OrcaSlicer Crashes on Startup Arch Linux #6471
Comments
|
Same issue, bambu-studio crashes too |
|
same....
|
|
Same experience here on Garuda Linux. Related bambustudio issue: bambulab/BambuStudio#4634 |
|
Same here, as said above it should be some package update in Arch because before updating it worked flawless. Still working Bambu Studio for me though |
|
Downgrading freetype2 to |
|
same here, Arch Linux. Downgrading freetype2 also solves the issue. |
|
Can also confirm that downgrading freetype2 fixes this issue! |
|
How can such an update cause an entire application to segfault? I don't know if freetype follows semver, but if they do, this little patch update shouldn't break anything. The question is, who is responsible for the breakage in the first place? Is it freetype, arch or orca? I can confirm downgrading freetype2 fixes the issue. |
|
Trying to assign blame like that isn’t helpful. Sometimes things break and it’s no one’s fault, esp since arch isn’t even officially supported. the freetype project dates back to 1996, so they may follow some type of semver but it’s not a perfect system, and you don’t always know when a change is breaking. freetype2 is from the main arch repos, and it’s gonna be tested against other stuff in the main repos, but it can’t and shouldn’t be tested against every aur package. |
|
Looking at the gdb output I see that the crash is caused by a return value from Cairo and/or Pango while getting information on a font with a 'COL' v1 table (most likely an Emoji font). You should thus first contact the Cairo people asking for help – of course it is also possible that this is a problem of OrcaSlicer (I have no idea). To do that efficiently you should build the crashing program without any optimizations to get better stack traces – in particular, the topmost frame (#0) doesn't give any useful information, which means it is not clear at all where the program crashes. |
Running Arch and other bleeding distros is nice because it gives us the latest software, but it's called bleeding edge because we get cut sometimes. It's nice though because we find these issues and can get this information to the developers that build our software, so they know to fix it. We are the canaries of the software coal mines. It's not glamourous, it's frustrating, but somebody has to do it. Regards, |
I understand and I'm not trying to assign a blame to anyone. For lack of better words my comment turned out to appear negative. I apologize for that. I'm just trying to understand which component is causing the segfault – is it freetype2 with a bug or breaking change or is it Orca for using something in freetype2 in a way that didn't cause issues before but now does, or is it an Arch specific issue. It's really just a question. |
I agree with that. It gives us the opportunity to fix things before they arise at users more downstream or with non-rolling distros. |
|
Don't despair, improve the backtrace. |
So let's wait for that 🙂 |
|
While imperfect, this should be a slightly improved backtrace. |
|
Thanks, but the most crucial information is still missing, namely FreeType's file and line number details for frame #0. If possible, please compile FreeType manually without optimization and temporarily replace the OS's version. |
|
|
Yes, it opens when I downgraded freetype2 to |
|
This might be a multithreading issue. FT_Face can be safely used within a single thread, but I see IA__FcFontSetSort in a separate thread. |
|
I'm seeing this problem in the appimage same as the OP and others... but I do not get the segfault when I compiled it using an Arch Linux PKGBUILD - specifically, I took https://aur.archlinux.org/packages/orca-slicer, updated the tagged version to 2.1.1, commented out some of the patches that were no longer relevant, and than a (it could still be a threading issue of course - by compiling it that way I might have just resulted in some changed linking order that works around whatever race condition is the problem. But at least it got me a working orca slicer on Arch...) |
|
same issue for me. downgrading freetype2 to 2.13.2-2 worked to open the app, and backup my configs, but I have no view in 3D area. also the backup is impossible cos I can't expand the windows to select my settings to export |
|
Same issue, after some testing, I can start OrcaSlicer successfully with the following command on ArchLinux |
Same issue for me. Package orca-slicer-bin-2.1.1-1 work fine with freetype2-2.13.3-1. |
|
Edit: Managed to make OrcaSlicer run properly by upgrading to the beta (2.2.0-beta2), downgrading to freetype 2.13.2-2, and using only |
The crashes can be worked around by editing the connection settings and changing the Device UI to something it can't reach: It was crashing for me otherwise with the same symptoms as you described. |
|
Looks like the latest beta has fixed this issue, perhaps simply through a build system update. Used |
|
For me installing Webkit2gtk, gst-plugins-base, and gst-libav from AUR fix the crash. The last one is only needed |
|
Can you try flatpak from #3949 (comment) ? |
|
This issue was resolved for me across two different computers (Arch and EndevourOS) with orcaslicer 2.2.0 and freetype2 2.13.3-1. Previously, I experienced the same crash on both computers With orcaslicer 2.1.1, and downgrading to freetype2 2.13.2-2 resolved the issue on both. |
Is there an existing issue for this problem?
OrcaSlicer Version
2.1.1 (AppImg)
Operating System (OS)
Linux
OS Version
Arch Linux
Additional system information
Printer
N/A
How to reproduce
I try to run orca-slicer from either my desktop, or the terminal
/opt/orca-slicer/bin/orca-slicer.Actual results
Both techniques of starting the app are terminated abnormally with signal 11/SEGV aka a segfault.
Expected results
Orcaslicer should start normally.
Project file & Debug log uploads
Debug Backtrace (missing orca-slicer symbols, but has library symbols, I can't get compiled version to build on my machine).
Journalctl Logs:
Checklist of files to include
Anything else?
I have reinstalled OrcaSlicer, and deleted my entire configuration. No printer configuration or project files are involved with the crash. There are no log files that I can find relating to OrcaSlicer, I think it might crashing before the log writing process.
It was working yesterday, however I did a system update today, so I assume it is some library compatibility issue. All signs point to a font library issue.
I have uploaded a full coredump as well here:
orca-slicer-coredump.gz
You can also find all libraries linked to OrcaSlicer here:
The text was updated successfully, but these errors were encountered: