wpcarro/url-blocker

Block URLs when you want them to be blocked

url-blocker

url-blocker blocks the URLs that you want to block when you want it to block them.

Let's say that you don't want to visit Twitter during the work week. Create the file /etc/url-blocker/rules.json with the following contents and url-blocker will take care of the rest.

# /etc/url-blocker/rules.json
[
  {
    "urls": [
      "twitter.com",
      "www.twitter.com",
    ],
    "allowed": [
      {
        "day": "Saturday",
        "timeslots": [
          "00:00-11:59"
        ]
      },
      {
        "day": "Sunday",
        "timeslots": [
          "00:00-11:59"
        ]
      }
    ]
  }
]

Usage

This project builds with the latest version of nixpkgs-unstable, (i.e. https://github.com/NixOS/nixpkgs-channels at commit ae6bdcc53584aaf20211ce1814bea97ece08a248).

1. Create and populate the /etc/url-blocker/rules.json file:

$ sudo vi /etc/url-blocker/rules.json

2. Ensure that Nix can resolve <unstable>. If you do not have this set up, the following is what I use:

$ git clone git@github.com:NixOS/nixpkgs-channels ~/nixpkgs-channels
$ export NIX_PATH=unstable="$(realpath ~/nixpkgs-channels)"

3. Build and run `url-blocker:

$ nix-build
$ sudo ./result

This is not my ideal workflow. See the Shortcomings section for what I would prefer but have not setup.

How does it work?

systemd is intended to run url-blocker once every minute. url-blocker will read /etc/hosts and map the URLs defined in rules.json to 127.0.0.1 when you want them blocked. Because systemd runs once every minute, /etc/hosts should be current to the minute as well.

I have not setup the systemd units, but I encourage anyone who may be interested in this to create a pull request.

If an example is worth 1,000 words, here is an example. The following rules.json file...

# /etc/url-blocker/rules.json
[
  {
    "urls": [
      "facebook.com",
      "www.facebook.com",
      "instagram.com",
      "www.instagram.com",
      "twitter.com",
      "www.twitter.com",
      "youtube.com",
      "www.youtube.com",
      "wsj.com",
      "www.wsj.com",
      "nytimes.com",
      "www.nytimes.com"
    ],
    "allowed": []
  }
]

...will create the following /etc/hosts file:

127.0.0.1	localhost

################################################################################
# Added by url-blocker.
#
# Warning: url-blocker will remove anything that you add beneath this header.
################################################################################

127.0.0.1	twitter.com youtube.com www.wsj.com www.twitter.com www.youtube.com nytimes.com www.instagram.com wsj.com facebook.com instagram.com www.nytimes.com www.facebook.com

Shortcomings

I would like to move the /etc/url-blocker/rules.json file to ~/.config/url-blocker/rules.json. I currently do not know if it is possible to define user-specific rules. If Linux permits this, systemd can run this as a --user module, which I would prefer.

I read that some people have successfully created user-specific rules by setting the HOSTALIAS environment variable to some file (e.g. ~/hosts). I could not get this to work on my end, but if other people can successfully use this, I would prefer supporting that instead.

Considerations

When I use this, I'm often adding duplicate entries in the form of website.com, www.website.com. It might be nice if url-blocker automatically prefixed non www. entries with www.. However, I am unsure how I would like to handle websites that are not prefixed with www., so for now I'm keeping any implicit behavior explicit until I have stronger opinions about this.