MichaelHatherly / CuratedSystemImages.jl

Julia package providing curated system images

CuratedSystemImages.jl

Note

Windows users need to enable "Long Paths" since the bundled artifacts provided by this package often have longer path names than the default limit allows. See this Microsoft document for the steps required to enable this feature.

This Julia package provides access to several pre-built system images containing collections of third-party Julia packages. It requires that your julia install is managed using juliaup.

It is an unregistered package and will remain this way since no package should include this package in its dependencies. To install use Julia's Pkg mode in your global project environment:

julia> import Pkg; Pkg.add(url = "https://github.com/MichaelHatherly/CuratedSystemImages.jl")

Then import the package and run the installer:

julia> import CuratedSystemImages

julia> CuratedSystemImages.install()

This will open an interactive picker where you can select the individual bundled system images that you would like to install. Once installed you'll need to close your julia REPL and return to the terminal. You can then use juliaup status to see the names of the newly installed channel names that can be used to launch julia with one of the installed system image bundles.

Now from the terminal run julia with one of the custom channel names that were installed in the previous steps, e.g if you installed a system image for DataFrames in Julia 1.8.0 then run

julia +1.8.0/CuratedSystemImages/DataFrames

If you selected to install "short names" for some channels then you'll be able to do

julia +DataFrames

to start with the system image containing DataFrames instead.

Notes

Default channels

Do not set any of the custom channels as the juliaup default channel. This will result in an infinite loop where the default channel attempts to call itself. This may cause your system to hang.

startup.jl and latency

Please note that if you are importing any packages in your .julia/config/startup.jl that happen to use different versions of packages included in the system image that you launch then you will likely encounter some amount of startup latency. Running with --startup-file=no will mitigate this latency.

Installing packages when using a custom system image

It is not advised to install additional packages into the named environment that the system image launches with by default, e.g. @DataFrames for an image called DataFrames. If you need to install extra packages then ensure that you use --project= to start a custom project environment. It will still have access to the packages installed in the custom system image. Ensure that you install extra packages using add --preserve=all rather than the default behaviour which may upgrade dependencies to versions which are not included in the system image which may cause the image to load incorrectly or not at all.

Available Images

Currently the following system image bundles are provided:

• AlgebraOfGraphics, contains the Algebra of Graphics plotting package along with the Cairo-based Makie backend, and DataFrames.
• DataFrames, contains the DataFrames and DataFramesMeta packages.
• JuliaFormatter, contains JuliaFormatter, with latency reduced enough that users may consider using it as the basis for a command line formatting script for Julia source code.
• JuliaSyntax, contains JuliaSyntax and uses the provided precompile scripts to reduce latency and enable JuliaSyntax as the default parser for all code.

Requests for additional images can be made to curated-system-images repo where the manifests and build scripts are located for the above images.

How does it work?

We use the SystemImageLoader.jl package to define an installer and loader for the lazy Julia artifacts that contain the system image bundles. Installing a particular bundle sets up a juliaup channel that points SystemImageLoader at the right system image file and associated Julia depot folder containing the artifacts required by the chosen system image.