Home Assistant is strange. You can't install it "normally" on a Pi because Python is too old (it wants 3.8, Raspbian has 3.7). You can install it via Docker but then you can't use Add-Ons. If you want full functionality you're really left with one choice - install their entire OS image on a dedicated Pi. For a collection of Python scripts.

@8none1 does it not work with Ubuntu 20.04 or newer? Those have Python 3.8+.

@artfulsodger I imagine that it does, but I'm already running Raspbian on my Pi4 and I don't want to have to blow it away and start again.

@8none1 it makes sense to stick with your existing OS and setup if you don't want to spend more time on redoing it. Have you tried using 'pyenv' to install a supported python version, alongside the system python? It does require some work, but that should still be a lot simpler than nuking and paving.

Follow

@artfulsodger @8none1 every time I had worked some place that did python we always ran into this kind of problem when it came time to deploy products to real end users. Then it became do we restrict it to just a couple of very specific distros that had bundled correctly usable versions of python and dependencies? Do we also create python virtual environments as part of deployment? This is why I didn't use python in my own stuff...

@artfulsodger @8none1 ruby is a little better both in terms of more reliable range of backward compatibility offering a wider range of pre-packaged distros that can be used as-is, and in that ruby bundler feels somewhat simpler to use in deployment scenarios than python virtual environments. But it is only easier in degree. However, it was enough for me to prefer ruby over python.

@artfulsodger @8none1 there were other reasons I found a preference for ruby over python, but production deployment with ruby bundler and Gemfiles vs python venv and requirements.txt was among them.

Sign in to participate in the conversation
Fosstodon

Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.