The final solution of dot-file management for Unix systems

Project N₂ aims to solve all the pain points in general dot-file management, such as version control, modularization, etc. It also provides an informative and customizable bash/tmux UI.

cd "$HOME"
git clone [email protected]:hengyang-zhao/n2.git .n2
.n2/install.sh

The installation tries to be stupid. It just appends lines to your existing dot-files. It does not soft-link, backup, or overwrite your original files for you, which makes it less obvious in case you wish to uninstall it later.

You can also install N₂ to a playground outside of your HOME:

PLAYGROUND=yes .n2/install.sh

or (to save some confirmation typings)

AUTO_CONFIRM=yes PLAYGROUND=yes .n2/install.sh

Note
AUTO_CONFIRM=yes works in the regular (non-playground) mode too.

Note
The cloned directory can be named and placed arbitrarily. It doesn't have to be named .n2 or placed directly under your HOME. But for M₂ directories (optional feature, mentioned below), they have to follow the predefined pattern.

After installation, you can just do man n2 to pull up the N₂ reference manual.

You must already have some configs in your dot-files. Move them into M₂ directory(s) such that they can be easily verson-controlled.

Note
You still have the freedom to keep your configs in their original places, i.e., ~/.bashrc, ~/.vimrc, etc. You can skip this section if you wish to do so.

An M₂ directory is a directory under your HOME and named like .m2*. When bash is initializing, N₂ will enumerate all the M₂ directories and source the configurations under them. A typical M₂ directory looks like this:

~/.m2
├── bash
│   ├── profile.d
│   │   ├── 10-my-profile.sh
│   │   ├── 20-another-profile.sh
│   │   └── not-starting-with-number-is-ok.sh
│   └── rc.d
│       ├── 10-my-rc.sh
│       ├── 50-another-rc.sh
│       └── not-starting-with-number-is-ok.sh
├── exec
│   ├── my-exec
│   └── another-exec
├── git
│   └── config
├── tmux
│   └── conf
└── vim
    ├── 10-my-config.vim
    ├── 20-another-config.vim
    └── not-starting-with-number-is-ok.vim

Then you know where to put your old configs.

If you don't like creating an M₂ directory from scratch, N₂ can automatically create one:

n2 create-m2

The demo M₂ dir is a good start point. It already has several files to help you customize N₂ and add personal configs.

Tip
Version-controlling your M₂ dir is often a good idea.

You also have the freedom to have multiple M₂ directories. This becomes useful when you want to separate your M₂ dirs for personal use and work. If this is the case, a typical home directory will look like this:

~
├── .n2
├── .m2-10-personal
├── .m2-20-work
└── MyOtherStuff

Note
The M₂ directories are discovered in lexical order. Those odd-looking infixes -10- and -20- are just to control that order.

For more details, see man n2.

By default, N₂ has a rich bash PS1 prompt. In addition to a colorful user@host and current working directory, it also has

• a git repo/branch indicator;
• the permission bits of the current directory if it's not readable or writable;
• the nesting level of the current bash, if it's not the outermost one;
• number of processes running in the background;
• the physical cwd, if the appearing cwd is a symlink;

and some less frequent ones

• the nice value of current bash if it's not 0;
• a chroot indicator honoring debian_chroot;
• the session name if in a GNU screen session;
• IFS value if not default (\x20\x09\x0a).

To checkout the current prompt:

echo "$PS1"

To customize this, just overwrite PS1 in your M₂ configs.

The status bar shows

• hostname and session name;
• the current TTY path if the session is multi-attached;
• window list;
• system load / number of cores;
• date and time.

Command expansion is the lines starting with [#] -> XXX, for example

me@laptop ~
$ ls
[1] -> /opt/homebrew/opt/coreutils/libexec/gnubin/ls --color=auto (MM/DD/YYYY HH:MM:SS)
-- OUTPUT SKIPPED --

or

me@laptop &1 ~
$ grep pattern < file | wc
[1] -> /usr/bin/grep --color=auto pattern < file (MM/DD/YYYY HH:MM:SS)
[2] -> /opt/homebrew/opt/coreutils/libexec/gnubin/wc (MM/DD/YYYY HH:MM:SS)
-- OUTPUT SKIPPED --

or

me@laptop ~
$ echo hello && echo world
[1] -> builtin echo hello (MM/DD/YYYY HH:MM:SS)
hello
[2] -> builtin echo world (MM/DD/YYYY HH:MM:SS)
world

Features include

• telling if the command was an external command (by expanding its true path), or if it's a shell builtin;
• timestamping the commands right before the command is executed;
• breaking up commands by pipe operators and logical operators;
• making sure that once a expansion is printed out, the command is starting to execute --- this is especially useful when you copy-paste a command with a trailing line-break character into bash, but you just don't know if it's already executing or waiting for ENTER.

Automatically prints the status code and timestamp after a user input is completed. If it's a piped command, it prints out status code of each pipelet. The timestamp is often used together with the ones in command expansion to figure out how long a command was running for.