The Longest Journey

mitu

I'm not sure why there's no trace, did I miss a step?

Maybe there's no crash and residualvm just exits. What's the content of /dev/shm/runcommand.log ?

GeekOB

This is the content of the runcommand.log file @mitu ...

Parameters: 
Executing: bash "/home/pi/RetroPie/roms/residualvm/+Start ResidualVM.sh"
WARNING: Invalid joystick: 0!
Virtual keyboard pack 'vkeybd_default' loaded successfully
WARNING: No hardware input were defined, using defaults!
User picked target 'tlj-win' (gameid 'stark')...
  Looking for a plugin supporting this gameid... Stark Engine
  Starting 'Stark Game'
INFO: OpenGL Vendor: Broadcom
INFO: OpenGL Renderer: VideoCore IV HW
INFO: OpenGL Version: OpenGL ES 2.0
INFO: OpenGL Red bits: 8
INFO: OpenGL Green bits: 8
INFO: OpenGL Blue bits: 8
INFO: OpenGL Alpha bits: 8
INFO: OpenGL Z buffer depth bits: 24
INFO: OpenGL Double Buffer: 1
INFO: OpenGL Stencil buffer bits: 8
INFO: GLSL version: OpenGL ES GLSL ES 1.00
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
WARNING: Stem darkening is not available with this version of FreeType!
ALSA lib pcm.c:8306:(snd_pcm_recover) underrun occurred
ALSA lib pcm.c:8306:(snd_pcm_recover) underrun occurred
WARNING: Stark::XARC Member: "map inv.xmg" has unknown=2 with unknown meaning!
WARNING: Stark::XARC Member: "map cursor.xmg" has unknown=2 with unknown meaning!
WARNING: Stark::XARC Member: "april_crouch_pickup.ani" has unknown=2 with unknown meaning!
WARNING: Stark::XARC Member: "april_run.ani" has unknown=2 with unknown meaning!
WARNING: Stark::XARC Member: "april_swimming_idle.ani" has unknown=2 with unknown meaning!
WARNING: Stark::XARC Member: "april.xmg" has unknown=3 with unknown meaning!
WARNING: Stark::XARC Member: "arkadianmale_sitting.ani" has unknown=2 with unknown meaning!
WARNING: Stark::XARC Member: "arkadiafemale_sitting.ani" has unknown=2 with unknown meaning!
No uniform named 'boneRotation'!
No uniform named 'boneRotation'!

mitu

@GeekOB There doesn't seem to be a crash logged in there, maybe that's why gdb doesn't show anything. You can try starting residualvm directly from the debugger, outside of Emulationstation, it may be easier to capture the crash. Just run

gdb 
(gdb) file /opt/retropie/emulators/residualvm/bin/residualvm
(gdb) run --fullscreen --joystick=0 --extrapath="/opt/retropie/emulators/residualvm/extra"

then reproduce the crash and try to get a backtrace.

GeekOB

thanks @mitu , I had to run the gdb without the extra flag as it was giving me a different error, this is the output I received when I "crashed" the game:

GeekOB

The ResidualVM team has recommended the following changes to the script to build from source:

first, change

--disable-debug

to

--enable-debug

in the

build_residualvm

function, recompile ResidualVM. Then, modify the +Start ResidualVM.sh script to call

ulimit -c unlimited

so that it produces a core dump on crash.

Do you have anything to add, @mitu ?
I tried the instructions above, the game quits as expected, where would I find the core dump? I checked /var/lib/systemd/coredump but it's not there.

Thanks

mitu

@GeekOB A core dump is created if the game crashes. From the gdb output you posted (just like the previous time) I see the game does not crash, just exits:

[...]
Inferior 1 (process 817) exited with code 01
[...]

So there's no crash, hence no backtrace or core dump.

GeekOB

Thank you very much @mitu
A bit more from the ResidualVM dev team:

This can be caused by too many uniform slots used in the shader.
The Raspberry PI GPU has a max of GL_MAX_VERTEX_UNIFORM_VECTORS = 136.
The vertex shader for actors blows through that budget with just the bone uniforms (140 vec4 uniforms).

As a quick test, can you try editing the stark_actor.vertex file and lowering the number to 30 in the line

const int maxBones = 70;

If that works, it will very probably glitch some actors pretty badly but the game should play.
You can try ramping that number back up, but the real solution is to rework the bone uniform passing, I'm afraid.

Can you help me find where this file would be located in RetroPie?

mitu

@GeekOB said in The Longest Journey:

Can you help me find where this file would be located in RetroPie?

Maybe in /opt/retropie/emulators/residualvm/share/residualvm/shaders/stark_actor.vertex ?

GeekOB

I tried these changes @mitu but it still doesn't work.
Now I'm just wondering if the Raspberry Pi 3b, which is the model I'm using for this testing, is actually capable of handling this port properly. Escape from Monkey Island also has some issues when I try to run it at full screen.
I'm not sure if you are going to test ResidualVM with the Pi 4 model.
I'm happy to continue trying to test but, as my knowledge with compiling from source is limited I might as well be shooting in the dark here.

What do you guys think?

mitu

@GeekOB said in The Longest Journey:

I'm not sure if you are going to test ResidualVM with the Pi 4 model.

ResidualVM works on the PI4 model, alas I don't have the game in my GOG library and didn't have the time to test this game in particular.
Is the normal game (i.e. non-HD remake) working ?

GeekOB

No @mitu, it is not. If I disable the HD mod, the main menu:

will display properly but the game will not run and exit back to the ResidualVM launcher.
If the HD mod is enabled the main menu screen is all black with only the mouse pointer, ©1999 Funcom, and the names of the options showing up when hovered over. Same behavior when trying to run the game.

mitu

I tried TLJ on the PI4, but it exits right when initialising the game. I'll try to get more info and maybe submit an issue to the residualVM project directly.

GeekOB

Thanks @mitu, that's probably the exact same issue I'm experiencing.
Let me know if there's anything I can do to help test/troubleshoot. I'll keep the game in my SD card just in case.

mitu

@GeekOB said in The Longest Journey:

Thanks @mitu, that's probably the exact same issue I'm experiencing.

Not quite the same issue, but it's quite similar in that's still a shader error.
While scouring for information on the project's Github repository, I stumbled upon this new addition which might help to run the game on a RPi3. You might want to keep an eye on it and - when the changes are merged - to update residualvm from source.

GeekOB

I'll definitely do that and let you know how it goes.
Thanks!

GeekOB

Hi @mitu,
I gave this a try again today as I saw the 0.4.0git version had some new commits, but now not even the main menu comes up, the game goes back to ES right after the runcommand dialog.
Can you see if the same is happening on your Pi4?

mitu

@GeekOB said in The Longest Journey:

Can you see if the same is happening on your Pi4?

I'll give it a shot, but the GLES improvement patch is still not yet merged.

GeekOB

@mitu I just wanted to put this final post here to close the loop.
I was able to test this game on a Raspberry Pi 4 using testing version of RetroPie, works beautifully!

mitu

@GeekOB That's good news ! Thanks for the follow-up.