Is is correct that right now if I want to upgrade my Julia I would have to install it from the scratch? This is a problem, because the users would have to change settings for Juypter and Atom Juno, etc. in order for them to point to the newer version of Julia.
Would you please consider making Julia upgrade a simple command like the below?
upgrade Julia to v1.7.3
That way, the new version will automatically replace the old version and there is no need to reconfigure anything.
That would break user’s behaviours (including mine) to test different version of Julia against each other. Do you have a proposal how to do that then? (IDE like Atom have settings to allow to select the currently active version)
I’m not really talking about disabling that approach. Users who have special needs like you just mentioned should always have the option of not running the above upgrade commands, but instead install a version parallel to the old version from scratch. My guess is that the majority of the users just want to have one version.
Ah, that sounds fair enough.
Have you tried juliaup. It is currently available for Windows, but it´s developing for other OSs.
In linux or macos using symlinks works perfectly.
I have all versions on ~/lib/julia/ 1.x/ and then the symlinks ~/bin/julia1.1 ~/bin/julia1.2 etc and finally for the laters version I have the symlink ~/bin/julia pointing to ~/bin/julia1.6
It takes 10 seconds to manually create these symlinks when “updating” julia and then I can just start the version of julia I want…
Quite brutal for a windows user;)
If you are a Linux user, you can use jill.
Just to follow up on that: if you are on Windows and install Julia from the Windows Store, you automatically get that. That version is fully supported and functional and provides both 1) update features for Julia and 2) a version multiplexer. The Julia VS Code extension also picks up multiple Julia versions installed with
juliaup for the integrated notebook story. You can get a feature overview here.
Support for other platforms is making progress
If you have access to python, jill.py may be a good cross-platform choice which I use this personally.
It’s awesome! I already have Julia in the system (installed manually, not from Windows Store), and how could I install Juliaup for future version management?
The best strategy is to completely remove any Julia version that was not installed from the Windows Store, and then only install Julia from the Windows Store, and use
juliaup to manage the specific versions you have on your system.
Great. Thanks for the advice.
I do the same thing, and it works great. But also we need to keep in mind that there is a large group of people who use Julia on Linux and who have no idea what a symlink is.
Such is the price of becoming a popular language used by scientists with limited computer affinity.