How to handle bloat on GNU/Linux?

[f33dm3bits]

Bloat in Linux us just the unneeded files on the drive. If you talk about having less stuff loaded on the system in memory, then that is relatively easy to manage.

As mentioned earlier you might want to have a look at Alpine Linux then, not many people use it on the desktop but it's design is simple, small and secure. It is used a lot in container images but also runs on embedded devices: routers, servers and nas devices. I just installed it in a vm to check for you how mucsh space a clean install takes without DE, it takes up 154M of disk space and uses 58M memory.

I regret installing programs from the AUR manually because now you can't update them automatically. And I have lots of unmanaged data which is my fault.

I tried to limit the amount PKBUILDS I install from the AUR or else your system will become a mess. If you installed them manually you can install reinstall those same packages with an AUR helper and then have them updated with that. The only thing you would have to figure out is which build packages were installed for the packages you manually built, maybe if you removed all the packages that use specific build packages that they would be seen as orphans and that you can remove them that way. I use paru as my AUR helper and have it configured that if I do install a PKGBUILD from the AUR and it installs buildeps that after having built the package it should also remove the build deps it installed. I hope that helps.

 

[Debian_SuperUser]

I just found out that I have kernel modules for the kernels I don't even have anymore. Pacman doesn't uninstall them. That's bloat.

 

[dos2unix]

For rolling distro's, (in my opinion, even for LTS) it's always a god diea to have 1 or 2 previous kernel versions to fall back to, in case the "latest and greatest" causes problems.

If you have more than 2 or 3 previous versions, I would think there is a configuration somewhere to set this.