GNOME Drops X11, Debian Adds Rust Dependency, Ubuntu Plans  ## Metadata - Author: [[Linux & Open Source News]] - Full Title: GNOME Drops X11, Debian Adds Rust Dependency, Ubuntu Plans - Category: #podcasts - URL: https://share.snipd.com/episode/7d72048f-6168-4c64-8f90-f41ff64712a3 ## Document Notes ## Highlights - **GNOME Ends X11 Sessions** - GNOME's Mutter removed X11 support so GNOME 50 will not offer an X11 session. - Users must run Wayland or choose a different desktop if they need native X11 support. Transcript: Nick So it is finally done. Mutter, Gnome's compositor and window manager has now ditched the X11 backend entirely. This merge request is four months old and it was merged this week, meaning Gnome 50, releasing in March 2026, will simply no longer support an X11 session at all. X-Wayland will still work, of course, apps that don't have native Wayland support will still run without problems, or at least without significant problems. XWayland is almost perfect, not fully perfect, but almost, but if you did not want to use Wayland, then you're going to have to move to another desktop, which I would assume most people Already did a while back because Gnome was very clear on their intention to drop that display server. A few things still need to be merged to adapt to that change. For example, the GNOME kiosk, the modified GNOME environment for fixed single-purpose computers, has a few things to remove, some conditions on detecting X11 or Wayland. I'm sure other places in GNOME's codebase have some stuff that needs to be pruned or to be removed as well. But yeah, this is going forward. Now, we all knew this was happening. I personally think it's fine. I've been a strong proponent of Wayland. I've been using it for years now without any real issue whatsoever. X-Wayland works when you need it. And honestly, if you did not like Wayland at all you probably were not using GNOME as well because well they've been very clear on the fact that they really wanted to ditch X11. You still have plenty of choices but it is clear that in the future these choices will just shrink considerably because Plasma will remove X11 with Plasma 7, XFC is working and finalizing Their own Wayland session. Even Mate is working at it. I can't see Cinnamon or Elementor iOS sticking with X11 for more than a few years after their fully featured Wayland session is released. So this is something that will happen to almost every major desktop. I'm sure some of them will keep X11 compatibility somewhere, but the major ones, you're just going to have to move to Wayland to use them. ([Time 0:01:44](https://share.snipd.com/snip/52238407-7bb0-4801-b5c7-6fc4e8802a15)) - **Debian Makes Rust Mandatory For APT** - Debian will make Rust a hard requirement for apt starting May 2026 to improve memory safety. - Some unofficial or niche ports may be unable to support apt without a working Rust toolchain. Transcript: Nick Now, something that is sure to annoy some people, Debian's package manager, apt, will have a hard requirement on Rust starting in May 2026. Debian developer Julian Andres-Clauder posted on the Debian developer mailing list that he intends to add hard dependencies on Rust in apps, notably the Rust compiler and the standard Library. The goal is to have the code that parses HTTP signatures, that parses.deb, and.tar files to be more memory safe, which is one of the main advantages of Rust. He also says that if someone maintains a Debian port that doesn't currently have a working Rust toolchain that would support this dependency, they need to build one or they need to sunset The ports they were working on, meaning they need to stop working on it. As he says, Rust is already a hard requirement for all officially supported Debian ports. There are seven of those, but there are also 11 unofficially maintained ports, which probably don't have all the required Rust support. They might have, but they also might not. These include GNU Herd, 64-bit and 32-bit, PowerPC, and its 64-bit variant, and more notably, a few 32-bit only architectures that might not have the required Rust toolchain and might Not even be able to have it. ([Time 0:03:54](https://share.snipd.com/snip/912add0e-de52-4aa8-b0be-1aabb609de54)) - **Ubuntu Aims To Streamline Development** - Canonical plans to simplify Ubuntu, consolidate communication channels, and automate builds. - They recommend moving to Matrix and increasing Rust tooling to reduce memory-related issues. Transcript: Nick Now this week Ubuntu discussed their plans for the desktop. Both Mark Shuttleworth and John Seeger, the VP of engineering at Canonical, shared their vision for the future with ZDNet and it looks like Shuttleworth still believes that there's A place for the Linux desktop. He says he would like to be part of open source winning the desktop space, and he thinks that Linux has the potential to deliver a desktop with what he calls universal appeal. But he also says that the Linux community needs to understand that desktops should not be built for engineers. Things should be simple and just work, which is something I agree with, but also it already isn't true of the other proprietary desktops like macOS and Windows. They're neither simple nor do they just work if you're interacting with them for the first time. The plan for the future of Ubuntu is, thus, first to make the Ubuntu building process more transparent with documentation and communication channels being consolidated, Matrix Becoming the primary communication channel. They're also looking at automating the building process to the maximum and they want to move to more Rust-based tools to limit memory-related issues. As per Snaps, they're going to be adding their permissions framework in 26.04. It is already available, it has been for a few releases, but it was experimental and disabled by default. It's basically like Flatpak permissions, but it's also kind of better because the user is being asked if they want to grant that permission for how long they want to grant it when the Permission is actually required. When Flatpak would just block the action and the user would never really know why that failed. Now, I do agree with Shuttleworth on this. We do need Linux desktops that aren't built by engineers for engineers. They need to be built by engineers, but they need to be built for the general user in mind. But I would argue that's really kind of the case right now. If you look at Cinnamon, if you look at GNOME, if you look at KDE, UX and UI is at the core of what they do now. There are still some features that are being developed in a first version with a more haphazard UI, but generally what is being released is already extremely good quality visually And UX-wise. You might not agree with the UX decisions that are being made, but generally UX is taken into consideration when those features are being built. This was not necessarily the case years ago, but today I don't see many desktops building features that don't have a nice, understandable user interface where text is thought upon, Where the button sizes, the button placement, the button labels aren't really thought out. So I don't think that's really a problem today. Packs, but the permission systems that they're building for snaps is just better than the flat pack one. There's just no debate. The application asking you, hey, do you want to let the app access this folder it doesn't have access to? And for how long is just better than the app refusing to access anything from that folder and not telling you why, and you having to find a graphical user interface to enable that permission By taking a checkbox among 20 different checkboxes. Like, this flat pack permission system is just not as good as what snaps are building right now. We also have more on Ubuntu's plans with an interview given to the register. Questions asked were a bit more incisive here, notably on Ubuntu Core, the snap-only version of Ubuntu that has some very interesting concepts like shipping every part of the system As a snap so you could swap out an old desktop environment for a brand new one while keeping everything else on the system extremely stable. Or you could get a much newer kernel on what is the equivalent of an LTS, for example. Canonical says that this new distro will not be the default for 26.04, for 28.04, it's probably 5 to 10 years before Ubuntu Core becomes the default experience and the current classic Ubuntu becomes an alternative download. But they do have plans to make Ubuntu Core the default product in the future. They say progress is slow. They've revamped the architecture of it because swapping components wasn't as modular as they wanted, they're also waiting on the permission stuff that they're working on for snaps And on the TPM-backed encryption as well. ([Time 0:06:00](https://share.snipd.com/snip/6869409d-cecc-438e-bddc-4f2030b87584)) - **Ubuntu Core Is A Long-Term Goal** - Canonical envisions Ubuntu Core (snap-first) as the future default but not for several releases. - They will use snap permissions, TPM-backed encryption, and component modularity to reach that goal over years. Transcript: Nick They're making progress with each Ubuntu release, for example in 26.04 they want to try and ship Pipewire as a snap, which would let them test things out for Ubuntu Core in the future. They also talked about Flatpak, saying it is not a priority for them, although they have put a series of tests in place to ensure that they don't break Flatpak support entirely in the Future, like what happened in Ubuntu 25.10. They also argued in favor of the snap model with a single store, saying that multiple stores basically give more control to other people on your computer, which I would argue is true For dev packages, because yes, Debian packages will be installed as root, and they can run scripts post-install with root permissions, but that's just not true for Flatpaks. You don't install Flatpaks as root and they don't run scripts in the background. They also say that the current Linux desktop is not polished in many ways compared to Mac and Windows, which I would personally disagree with. And they say it's likely due to always wanting extensibility, plugins, and the like. Now this whole interview read a bit arrogant to me, you'll read it and tell me what you think in the comments, but what I got out of it is Linux needs to grow bigger, but the only way to do so Is to make Linux more like macOS and Windows in terms of being more locked down, fewer choices, and that's not what I want at all. I do want Linux desktops to be more polished, I do want them to be more accessible to more people. And I do want them to be an actual choice that people can get without really thinking about it. But if that comes at the cost of extensibility, of plugins, of choice, of modularity, then I don't want it. Those are the things that make Linux great as a system. It's the things that make Linux different from Windows and Mac OS and if we remove that then it's no longer Linux and I don't really see the point of pushing that system to more people. Now fortunately just because Ubuntu is pushing for that or trying to move towards that doesn't mean old distros will. ([Time 0:10:24](https://share.snipd.com/snip/b0913b4c-3918-424c-bca5-96b7af998b4b)) - **Linux Crosses 3% On Steam** - Linux gaming share on Steam surpassed 3% for the first time, driven by the Steam Deck era. - Growth suggests more developers may start addressing Linux as a meaningful market. Transcript: Nick Now Linux passed the 3% threshold for gaming on Steam this week. As always, this number will move up and down as numbers from the Chinese market, which is predominantly on Windows, are tallied up or how they are being counted. Those can fluctuate a lot, but it is the first time Linux passed the 3% mark. And no matter the rises and falls in that progression curve, it is clear that overall Linux is gaining ground. Ever since the release of the Steam Deck, market share has tripled, which is not bad, even though, yes, it's still 3%, and the growth is accelerating quite a bit over the past few years. As per distributions, SteamOS is still dropping a little bit, and Arch-based distros and Ubuntu-based distros are rising. Ubuntu is actually clawing back some market share right now. ([Time 0:12:40](https://share.snipd.com/snip/673b7877-e60d-4a9c-bd10-a6b008b7164d)) - **Mozilla's SumoBots Trigger Community Exit** - Mozilla deployed SumoBots which auto-edited and auto-approved Japanese support articles without human oversight. - The community leader disbanded the Japanese Sumo team in protest over erased translations and poor AI handling. Transcript: Nick Now another week, another AI-induced disaster as the Japanese branch of Mozilla's support community called Sumo for Support Mozilla, where that branch has officially disbanded After two decades of work translating and supporting Mozilla products. Why did they disband? Well it's because Mozilla introduced SumoBots for their community at the end of October apparently. It's an AI based tool that edits and approves Japanese knowledge based articles all without any community or human oversight, unfortunately. The bot apparently disregarded the translation guidelines, so it was inaccurate and way too literal. It also erased existing translations that had been made and approved by the community. It automatically approved machine-translated articles and content for archived articles in a time frame that is way too short for human contributors to review those changes, so They were automatically applied, and it just didn't cede control at all to the community, rendering it basically powerless. Apparently 300 articles were overridden without any prior test, and the community calls it mass destruction of previous work and an explicit violation of the Mozilla mission. The leader of that team thus decided to stop contributing. He says he prohibits all of his translations being used as learning data for that bot, and he requests the removal of all his translations from the bot's datasets, something that is Very unlikely to happen. Now, judging from the comments under his post saying that he was leaving the community, some other communities seem to have fewer issues with this, so maybe it's more on a language per Language basis, but what is sure is that this is an AI tool that has permissions to automatically do stuff that a human should always review first. If you look at the Fedora guidelines for AI, they have the right way of doing it. Yes, anyone can use AI in any part of the workflow, but the human still makes the decision and the human is held responsible for those decisions. You never want to implement an AI system that decides anything without human supervision because those tools are just not capable of doing the things companies selling them say they Are capable of. They're just not good at those things. They can be very good, but they can also be very hit or miss. And if you don't have a human reviewing this stuff and approving it, then it leads to disaster like it has for the Japanese community for Mozilla. Deploying that bot was a huge blunder by Mozilla. ([Time 0:14:18](https://share.snipd.com/snip/1b8bb69e-e82d-4b30-961f-0fed016f4a58)) - **Ship Graphical System Tools** - Provide graphical system information and administration tools to match user expectations from Windows and macOS. - Linux distributions should ship easy, visual tools for hardware, crash reports, and GRUB settings. Transcript: Nick Now the Linux Mint team showed off a few new changes coming to Mint and Cinnamon. First are some new and improved troubleshooting tools. So the System Reports tool, for example, has received a lot of love. It was rebranded System Information. It now shows a lot of technical details about your hardware and your software, including System Reports and Crash Reports. Think System Information tool from KDE, but without as many entries in the sidebar. It still gives you a view of plugged-in USB devices, of your GPU feature support, of PCI devices that are detected, and some BIOS information. All of this is graphically handled, unlike KDE's tool, which often just prints some text without sorting it in a nice little graphical list. A new tool is also coming called System Administration. It, for now, only handles the boot menu settings, so the grub settings. It lets you hide the grub menu, it lets you change the duration if you want to see it, and it also lets you add boot parameters. On top of that, Cinnamon also gained more symbolic icons used by their ex-apps, and the new menu design has more config options, notably to move the search bar to the bottom of the menu And to move the system power buttons in the sidebar instead of the main panel. Now those are important changes, those graphical system information system management tools are absolutely needed, people coming from Windows or Mac OS expect those things to be Graphical and yes, the command line lets you do so, but having a graphical way of looking at all of this is crucial. ([Time 0:17:06](https://share.snipd.com/snip/19f12796-7806-4748-a402-f069417ac5b3)) - **Flatpak Development Resumes** - Flatpak regained momentum with version 1.17 and work on permissions, OCI support, and session restore. - Backward-compatible permissions and a systemd query daemon aim to ease developer and distribution integration. Transcript: Nick Now the Flatpak packaging system is making progress again. After a worrying situation earlier this year where maintainers had apparently left the project and pull requests were not reviewed anymore, the project is back on track with the release Of Flatpak 1.17 this week, the first one in six months. They've made progress on a bunch of future stuff, for example defining which Flatpak apps are pre-installed on a system, with the Flatpak tool installing and removing stuff according To that well-defined list. It can also allow for installation of OCI images and remotes. There's also a backwards compatible permission system, meaning apps can move to the new permission system that's more restrictive, without breaking the application on older systems That don't have the latest version of Flatpak, which is good because Flatpak is supposed to be packaged once, released everywhere, so you do not want a backward compatibility problem. Documentation has also been updated to reflect all of that. There are also plans to work on a systemd daemon that would let the system query running app instances so they can be authenticated. This would, as far as I understand it, allow to retire the older socket system that is apparently a burden on developers. The application will be able to communicate its identity more clearly to the system and stuff like Wayland compositors, pipe, and other things could just plug into that more easily. There's also some progress on the Session Save and Restore portal that applications could use to restore their sessions. It's good to see the format evolve a little bit. I had no idea that the project had lost a bunch of maintainers and was basically stagnating, so it's good to see it resume some work. I personally am a big proponent of Flatpak. ([Time 0:18:49](https://share.snipd.com/snip/3da57b7c-f32c-48ee-9eda-d6bbe1079b60)) - **Rust Maintainers Fund Launched** - The Rust Foundation launched a Maintainers Fund to support language development and tooling. - Funding targets the language itself, aiming to improve docs, integration, and long-term maintenance. Transcript: Nick We do need the remaining issues with it to be fixed, and so having more people working on that is good. The Rust Foundation also announced some funding efforts with the Rust Maintainers Fund. The goal is to provide support long-term for the developers who work on the Rust language. It's just been announced, so it still needs to be structured to get some contributions that they can redistribute, so they need to find some companies that would give them money that They could use. It's just the beginning of this thing. But it's good to see some efforts on that front as well, especially since Rust is starting to be used more and more in critical projects across the entire Linux ecosystem. The Rust Leadership Council and the various projects team will identify the priorities and they will push financial support towards the developers tackling these priorities with As transparency as possible. The fund is operated as a non-profit and of course the money will come from companies who use and want to support Rust's development. Now this funding will go specifically towards Rust as the language itself, not to projects being built using Rust. It's to evolve the language, solve the various problems that people could have with it, and implementing stuff that the language cannot do right now and will be able to do. Maybe even some bindings with other stuff. ([Time 0:20:43](https://share.snipd.com/snip/6434c8d1-4a59-4252-a42a-0f8bbcfbe07b)) - **elementary OS Adds A Rich System Monitor** - elementary OS added a detailed GTK4 system monitor and many UI refinements ahead of 8.1 and future 9. - Their focus on polish plus more settings aims to broaden appeal while keeping the platform cohesive. Transcript: Nick They dropped their own system monitor, which will be a default app going forward. It is built with GTK 4 and it looks really, really good, with beautiful graphs but also tons of information. It also of course has the usual list of processes with details about the process you selected appearing in a sidebar letting you quit or terminate processes seeing the list of opened Files for that process. I honestly wasn't expecting an app that complete and detailed from elementary. They generally tend to stick to much simpler design and this one manages to look clean and also show plenty of information so well done to the developers. They've also published many bug fixes to their window manager, which fixes high DPI issues, animations, menu positioning, and also gives better performance. There's also a new blur effect in the alt-tab switcher. This will also be applied to the dock and the notifications in the future. Folder icons got a redesign a bit more rounded to match the rounded corners of the elementary OS applications, and other icons were revamped as well in the system settings. When apps are missing certain icons, they will automatically fall back to Odweta now, so there are no missing parts in the UI. Their code text editor slash IDE now has notifications to let you know when cloning a Git branch has completed. The file manager now lets you configure the date and time format that is being displayed in the file manager. And there are some small updates to the system settings and lock screen as well. As per what's coming, they're going to publish a complete package, which is elementary OS 8.1, with all the improvements since the release of elementary OS 8, plus probably the hardware Stack, the latest one at least, from Ubuntu 24.04 LTS and so all big new features will now be targeted for elementary OS 9. It doesn't have a test image yet, it will be based on Ubuntu 26.04, but it likely won't arrive for months after the LTS release, because elementary OS tends to have their major versions Quite late after the Ubuntu LTS drops. ([Time 0:22:34](https://share.snipd.com/snip/335753b2-b9d4-4c9e-8833-a1e3843e6a7c))