Om-welcome dies after latest update (rome)

Support
1

Hello,

  • OpenMandriva Lx version:
    can’t find the version

  • Desktop environment (KDE, LXQT…):
    KDE

  • Description of the issue (screenshots if relevant):
    After installing from the ISO, then running the update with 900+ files, and then rebooting, om-welcome pops up then closes. I tried to run it from the CL and got the following. Feels like something got missed? But then, everyone would have this issue, soooo, I dunno.

[steve@professor ~]$ om-welcome
GBM is not supported with the current configuration. Fallback to Vulkan rendering in Chromium.
The following paths were searched for Qt WebEngine locales:
  /usr/lib64/qt6/translations/qtwebengine_locales
  /home/steve/.OM Welcome
but could not find the translation file for the current locale: en-US.pak
You may override the default search paths by using QTWEBENGINE_LOCALES_PATH environment variable.
Translations WILL NOT be correct.

[0327/142909.766755:WARNING:resource_bundle_qt.cpp(83)] locale_file_path.empty() for locale 
The following paths were searched for Qt WebEngine locales:
  /usr/lib64/qt6/translations/qtwebengine_locales
  /home/steve/.QtWThe following paths were searched for Qt WebEngine locales:
  /usr/lib64/qt6/translations/qtwebengine_locales
  /home/steve/.QtWebEngineProcess
but could not find the translation file for the current locale: en-US.pak
You may override the default search paebEngineProcess
but could not find the translation file for the current locale: en-US.pak
You may override the default search paths by using QTWEBENGINE_LOCALES_PATH environment variable.
Translations WILL NOT be correct.

ths by using QTWEBENGINE_LOCALES_PATH environment variable.
Translations WILL NOT be correct.

[0327/142909.805883:WARNING:resource_bundle_qt.cpp(83)] locale_file_path.empty() for locale 
[0327/142909.805883:WARNING:resource_bundle_qt.cpp(83)] locale_file_path.empty() for locale 
Segmentation fault (core dumped)
  • Relevant informations (hardware involved, software version, logs or output…):
    R7, rtx3060ti, 16GB
1 Like
2

Steve, it looks like the locale did not get set. I could be wrong though. Please post the output of

locale
3 
[steve@professor ~]$ locale
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC=en_US.utf8
LC_TIME=en_US.utf8
LC_COLLATE="en_US.UTF-8"
LC_MONETARY=en_US.utf8
LC_MESSAGES="en_US.UTF-8"
LC_PAPER=en_US.utf8
LC_NAME=en_US.utf8
LC_ADDRESS=en_US.utf8
LC_TELEPHONE=en_US.utf8
LC_MEASUREMENT=en_US.utf8
LC_IDENTIFICATION=en_US.utf8
LC_ALL=
[steve@professor ~]$
1 Like
4

That looks right

5

@bero might need to look at this.

6

Does everything else work? Wayland or X11?

7

I noticed it’s looking for /home/steve/.OM Welcome but the directory I have is .om-welcome. Not sure if this matters, bu the case is different and the dash.

8

Mine is a file named .om-welcome

9

Wayland. I have to auto-login because it sticks loading sddm. I did notice that in one of the many iterations of re-installing that after installing the nvidia drivers, sddm worked (after I disabled auto-login). So there’s that. :wink:

10

oops - I meant file, not directory

11

I just tried, can’t reproduce this… I see neither the warnings about webengine locales nor the crash.

12

empty qtwebengine_locales directory…

[steve@professor ~]$ ls /usr/lib64/qt6/translations/qtwebengine_locales/
[steve@professor ~]$
13

I should also mention, I installed from the ROME Plasma6.wayland AMD CPU (newer than 2017) iso. Formatted my drive, so a spiffy clean install, then immediately updated using the link in om-welcome, which updated 900+ files, then immedially rebooted. Nothing else in between.

14

Does

dnf install qt6-qtwebengine

Fix it for you? (That’s what contains the files in /usr/lib64/qt6/translations/qtwebengine-locales)

15

It did install, as opposed to upgrade. I verified the pak files. But it still crashes with

[steve@professor ~]$ om-welcome
GBM is not supported with the current configuration. Fallback to Vulkan rendering in Chromium.
Segmentation fault (core dumped)
16

Well here’s an unusual twist. I eventually installed the nvidia drivers from whchever of the additional repos after activating them and rebooted. After auto-logging in, the Welcome appeared without issues. I’m thinking something to do with the “Fallback to Vulkan Seg fault…” got fixed with the install of the proprietary drivers.

I also deactivate auto-logon and I finally see the sddm login prompt. Couldn’t see it before the driver install.

An unfortunate side-effect is now my display is super sluggish. When dragging around a window it skips around. It’s like playing a game at 2 FPS. Yikes! Before the install of the drivers everything was super smooth. Guessing a driver conflict, maybe. Not sure - way outside of my tech knowledge.

1 Like
17

I’m using nouveau driver with an nvidia card (only nvidia-firmware installed, none of the other nvidia packages) and om-welcome segfaults for me but for a different reason, I immediately get this on terminal:

[33998:34022:0412/195113.223522:ERROR:angle_platform_impl.cc(44)] ImageEGL.cpp:112 (operator()): eglCreateImage failed with 0x00003009
ERR: ImageEGL.cpp:112 (operator()): eglCreateImage failed with 0x00003009

0x00003009 is EGL error “EGL_BAD_MATCH: Arguments are inconsistent (for example, a valid context requires buffers not supplied by a valid surface).”

i also have an iGPU, maybe I can try disabling it on next boot or something to see if that works around the issue, seems like a bug in om-welcome though.

started happening after I distro-sync’d to the Rome 25.04, which I did via:
$ sudo dnf clean all ; dnf clean all ; dnf repolist
$ sudo dnf distro-sync --allowerasing

1 Like
18

Welcome! We are glad to see you and hope you will decide to make this your home.

It does appear that you did everything correctly. If you reboot and select an older kernel in GRUB, does it still fail?

19

@yorrbu
welcome1

1 Like
20

This looks to be one of those “only on my system” kind of problems that you’d need a fully staffed hardware testing lab to find.

From what I can tell it’s a combination of 6.13.9 and having both an nvidia gpu (using nouveau in this case) and an integrated intel graphics which was set specifically to always on in the BIOS. It’s an old i7 7700, when I set the igpu to “auto” in the BIOS the problem stopped happening even on 6.13.9, I’m guessing that when I power cycled it with the igpu set to auto it detected my nvidia card and disabled the igpu.

Anyway looks like (maybe) it was an issue with om-welcome trying to do something with EGL on the igpu instead of the one that was currently driving the display. Maybe not, but seems plausible.

1 Like