dot is a simple json based dotfile manager

(it can also be used as a very basic system bootstrapper)

Key features:

• Install packages on your system
• Create symlinks for all your dot files
• Refresh your configs from your dot.json (see Limitations)

How to

First setup

1. Fork this repository or use it as template

2. Edit dot.json. My personal example can be found here

3. Run bootstrap.sh it will walk you through the process and you can rerun it all the time

4. Commit your changes and push them to your personal git repository

Integrate changes from the main dot framework

1. Run git remote add dot https://github.com/cschlosser/dot && git fetch --all

2.1 If you forked the repository run git merge dot/master

2.2 If you used the repository template run git merge dot/master --allow-unrelated

3. Resolve merge conflicts if there are any

Updating your dot.json

If you want to add new packages or add new configs to existing packages you can do so easily by editing dot.json and rerunning bootstrap.sh. The shell scripts generated are aware of changes to some extent.

dot.json syntax

Everything starts with a json object containing a version object and a packages array.

The version has to match the version of dot you're using.

{
  "version": 1,
  "packages": [
  ]
}

Now we can start adding packages

{
  "version": 1,
  "packages": [
+   {
+     "name": "neovim"
+   }
  ]
}

This will install the neovim package from your system package manager specified in providers.json. If you're missing the package manager for your platform feel free to create a Pull Request for adding it to the file for all users of dot.

If your package is only available on one platform and you intend to use this repository for different ones you can specify them with the os object.

{
  "version": 1,
  "packages": [
+    {
+      "name": "herbstluft",
+      "os": "Linux"
+    }
  ]
}

The value of the os object should contain the string from Python's platform.system().

If the package is named differently for a package manager you can provide an alternative name

{
  "version": 1,
  "packages": [ 
    {
      "name": "herbstluft",
-      "os": "Linux" 
+      "os": "Linux",
+      "providers": [
+        {
+          "apt-get": "herbstluftwm"
+        }
+      ]
    }
  ]
}

If you have to run some command prior to installing your package you can put it into the pre_install array where each object should have a key with the id of your package manager and the value is the command you want to run.

{
  "version": 1,
  "packages": [
+    {
+      "name": "fish",
+      "pre_install": [
+        {
+          "apt-get": "apt-add-repository ppa:fish-shell/release-3"
+        }
+      ]
+    }
  ]
}

Now for the config files:

{
  "version": 1,
  "packages": [
    {
-      "name": "neovim"
+      "name": "neovim",
+      "config": [
+        {     
+          "init.vim": "$HOME/.config/nvim/init.vim"
+        }
+      ]
    }
  ]
}

The config files have to be stored in your repository in a folder with the same values as the the name key of the object in your packages array.

.
├── LICENSE.txt
├── bootstrap.sh
├── dot.json
├── gen_configure.py
├── gen_install.py
├── neovim
│   └── init.vim
├── providers.json
└── third_party
    └── jinja
...

Each object in the config array is going to be linked from the value to the key with a relative path.

Let's say you store your personal dot file repository in $HOME/dot-files.

For our example above it will create a link for $HOME/.config/nvim/init.vim pointing to ../../dot-files/neovim/init.vim. The script will also create parent directories if necessary and a backup if a file with the same name already exists.

You can specify folders and files as key/value for objects in the configs array.

Limitations

• There is no way to detect removed packages/configs
• No post_install command
• No pre/post_configure commands
• Dependencies between packages can not be modelled