February 2020 in NixOS

As with last month’s review, this focuses only on merged PRs. Any change sent directly to master will not be seen. And again, this is merged PRs, so upcoming changes are not relevant yet.

Note that all the PRs listed here are arbitrarily chosen by entirely ill-defined biases. If you don’t see PRs you found interesting, or even, your PR, please share it with us!


This February, we merged 1865 PRs. Of these, 504 were automated. Giving us 1361 mainly handcrafted PRs.

(I’m sorry, starting here GitHub doesn’t give the right tooling to dig more deeply with links here.)

Of the 1246 targeting master or staging, 123 are new packages , and 1003 are package updates.

Special mention: 110 automated PRs part of #79975 by @bhipple.

Notable PRs

Package upgrades


Disclaimer: 3 PRs got caught in an incident that made them lose their original title, this affects some of the stats.

Finally, saying it again, I have some homework for you: I want you all to tell me what caught your interest in February! Reminder: what was merged in February!

Note: as a courtesy, you may want to use the official GitHub link shortener so the PRs are not swamped with useless Discourse messages.


oh maan, did you look through all 1361 PRs?

Kind of.

I had some tooling to dig in GitHub PRs ready already, which I augmented last month to do the first one of these.

The intro is kind of my methodology. I batch it. I cut all automatic updates. Then, I look all obvious updates (-> and ). Then all init at. The final step is looking at all the other PRs.

Now, when I say “look”, I mean “look at the title”, and if the title is interesting, I open the PR, see what it is, then add it to the list. It goes quite quickly.

I think it took about 2 hours total. This is definitely not a waste as it also allows me to keep tab on cool things. This month there were fewer cool things than last, but still I noticed this cool test, though this was not added to this particular list. 2 hours monthly to know what happened, what big updates we’ve had, it’s not much. And there are still optimizations that can be done.

Furthermore, I may start cheating and gather data weekly, while still publishing monthly. This would reduce the “chore” some, and I would keep better tabs on new cool stuff as it happens.