opensnitch: lessons learned in summer 2026
For the past month (~6 weeks), I've been working on migrating opensnitch from 1.6.9v to 1.8.0v release. There have been 5 releases between the two, and 1.6.9v is not on the list of major releases for opensnitch.
As of now, the current 1.8.0v source builds on Debian unstable.
Why I did this was to solve a few bugs 1131582, 1135540, and 1135591: all related to updating opensnitch, which I found as an RFH. There are still a fair number of bugs, some that could be fixed with a stable 1.6.9v patch, but I've wanted to work on 1.8.0v release candidate and get it finished before moving to different patches.
Some lessons learned in an unordered list:
- When working on GitHub sourced packages, your colleagues may weigh your work based on gbp buildpackage viability, so test this first before pushing a commit.
- Upstream tests may require root permissions, which fails running a Debian build, so anticipate creating rules to by-pass the noncompliant tests (or run as root, just for testing).
- Create the upstream branch on your Salsa repository rather than your local git.
- Review can take a lot of time.
I intend to write a more technical post on how I did my work, but with how little time I have I'm unsure how comprehensive I can be. In the past I thought it was odd that people wouldn't write extensive documentation about their work, but now I can see how obscure development can be when working in a mess.