-
-
Notifications
You must be signed in to change notification settings - Fork 2.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Cura 5.0 AppImage won't launch (summary of worakrounds are in the end of the first message) #11856
Comments
Ah, Cura 4.13.1 is working fine. |
Could you try if LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libstdc++.so.6 ./Ultimaker-Cura-5.0.0-beta+1-linux.AppImage Fixes it? We already had one report that this would resolve it. It's not a proper fix, but it might work as a workaround |
I just tried the LD_PRELOAD and received this error: |
This is an issue with library incompatibilities between the graphics drivers and libstdc++ that comes with the AppImage. FreeCAD AppImage (any AppImage with applications requiring shared library loading between the App and your X11/Wayland environment) has the same issue. As Ubuntu 22.04 was released today only, it's only partially surprising that it is not fully supported IMO. The LD_PRELOAD workaround above worked for me on pop!_os 22.04 which is based on Ubuntu 22.04 (and also still in beta). You may need to find the right path to the libstdc++ your graphics drivers use to make it work. This issue is inherent with AppImage. According to the AppImage best practice you should package the standard libs that cover the oldest OS release you want to support. That this breaks shared library loading for people that use current graphics hardware and driver libraries seems of little import to them. From my point of view, Ultimaker has two options, neither of them compelling:
(and to cut to the chase, neither flatpack nor snap are free from issues either, so switching the packager is not a solution IMO). |
@nallath the following process does not solve for me.
|
@oliof cool, these are the libs I could find. None of them are working with Cura 5.0, however, Cura 4.13 does work without any special setup. Maybe something else is missing in this AppImage. And yes, I saw the same issue with the Appimage of FreeCad
|
Ah, it is related with the graphics card. Alright. I will have to check how it works because I am using NVIDIA drivers. |
Weird that my lib has a reference to this ➜ Applications strings /usr/lib/x86_64-linux-gnu/libstdc++.so.6 | grep GLIBCXX
GLIBCXX_3.4
GLIBCXX_3.4.1
GLIBCXX_3.4.2
GLIBCXX_3.4.3
GLIBCXX_3.4.4
GLIBCXX_3.4.5
GLIBCXX_3.4.6
GLIBCXX_3.4.7
GLIBCXX_3.4.8
GLIBCXX_3.4.9
GLIBCXX_3.4.10
GLIBCXX_3.4.11
GLIBCXX_3.4.12
GLIBCXX_3.4.13
GLIBCXX_3.4.14
GLIBCXX_3.4.15
GLIBCXX_3.4.16
GLIBCXX_3.4.17
GLIBCXX_3.4.18
GLIBCXX_3.4.19
GLIBCXX_3.4.20
GLIBCXX_3.4.21
GLIBCXX_3.4.22
GLIBCXX_3.4.23
GLIBCXX_3.4.24
GLIBCXX_3.4.25
GLIBCXX_3.4.26
GLIBCXX_3.4.27
GLIBCXX_3.4.28
GLIBCXX_3.4.29
GLIBCXX_3.4.30
GLIBCXX_DEBUG_MESSAGE_LENGTH But it looks like when I run Cura, it uses a lib from inside the AppImage. ➜ Applications LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libstdc++.so.6 ./Ultimaker-Cura-5.0.0-beta+1-linux_c714ec65293e8f948ace6fd69440b1d4.AppImage
PyInstaller/loader/pyimod03_importers.py:495: DeprecationWarning: the imp module is deprecated in favour of importlib and slated for removal in Python 3.12; see the module's documentation for alternative uses
libGL error: MESA-LOADER: failed to open iris: /tmp/.mount_Ultimap8B3jp/libstdc++.so.6: version `GLIBCXX_3.4.30' not found (required by /lib/x86_64-linux-gnu/libLLVM-13.so.1) (search paths /usr/lib/x86_64-linux-gnu/dri:\$${ORIGIN}/dri:/usr/lib/dri, suffix _dri)
libGL error: failed to load driver: iris
libGL error: MESA-LOADER: failed to open iris: /tmp/.mount_Ultimap8B3jp/libstdc++.so.6: version `GLIBCXX_3.4.30' not found (required by /lib/x86_64-linux-gnu/libLLVM-13.so.1) (search paths /usr/lib/x86_64-linux-gnu/dri:\$${ORIGIN}/dri:/usr/lib/dri, suffix _dri)
libGL error: failed to load driver: iris
libGL error: MESA-LOADER: failed to open swrast: /tmp/.mount_Ultimap8B3jp/libstdc++.so.6: version `GLIBCXX_3.4.30' not found (required by /lib/x86_64-linux-gnu/libLLVM-13.so.1) (search paths /usr/lib/x86_64-linux-gnu/dri:\$${ORIGIN}/dri:/usr/lib/dri, suffix _dri)
libGL error: failed to load driver: swrast See, it is using The one I am preloading exists on my system. See: ➜ Applications ls -lsa /usr/lib/x86_64-linux-gnu/libstdc++.so.6
1 lrwxrwxrwx 1 root root 19 Mar 19 07:36 /usr/lib/x86_64-linux-gnu/libstdc++.so.6 -> libstdc++.so.6.0.30
➜ Applications ls -lsa /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.30
1137 -rw-r--r-- 1 root root 2260320 Mar 19 07:36 /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.30 |
Yes, that is how AppImage works, it collects the runtime and standard libraries from the system it is built on. and GLIBCXX_3.4.30 is not covered in the one that's included here because it wasn't part of Ubuntu releases prior to 22.04 which wasn't a stable release until today. No idea why the LD_PRELOAD workaround does not work on your machine, since I have the same layout. Here is the wrapper script I use to reference in my .desktop file (but it also can be used on the command line of course): #!/bin/bash
LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libstdc++.so.6 <path_to>/AppImages/Ultimaker-Cura-5.0.0-beta+1-linux.AppImage I am using an absolute path, but I have confirmed that it also works when I use a relative path. Since your shell prompt looks slightly non-default in your examples, maybe you are not using bash? running it yields the following output, and Cura 5 starts up fine: PyInstaller/loader/pyimod03_importers.py:495: DeprecationWarning: the imp module is deprecated in favour of importlib and slated for removal in Python 3.12; see the module's documentation for alternative uses
Cyclic dependency detected between "file:///tmp/.mount_UltimaLZurP1/share/cura/resources/qml/Actions.qml" and "file:///tmp/.mount_UltimaLZurP1/share/cura/resources/qml/Actions.qml"
QQmlEngine::setContextForObject(): Object already has a QQmlContext |
FWIW I went through building an AppImage using Cura Build Environment on pop!_os 22.04 and as expected that launches without error. I do not have access to older environments to check whether this would run on 20.04 / 18.04 releases. It misses some built-in plugins, but that's more likely due to my lack of experience with the build system than anything else. |
could you share your appimage so I can test it on my system @oliof ? |
I was able to run it. It looks like preloading the system lib with AppImage=./Ultimaker-Cura-5.0.0-linux_07ecc3f54905e865167d2bcd7cfe459c.AppImage
# Extract the package
${AppImage} --appimage-extract
# Remove (rename) libstdc out of the sight...
mv squashfs-root/libstdc++.so.6 squashfs-root/libstdc++.so.6.removed
# Recreate the AppImage file (thanks luzpaz)
./appimagetool-x86_64.AppImage squashfs-root
# Rename generated AppImage to the original name
mv -f Ultimaker_Cura-x86_64.AppImage ${AppImage} Edit: Added a script to fix the AppImage since it is not launching at all on Ubuntu 2022.04. |
you need to use LD_PRELOAD, not LD_LIBRARY. But I've had a discussion with the AppImage Core developer on Twitter and they've pointed nallath towards a solution that will allow dynamic loading of of system libstdc++ if needed.
Gesendet von ProtonMail mobile
…-------- Original-Nachricht --------
Am 25. Apr. 2022, 20:56, Leandro Heck schrieb:
I was able to run it. It looks like preloading the system lib with LD_LIBRARY is doing nothing on my system.
This is what I've done to run it once.
# Extracting the package
./Ultimaker-Cura-5.0.0-beta+1-linux_c714ec65293e8f948ace6fd69440b1d4.AppImage --appimage-extrac
# Going inside the extracted package
cd squashfs-root
# Remove incompatible lib that is not allowing system lib work
mv libstdc++.so.6 libstdc++.so.6.removed.by.me
# Then I could launch it with ./Ultimaker-Cura
—
Reply to this email directly, [view it on GitHub](#11856 (comment)), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AAJMOIYRDJMH3JANWDH5OD3VG3THLANCNFSM5UASMVNQ).
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
Please, looks this comment here: #11856 (comment) |
Thank you all for reporting! |
Extra things I saw in this release
|
the author of the Octoprint plugin has released a prerelease version that provides basic functionality (connectivity and upload of files), but requires more work. This is not surprising as all plugins need updating for Cura 5 due to the library and API changes. The icon works in the task switcher, I don't see it anywhere else but I start Cura through a script so I don't know how that'd work |
Cura does not open .stl files too. I think it is not usable on Linux yet. Actually it does not do anything. It does not launch the file browser to load a file, it does not launch the popup window to login Ultimaker's website. It does nothing. |
it does open STL files fine for me. Maybe its best to open another issue for this topic?
Gesendet von ProtonMail mobile
…-------- Original-Nachricht --------
Am 2. Mai 2022, 19:45, Leandro Heck schrieb:
Cura does not load .stl files too. I think it is not usable on Linux yet.
—
Reply to this email directly, [view it on GitHub](#11856 (comment)), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AAJMOI7SM2PRFHLSRHBUHGLVIAIEVANCNFSM5UASMVNQ).
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
I would love to open another issue @oliof , but there are too many fields and required things. Too much text I have to read.. seriously, this template to report issues, sucks all the energy I have to create the bug reports. I have already started a new issue here at least 4 times this week. But I dropped the idea after some time since I have to fill all those required things. It is crazy that these things are getting horrible with time and are a show stopper for reporters like me. I do these things while I am working. So, it has to be quick and easy. And it can be improved later. This repo could at least have an alternative blank template to report issues, seriously. |
You can use a blank one, but we would rather not have people use that. The amount of poor reports that we get without it is simply staggering. We also have to do this during work, so we want to make sure that we can do this somewhat efficient. With a blank template, we got people reporting "cura crashed" as the full set of information that they gave us (and if we were lucky, we got a very unhelpful screenshot from windows saying something like "Cura stopped responding" I understand that it's a balancing act for both sides, but filling in the template takes less time than it does to write the post you just did ;) |
Personally as an end user I think you're just lacking the courtesy to your fellow community members trying to follow specific issues that are being tracked by word salading everything in here. Contrary to you I made an effort to understand the issue and helped find a fix for future releases, and I'm only interested to read about progress on this issue here.
So besides the fact that you're incredibly rude to the developers on this project, you're also making it hard for me to follow this specific issue by insisting to treat this issue as if you owned it. You don't.
Gesendet von ProtonMail mobile
…-------- Original-Nachricht --------
Am 3. Mai 2022, 16:04, Jaime van Kessel schrieb:
> It is crazy that these things are getting horrible with time and are a show stopper for reporters like me. I do these things while I am working. So, it has to be quick and easy. And it can be improved later. This repo could at least have an alternative blank template to report issues, seriously.
You can use a blank one, but we would rather not have people use that. The amount of poor reports that we get without it is simply staggering. We also have to do this during work, so we want to make sure that we can do this somewhat efficient. I understand that it's a balancing act for both sides, but filling in the template takes less time than it does to write the post you just did ;)
—
Reply to this email directly, [view it on GitHub](#11856 (comment)), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AAJMOI72SAUNFPWYP62LKJTVIEW65ANCNFSM5UASMVNQ).
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
@nallath Well, this is your opinion. Not a fact. But that is fine.
@oliof I don't think you understand the point here. By the way, are you saying that you don't want to have users' feedback on issues and features? Because this is what I am getting here. Seriously. If you are saying that, those 1.6k issues that help to improve the platform are useless, right? Is this what you are saying? What is the problem with having a way to submit a blank issue? This gives flexibility to users like me to want to contribute but don't have time to spend filling out useless forms. I am not even asking you to fix these issues. I am just returning what I am seeing on the screen, so you can have your product improved. I am not asking or demanding anything. I am using my free time, to help you guys, to be aware of issues there your software is going to find when released. You should not treat users like that. Seriously. But moving one here... why don't we have a blank template for reporting issues instead of jailing all users, who want to contribute to the project, forcing them to fill a poor form? Gesendet vom Gmail Desktop |
@leoheck I am not saying what you claim and your bad faith argument is not buying you any points. As @nallath said, you can submit a blank template issue, but the developers ask you to use the template. You could consider being corteous, or ignoring the request and not use the template, or you can continue poisoning this issue. Either path reflects mostly on you. Also, if you think you earn any points by insulting me although I am just a user of the product -- like you, only with some empathy for the developers -- I think you are mistaken. Seriously (-: PS: To make it clear: I have no affiliation to Ultimaker or Cura other than being an end user. And I would like to follow this issue to see whether the library dependency issue on distributions derived from Ubuntu 22.04 will be resolved in a future beta release or not. I would like to follow other issues in, well, other issues. |
FYI: I'm on wayland on ubuntu 22.04 and I can start cura 5.0 beta AppImage with
but it fails when I configure for wayland:
|
On Crostini (Chromebook) running Debian based Linux, .
|
Devs see CURA-9516 for the ticket to make it one AppImage again |
I am having this exact same issue. |
Running Ubuntu 22.04, got this working. Not sure if this is going to break anything else.
Now, the AppImage should run without any additinal changes. I am not sure if moving the deprecated keys back to main causes any issues or breaks anything. anyone else if finds issue here please let me know. |
Just download the flatpak version instead, works perfect on all distros |
AppImage is also working too, at least for Ubuntu 22.04 (use the one that has Modern distro or something like this in its name) |
Hey. Issue return with 5.2.2 |
@dmitrygribenchuk really? there is 2 versions.. one has the word |
@leoheck oh, my fault. ~/Downloads$ ./UltiMaker-Cura-5.2.2-linux.AppImage ~/Downloads$ ./UltiMaker-Cura-5.2.2-linux-modern.AppImage |
@dmitrygribenchuk awesome. Cheers! By the way, this "modern" could be explained better on the Releases page. |
Does this solution also offer a workaround for us sad users stuck with GLIBC 2.27 for longer (I am not using a 18.04 derivate but a still supported distro with old GLIBC). Cura 5.4 appimage that works with GLIBC 2.27 systems? Would be nice. Why do we have AppImage if it can't fix this little hiccup? |
For me :
LD_PRELOAD, or remaking the image with the bash code below does not work.
The issue was the same here : But this one has been marked as duplicate, so im here and nothing seems to work.... Anyone to help me ? Thanks in advance. |
Hi @JordyProvost what is your Operating System and its version? |
Hi @leoheck Thanks for your answer. I'm on Debian 12.1 bookworm, but i'm using Mesa 23.1.4 from git. Regards, |
If this |
Omg, sorry, this was so obvious...its working flawlessly, thank you very much ! |
Awesome, have a good day dude. |
Guess they've just decided to stop releasing the -modern versions of the appimage. This should not be closed, it's still an issue. |
they stopped releasing the -modern AppImage because they found a way to solve the issue using a single AppImage.
Gesendet von Proton Mail für Android
…-------- Ursprüngliche Nachricht --------
Am 16.06.24 13:17 um Hatsune Cthulhu schrieb :
Guess they've just decided to stop releasing the -modern versions of the appimage. This should not be closed, it's still an issue.
—
Reply to this email directly, [view it on GitHub](#11856 (comment)), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AAJMOI2FOT5GXI4CYXKQFK3ZHVX5BAVCNFSM6AAAAABJMQLTKKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNZRGQ2DIMRRGE).
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
Clearly they didn't fix it, otherwise I wouldn't be here asking about the -modern builds. I just updated and now I'm having to use environment variables to get it to work in Wayland. |
Very correct, same issue in Arch with KDE for a while now. |
The problem appeared again on Ubuntu 24.10:
Cura is 5.8.1 |
Why is this closed?? The latest appImage is still broken |
You can find a workaround here: |
Cura 5.0 won't launch, see the console log.
Currently using
./Ultimaker-Cura-5.0.0-linux.AppImage
Possible workarounds
# Reported on KDE with Wayland and Intel graphics. QT_QPA_PLATFORM=xcb LD_PRELOAD=/usr/lib64/libstdc++.so.6 ./Ultimaker-Cura-5.0.0-linux.AppImage
The text was updated successfully, but these errors were encountered: