Closed rainfallmelonflysong closed 3 months ago
Need testing, but it's probably electron issue, maybe related to one of the flags set here that changed behaviours between the version bumps.
I cannot reproduce this on wayland+nvidia proprietary 550.54.14 drivers, would help sending more info on GPU, current driver, distro, window manager, display server, etc.
The relevant output of inxi -Fxxrzc0
should be enough, together with the method you used to install (AppImage or deb) to help isolate the problem.
Meanwhile, from discord discussion, running miru with miru --in-process-gpu
should work, tho I have no clue what implications this flag has on actual usage.
Indeed, it seems to work by using the --in-process-gpu
flag. I'm running arch with x11/i3/nvidia-dkms with no compositor, so not technically supported. I've also tried running the appimage with the same result.
no-sandbox
likely causes this, but since we've got no reproduction this is being labeled as a "linux config issue"
Can replicate this issue on my linux machine, crashes with identical errors as above unless I launch with the --in-process-gpu
flag, here's output from inxi -Fxxrzc0
:
This also occurs on my arch system. Some digging later I found this might be an electron issue or possibly even chromium itself. I ensured I have all of the appropriate libva drivers present on my AMD integrated graphics system with no avail. --in-process-gpu
was the only one that stopped the crashing.
@VictorGamerLOL @CaptainD54321 @rainfallmelonflysong I've released https://github.com/ThaUnknown/miru/releases/tag/v5.0.7 which is a beta, download and install manually, check if issue persists
@ThaUnknown Tried 5.0.7 and the issue sadly persists (same behaviour and same --in-process-gpu workaround), same with 5.0.10
@ThaUnknown works for me. The activities I usually do that induce the crashing, like getting the popup to choose a torrent, no longer do. And this was 5.0.7 not 5.0.10. I am using the Appimage version.
Some more info if it is of any relevance:
pacman -Q | grep libva:
lib32-libva 2.20.0-1
libva 2.21.0-1
libva-mesa-driver 1:24.0.5-1
libva-utils 2.21.0-1
pacman -Q | grep mesa:
lib32-mesa 1:24.0.5-1
libva-mesa-driver 1:24.0.5-1
mesa 1:24.0.5-1
mesa-utils 9.0.0-4
mesa-vdpau 1:24.0.5-1
opencl-clover-mesa 1:24.0.5-1
opencl-rusticl-mesa 1:24.0.5-1
lspci | grep VGA:
04:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Lucienne (rev c1)
Edit: 5.0.10 also does not crash.
@ThaUnknown works for me. The activities I usually do that induce the crashing, like getting the popup to choose a torrent, no longer do. And this was 5.0.7 not 5.0.10. I am using the Appimage version.
Some more info if it is of any relevance:
pacman -Q | grep libva:
lib32-libva 2.20.0-1 libva 2.21.0-1 libva-mesa-driver 1:24.0.5-1 libva-utils 2.21.0-1
pacman -Q | grep mesa:
lib32-mesa 1:24.0.5-1 libva-mesa-driver 1:24.0.5-1 mesa 1:24.0.5-1 mesa-utils 9.0.0-4 mesa-vdpau 1:24.0.5-1 opencl-clover-mesa 1:24.0.5-1 opencl-rusticl-mesa 1:24.0.5-1
lspci | grep VGA:
04:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Lucienne (rev c1)
Edit: 5.0.10 also does not crash.
I retract my words. Aggressively scrolling still induces a GPU process exited unexpectedly: exit_code=133
but not enough for it to be considered unusable and crash the entire app. Still stands.
And I retract them again because it seems like the main cause if this is hovering for too long on a torrent entry and getting a tooltip. The moment the tooltip appears the app turns white and 5 seconds later it closes.
Huh I swore I posted this reply the other day, but apparently github ate it: @ThaUnknown I tried both 5.0.7 and the latest version (5.0.12) and my issue (Miru crashing immediately on launch) persists if I don't use the --in-process-gpu
flag.
Preflight checklist
What app version are you using?
5.0.2
What operating system version are you using?
Linux
Expected Behavior
Miru launches with no issues
Actual Behavior
When attempting to launch miru v5 on linux, the app displays a blank screen for 3 secs and then crashes, giving this (electron?) error
Checking for beta autoupdate feature for deb/rpm distributions Found package-type: deb 19:20:59.399 › Checking for update [4595:0325/192100.022102:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 [4595:0325/192100.505817:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 19:21:00.564 › Update for version 5.0.2 is not available (latest version: 4.5.10, downgrade is disallowed). 19:21:00.564 › checkForUpdatesAndNotify called, downloadPromise is null [4595:0325/192101.030652:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 [4595:0325/192101.658375:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 [4595:0325/192102.281327:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 [4595:0325/192102.887335:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 [4595:0325/192103.176312:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 [4595:0325/192103.466594:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 [4595:0325/192103.763053:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=133 [4595:0325/192103.763068:FATAL:gpu_data_manager_impl_private.cc(448)] GPU process isn't usable. Goodbye. [1] 4595 trace trap (core dumped) miru
Screenshots
No response