I’ll admit it, I’m obsessed with being an early adopter of technology and I switch between Linux distributions frequently. Before Ubuntu 18.04 came out (it was just released on April 26th), I decided to test it out around the end of March, just a month before it’s official release date. There’s just one issue with that: it wasn’t the final version. Breakage can often happen when testing out a pre-release Linux distibution. For instance, a simply
sudo apt update; sudo apt full-upgrade can cripple a system with a bad package, resulting in a reinstall. I stumbled upon a Makefile by Julius Beckmann that can be run on a new Ubuntu system and immediately saw the benefit of such an approach. I can reinstall as many times as I like, and so long as I set up the commands correctly, I can have a system configured exactly how I like it within an hour. I can even run a specific make target on its own in order to install a certain set of programs or updates without rerunning the whole script. I decided to fork it, creating my own Ubuntu Makefile repository, adding many packages, and removing things I do not need.
I approached this as a development project at first, running Ubuntu 18.04 daily images in VirtualBox. VirtualBox is free and it has a very nice feature called snapshots. I would take a snapshot of a running virtual machine, test out the project, check the final state, and revert to the snapshot before it was applied, so that I could make corrections. After a few days of testing, I was ready to install this on real hardware. I backed up my desktop’s home directory, and installed the 18.04 daily build. As I add packages or software to my system, I make changes to the Makefile if I want that software to be reinstalled when I rebuild my system. Now, I don’t have to worry about wiping and reloading, and it is a huge timesaver. It also allows me to keep the software between my laptop and my desktop relatively in-sync since I can recreate the steps taken perfectly. Now, when I have to reinstall for whatever reason, the only data I have to worry about backing up is the
To take this one step further, as I love to test out newer versions than provided in the Ubuntu repositories, I am actually using Flatpak, Snap, PPAs, and some Git repositories to provide software. I’m actually removing the versions of things like LibreOffice and installing the flatpak version instead. As such, I need to check for Flatpak and snap platform updates in order to fully update my machine. Inspired by the
zypper dup command on OpenSUSE, I created an alias called
dup, which can easily update everything, including Flatpaks and snaps:
alias dup="sudo apt clean all; sudo apt update; sudo apt -y full-upgrade; sudo flatpak update; sudo snap refresh; sudo apt autoremove"
I’ve created a gist of the common aliases that I use. I then added this alias to my
~/.bash_aliases file By simply typing
dup into a terminal, I can now get a fully updated system with many bleeding edge packages.