The title says that because, until now, the core utilities in Ubuntu were not fully Rust. cp, mv, and rm had stayed on the GNU versions over compatibility issues. They have since moved to the Rust-based uutils implementation, taking the distro even further in its oxidation process.
Ubuntu 26.10, codenamed "Stonking Stingray", is an interim release, which means it brings newer software than the earlier LTS while only getting nine months of support, ending in July 2027.
Although our features roundup covered a lot of ground, there were some changes that it missed.
Ubuntu 26.10: What's New?
Well, we start with the kernel powering this interim release, Linux 7.3, which offers support for recent Snapdragon X2 Elite laptops along with several older Snapdragon X Elite and X Plus machines from Lenovo, HP, and Acer.
Gamers will be glad to know that the first chunk of Valve engineer Natalie Vock’s VRAM overcommit work has also landed in this kernel. Games on graphics cards with too little video memory should run better, as some of the data can now be offloaded to regular system RAM.
You already know that this Linux release is shaping up to be one of the most performance-focused releases in years.
Ubuntu is also moving to the NVIDIA 615 driver series, which replaces 595 and 610 and ships open kernel modules only, supporting Turing and newer GPUs.
The rustification we talked about earlier also extends to OpenPGP, where Sequoia PGP, a Rust-based implementation built by former GnuPG developers, has landed in the main archive. Its sq and sqv tools now sit alongside GnuPG's gpg and gpgv.
For the desktop environment, Ubuntu 26.10 has GNOME 51, with its range of upgrades such as a fingerprint manager in the Users panel, DNS search domains in the Network settings, and a count badge in Files when you drag multiple items.
There's also the Ubuntu Tiling Assistant that shows a Windows-11-like interface for the built-in snappable window layouts.
If you were hoping for a redesigned installer and first-boot onboarding flow, those are planned to come for the next long-term support release, Ubuntu 28.04 LTS.
Other Changes & Improvements
These are some of the other notable changes in this release:
OpenSSL 4 brings Encrypted Client Hello (ECH) and better QUIC support.
TPM-backed full-disk encryption now works without a hardware root of trust if a PIN or passphrase is set.
The RISC-V platform gets official RVA23 support for the SpacemiT K3 boards and SiFive's BigSky platform.
Twelve new wallpapers that were picked across the Mascot, Photography, and Digital/Abstract categories.
The toolchain upgrades consist of GCC 15.2.0, LLVM 22.1.8, Python 3.14.7, Rust 1.97.1, Go 1.27, and OpenJDK 25.0.4.
Download Ubuntu 26.10
ISOs for this interim release are available for AMD64, ARM64, RISC-V, and others from the official website; go for the "Desktop" ISOs if you are a regular user. Specific flavors like Server, Netboot, and WSL can be fetched from the releases portal.
What part of changing working, tested and very stable utilities with semi-compatible unstable ones is pretty good?
Modern, clean codebases that can attract rather than deter contributors. Sure, a lot of the coreutils are "feature complete" but they still need maintenance+security patches and someone will need to work on and release those.
Are you someone who's also against the deprecation of X11 for Wayland?
Are you implying that coreutils aren't supported or patched on time? That there aren't enough maintainers? That nobody works on them? State in which way current coreutils aren't up to par compared to modern and attractive reimplementation. Personally I don't see the point, there is nothing extremely good about them that would require a switch.
My stance on other projects is irrelevant
Literally no one is forcing you to switch. Just keep using coreutils.
Unless I'm an ubuntu user?
I’m sorry, I must have not said this loud enough. JUST KEEP USING COREUTILS.
UNLESS I'M AN UBUNTU USER
Omg dude, you can still use coreutils on Ubuntu:
You can also opt out before you upgrade:
/etc/apt/preferences.d/uutils:NO ONE IS FORCING YOU TO SWITCH!
Opt-out and not forcing are mutually exclusive. Obviously I can install GNU Coreutils.
Because rust is inherently memory safe, so these utilities will be less likely to have hidden memory errors that could cause security issues.
So the change brings real incompatibility in excange for potential safety?
If the tools are actually incompatible, you can always go back to the old tools. They’re in another package.
They're very open about not being 1:1 on feature matching.
sudo-rs breaks Ansible privilege escalation. They blamed the Ansible team.
The Rust coreutils package upgrade removes coreutils from your system and then fails to install the new ones because there aren't any coreutils installed. They blamed the dpkg team.
Does dpkg depend on coreutils??
They're saying "dpkg needs to put in the work to make an exception for our specific use-case because why should we change it"
My question would be why does dpkg need to except anything? Why do they depend on coreutils at all?
It's not that they depend on coreutils, it's that without coreutils, the operating system no longer has the tools to (for example) navigate and make changes to the file system.
No, the kernel does that. Coreutils are the tools for the user to do that through the terminal, and for shell scripts that can’t make kernel calls (because they’re scripts).
(Fun fact: you can absolutely call that kernel function I just linked to in a super unsafe way in C, leading to a vulnerable tool whose only job is to read a directory, like ls.)
Fair point. dpkg, being effectively some scripts, suddenly lose access to user-level utilities for managing files in the filesystem, like mv, mkdir, cp, ln, etc.
dpkg is mostly written in C and Perl. It is not “effectively some scripts”. I feel like you don’t actually know what you’re talking about.
My bad, I was just trying to meet you halfway and not suggest that you don't actually know what you're talking about, because that would be rude.
I don't really care what dpkg is written in but I do know two things for a fact: Ubuntu's coreutils upgrade broke Ubuntu because coreutils first gets removed, and Canonical's official response was "apt maintainers will have to clean up our mess": https://discourse.ubuntu.com/t/migration-to-rust-coreutils-in-25-10/59708
My memory was wrong on one point: dpkg is the workaround, apt is what the linked post blames.
I read through that whole discussion. None of it seems unreasonable. Considering Debian also wants apt to support replacing coreutils too, I don’t even know why you think it’s unreasonable. There is limiting behavior in apt, and both the Debian and the Ubuntu devs want it fixed. It’s a win-win.
There are probably edge cases and you might not know about them until you actually run into them by running a script or similar but then it is to late as people dont want things to break. There is 30 years of testing, and expected behavior, behind on what we have today.
Yeah man, we should never change anything.
Change => changing behavior etc. That is not a problem at all as long as it 100% compatible - which in this case is very hard to prove. The way to introduce change is therefore to say it is a new, different, standard. API version 2 or a new protocol, and let users migrate over to that by requesting the new version/protocol etc.
One difference is though that Ubuntu 26.10 is a new release(not LTS) so you would need to have a migration process anyway.
sudo-rs breaks Ansible. Imagine pushing this nightmare out to several hundred computers and then not being able to reverse it.
"I'm against change!"
Great, then don't change. No one is forcing you to. Just let others do their thing.
Should we change for the sake of change without any tangible benfits and obvious disadvantages on supposedly enterprise grade distros?
You seriously thing somebody woke up one morning and said "let's change it up" and set out to replace core-utils with uutils? No reason besides "I want change"?
You seriously think memory safety isn't a "tangible benefit"?
If this was a change just for the sake of change, you would totally have a point!
is this the one that AI has found 18 year old bugs and inconsistencies in?
replacing this C shit show guarantees safety and using new fuzzers to ensure correctness and fix bugs in the gnu version is a bonus
how is your non systemd distro going for you? have you heard of gui's? nothing like the old clickety clack of the vt100s eh?