Intel AppImage issues

[Miepee]

For some reason, intel iGPUs segfault on the AppImage. This makes it (nearly) impossible for such users to play the native AM2R Linux version:

Not exactly sure why this happens, as the intel drivers are included there. I don't have an intel machine to try this out on, sadly.

[ludvigng]

There's probably some problem with libgl, mounting the appimage with ./AM2R.AppImage --appimage-mount and then entering the mounted image in /tmp and manually running the app with LIBGL_DRI3_DISABLE=1 ./AppRun makes it start.

[Miepee]

So basically ./AM2R.AppImage --appimage-mount-and-run ? EDIT: Nevermind, just saw about that environment variable. Can you confirm, if the native binary has the segfault problem as well, and is fixable with that environment variable as well? Also, does the AppImage run, if you just run it like this LIBGL_DRI3_DISABLE=1 ./AM2R.AppImage?

[ludvigng]

Can you confirm, if the native binary has the segfault problem as well, and is fixable with that environment variable as well?

It didn't, but given the workarounds I had to do to get it running (libcurl3 and openssl 1.0) it's really unfeasable on ubuntu.

Also, does the AppImage run, if you just run it like this LIBGL_DRI3_DISABLE=1 ./AM2R.AppImage?

Yeah it does. Don't know why I didn't think of that at first.

[Miepee]

Just out of curiosity, can you give me a neofetch for spec information?

[ludvigng]

It's a dual GPU laptop, currently running on the intel card. Running the game with discrete graphics still segfaults though.

[Miepee]

Thank you! I'll try to get some people to test that fix!

[Miepee]

Also @ludvigng you don't happen to know if that's maybe an Ubuntu/Debian derivatives specific problem? So far, I only encountered these problems on these machines. Me and my "test subjects" mostly use Arch, and we never got any problems there.

[ludvigng]

I honestly have no idea, sorry. I can see if I can setup a test environment later, but no promises. :)

[Miepee]

It's understandable if it's not possible, but I still wish you good luck with that! In the meantime, I'll try this test on my "test subjects"

[ludvigng]

Did the simplest test possible, booted a live environment manjaro and just copied the AppImage over, and it worked fine out of the box. So yeah there's something going on with ubuntu it seems.

[Miepee]

Heya, @ludvigng can you try the intel-ubuntu-fix branch, and see if the issue is resolved there?

[ludvigng]

Tried it in a live image, following the new instructions (enable universe repo, installing python3, xdelta3, enabling multilib and installing libc6:i386 libncurses5:i386 libstdc++6:i386 and libopenal1:i386). It now runs!

There's still a problem with audio, but that's probably due to me trying it on an live environment and unrelated to this issue (I fixed it on my old system, but I don't remember how and have since switched distro, I suspect it's just a mismatch due to new lib32 stuff interacting with the old 20.10 live system). :)

[Miepee]

Thanks for testing! I'll PR it then! I assume the sound problem might've been me, I made a commit in master which would hopefully fix that. It's either a conflicting sound library in the appimage, or a missing libopenal1 install.

[ludvigng]

No prob!

I made sure libopenal1:i386 was installed in the live environment, but if I remember correctly how I made it work was to install libpulse0:i386 on my old system. I did also try that in the live env, but it was already installed, so I think it's just the live env being the way it is. Either way people probably won't patch this from a live env, so it should probably run fine on an actual system now.

[Miepee]

Yeah. In addition most forks actually have that library / option pre- enabled/installed, so I don't expect too many issues there.

[Miepee]

Closing due to merged PR 47fa4ea