SUMMARY This is from Dolphin, but presumably the issue is deeper in KDE. I can run a downloaded AppImage from Dolphin - when I double click, it pops up a dialog asking if I want to execute it. However, if the AppImage is electron and needs to mount the chromium sandbox(??), nothing will visible happen after clicking "execute". This is suprising and annoying. KDE should detect this situation, and give a choice (presented in a more explanatory way): a) Sudo to fix the perms on the sandbox b) Run without sandbox STEPS TO REPRODUCE 1. Download Quiqr AppImage (https://quiqr.org/) 2. Attempt to run it from Dolphin OBSERVED RESULT Nothing apparently happens, as described. Per Claude, the two alternatives are: 1) Run the AppImage with the --no-sandbox flag: ./your-appimage.AppImage --no-sandbox 2) Set the correct permissions as root: sudo chown root:root /tmp/.mount_quiqr_huIfd2/chrome-sandbox sudo chmod 4755 /tmp/.mount_quiqr_huIfd2/chrome-sandbox ./your-appimage.AppImage EXPECTED RESULT Containers (hopefully including AppImages) are part of the future. KDE should help remove the friction of using them. SOFTWARE/OS VERSIONS Operating System: Kubuntu 24.10 KDE Plasma Version: 6.1.5 KDE Frameworks Version: 6.6.0 Qt Version: 6.6.2 Kernel Version: 6.11.0-13-generic (64-bit) Graphics Platform: Wayland Processors: 16 × AMD Ryzen 7 5700G with Radeon Graphics Memory: 30.6 GiB of RAM Graphics Processor: AMD Radeon Graphics Manufacturer: ASUS
How should we detect that certain executable needs a random (from the system pov) flag set? The appimage author should make sure that their appimage can run correctly or the appimage itself
With respect, I don't think this addresses the issue. The issue is: running an AppImage can fail silently. That's unquivocally a bug. With regard to the flag being "random" - not sure I agree either, though this is more a question of how comprehensively KDE wants to support AppImages. a) AppImages can run in a sandbox or without a sandbox - that's a fundamental architecture - not a random flag. b) Electron is not exactly uncommon. And AppImages that were written in Electron all have this issue. It's detectable by KDE simply by looking at the stdout when you attempt to run it after the user clicks "Execute". It would be feasible for KDE to (as I suggested) pop up a second prompt explaining what happened to the user and offering one of two options. If it's not a goal for KDE to be able to run AppImages, then I'd suggest removing the current dialog, since it's half assed. Minimal fix here would be to at least show the output when it fails to run. I think a better fix is as I described.
(In reply to Reuben from comment #2) > With regard to the flag being "random" - not sure I agree either, though > this is more a question of how comprehensively KDE wants to support > AppImages. > > a) AppImages can run in a sandbox or without a sandbox - that's a > fundamental architecture - not a random flag. Just a correction but your LLM is wrong, --no-sandbox is from the application - AppImages do not have anything to do with that.
Oh. My bad! It appears to be an electron thing rather than AppImage thing then. https://www.reddit.com/r/debian/comments/hkyeft/unable_to_run_appimage_applications_without/ etc
Indeed. Which means the app itself is either mis-written or mis-packaged. If it needs that flag to run successfully, it's the app's responsibility to do so. Really this is no different from any other misbehaving app. If an app successfully launches but then doesn't show a window or quits immediately, it's the app's fault, and there isn't really anything we can do about it. Showing STDOUT or STDERR won't work because 1) there's no guarantee anything will be printed there and 2) there's no guarantee anything printed there would be related to the issue. Lots of app (including many of our own, unfortunately) print a lot of junk there that doesn't actually indicate a problem. So please report this to the app's developer or AppImage packager. Thanks!
> If it needs that flag to run successfully, it's the app's responsibility to do so. Again, it's every electron AppImage. Not just this single one. This is not a random or niche thing. They even exit with a specific non zero error code. But yet another half finished KDE thing I guess. Fine.
Then it's every Electron app's responsibility, or Electron's responsibility. Basically, if the app is written or packaged so badly that it quits immediately after launching, that's the thing that needs to be fixed. Anything we could do on our side would be an unreliable and potentially misleading band-aid.
So whoever is packaging these electron apps using AppImage needs to do some basic testing to make sure the app at least runs when launched using the GUI. That's literally the bare minimum.