rofi-lbonn-wayland coredumps on scaling factor higher than 1 #1964
Unanswered
siddharthkumarjha
asked this question in
(Possible) Issues
Replies: 2 comments 2 replies
-
|
Okay this is extremely mindbogglingly strange shortly after posting this I tried to use my system wide installation of rofi-lbonn-wayland-git again and it started working again.... I recompiled the rofi I had downloaded from github removing the workaround I had placed for stopping the scaling and tested it.... That too seems to work fine now... Really what is happening with this software. On a side note I apologize for my top post it seems I had jumped the gun when I said scaling is not supported... |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
@siddharthkumarjha thanks for gathering the details. It could be a problem with fractional scaling... Could you open an issue on the fork's bug tracker? https://github.com/lbonn/rofi/issues If you can, try with a scaling of 2 and see if it also happens. |
Beta Was this translation helpful? Give feedback.
2 replies
-
|
|
Beta Was this translation helpful? Give feedback.
-
|
There is one more thing. With the scale of 2 set in my hyprland config my rofi install goes out of screen it's too big to be contained. But that doesn't seem to happen when I have my scale set to 1.2 in hyprland config and as can be seen from above logs according to rofi it should be rounded off to 2 by wayland. Not sure I understand the why behind it or if it's even connected but I have noticed that the size of rofi remains same regardless of scale it's getting from unless I physically change the scale in hyprland config. Though that shouldn't really be happening given this, from |
Beta Was this translation helpful? Give feedback.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hey everyone,
So this is a recurring issue from discussion #1932 where I had marked this solved. However the circumstances this time are a tiny bit different.
In the last discussion my rofi kept stack crashing after a system wide update with error,
and no matter what I did I could not get it to work when suddenly rolling back my system a snapshot and re-updating my system did the magic. The issue was closed at the time and I paid no mind to it. But this time the circumstances were different.
My system was running fine and last I had performed a system updated it was on 26-feb-2024. Rofi was working fine and suddenly it started throwing the above error again. This error came despite the fact that literally nothing had changed in my system. I thought okay maybe an update will fix it but that didn't work out. So I began to debug this and git cloned the 1.7.5+wayland3 tag of the repo.
did a,
followed by,
It compiled and put all the binaries in target/bin of the repo I had cloned and running that I still had the same issue. So, with the help of gdb and running rofi with debug logs on I was able to figure out the issue is resolution.
I noticed the program was crashing in
wayland_rofi_view_updatefunction insource/wayland/view.cat line 357 whencairo_paintis invoked over the scaled version of rofi. To identify the cause I did what programmers have been doing for a millena... adding print logs. (I know ppl can do what I did with print statements using gdb but I am noob here 😓 please cut me some slack 😅 )Below are the logs I obtained with the additional print messages,
As you can see the program crashed for scaled display output. So I went to my hyprland config file and changed,
my scaling here from 1.20 to 1.00 and sure enough rofi works happily with this change. Go back and it starts crashing again.
I went through the code and saw that hidpi is currently marked as fixme and noticed that if I stop rofi from scaling like this,
(function in line 231 source/wayland/display.c)
Just changing
wayland->scale = output->current.scale;in the commented out if statement towayland->scale = 1works as well.Well whatever the issue (could be just wayland or cairo) is just using the default scale of 1 works. I am not sure why this issue took so long to materialize again but oh well software I guess.
Whatever the reason since scaling is currently unsupported why is the above code even in place? Put this in an experimental branch or something and remove it from master branch that everyone is using to build the software. If you want I can raise the git pull for this myself or if you want to fix this from your end please do it.
here is the gdb stack trace just in case
Beta Was this translation helpful? Give feedback.
All reactions