_ has joined #openscad
GNUmoon has quit [Remote host closed the connection]
<guso78k>
https://imgpile.com/i/DRNQ1F when creating an openscad linux release, openscad main window will not show the main graph window, but instead writes lots of warning messages to terminal
<guso78k>
when commenting line LIBGL_DRIVERS_PATH in the shell script, it will work. the LIBGL_DRIVER_PATH shows to a non-exising directory IMHO. is viable not to use this export ?
<teepee>
that's probably some old unsupported stuff
<teepee>
common linux release is appimage, what error did you get with linuxdeploy? for x86, I'm using the linuxdeploy official builds, only building for ARM64
misterfish has quit [Ping timeout: 245 seconds]
teepee_ has joined #openscad
guso78k has quit [Quit: Client closed]
teepee has quit [Ping timeout: 252 seconds]
teepee_ is now known as teepee
guso78k has joined #openscad
cart_ has quit [Ping timeout: 252 seconds]
<guso78k>
yes i launched ./scripts/release_common.sh without argument and it built for linux.
<guso78k>
did also some tests with linuxdeploy for app image but its hard to compile. finally i found an appImage for linuxdeploy
<guso78k>
how to create an appimage of openscad without using CI ?
<teepee>
you probably could also use the docker image I use for the official builds
<teepee>
in theory it's mostly "make install" + running linuxdeploy, just like the docker image does internally
<guso78k>
soo many new things: appimage, dockerimage .... i need a master plan '=D '=D
<guso78k>
yeahh, found this file. this is why i learned yesterday that i need linuxdeploy ...
<teepee>
I suppose we could make it even easier providing a docker image that can just be called from inside the source repo
<guso78k>
there are really poor programming errors in linuxdeploy like defining variables in header files which results in duplicated symbol linker errors. glad they offer AppImages of itself ..
<teepee>
just noting the main author can hear you ;-)
<teepee>
I did not have much issues building it in the past though
<guso78k>
i think the problem arose one month back or so, there is an already an "issue" on the github page about that ....
<teepee>
ah, ok. people are busy sometimes
<teepee>
that could be gcc-13 being more strict in some checking
<teepee>
I've seen a number of those type of issues
<teepee>
thing is, appimage is specifically targeting older platforms as only because of that it's possible to run the resulting apps on various systems
<guso78k>
is there an environment variable with can set an old standard for gcc ?
<teepee>
not that I know of, some distros have multiple versions, but usually the official one is the oldest
<teepee>
hmm, actually it seems my debian still has gcc-10 too?
<teepee>
so if you can install an older package in addition you can set CC=gcc-10 CXX=g++-10 or so
<teepee>
assuming the build system is not breaking that, for OpenSCAD itself, it works, so you can compile with clang instead of gcc with "CC=clang CXX=clang++ cmake ..."
<teepee>
debian bookworm shows oldest gcc-11
<teepee>
that would be the benefit of a build docker-image, you can select a distro, install the needed stuff and have that image ready even if the host system upgrades
<guso78k>
a build docker image sounds promising, suppose i need to start with docker itself, first ....
<teepee>
if your distro has it, it's as easy as
<teepee>
apt install docker.io
<teepee>
docker run -it debian:bookworm-slim
<teepee>
and there you have a shell inside the bookworm container
L29Ah has quit [Ping timeout: 240 seconds]
<guso78k>
sounds really easy, maybe i should give it a try
<guso78k>
whats space requirement for using docker ?
<teepee>
of course there's lots of much more complicated stuff, but for some simple things, it's not too bad
cart_ has joined #openscad
<teepee>
lots of space as it builds operating system images, I'd say maybe a couple of GB is probably minimum
<teepee>
bookworm-slim is only 80MB, but installing all the build stuff and doing a couple of builds is probably 2 or 3 GB
<teepee>
the full MXE build stuff all together might be even closer to 10 GB
<guso78k>
this is what i was afraid off. my linux itself is a virtual box.
<teepee>
there is docker for windows too
<teepee>
running the containers somehow in VMs
<teepee>
never used that though
<teepee>
oh, it can do hyper-v or WSL
misterfish has joined #openscad
<teepee>
also I don't know how their licensing works for windows
<teepee>
right: "Docker Desktop is free for small businesses (fewer than 250 employees AND less than $10 million in annual revenue), personal use, education, and non-commercial open source projects."
<teepee>
sounds like using docker desktop on windows would be ok
<buZz>
how can docker even work on windows, it has no namespaces in NT kernel?
<buZz>
or you mean docker-on-linux-on-windows
<Scopeuk>
docker on windows is weird
<buZz>
s/docker on //
<Scopeuk>
you have to chose windows or linux base with you set it up and I forgot far more details that should have existed
<Scopeuk>
it mostly hands off to hyper v infrastructure like everything else from what I remember
<buZz>
but hyperv requires a full kernel
L29Ah has joined #openscad
<teepee>
yes, I suppose it's docker inside separate VMs
<guso78k>
got docker installed at least now ...
<guso78k>
in fedora there is a separate repo to source it from
<teepee>
so run: docker run -it debian:bookworm-slim
<teepee>
on that prompt a "cat /etc/os-release" should give you debian bookworm
<guso78k>
yep, i eager to try,
<guso78k>
need to await my next free timeslot
<teepee>
I'm out for a bit too, we can get stuff running later
lostapathy has joined #openscad
mmu_man has joined #openscad
<guso78k>
yeahh, bookworm arrived in my computer,
<guso78k>
but the docker images are not persistent. files which i create are lost on next start
bozo16 has quit [Quit: Leaving]
bozo16 has joined #openscad
mmu_man has quit [Ping timeout: 258 seconds]
bozo16 has quit [Quit: Leaving]
bozo16 has joined #openscad
guso78k has quit [Ping timeout: 245 seconds]
mmu_man has joined #openscad
L29Ah has left #openscad [#openscad]
<ndnihil>
lol
<ndnihil>
I just found a fun bug
<ndnihil>
locks openscad right the fuck up
<ndnihil>
hrm, seems to only do it if $fn is high
<ndnihil>
WARNING: Scaling a 3D object with 0 - removing object
<ndnihil>
that's the only warning
<InPhase>
ndnihil: I bet you have undef values propagating.
bozo16 has quit [Quit: Leaving]
<ndnihil>
with $fn at default it's still laggy af
<InPhase>
ndnihil: undef will slow everything down wildly. So if you make a mistake that makes a lot of undef, you'll end up with some REALLY slow code.
<ndnihil>
might be the case
bozo16 has joined #openscad
J23k has quit [Quit: Client closed]
<InPhase>
I like to address this with a lot of echo statements added earlier, sometimes commenting out the later bits and working from the top of the file down, and try to track where values go bad.
J23k has joined #openscad
<ndnihil>
eh, I'm like 2 lines away from being done, and it's just for a shower drain insert
<ndnihil>
it can be slow
<InPhase>
Well if it actually completes then you probably don't have an undef problem.
<InPhase>
undef propagation will usually give you no render or mangled garbage geometry.
<InPhase>
I had a drain where the screw hole rusted out, so I designed one that could be pulled out easily by that little raised bit in the middle, and cleaned out.
<ndnihil>
that's what I'm doing, separate basket will be beneath that
<ndnihil>
it's a weird shower pan
<InPhase>
Well that's weird looking, but presumably you have a weird looking shower if that fits it. lol
<ndnihil>
it nestles in there perfectly
J23k has quit [Quit: Client closed]
J23k has joined #openscad
<ndnihil>
last version had some random hexagon cutouts, but it broke because I did a test print with fuck all for infill
<ndnihil>
and stepping on it for several months made it crack
<ndnihil>
so I radiused the throat on this one and did nested hex cutouts
<ndnihil>
because hexagons
<ndnihil>
and going to do it out of ABS, which will flex a bit more without being so brittle
<ndnihil>
your basket does look a bit like the lampshades I did for the kitchen peninsula thing though
<ndnihil>
cylinder with circular cutouts
<ndnihil>
and I lined it with paper as a diffuser
<InPhase>
I used PETG for my drain for the flex but still easy printing.
<ndnihil>
haven't even done pet on the voron yet
<ndnihil>
pretty sure I have some
<InPhase>
Also for a diffuser problem, I found this T-Glase filament which is amazing. I probably spent almost as much on it as a new light would have cost, but I used it for a bunch of other things later. Photo of the diffusion in image 3 here: https://www.thingiverse.com/thing:3024518
<othx>
InPhase linked to "Ceiling Light Cover / Ceiling Lamp Globe by rcolyer" on thingiverse => 6 IRC mentions
<ndnihil>
yeah pett/t-glase
<ndnihil>
that's the pet I have in red/blue/green
<ndnihil>
ooh, and apparently I ordered black, grey, starry night, and khaki petg when I did that last atomic order
<InPhase>
I was actually disappointed with how cloudy it kept printing for me. But it worked very well for that specific purpose. My prints of the T-Glase clear have never looked like their demo prints.
<InPhase>
Oh. I see they've upgraded their Amazon listing to read, "With proper optical coatings, the translucent T-Glase colors can appear as clear as a glass part."
<ndnihil>
yeah the t-glase I used for the cookie cutter for the SIL was kinda cloudy
<InPhase>
No idea what "proper optical coatings" they mean. Probably they are smearing on some index of refraction matching goo.
<InPhase>
Ah! Glossy shellac as a varnish should do it. That's almost the exact right index of refraction.
<InPhase>
I think I would try shellac before that 2 part epoxy they are recommending.
<InPhase>
Lower fumes and less of a pain. You can also get it in a spray if you want.
bozo1 has joined #openscad
bozo16 has quit [Remote host closed the connection]
TheAssassin has quit [Ping timeout: 252 seconds]
bozo1 has quit [Max SendQ exceeded]
TheAssassin has joined #openscad
bozo1 has joined #openscad
<ndnihil>
lol, I'm a dumbass
<ndnihil>
forgot to set the infill higher
<ndnihil>
about halfway through the print
<ndnihil>
I'll do it over in another 6mos
mmu_man has quit [Ping timeout: 246 seconds]
bozo1 has quit [Remote host closed the connection]
bozo1 has joined #openscad
bozo1 is now known as bozo16
califax_ has joined #openscad
califax has quit [Ping timeout: 252 seconds]
califax_ is now known as califax
L29Ah has joined #openscad
<gbruno>
[github] thehans closed issue #4749 (MacOS build failure: github/openscad/build/objects/lexer.cxx:955:2: error: ISO C++17 does not allow 'register' storage class specifier) https://github.com/openscad/openscad/issues/4749
<guso78>
Inphase i also Used to use Petg Filters in my washbasin sink but i Had to replace IT all 6 months
<teepee>
guso78: yep, if you restart the container you'll get a clean image again. for persistent data you can mount volumes, easiest current folder via "-v $(pwd):/data" this makes the current directory show up as /data inside the container
guso78k has joined #openscad
<guso78k>
teepee. this is actually very easy. i will try the linuxdeploy thing again first ...
<teepee>
for permanent stuff you can generate your own images with the compiler and other build things installed already
<teepee>
that's those Dockerfile descriptions which are pretty easy to get started too