I don't... when setting up a new system I just copy what I need from any random machine I have logged into at the time. As I need different config for different systems it doesn't really make sense to have one perfect config on all either..
Besides. What is there to really mange. There are only a few that one are likely to change. Every thing else is in /etc. Besides all of thia is in whole system backups and snapshots anyway.
Gnu stow and a git repository.
stow creates bulk symlinks according to the folder structure.
This is what I use and it's been a godsend in my home lab.
Nix, but I’d only recommend it if you share my same brand of mental illness
This is the way.
chezmoi does everything I need. It's really nice; would recommend.
Chezmoi with auto-push is the way
GNU Stow, backed up to a git repo.
For those who might not see the purpose, it’s for keeping your profiles on different machines in sync, especially if you use the terminal a lot.
Same here.
I used to have everything in a git repo, but nowadays I use Nix with Home Manager, and I don't want to look back.
Same! I even manage three different configs with the same flake!
How many dot files are you changing to need some sort of manager?
For me it’s more about keeping multiple systems in sync and working as intended. E.g. I have my laptop, a home server with couple different distros and a few cloud VMs. Whenever I change a keybind in neovim or tmux or whatever, I just save the changes in a slightly different way and now every machine has those changes so I don’t need to update each machine manually.
ln exists.
I keep most of them in a git repository that I check out at ~/.config. Some software doesn't follow the XDG Base Directory Specification out of the box but can be configured to do so; the Arch Linux wiki has an article about it. For software that cannot be configured to follow the standard I have dedicated repositories, like for OpenSSH.
I'm using stow, and then git for versioning. The only question I'm currently facing is whether to keep my stow packages as individual got repos (so I can switch branches for radically different configs or new setups) or treat the whole lot as a big repo, and set the others up as subtrees.
IMO you shouldn't use branches for switching between configs. It makes it hard to compose and reuse your configuration.
Just imagine you have a new config file that you now want to add to all of your machines. You'd now have to commit/cherry-pick on all of your branches. Change that file? Now you have to distribute the changes again and might even get conflicts.
Additionally, many programs allow you to change your configuration directory via environment variables (usually XDG_CONFIG_HOME , I think).
So you can even keep all machine-specifig configuration files in a single repo under different prefixes for those.
I use home-manager to manage most of my dotfiles and store them in a git repository. Additionally, I just symlink the configurations I'm too lazy to write as home-manager options (yet), which in turn is a decent way to make migrating to home-manager options less upfront.
I use yadm which is basically git repo with automated path handling and alternate files support. I heard that gnu stow is also a good choice, but I haven't tried that
Ok I’m going to look into this one for sure.
I use a bare git repo. After the initial set-up its just the basic git commands but invoked with a gitdf alias. I wrote a (non-monetized) blog post here about it.
If still needing a tui you could write a simple helper script to call the commands.
I use Syncthing with a folder called "Configuration Files" where I drop all my sync-able configs and dotfiles and I simply symlink them as needed.
So there's at least the two of us doing it this way. Syncthing is an absolutely fantastic tool.
Yeah, it's fantastic. It's more effective at this than I thought because even though I use git a lot for programming and projects, I've admittedly forgotten far too often to commit and push/pull trivial changes on devices and in most cases, all I need is just for dotfiles to sync across devices anyway.
I am syncing a whole lot more than just dotfiles and have done so for years. Admittedly, I am running an always-on-instance (multiple, in fact…). Syncthing hasn't put a foot wrong in over a decade.
I am donating a small amount to the project every month and would like to encourage others to do the same.
I just wrote a bash script copies the relevant files or directories I want to back up from either ~ or ~/.config and places them in a local git directory which I then push to a private repo on Codeberg. Super janky and manual but I update dotfiles so seldom that it works for what I need.
I'm using VCSH, which allows me to split the home directory into multiple git repos without worrying about symlink consistency (like with GNU stow) or apps that have problems with symlinks.
The syntax is exactly the same as git, only that I do (for example) vcsh emacs add -p instead of git add -p.
I keep them in a Fossil repository (with a few private log-in data stored in a SyncThing folder instead) and I just ln them where I need them.
Chezmoi
It’s the only one that just works and has enough features for me, but not too many.
I use YADM which is a thin wrapper around a bare git repo but still has some creature comforts like per-machine configs and templating.
Since you still need to interact with Git, I pair it with Lazygit. Love that software, I do everything Git with it now!
Unfortunately it is a little jank due to the way Lazygit handles bare repos, thankfully there's a command that sets up the needed Git variables for it to work correctly: yadm enter lazygit
I send the repo to my own Forgejo instance. Kinda overkill but I was already self-hosting other services so I thought "Fully private Git server just for me? Why not?"
Now, this is only for managing user level files. For managing system level configs I haven't setup anything yet.
YADM has a solution for this but it seems a little jank, maybe I'll learn Ansible instead, dunno.
I do a git repo for my dot files with an installer that configures it based on whether I’m using Linux, macOS, or FreeBSD; a server or desktop; and whether I’m in bash or zsh. It also includes a bunch of functions and aliases that I find useful. It’s not always pretty because I also use it as a practical place to try new shell script bits when I have time. I’m hoping to change some things around soon thanks to some ideas from Dave Eddy’s bash course at ysap.sh.
I don't necessarily sync my dotfiles across machines, at least not in a blanket fashion. But I use Seafile to keep all sorts of directories in sync across machines. It's basically a self-hostable Dropbox.
I never did until I started using hyprland.
Now I have those in a git repo: https://git.jeena.net/jeena/hypr-dotfiles
I push them to Git, secrets are masked with Mozilla SOPS.
Nothing at all. If things go south on the install I'll have a clean slate. Really the only thing I have backed up are keys, everything else is nonessential.
Glad I'm not the only one.
I created a file tree that looks similar to my system's file tree, except it only contains all the files that I modified or added and only their respective directories. From there I just use rsync to sync those files/file tree to the system's /.
It's convenient to see what changes I currently have but it requires a bit of manual maintenance. I only really started doing it that way because I was learning how to use rsync and I just kept going on with it because it was working for me.
I'm only working with my laptop, android phone and two Raspberry Pi's so I can get with my little rsync based setup.
I do manage them via git. But I only do it so have settings (and their changes) synchonised between 2 PCs and a laptop.
With just one main device I don't even see a reason to "manage" anything... a basic backup strategy completely independent of just dotfiles aside.
Neovim and scp. I should clearly test some things.
Don't think too hard on it. Just use git. For example, I have a repo called handy-scripts that hosts all my dotfiles. I just check that out into ${HOME}/projects/handy-scipts and then symlink everything from where it's expected to its corresponding place in the repo.
As you make modifications, remember to occasionally do a git pull --rebase && git commit -m WIP && git push so that all your devices are synced up.
For ZSH, I have a script that automatically installs oh-my-zsh and a bunch of zsh plugins, and changes my default shell to ZSH.
For everything else, some files will just get copied over, others left behind. I only "manage" them when I need to view/edit them.
I use Chezmoi and a git repo for keeping mine in sync with each other across several machines.
Linux
From Wikipedia, the free encyclopedia
Linux is a family of open source Unix-like operating systems based on the Linux kernel, an operating system kernel first released on September 17, 1991 by Linus Torvalds. Linux is typically packaged in a Linux distribution (or distro for short).
Distributions include the Linux kernel and supporting system software and libraries, many of which are provided by the GNU Project. Many Linux distributions use the word "Linux" in their name, but the Free Software Foundation uses the name GNU/Linux to emphasize the importance of GNU software, causing some controversy.
Rules
- Posts must be relevant to operating systems running the Linux kernel. GNU/Linux or otherwise.
- No misinformation
- No NSFW content
- No hate speech, bigotry, etc
Related Communities
Community icon by Alpár-Etele Méder, licensed under CC BY 3.0