Controlled installation of Lean and mathlib on Debian/Ubuntu #

This document explains a more controlled installation procedure for Lean and mathlib on Linux distributions derived from Debian (Debian itself, Ubuntu, LMDE,...). There is a quicker way described in the main install page but it requires more trust. Of course you can get even more details about what is going on by reading the bash script that will be downloaded below: elan_init.

• Lean itself doesn't depend on much infrastructure, but supporting tools needed by most users require git, curl, and python. So the first step is:

  sudo apt install git curl python3 python3-pip python3-venv
  

• The next step installs a small tool called elan which will handle updating Lean according to the needs of your current project (hit Enter when a question is asked). It will live in $HOME/.elan and adds a line to $HOME/.profile.

  curl https://raw.githubusercontent.com/leanprover/elan/master/elan-init.sh -sSf | sh
  

• There are two editors you can use with Lean, VS Code and emacs. The recommended choice is Visual Studio Code.

  wget -O code.deb https://go.microsoft.com/fwlink/?LinkID=760868
  sudo apt install ./code.deb
  rm code.deb
  code --install-extension jroesch.lean
  

  Now open VS Code, and verify Lean is working, for example by saving a file test.lean and entering #eval 1+1. A green line should appear underneath #eval 1+1, and hovering the mouse over it you should see 2 displayed.

  The alternative is to use Emacs, and its lean-mode.

• Then we install a small tool called leanproject that which will handle updating mathlib according to the needs of your current project. We use pipx to perform the installation.

  python3 -m pip install --user pipx
  python3 -m pipx ensurepath
  source ~/.profile
  pipx install mathlibtools
  

Lean Projects #

You can now read instructions about creating and working on Lean projects