JILL.py

The enhanced Python fork of JILL -- Julia Installer for Linux (MacOS, Windows and FreeBSD) -- Light

Why jill.py? Distro package managers (e.g., apt, pac) is likely to provide a broken Julia with incorrect binary dependencies (e.g., LLVM ) versions. Hence it's recommended to download and extract the Julia binary provided in Julia Downloads. jill.py doesn't do anything magical, but just makes such operation even stupid.

Using Python to install Julia? This is because Python has became one of the main tool for sys admins and it's available in all platforms; this makes a cross-platform installer for Julia possible.

Is it safe to use this? Yes, jill.py use GPG to check every tarballs after downloading. Also, *.dmg/*.pkg for macOS and .exe for Windows are already singed.

Features

let's make a simple and stupid julia installer

download Julia release from the nearest mirror server.
immediately knows once there's a new Julia release.
across multiple platforms.
manage multiple julia releases.

Installation

First you'll need to install jill using pip: pip install jill --user -U

Note that Python >= 3.6 is required. For base docker images, you also need to make sure wget and gnupg are installed.

Usage examples for most users

Basic usage:

jill install [version] [--confirm] [--upstream UPSTREAM] [--unstable] [--reinstall] [--install_dir INSTALL_DIR] [--symlink_dir SYMLINK_DIR]

For the first-time users of jill.py, you may need to modify PATH accordingly so that your shell can find the executables when you type julia.

installation demo

When you type jill install (the simplest usage), it does the following things:

query latest stable release, it's 1.6.1 at the time of writing.
download, verify and install julia 1.6.1
make alias: julia, julia-1, julia-1.6
- for nightly build, it only bind alias to julia-latest

Valid version syntax:

stable: latest stable release
1: latest stable 1.y.z release
1.2: latest stable 1.2.z release
1.2.3/1.2.3-rc1: as it is
nightly/latest: nightly builds

At the time of writing, "1.6.1" and "1.7.0-beta1" are released, while "1.7.0" isn't yet. There are two ways to install an unstable release version 1.7.0-beta1:

jill install 1.7.0-beta1, as you explicitly requested, installs "1.7.0-beta1" for you.
jill install 1.7 --unstable will give you the latest 1.7.x version including those unstable ones. Because "1.7.0" is not released yet, it will gives you "1.7.0-beta1". The result may change when more releases are coming. Note that it won't include the "nightly" builds.

Here's a list of slightly advanced usages that you may be interested in:

download only:
- latest stable release for current system: jill download
- specific system: jill download --sys freebsd --arch x86_64
- download Julia to specific dir: jill download --outdir another/dir
install Julia for current system:
- (linux only) system-wide for root: sudo jill install
- upgrade from older julia version: jill install --upgrade (copy and paste the root environment folder)
- don't need interactive promopt: jill install --confirm
upstream:
- from specific upstream: jill install --upstream Official
- find out all registered upstreams: jill upstream
- add a private upstream: make a modifed copy of public registry at:
  - Linux, MacOS and FreeBSD: ~/.config/jill/sources.json
  - Windows: ~/AppData/Local/julias/sources.json

You can find a more verbose documentation using jill [COMMAND] -h (e.g., jill download -h)

For Julia (>= 1.5.0) in Linux with musl dependency, you can

install it normally, i.e., using jill install; jill knows that you're using musl.
download it by passing --sys musl command. In the meantime, --sys linux will give you Julia binaries built with glibc dependency.

Setting environment variable JILL_UPSTREAM will disable the fancy "find-the-nearest-upstream" feature of jill and give you a faster download experience if you just know which upstream is the fastest. (It has lower priorty than --upstream flag.)

The Python API

jill.py also provides a set of Python API:

from jill.install import install_julia
from jill.download import download_package

# equivalent to `jill install --confirm`
install_julia(confirm=True)
# equivalent to `jill download`
download_package()

You can read its docstring (e.g., ?install_julia) for more information.

Example with cron

If you're tired of seeing (xx days old master) in your nightly build version, then jill can make your nightly build always the latest version using cron:

# /etc/cron.d/jill
PATH=/usr/local/bin:/usr/sbin:/usr/sbin:/usr/bin:/sbin:/bin

# install a fresh nightly build every day
* 0 * * * root jill install latest --confirm

Similarly, you can also add a cron job for jill install --confirm so that you always get a latest stable release for julia. jill knows the existence of a new version of Julia once it's released -- you don't even need to upgrade jill.

For who are interested in setting up a new release mirror

Please check out register new mirror.