teepee changed the topic of #openscad to: OpenSCAD - The Programmers Solid 3D CAD Modeller | This channel is logged! | Website: http://www.openscad.org/ | FAQ: https://goo.gl/pcT7y3 | Request features / report bugs: https://goo.gl/lj0JRI | Tutorial: https://bit.ly/37P6z0B | Books: https://bit.ly/3xlLcQq | FOSDEM 2020: https://bit.ly/35xZGy6 | Logs: https://bit.ly/32MfbH5
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mtm has quit [Ping timeout: 260 seconds]
mtm has joined #openscad
<church_> may i propose to add button to start dialog 'Backups'? (should open appropriate dir, preferably sorted by modification time?) Of course, not that often needed (and due that i didn't remember location (or rather last time i needed was on macos, now - on windows)). It was not that hard to find docs on that path by googling, Help>Library info:Backup Path .. but would be nice to have button in open dialog :)
<teepee> yeah, we could have "Open Location..." instead of the show library folder
mmu_man has quit [Ping timeout: 260 seconds]
<church_> Also imho "open recent" button is redundant. why not - "if user has clicked on any file in "Recent list", then [Open] would open it, instead of calling open file dialog"
teepee_ has joined #openscad
teepee has quit [Ping timeout: 264 seconds]
teepee_ is now known as teepee
<gbruno> [github] kintel closed pull request #5644 (also reset camera translation in viewAngle* to make one axis disappear) https://github.com/openscad/openscad/pull/5644
<teepee> church_: and the examples?
J25k19 has joined #openscad
kintel has joined #openscad
<kintel> Looks like one of the Python PRs broke non-experimental builds: https://github.com/openscad/openscad/pull/5691
<kintel> Probably just a missing experimental ifdef, but didn't look into it yet
<teepee> I see what the problem is
<kintel> Perhaps we should promote one of the non-experimental builds to a pre-merge check
J25k65 has quit [Ping timeout: 240 seconds]
<teepee> yeah, that would be usefull, still waiting for MS to complain about our build usage :)
<teepee> but then I guess it's nothing compared to AI stuff
<kintel> yeah, and just a build without running tests is probably a good start
<teepee> argh, why is it building Python-3.8.2?
<teepee> omg
<teepee> aha! PYTHON_SOURCE=https://www.python.org/ftp/python/3.10.12/Python-3.10.12.tgz
<stealth__> python3.8 has reach end of life.
<teepee> argh, cannot access 'python?.?': No such file or directory
<teepee> well, yes, the 2nd is 10
<teepee> :(
<teepee> that plugin only works for single digits
<teepee> yeah, looks like the plugin was abandoned 3 years ago
TheCoffeMaker has joined #openscad
lockywolf_ is now known as lockywolf
snaked has joined #openscad
<church_> teepee: i browsed those just once to check out what's that, but usually just google how to do specific stuff / or checking cheatsheet or guide about specific functions. But that's just me, i cannot speak for all. Suggested only that imho quicker finding backups might be useful, and that mostly i use Open Recent, but not sure it needs "separate" button, when one button can serve two roles
<church_> Even more so that [Open Recent] is greyed out prior any of files in recent list are selected. And double click in list also works. So - open w/o selecting - file open dialog, selecting, then [open] - same as ex. [open recent], no?
guso78k has quit [Ping timeout: 240 seconds]
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Ckat has quit [Ping timeout: 248 seconds]
Ckat has joined #openscad
J25k19 has quit [Quit: Client closed]
J25k has joined #openscad
stealth__ has quit [Quit: Leaving]
misterfish has joined #openscad
<gbruno> [github] damienmarchal opened issue #5695 (RubberBand problem when switching windows using the "Next window" menu entry, the window stays "rubberband selected". If I use Cmd-k, it works as expected. This is on macOS.) https://github.com/openscad/openscad/issues/5695
<gbruno> [github] damienmarchal edited issue #5695 (RubberBand problem when switching windows using the "Next window" menu entry.) https://github.com/openscad/openscad/issues/5695
J25k has quit [Quit: Client closed]
J25k has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 264 seconds]
teepee_ is now known as teepee
guso78k has joined #openscad
joso has quit [Quit: joso]
joso has joined #openscad
J25k has quit [Quit: Client closed]
J25k has joined #openscad
<guso78> Teepee its relevant to Use Same Version which IS Used to Compile and Link Openscad
<guso78> Use exactly one defined Version all over Openscad. I believe this reduces incompatibility Error sources.
mmu_man has joined #openscad
<buZz> ooo, i like the idea to have the auto-backups easily visible from a 'open recent failures to save' menu or something :)
mtm has quit [Ping timeout: 244 seconds]
<guso78k> me too. its a little bit combersome to restore  backups. was almost writing a script ;)
<buZz> heh well, i mostly just usually forget where the f they are
<buZz> or that they exist at all :P
mtm has joined #openscad
<guso78k> +1
<InPhase> I forget about them because it's a non-feature when running in external editor mode. The backups are never generated, as nothing ever previews or renders unsaved.
guso78k has quit [Quit: Client closed]
guso78k has joined #openscad
snaked has quit [Remote host closed the connection]
Guest48 has joined #openscad
<Guest48> اه
<Guest48> hi
<Guest48> Is there any standard machinery component library ?
<Guest48> bye
Guest48 has quit [Client Quit]
hyvoid has quit [Ping timeout: 246 seconds]
hyvoid has joined #openscad
guso78k has quit [Quit: Client closed]
misterfish has quit [Ping timeout: 276 seconds]
bertronika has joined #openscad
Jerr has joined #openscad
stealth_ has joined #openscad
joso has quit [Ping timeout: 248 seconds]
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
misterfish has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
voxpelli has quit [Read error: Connection reset by peer]
voxpelli has joined #openscad
dingodoppelt has quit [Remote host closed the connection]
dingodoppelt has joined #openscad
ubuntourist has joined #openscad
ubuntourist has quit [Remote host closed the connection]
J25k has quit [Quit: Client closed]
J25k has joined #openscad
<gbruno> [github] amatulic opened issue #5696 (Colored console text output) https://github.com/openscad/openscad/issues/5696
<gbruno> [github] amatulic edited issue #5696 (Colored console text output) https://github.com/openscad/openscad/issues/5696
<teepee> guso78: yep, that's what I'm trying, but now I'm getting a build that's just crashing :(
<stealth_> Me and InPhase where just talking about Zig being a good alternative to OpenScad C language to fix build issues, lol
<teepee> standards?
<othx> standards is https://xkcd.com/927/
<stealth_> with zig its just own standard, itself.
<teepee> so we just rewrite everything in a different language?
<stealth_> no, you can just include your C header and use it in Zig, and it even comes with a C, ... compilers builtin.
<teepee> and that helps with crashes in Qt how exactly?
<othx> stealth_ linked to YouTube video "Zig in 100 Seconds" => 1 IRC mentions
<teepee> other than introducing even more chances of breaking things
<stealth_> teepee, i am talking about future improvement not fix issues atm.
<teepee> c++ has all that except the integrated tests
<teepee> the problem is not simple code, the problem is dependencies
<stealth_> you have to at least watch more videos or read about zig, its a totally different.
<stealth_> dependencies like?
<teepee> cairo for PDF
<teepee> lib3mf for read/write 3MF files
<teepee> polyclipping2 for 2D geometry
<teepee> Manifold for 3D geometry
<teepee> freetype/fontconfig/harfbuzz for font handling
<teepee> ...
<teepee> ...
<teepee> ...
<stealth_> clearly all those features wont exist in zig since its new kid on the block, but we are talking future here though.
<stealth_> sorry if i upset you teepee
<teepee> it's 10 years old?
<teepee> I'm not upset
<teepee> I just question the wisdom of chasing the new kid that will mean reinventing the wheel without benefit
<stealth_> there is no official 1.0 release yet, so its still in development but the results so far are super good. since it was written from ground up. like no middle man.
<teepee> like webbrowsers invent operating systems
<stealth_> this will be a good change for the future vs C and its unending problems.
<teepee> same goes for swift, dart, go
<teepee> rust
<stealth_> ya, but i wouldn't have recommend if it wasn't that good.
<teepee> I'm not going to defend C, I agree that should be avoided
<teepee> all those things I heard about go 5 years ago
<teepee> now they have all those features they claimed to not need or implement because they make the language complicated
<teepee> I'm not saying it's bad, I'm very likely going to watch that as I am interested in a general way
<stealth_> i never used go and will use it. so can't talk about it.
<teepee> but chasing the next thing is not how real work is done
<teepee> for that, the cppcon fireside chats are pretty interesting
<stealth_> thanks, that's all i was trying to say, just give it a look and see what up, maybe keep it in the back of the mind for years if/when openscad plans a move. you want to actually
<stealth_> know a bit about the language thats out there.
<teepee> well, if browsers will be your operating system, it's going to be typescript ;-)
<teepee> after all they just invented threads
<teepee> they just need to make it as stable as real operating systems
<stealth_> well browser are good for many things but i don't know about using that as your operating system.
<stealth_> those seems like a hype to see something.
<teepee> wasm + webgpu + webusb is a clear path into that direction
<stealth_> yes/no, its mainly to find compatibility between operating system, since big tech kept rewriting their software for every OS/browser, it was tedious and wasted $$$
<stealth_> You see software like Atom use browser based engine and now I think they moved to using Rust. since the speed difference is huge.
<stealth_> you can get really good speed with wasm but who in their right mind is going to code in wasm? you need language like zig, rust, ...
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openscad
<InPhase> "<stealth_> Me and InPhase where just talking about Zig being a good alternative to OpenScad C language to fix build issues, lol" Hey now... InPhase was saying, "No, I don't think so." on that. :-P
<stealth_> InPhase, :p we still "talked"
<InPhase> Not exactly a ringing endorsement, however.
<stealth_> it takes time to learn a language too, so all that factors needs to be considered.
<InPhase> Or more specifically, I was not talking about it being a good alternative.
<gbruno> [github] gsohler opened pull request #5697 (separate Python Target) https://github.com/openscad/openscad/pull/5697
<gbruno> [github] amatulic edited issue #5696 (Colored console text output) https://github.com/openscad/openscad/issues/5696
<teepee> I don't know what that tries to be
<teepee> a game engine, a gpu api, a gui framework?
<gbruno> [github] amatulic edited issue #5696 (Colored console text output) https://github.com/openscad/openscad/issues/5696
<stealth_> looks like all of it? and such things can be done in zig... why its good.
<teepee> I'm not seeing much, the videos can be done on shadertoy
<gbruno> [github] amatulic edited issue #5696 (Colored console text output) https://github.com/openscad/openscad/issues/5696
<teepee> it's a huge project and maybe it will produce what is claimed, but that seems many many years away
<stealth_> ya its in early development.
<teepee> unless they find someone paying for multiple full time developer
<teepee> I would appreciate a good gui framework, the current options seem not ideal
<stealth_> if openscad has $ to support that project that might help them as well in turn openscad would get a new GUI framework and 3d engine.
<teepee> we do have donations to support running the servers and services
<gbruno> [github] amatulic edited issue #5696 (Colored console text output) https://github.com/openscad/openscad/issues/5696
<stealth_> teepee, https://github.com/sponsors/emidoots#sponsors you can join the club of supports.
<teepee> if we would look at sponsoring other projects, that would likely go to stuff currently used
<teepee> my personal sponsoring goes to other places
<gbruno> [github] amatulic edited issue #5696 (Colored console text output) https://github.com/openscad/openscad/issues/5696
<gbruno> [github] nomike opened issue #5698 (Shebang missing in check-dependencies.sh) https://github.com/openscad/openscad/issues/5698
<gbruno> [github] amatulic edited issue #5696 (Colored console text output) https://github.com/openscad/openscad/issues/5696
misterfish has quit [Ping timeout: 248 seconds]
<stealth_> As i said zig is very good, cross-compilation right from the language itself with bare metal speed. What you do after is your choice!
<gbruno> [github] nomike opened pull request #5699 (Fix check dependencies.sh) https://github.com/openscad/openscad/pull/5699
nomike has joined #openscad
<nomike> Hi!
<othx> Welcome nomike!
<nomike> I just submitted #5698 which is about a bug in the check-dependencies.sh script and PR #5699 which fixes it.
<teepee> hi! yep, CI already aproved, thanks
<teepee> Looks good to me, I can't judge the macos cases though
<nomike> I have a mac lying around somewhere, but it's a late 2009 macbook. I'm afraid that won't do it ;-).
<teepee> nope, that's not going to work, I have a similar old thing here and that got broken by apple/qt long ago
lastrodamo has joined #openscad
<nomike> I spotted something else though, which isn't a bug, but perhaps something worth thinking about:
<nomike> I'm on version 2024.05.24 at the moment (currently compiling the latest one) and in the preferences->features, I could enable the "manifold" feature. When I exit out of openscad, and open it again, the feature is still enabled. Rendering a design uses the manifold library. As would be expected.
<nomike> However, when I now run `openscad -o test.stl test.scad` it doesn't use the manifold feature.
<nomike> I know, I could add `--feature manifold` to the command line, but the point is, that depending on whether I use the GUI and hit "F6" or I just use the CLI interface, OpenSCAD behaves differently. And I'm not sure if this is a good thing or not.
<nomike> What do you think?
<teepee> I think it makes for a more stable behavior on command line but it's also not easy to change as the gui options depend on Qt
Ekho has quit [Quit: CORE ERROR, SYSTEM HALTED.]
<nomike> The thing is, if you use the "--feauter manifold" option on the CLI, it's temporary, for this run only, while the GUI setting is permanent. There is no temporary way to set this in the GUI. So I guess the current behavior is good.
<nomike> I wonder though what happens if you specify `--feature manifold` and use the GUI?
<teepee> I think it's ignored
guso78k has joined #openscad
guso78k has quit [Quit: Client closed]
lastrodamo has quit [Quit: Leaving]
Ekho has joined #openscad
guso78k has joined #openscad
<guso78k> teepee, did you succeed with the AppImage ? Did die AppImage work in your computer ? How can you check in your place, if the appimage would work in somebody else's computer ?
<teepee> still getting crashes :(
<guso78k> real segffaults ? do they happen once you evaluate(f5) python code ?
<guso78k> maybe start the appimage and use gdb to hook onto the running process ?
<teepee> nope, already when loading the app in QtCore
<teepee> it's not even near running the python stuff I think
bertronika has quit [Quit: Konversation terminated!]
<guso78k> but the appimage did definitiely work prior the python pr's .  My personal strategy for addressing such unexpected /weird issues is binary search. go back half way to see if app still crashes. maybe just create appimage with ENABLE_PYTHON=0 first ?
SamantazFox has quit [Quit: Bye]
<teepee> just rebuilding with different appimage plugins changes things already
<teepee> the python plugin seems to break things even when just installed, so I'm probably messing up something there
<guso78k> i doubt that you actually need the python plugin. i think its sufficient to offer the python stdlib (.py files) somew
<guso78k> somewhere, where python_path points to in the appimage
<teepee> probably, it mainly does the python compilation which would be nice if it would work :)
<guso78k> you could compare directory trees before and after the python plugin to learn, whats actually changed
<teepee> funny, I was thinking the difficult case will be Snap and Flatpak and AppImage the easy one
<guso78k> how can *I* test your flatpak image ?
<guso78k> no experience with FlatPak yety
<teepee> beta version
<teepee> well, if you have fedora, you should have flatpak :)
<guso78k> will try to use it
<teepee> it's probably still on cube only commit
<teepee> let me update that. unfortunately it's not fully automaitc
<guso78k> np,  i am interested only on flatpak. my fully featured enginge is at home :)
<ali1234> error: The application org.openscad.OpenSCAD/x86_64/beta requires the runtime org.kde.Platform/x86_64/5.15-23.08 which was not found
<teepee> ¯\_(ツ)_/¯ it should install dependencies
<teepee> 23.08 sounds too old though
<ali1234> i don't know about flatpak
<ali1234> i can paste instructions though
<teepee> hmm, no runtime-version: 5.15-23.08
SamantazFox has joined #openscad
<teepee> aha Required runtime for org.openscad.OpenSCAD/x86_64/beta (runtime/org.kde.Platform/x86_64/5.15-23.08) found in remote flathub
<teepee> I guess that means you still need the "normal" repo too
<ali1234> it's not looking for dependencies in other remotes unless they are already installed?
<ali1234> i do have flathub
<teepee> flatpak remote-add --if-not-exists flathub https://dl.flathub.org/repo/flathub.flatpakrepo
<teepee> it asked me if I want to install that ^
<teepee> Installation complete.
<ali1234> installing regular openscad asks if i want to install the kde platform
<ali1234> installing the beta just says it can't find it
<ali1234> so i guess it doesn't know how to find dependencies across remotes
<ali1234> hmm nope even after installing stable it still won't install beta
<teepee> and why does it work for me then?
<teepee> I'm not doing anything different as far as I can tell
<guso78k> error: The application org.openscad.OpenSCAD/x86_64/beta requires the runtime org.kde.Platform/x86_64/5.15-23.08 which was not found
<guso78k> same issue here
<teepee> did you add the repo with --user?
<ali1234> i pasted the commands from the github
<teepee> maybe:
<teepee> flatpak remote-add --user --if-not-exists flathub https://dl.flathub.org/repo/flathub.flatpakrepo
<ali1234> yes that fixed it
<guso78k> yes!
<teepee> right the user and system/root setup is strictly separate
<teepee> and I have an AppImage running, but that's with the python init error
<ali1234> okay it's running... now what?
<guso78k> its running, but there is no preferences to turn on python ?
<ali1234> i don't know. where would i find them?
<ali1234> i can see --trust-python on the cli options
<teepee> guso78: did you install beta?
<teepee> what is it showing in status bar?
<guso78k> i think so. it it was not beta, openscad would not install .py as file extension
<guso78k> burt preferences is missing in Edit menu
<guso78k> suppose i have to hack qt registry :)
<teepee> huh?
<guso78k> would not *allow* py file extension
<teepee> no preferences? how did that happen?
<ali1234> "python-engine"
<ali1234> preferences works for me
<ali1234> okay this whole thing works for me
<guso78k> hmm, need to try harder
<guso78k> yes, now is see my cube !
<guso78k> so flat
<guso78k> fo flatpak also works outside of your computer
<ali1234> as you would expect, it is completely isolated from the host system
<ali1234> appimages are not
<teepee> yep, flatpak and snap are container
<teepee> well, not sure how --classic snap works but we do strict confinement
<ali1234> even classic snaps should never be loading .so from the host
<ali1234> a fundamental difference between snaps and flatpak is the internal container structure
<teepee> ok, the crash might be an old patchelf from the python plugin
<ali1234> snap apps link against something like "/snap/lib" where as flatpaks link against "/usr/bin"
<ali1234> so you can mount the host /usr inside a snap container, and that is what classic mode does
<ali1234> you can't do this with flatpak because it would shadow the container /usr, which is why flatpak doesn't have a real classic mode
<ali1234> but mounting /usr inside the snap container doesn't affect the snap app, because it doesn't look for its libs there anyway
<ali1234> hmm i mean to write: flatpaks link against "/usr/lib"
<ali1234> you can give permission to a flatpak to access any other folder than host /usr, and you can even have host /usr be mounted in some weird place like /var/host/usr or something, but these are not enough to eg make your C/C++ IDE work properly
<ali1234> and this is why jetbrains IDEs have official snaps but no official flatpaks
<teepee> yeah, I have codium as snap and that does not work well with c++ debugging
<teepee> I might go back to NetBeans :)
<ali1234> yeah? my experience is the exact opposite. IDEs installed with snap work flalessly as long as they are classic
<ali1234> when installed with flatpak, everything breaks
<teepee> jetbrains yes
<teepee> actually NetBeans is a snap too, hmm...
<teepee> latest upstream commit as test flatpak
<teepee> hmpf: NameError: name 'openscad' is not defined
<teepee> oops, my fault :)
<guso78k> hmm, openscad module should be made available internallly
<teepee> I was messing with the imports
<teepee> and left an openscad.show() in there
<teepee> it works
<guso78k> (y)
<guso78k> hope you have more than a cube. i can only vary the dimensions ...
<ali1234> i only see cube and show on dir(openscad)
<guso78k> BTW: shall we change the behavior such as multiple show's in the code behave like lazy unions ?
<teepee> lazy union is bad :)
<guso78k> right now only last show() wins
<ali1234> i think it should update the preview everytime you call show() so you can do animations :)
<ali1234> but currently UI freezes until python finishes
<guso78k> yes, exactly. i am using customizer and static python variables to perform the same
<ali1234> but i think it would be more useful if that didn't happen. you could have interactive scripts that pop up a message box eg "does the preview look okay? y/n"
<othx> guso78k linked to YouTube video "PythonSCAD games" => 3 IRC mentions
<guso78k> yeah that would be cool. but it would require very good synchronization across threads '=D
<ali1234> look at how python is integrated in blender
<ali1234> it can do all this kind of stuff, it can take control of the UI and add buttons to it
<ali1234> that's the ideal... but it will be a lot of work of course
<guso78k> pythonscad can also add its own custom menus ....
<ali1234> from python code?
<guso78k> (its just very poorly documetned)
<guso78k> yep
<ali1234> cool
<guso78k> this is a feature dedicated to Will Adams . i feel his gcode preview should start from the Menu, not from F5 :')
<ali1234> i would like a simpler export workflow that remembers the last set of options you used and just does it again without asking
<ali1234> instead of render -> export -> choose filename
<ali1234> you just click "export again"
<ali1234> kind of like the one in gimp
<guso78k> you can script your export  in pythonscad , so EVERYTHING is defined and repeatable. of course it would never merge ...
<ali1234> of course
<ali1234> this is also really nice
<guso78k> and for 3mf you can export multiple geometries into the same file
<guso78k> ali1234, where are you from ?
<ali1234> the UK
<guso78k> i am located at lake of constance
<guso78k> germany
<teepee> oh, promising
<guso78k> ??
<teepee> python plugin installed (but not yet run) and no crash
<teepee> now building with python plugin activated
<guso78k> (y)
<guso78k> likely you will fork the python plugin as its not maintained anymore :')
<teepee> yeah, it's broken for versions > 3.9 not expecting 2 digits