Core dump on 0.3
Closed this issue · 23 comments
start-newm -d core dumps from tty after yay -s install
@jbuchermn says something about libseat
It is possible that you are missing a package that has already happened to me
Please post logs, otherwise there's nothing we can do. There have been libseat issues in the past (it's kind of a mess...) so I think we should be able to fix it.
BTW, aur has v0.3
As such it doesn't have v3 you can build v3 by modifying the PKBUILD
by the way the the fuzz library is not available in the repositories it has to be installed with pip
As such it doesn't have v3 you can build v3 by modifying the PKBUILD
No. Check the PKGBUILD.
by the way the the fuzz library is not available in the repositories it has to be installed with pip
That's an important piece of information, can you add that to the doc?
@CRAG666 shouldn't libseat be a dependency in the PKGBUILD?
That's a good point, however on nixOS it's not necessary...
There's an easy way to test this - install libseat, try again, does anything change?
@jbuchermn I'll get to that now
As such it doesn't have v3 you can build v3 by modifying the PKBUILD
No. Check the PKGBUILD.
by the way the the fuzz library is not available in the repositories it has to be installed with pip
That's an important piece of information, can you add that to the doc?
Of course I will do a documentation pr
@jbuchermn that doesn't help ...seatd is the package , right?
LOG NOTICINGS:
-
brightnessctlerrors untill I manually install it , it should be in thePKGBUILDwithseatd, can I submit a PR? -
python says
traceback error : ValueError: invalid literal int() with base 10showing lines 99 and 24 in file helper/backlight_manager.py.Is this a syntax error or just something I did? -
the log has a lot of libseat errors and a final failed assertion from pywm/newm
-
OTHER STUFF:
if I use
dbus-run-session start-newm -d
it kinda works....(possible screen freeze)
following sway suggestions helps , i've noticed
if I use
dbus-run-session start-newm -d
it kinda works....(possible screen freeze)
following sway suggestions helps , i've noticed
@jbuchermn that doesn't help ...
seatdis the package , right? LOG NOTICINGS:
brightnessctlerrors untill I manually install it , it should be in thePKGBUILDwithseatd, can I submit a PR?- python says
traceback error : ValueError: invalid literal int() with base 10showing lines 99 and 24 in file helper/backlight_manager.py.Is this a syntax error or just something I did?
brightnessctl is an optional dependency, only if you use BacklightManager. However these errors should not be fatal, just a lot of log output, but this can't cause the core dump.
Nope, it's nothing you did, probably fixed once brightnessctl is working.
- the log has a lot of libseat errors and a final failed assertion from pywm/newm
- OTHER STUFF:
if I use
dbus-run-session start-newm -d
it kinda works....(possible screen freeze)
following sway suggestions helps , i've noticed
It's most certainly related to libseat / seatd. Possible cause could be seatd not running and libseat refusing to use systemd-logind (which is something I've seen in the past on arch linux). If you would post a log, we could tell. Does sway start up properly?
Probably @CRAG666 knows better, how to fix this, as I'm not on arch anymore.
@jbuchermn it's in a VM , I can't get the log.
anyway ,
- sway doesn't start either with the same result as
dbus-run session start-newm -d, the screen hangs - "Possible cause could be seatd not running and libseat refusing to use systemd-logind (which is something I've seen in the past on arch linux)." - How did you fix it? I'll try using
elogind - you disable
BacklightManagerthrough the config , right?
@jbuchermn it's in a VM , I can't get the log.
anyway ,
- sway doesn't start either with the same result as
dbus-run session start-newm -d, the screen hangs
Okay, probably also a good idea to look for help there as it's a lot bigger of a community.
- "Possible cause could be seatd not running and libseat refusing to use systemd-logind (which is something I've seen in the past on arch linux)." - How did you fix it? I'll try using
elogind
Using seatd.
- you disable
BacklightManagerthrough the config , right?
Yup, just remove it from your config.
@jbuchermn can you send me a example newm config without backlight manager and i'll try it?
Just remove it from your config.
But as I said above, that's not the cause of your coredump. It's most certainly a libseat issue.
@jbuchermn yes , but its easier to disable for a easier to read
log
The problem is easier to fix than it seems, it's just a bit tedious. First make sure you have all the dependencies that newm uses installed, don't rely on pkbuild. Second once compiled and installed in a tty section start newm, it is very likely that you use systemd so set the environment variables correctly. You were able to check my dotfiles, I have a folder called waylando where I have a group of scripts to be able to execute any composer in wayland without problems.
@CRAG666 thanks i'll try this on the VM
Getting segfaults here after 1-2 minutes of the program being open. A bit of a pattern I've noticed is that normally, an animation is played at some point between animation queues. Whenever there's a segfault, there are two animations queued with no animation played between. I uploaded this part of the log here.
Mesa: 21.3.7-2 on Intel Tiger Lake (Xe) Graphics
Getting segfaults here after 1-2 minutes of the program being open. A bit of a pattern I've noticed is that normally, an animation is played at some point between animation queues. Whenever there's a segfault, there are two animations queued with no animation played between. I uploaded this part of the log here.
Mesa: 21.3.7-2 on Intel Tiger Lake (Xe) Graphics
Thanks for the logs. This however appears to be unrelated to the particular issue here, but a pywm bug triggered by closing windows. I'll open an issue there and take a look.
Closing this for now.