@sandro@c3d2.social avatar

sandro

@sandro@c3d2.social

Some lefti :AFD:, NixOS :nixos:, Linux :tux:, Chaos :fairydust:

Don't be afraid of colorful flags :progress_pride: 🏳️‍🌈🏳️‍⚧️🚩🏴

This profile is from a federated server and may be incomplete. Browse more on the original instance.

sandro, to random
@sandro@c3d2.social avatar

The entire nixfmt project recently has been a big disappointment for me. nixpkgs-fmt would have been a good starting point, having a balance between enforcing rules and allowing some freedom. While nixfmt initially started with outdated ideas like 80 line length and ill fitted tries to enforce that. Adding new lines before long strings like URLs, sometimes nudging people to split them with a + which makes greping for them unnecessarily hard. They didged that which is good.

sandro,
@sandro@c3d2.social avatar

Then that people are sprinting ahead and reformatting single packages in otherwise straight forward PRs is not improving the situation in any way and is just creating lots of noise. Most often those commits are also not added to the ignore refs file and clutter git blame.

Then also issues like mentioned above might still get fixed in the future which then could result in yet another round of formatting changes because of the now strictness of nixfmt.

sandro,
@sandro@c3d2.social avatar

But at the same time adding rules, especially around lib.optional*, which often results in ugly formatted code that has strange line breaks and the general placement of the entire structure feels wrong.

Also that just appending something to a list like structure can result in the reformat of the entire thing which can make a simple one line change be blown up in the diff to many more lines.
At least an issue for that exists after arguing with them for a bit.

sandro,
@sandro@c3d2.social avatar

Does it matter in the end? Probably not that much but it will create a lot of noise on the way and will make my live of cherry-picking specific changes from master to the stable branch probably a lot more miserable and with lots more merge conflicts to solve.

Usually people are very focused to reduce churn but looking at the formatting topic some people seem to completely thrown that mindset overboard.

hexa, to NixOS
@hexa@chaos.social avatar

#NixOS 24.05 branch-off happened. Snuck in a small reformat of the python packages at the last minute.

https://github.com/NixOS/nixpkgs/pull/313628

Currently, evaluations are running, nixos-24.05-small has already completed eval and is building.

https://hydra.nixos.org/jobset/nixos/release-24.05-small

It will be mostly cached from nixos-unstable-small anyway, so its channel should be available within the next two hours.

sandro,
@sandro@c3d2.social avatar

@hexa Which also created hundreds of merge conflicts and lots of extra work for contributors... 😮‍💨

sandro,
@sandro@c3d2.social avatar

@hexa We should have waited another 6 months. There are still glaring issues around lib.optional/s wild formatting.

pid_eins, to random
@pid_eins@mastodon.social avatar

Kernel breaking userspace episode nr. 43858262:

https://github.com/torvalds/linux/commit/a1912f712188291f9d7d434fba155461f1ebef66

Hacking around it:

https://github.com/systemd/systemd/pull/32892

sandro,
@sandro@c3d2.social avatar

@pid_eins I always read not breaking userspace as in compiled programs don't suddenly crash because of eg. removed syscalls.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • megavids
  • kavyap
  • DreamBathrooms
  • InstantRegret
  • magazineikmin
  • osvaldo12
  • everett
  • Youngstown
  • khanakhh
  • slotface
  • rosin
  • thenastyranch
  • ngwrru68w68
  • Durango
  • JUstTest
  • normalnudes
  • ethstaker
  • GTA5RPClips
  • modclub
  • cisconetworking
  • mdbf
  • tacticalgear
  • cubers
  • provamag3
  • tester
  • anitta
  • Leos
  • lostlight
  • All magazines
    •