Krita 5.3.0 alpha appimage crashes on startup

I reported a bug a couple of weeks ago, which still hasn’t been triaged, so I thought I’d ask here about a problem I’m having. Krita 5.3.0 alpha crashes on startup on all my AMD machines (25 of them, plus a couple of laptops). While for some strange reason, it works fine on my few intel machines (Xeon and i9-14900-based machines). I just get a core dump after the ‘Loading Main Window…’ message. Funny thing is that Nightly Builds of 5.3.0 from a few months ago seemed to work. I don’t know what changed.

I get messages in journalctl I get a range of errors that look like:

Module /tmp/.mount_krita-icceiH/usr/lib/libxcb-shm.so.0 without build-id.

In fact, there’s a lot of the ‘… without build-id’ errors. That’s an AppImage thing, I should think. These errors do not exist when running Krita 5.2.14 or the nightly builds of the stable branch.

Any ideas?

Hi @hpetrus

I moved your comment out of the support category (that is for versions released as stable) and moved it into Develop: Artists Feedback & Testing.

Thanks for testing 5.3.0. (I’m using it too, but on Android.)

Thanks, @sooz . I wouldn’t have known. I’ll remember that next time.

Yeah, we’re doing a lot of testing on 5.3.0, but can’t run it on almost 30 of our workstations (Fedora and Arch) for unknown reasons. If I knew more about AppImages, I’d have a clue, but I don’t. Again, funny how 5.2.15 (nightly builds) work just fine on all systems, though.

1 Like

There was a recent problem topic for Next on Debian 13. This may be related:

I’m happy to report that, as of today, krita-5.3.0-prealpha-1d2c802e14-x86_64.AppImage starts up as expected on all 28 of our Linux production systems. The bug report has been unceremoniously removed from bugs.kde.org, so I don’t know what was done to correct it, but YAY! It’ fixed!

Thanks, Krita dev team! Now, we can test 5.3.0 features effectively.

Oh, it wasn’t removed, I just marked it as a duplicate because we already had a bug report for it that came in earlier that turned out to refer to the same thing: https://bugs.kde.org/show_bug.cgi?id=513241

Not sure if we really did anything to fix it though, it kinda seems to have disappeared as mysteriously as it appeared. I’ll ask the others if they maybe know what’s at fault.

This topic was automatically closed 4 days after the last reply. New replies are no longer allowed.