OSnews

About five months after the previous release, the MorphOS team has released MorphOS 3.21. It brings the usual list of big fixes and smaller improvements, but there's a few big ticket items in here as well. MorphOS 3.21 introduces BasicWorks, an office suite made up of CellBasics and PageBasics. It offers document and spreadsheet authoring and loads and saves Microsoft Office and OpenOffice documents. Also new are NVMEMount, an auto-mounter for booting from and mounting NVMe partitions, and the Marbles and Cyber 3D blankers. Graphics performance has been improved, with a faster Radeon driver on cards with little video memory thanks to GART backed bitmap transfers, quicker TinyGL texture copies, hardware accelerated antialiased text in TTEngine and much faster color SVG glyph rendering in Cairo. The Bcm43xx wireless driver has been substantially reworked, with automatic transmit rate selection, updated firmware and numerous connection fixes. PowerBook and iMac hotkeys are now fully configurable. MorphOS 3.21 release announcement The BasicWorks office suite has been in the works for a while now, and more information on it can be found in a thread on the MorphZone forums and on the website for the spreadsheet component of the suite, CellBasics. The formula loaders and xlsx loaders are ports of open source projects, but the rest of the application is new code based on ObjectiveC++ and Cairo/Harfbuzz. It uses its own file format, but can import csv and xslsx files, so I'm assuming the same more or else applies to the word processor (but with docx, of course). Alongside MorphOS 3.21 also comes a new release for Wayfarer, the web browser the Amiga-inspired operating system uses. Wayfarer 12.0 moves to WebKitGTK 2.52.5 (which was released in July of 2026, so it's quite up to date), and now uses the Skia graphics library. R300-R500 Radeons will render websites with Skia-GL for some massive performance improvements, helped further along by a slew of additional performance optimisations. Adblocking has been improved as well, YouTube should work a lot better now, and much more. The full release notes for MorphOS 3.21 give you all the details.

Ask any program to listen on port 0 and it will hand you a socket on port 47193...or 58037...or whatever the kernel felt like that day. Port 0 is not a port you can have. It is the number you pass when you want the operating system to pick one for you. But a TCP header is just bytes. The destination port field is 16 bits wide, and 0x0000 fits in 16 bits as well as anything else does. Nothing on the wire stops a packet from being addressed to port 0. Your tools will refuse it, your browser will block it, and your scanner's default port list skips it, because everyone agrees it isn't a real port. Which makes it a pretty good place to hide things, whether anyone meant to or not. So we did some TCP cowboy shit & port scanned the entire fucking internet for open ports on port 0, which uncovered some interesting unexpected findings. acidvegas An awesome and entertaining read. Don't skip this one.

A little tidbit of news regarding the nest major update to Windows, known as Windows 11 27H2. Apparently, it's going to come with some more serious changes to the codebase, therefore necessitating a full operating system swap instead of a mere enablement package, much like Windows 11 24H2 was. For those who need a refresher, Microsoft has long been rumored to be running on a three-year release cadence and for a while, it was even thought that the company would launch Windows 12 in 2H 2024. That evidently didn't happen, as instead we got Windows 11 24H2, but it was a major upgrade, a full OS swap compared to Windows 11 23H2 and 22H2. Essentially, it looks like every three releases, Microsoft is performing a full swap with major changes in the underlying codebase, and it seems to line up with Windows 11 27H2 as well. At the moment it's not known if system requirements will change or not. Sayan Sen at Neowin.net Microsoft is currently testing adding support for Intel's Advanced Performance Extensions, which should yield improved 64bit performance and reduced power consumption. It makes sense to assume this support will be coming with Windows 11 27H2, which should hopefully lead to tangible performance improvements across the board. Also, there's going to be a new default wallpaper. Be still my beating heart.

Ubuntu has been at the forefront of replacing the traditional versions of basic tools like sudo with replacements written in Rust, but aside from worries about replacing battle-tested code with brand new, relatively untested code, there's new things to worry about it. As it turns out, sudo-rs, the Rust replacement for sudo, is intentionally incompatible with the classic version in dangerous ways, as Chris Siebenmann discovered to his dismay. Basically, the classic sudo supports the use of wildcards defined using square brackets, so that if you allow a user to run, say, /usr/sbin/exim -bp, without a password, they can run exim -bpa, exim -bpc, exim -bpr, or exim -bpu without having to supply a password. This allows for very specific, fine-grained access control, which is quite a nice capability to have if you're managing many systems and users. Sadly, sudo-rs intentionally does not support this, opting to only allow you to use the generic * wildcard. Worse yet, it doesn't really make this clear, either. Chris Siebenmann I do not really have words for this. The Rust sudo has apparently opted to keep the single most dangerous wildcard while disallowing all of the wildcards that allow one to be narrow and specific about what you're allowing, and it buries this at the end of a section that devotes a bunch of words to describing functionality it doesn't have and doesn't support, with absolutely nothing else in that section to say this (either up front or in the descriptions of how all the wildcards would work if sudo-rs supported them, which it doesn't). If you're going to replace such vital, core tools like sudo that god knows how many people rely on every single day, you better be damn sure you're feature-complete and can act as a drop-in replacement. If someone switches from sudo to doas, nobody expects the experience to be identical. However, if you're consistently being told that sudo-rs is a drop-in replacement and you shouldn't expect any disruptions, you better be able to make true on that promise. Ubuntu isn't some also-ran small hobby distribution, and its developers should act like it. What a clown show.

The second desktop environment release I've got for you today is LXQt 2.5.0. This new version excises its last LXDE dependency, and QTerminal and QTermWidget received considerable improvements, like new background image modes, a visual bell, support for OSC-8 hyperlinks, and more. There's a ton of smaller changes and improvements as well, but do note that due to a bug, 2.5.0 broke LXQt's Monitor Settings on X11. The project intends to fix this as quickly as possible and push a point release to address this issue.

It's desktop environment Thursday, as I've got two new releases for you today. First, Cinnamon, the desktop environment primarily associated with Linux Mint, has hit version 6.8. This version will officially ship with Linux Mint 27, scheduled for around Christmas of this year. Highlights of Cinnamon 6.8 include hibernation support via a Suspend action in the main menu, modern one-to-one trackpad gestures for both Wayland and Xorg sessions, support for the oo7 Secret Service provider, support for multiple instances in the Sound applet, and a more unified Network applet that merges the WireGuard and VPN sections. Marcus Nestor at 9To5Linux While Cinnamon is first and foremost a Linux Mint project, it's available on countless other distributions as well, including dedicated variants for popular distributions like Fedora and Ubuntu. Its Wayland support is finally materialising too, which is long overdue.

The first point release for NetBSD 11 has been released. As the version number suggests, there's no major changes here. This release includes security updates for mozilla-rootcerts, brotli, xorg-server, hdaudio(4), compat_linux(8), unbound(8), and ipfilter. Among the non-security related changes are improved USB device stability on Raspberry Pi, fixed hot-plugging bugs in various USB Wi-Fi drivers, support for battery monitoring on the Microsoft Surface Go 3, fixed thread-local storage on ARMv5, fixes for running the MIPS Malta and PA-RISC ports under QEMU, support for the Zhaoxin KX-7000 CPU, and support for sensors, i2c, real-time clocks, and fan control on Sun Ultra 25/45 and Sun Fire V215/V245 machines. Nia Alarie on the NetBSD blog The full release notes provide more details.

Another interesting post from Raymond Chen, this time about the difficulty of dealing with a system that's become CPU-starved. The question is simple: why not just run Task Manager and important parts of the UI like the taskbar at a higher priority? The catch with running anything at high priority is that unless you have been very careful to avoid it, you will eventually run into a shared resource that you will have to wait for, and you are now stuck. Your priority checkout pass gets you to the front of the line, but if the person at the register has a large shopping cart full of groceries, you just have to wait until they are done before you get your chance. Even if the code that protects the shared resource is careful to avoid priority inversion by transferring the priority of the waiter to the thread that owns the resource, you are still stuck waiting, and that owning thread might be doing something that takes a long time, like, say, accessing a file over the network. Raymond Chen If a system has reached a state where even accessing or using something like a task manager becomes a problem, I personally just bonk my computer over the head with a forced shutdown using the power button. Honestly though, I can't remember the last time I've had to do this, as my Linux machines all seem to handle problematic applications just fine. That being said, I'm sure people managing tons of servers and/or workstations and developers debugging their own running code run into catastrophic problems like this far more often, making forced shutdowns not as practical. Have a terminal window open with top and easy access to the kill command, or make sure you can just ssh into it remotely if the UI has gone completely nuts.

What if you want to change the time in your car's infotainment system, but it's so locked down that you can't even open the time settings menu without a very specific, DRM'd SD card? Well, you try to jailbreak it, of course, but the amount of effort required can kind of spiral out of control. And then, all the way at the end, it's just an offset. The first thing I did was use this to create a full backup of the internal eMMC using dd. With a more complete dump of the system, I pretty quickly figured out that the date & time settings that I'm interested in are in /mnt/flash/settings.dat, which is a custom binary format. After some more reverse engineering, it turned out the unit doesn't really have a concept of timezones at all. It gets the time from GPS and stores the local offset as a number of minutes. Mine was set to 180. I changed it to 120. That is how I was finally able to fix my car's timezone. Rasmus Moorats This harrowing story involved an older car - a 2014 Renault Latitude - but with the number of screens and the amount of user-facing software only increasing every year, requiring jailbreaks to get your car to do what you want it to do are only going to become more common. I wouldn't be surprised if whole generations of cars are going to become much harder to sell in the near future, cratering the used market, because people are not willing to put up with software problems nobody wants to or can fix. He just wanted to change the goddamn time.

Enlightenment, the long-running desktop environment and associated libraries and frameworks, just released version 0.28.0. It brings initial support for NetBSD, support for Network Manager and a whole slew of related changes and fixes to make that possible, as well as native PipeWire support. There's also a brand new clipboard manager, which brings with it modernisations and bug fixes for related UI elements. This release also reworks a ton code related to battery management, AC polling, and more, adding features like setting battery thresholds, ACPI power profile support, and so on. Of course, there's also the long list of smaller fixes and changes. You can download the release and build it yourself, or wait for your distribution of BSD variant to package it.

A little while ago, we talked about 9front's Good Enough File System being ported to OpenBSD. A lot of progress has been made on this front lately. Aside from fixing bugs and porting over fixes from the 9front side, there's also been a ton more substantial changes. Some userspace tools have been done (you can now mkfs on openbsd!), but snapshot control hasn't. I'm thinking we have a userspace gefs' program that lets you handle this, trying to keep the interface similar to the 9front ctl file. Ori Bernstein There's still some major hurdles to overcome, some of which have already been addressed on the 9front side, but the progress is clear. Considering OpenBSD desperately needs a more modern file system, this is great news. A fun sidenote is that I'm actually posting this news item from OpenBSD.

From a global allocation perspective, actually, it's quite terrible to have everybody have a bunch of the world's limited supply of DRAM sitting in their own computer and being used 5% of the time," Altman said. You want maximum utilization in the cloud if you're trying to deal with this very supply-constrained environment. And somehow many people's intuition is the opposite of that." Andrew E. Freedman at Tom's Hardware I'm installing a 32bit operating system without support for physical address extension on my laptop with 64GB of RAM, just because I can.

Sculpt OS 26.10 brings quite a few new features, like scalable cursors, support for UK keyboards, better webcam support, and more, but the real meat of this release is much tastier. First, Sculpt 26.10 is the first version based on Genode's custom microkernel instead of relying on a 3rd-party kernel. This kernel - designed from the ground up for the needs of Genode and Sculpt OS in particular - is not only more micro" than any other of the microkernels supported by Genode, it also solves long-standing deficiencies concerning CPU scheduling and kernel-memory management. Sculpt OS finally stands on Genode's own footing! Finally, Sculpt OS 26.10 pays tribute to this year's overall theme of building bridges" by featuring our software development kit (Goa SDK) running directly on Sculpt OS without virtualization. This entails a wealth of tooling including GCC, Binutils, Git, Sequoia, Expect, Curl, man, Make, and CMake having landed on Genode. Although still young and not time-tested yet, it shows what's possible today and allows us developers to gradually shift more and more development routine into this self-hosted environment, stressing the general-purpose qualities of Genode and motivating further optimizations. Sculpt OS release 26.10 announcement This new release also completes the transition from XML to Genode's own HID language for configuration, moves from VirtualBox 6 to 7, and upgrades a whole slew of its Linux-based drivers

Raymond Chen published another story from the Windows development trenches, this time about what happens when you hold down the power button. If you hold a PC's power button for 12 seconds, it will turn off at the hardware level without notifying the operating system. This is a safety feature so that you can turn of the computer even if the operating system's Shut down" command isn't working. After eight seconds of the power button being held down, the Windows kernel realizes that the user is trying to do a force-power-off of the system, and it quickly takes a kernel dump that can be uploaded to the Windows Error Reporting service the next time you turn the system on. Raymond Chen I'm a bit confused by these numbers, because I have never seen a PC take this long to power off when holding down the power button; usually it takes only a few seconds to do so, meaning Windows would never generate said kernel dump. I tried reading through the various ATX specification documents to find more information on this, but then I remembered I'm not an electrical engineer so I have no idea what I'm looking at. The baseline is that ATX always provides standby power to the motherboard, and pressing the power button shorts the power button pin to ground, telling the PSU to power on fully. Pressing the power button if the computer is running tells the operating system to initiate a shutdown (although you can change this behaviour), but if you hold the button down for a certain number of seconds, it forces a power-off. So far, so good; basic information. It seems that the length of time you need to hold down the power button to force a power-off is dependent on the motherboard in question, handled by a microcontroller or some other IC. Virtually every source I can find says this timing is set to about 4-5 seconds, which is quite a bit shorter than the number of seconds given by Chen. Perhaps motherboard manufacturers consider the 12 seconds to be too long, and just do their own thing, regardless what the ATX specification suggests? What further complicates matters is that Chen details how the kernel dumps created after pressing the power button for 8 seconds do, in fact, get sent to Microsoft once the computer turns on again. In fact, he is one of the people who gets to see these dumps and respond to them, so obviously they are indeed rolling in. Does this mean there's a huge number of computers out there that require you to hold down the power button for 12 seconds to force a shutdown that nobody's ever seen? These emergency kernel dumps are internally classified as long power button hold" (LPBH) failures. One of the pieces of information recorded in these emergency kernel dumps is whether the screen consists entirely of black pixels. Those failures are sub-classified as black screen LPBH", and they go to a special virtual team for analysis. (I happen to be a member of this black screen LPBH team.) Raymond Chen What's going on here? Obviously, I'm not going to doubt Chen, so there must be something here I don't quite understand or that I'm missing. The ATX specification might suggest 12 seconds, but could Windows perhaps adapt its kernel dump behaviour - dumping much sooner than 8 seconds - based on the motherboard it's running on, maybe through information provided by the chipset drivers? I would love it if someone more knowledgeable than me can help us unravel this small mystery.

From curl 8.12 through 8.16, config files could slice variable values and include other config files without an explicit depth cap. Together, these features let curl simulate any two-counter Minsky program. Under the usual idealization of unbounded strings and resources, this construction was Turing-complete. Curl 8.17.0 added a five-level include cap that breaks this reduction. Nicolas Seriot Neat.

AlpineBSD is an operating system based on the NetBSD kernel with a userspace inspired by Alpine Linux and the objective of making it possible to use as little AI-generated software as possible. There is no LLM output included in the base system or main package repository (AlpineBSD AI policy). AlpineBSD's website AlpineBSD uses the NetBSD kernel, libc, and essential tools, and combines them with Alpine's apk-tools package manager, busybox, and generlaly preferring GNU tools over their BSD equivalents. It uses the s6 init system, but the project intends to add OpenRC as an alternative. The project is in its very early stages, so don't expect things like an installer; you're on your own, basically.

The Nintendo Wii (and its related consoles) were already made to run the PowerPC version of Windows NT as well as Mac OS X, but now we can add another milestone to the list: classic Mac OS. So today I reached a nice milestone in one of my projects: my work porting classic Mac OS to PowerPC systems with what I affectionately call ArtX chipsets (Nintendo GameCube, Wii, eventually Wii U; the GameCube's Flipper integrated chipset was very much designed by ArtX as it was effectively finished when ATI acquired the company, and the Wii's Vegas (and its later revision Bollywood) and Wii U's Latte are effectively straight successors) reached the desktop for the first time, most likely the first time since the CHRP clone era that someone outside of Apple brought up classic Mac OS on a new system to the desktop. Despite still being nowhere near ready for a public release, I want to talk about the challenges I faced with this project so far. Rairii at Virtuallyfun Various versions of Mac OS boot to the desktop, including Mac OS 9.2.2, the last ever version, but there's still a ton or work to be done before this is ready for release. Still, with Rairii amazing track record of porting all kinds of things to various PowerPC machines, including getting the PowerPC version of Windows NT to run on various PowerPC Macs, I have little doubt they'll manager to get this to a releasable state. I absolutely love this kind of work.

I think it's safe to say that Windows' Modern Standby feature isn't exactly free from issues. The biggest problem seems to be that it's not very good at actually determining when to transition between regular sleep and hibernation, often leaving users with an empty battery as they open their laptop's lid. Microsoft has now released an update to hopefully address this long-standing problem. This update improves the adaptive hibernate policy to provide more instant-on resumes from Modern Standby while helping preserve battery life when the charge is low. Noticeable changes include less hibernation while the battery is above 80% and hibernating sooner when the battery is below 10%. KB5127936 Among other new features and changes this update brings are Cloud Rebuild, a way for devices to install Windows without needing a local image. Cloud Rebuild can be started form within an existing installation, remotely by administrators, or through the Windows Recovery Environment. Obviously, a working internet connection, wired or wireless, is required for this feature to work. This is the same update that introduces CLAT, which we talked about yesterday, to smooth the transition from IPv4 to IPv6.

Flag LinuxNetBSD made a seemingly small, but very welcome announcement today: they're adding a more user-friendly, easier to use command-line text editor to the base install. To ease the transition for new users, NetBSD -current now includes an alternative editor to vi(1) and xedit(1) in its base system, atto(1). It is a modeless screen-based editor similar to nano, pico, xedit, or emacs. It is only 48 kilobytes, making it suitable for inclusion in the rescue system. NetBSD Foundation on Fedi Most guides and how-to article for operating systems like NetBSD assume the use of vi, and while I don't want to ignite a flamewar, vi isn't exactly the most approachable or friendly piece of software out there. My occassional text editing in the command line is all done through nano, but now I'm definitely curious to give atto a go. While other tech news websites are plastered with articles about a PC maker adding a touch screen to one of their laptops, we're talking about a new CLI text editor in NetBSD's base system. Priorities, folks.

Windows 11 25H2, starting with build 26200.9550, now comes with something called CLAT. As organizations move toward IPv6-only and IPv6-mostly networks, some applications still depend on IPv4-only applications or services. Windows CLAT helps those applications continue to work when a network does not provide native IPv4 connectivity. CLAT is the client-side component of464XLAT. It statelessly translates IPv4 traffic from the Windows client into IPv6 for transport across the network to a provider-side translator (PLAT), which provides stateful NAT64 translation to IPv4. Jim Alumbaugh at Microsoft's Networking Blog Yet another important component of the long, neverending transition to IPv6. Like endlessly cutting a cucumber in half, you'll never quite run out of cucumber to cut.

Do you want a hobby operating system with a desktop for a microcontroller-based computer? Sylphide OS is a new project aiming to build just that. For the last few weeks, I've worked on a new pet project: an operating system for computers based on microcontrollers. There aren't many yet, but I expect them to be more and more common. Why MCU computers ? You get a relatively very simple platform, booting straight to your code (or almost), and you can understand the whole architecture. Exactly as it was with home computers long ago. And not at all what you get with modern computers. Even booting involves lot of very different systems, black magic, and even blog/proprietary software. Thomas Capricelli It's very early in development, but by building it atop FreeRTOS, Capricelli had a head-start. It's got a basic desktop, one that might need some changes to avoid the lawyers. The code's open source, of course.

What's the oldest Windows GUI application you can still run on Windows 11 today, without any modifications? For now, the best answer Boris Popik can come up with is an innocuous application from build 347 of Windows NT 3.1, an early prerelease: CDPLAYER.EXE. You can't go any older than that, because earlier prerelease builds had not yet finalised its PE format and calling convention. Neat find. Popik also mentions a few other contenders, but I won't spoil those here.

Another month has crumbled to dust under the oppressive boots of time, so it's time for another month of Redox improvements. This past month, Redox' build system gained support for Clang, SDL2 saw performance improvements, and WebKitGTK3's HTML layout rendering was fixed. Users of Intel Tiger Lake processors can now enjoy external display hot-plugging, and overall system performance has been improved in a variety of ways. And as always, there's the long lists of kernel, relibc, and driver improvements and fixes as well. Redox keeps making big strides every month, and it's definitely one of the few alternative operating systems that actually feels like it might, one day, be an actual contender for general purpose use.

Thew day ends in a Y, so Microsoft is launching a new way to build Windows applications. Get excited, Windows application developers! If you caught our session at Microsoft Build 2026 or you've been watching the repo, you've seen this one taking shape in public for a while now, and a lot has changed since Build. Here's the longer tour: Microsoft.UI.Reactor, an experimental C# framework that brings the MVU (Model-View-Update) pattern to WinUI desktop apps. No XAML files, no code-behind, no ViewModels. Two things before the fun part. Reactor is experimental: no support policy, no API guarantees, published this early on purpose because we'd rather hear we got the design wrong now than once people have shipped on it. And it is not a replacement for XAML or MVVM, which we'll come back to. Alexandre Zollinger Chohfi at Microsoft's Dev Blogs At multiple points in the linked blog post, Microsoft underlines that no, Reactor is not going to replace XAML. I'm sure Windows developers who learned Win32, Silverlight, WinRT XAML, WPF, UWP, WinUI 2 and 3, and all the others, over the past two decades will feel deeply comforted by Microsoft's insistence that Reactor will not replace XAML. Microsoft repeats: this won't replace XAML. They swear.

We've been following along with Carmen Kaiser's Apple Network Server exploits, and by now things are really starting to come together for owners of this machine. When we last left off, the ROM to allow Mac OS to run on the machine which normally runs IBM's AIX, but there were still a number of bugs, limitations, and issues to be worked out. A lot of progress on that front has been made. The good news is, many of the issues with the preproduction ROM in Mac OS can be fixed or worked around in software. So I've fixed or worked around many of them in software: the MACE Ethernet is fixed, the on-board video is now supported, and the ANS can restart properly back into Mac OS. Plus, because it's the preproduction ROM, you can still boot AIX, thus preserving the entire historical reason for Shiner's existence. Here's the first cut of that software and how it was developed. Cameron Kaiser There's still two issues Kaiser wishes to address - the internal SCSI bays don't work, and the ROM doesn't properly register parity RAM - but the progress here is amazing.

Thomas Adam, one of the core developers of FVWM, has been working on CoW for a while now. CoW can best be described as FVWM for Wayland", which uniquely uses the River compositor. River, in turn, is unique in that it does not combine the compositor and window manager into a single program, allowing for a more traditional X11-like approach. Adam released version 0.4 today, which comes with a load of fixes, changes, and minor improvements, as well as one bigger-ticket item: container windows. This PR introduces the concept of a top-level container window. This means a window can be subdivided into columns and rows (called panes) and can be resized independently of one another. The idea for this has been unashamedly stolen from shod. Thomas Adam I've been keeping an eye on CoW ever since I learned about it when it was first released, and at some point soon I want to build a desktop around it for myself.

Speaking of games, if you've ever heard anything about or even played Sonic '06, you know this wasn't the most polished entry in the hedgehog's franchise. In fact, it's probably one of the worst, and considering the average quality of Sonic games, that's quite an achievement. The team that's working on decompiling the game to make it run on the PC obviously had to dive deep into the game's inner workings, and the process learned a lot about why, exactly, the game is so notoriously bad. It really makes you wonder what exactly went wrong? If you look at what limited footage we have of beta content, we can see many effects that aren't present in the final game. Kingdom Valley had full bump maps, Castle Town had more complex geometry and plants, and New City's collision implies a very different building layout at some point in development. It turns out, much of the cut content is actually still in the game, but is simply not enabled for performance reasons in vanilla. Even with the cut-back visuals, the game struggles to reach target FPS on real hardware. It's not surprising that given the crunch Sonic Team were under to get the game out the door, these features were among the first to go. Isaac Marovitz The article details what, exactly, was turned off for the final release, and it's quite an interesting read. Of course, the team intends to try and address many of these issues, and they think they can push the game and its engine to much greater heights.

When thinking about impressive technical achievements in retro games, most people will first point to graphical milestones set by action games like Doom or Quake. Rightfully so, but at least I am personally far more impressed by the early examples of the city builder genre. Simulating whole economies on a large scale with a fraction of today's power seems impossible if modern builder games even struggle on modern machines. Of course this didn't stop developers in the past to develop and run city builder games on significantly weaker hardware. The original SimCity for example lists Intel's 386SX as the CPU in it's system requirements. A CPU which at the time of it's launch ran at magnificent 16 MHz. Lars Thieen What's even more impressive are some of the other systems the origin al SimCity was ported to, which had even more limited hardware than a 386SX. The Super Nintendo got a SimCity port, which even included fun extras like a Mario statue and Bowser taking up the mantle of the Godzilla disaster, and additional systems the PC version didn't have. Hell, they even worked on a port to the NES, but it was cancelled and thought lost, until a relatively complete and working copy was discovered in 2017. Even this NES version contained features not available in the PC version. Hardware limitations really did breed a lot of innovation back in those days, and sometimes I wonder if programmers and developers could benefit from tasking themselves with working within such limitations today, if only to become a bit more prudent with resources.

Earlier this year, I needed to learn PowerShell scripting for one Windows-related project, and I have tried it out a few more times on Mac. It has quickly become one of my favorite tools for automation and scripting-taking on jobs that I would have otherwise tried to do with Node.js, Shortcuts, macOS Automator, or heaven forbid, Bash scripts. I get it now. PowerShell is great, even if you don't use it as your shell, and even if you're on Mac or Linux. Corbin Davenport I don't use the shell often enough on any of my main machines to have much of an idea about the strenghths and weaknesses of PowerShell compared to more traditional shells, but this is definitely not the first time I've heard a lot of praise for it from people otherwise knee-deep in the UNIX and traditional shell world. There really must be something to this, and since PowerShell is also available on Linux and macOS, it's pretty straightforward to give it a go without having to deal with Windows.

Earlier this year, Garmin closed its Connect API developer program because it was being flooded with AI"-generated applications to the point it couldn't handle them anymore. What made it worse was that half of the applications were never actually used, not even once. It also didn't help that Garmin treated every single application to the program the same way, whether it was from a hobbyist just wanting access to their own data, to a massive corporation developing a popular service used by god knows how many people. Today, the company has relaunched the program, and it's been considerably overhauled. While hobbyist use of the API is going to be free - which is way better than it was before - once you hit a thousand users, you're going to have to start paying. Still, Garmin's shift here towards charging for API access will set a dangerous precedent for the industry. Especially for a company that bills itself as a (or the?) premium sports hardware provider. On the flip side, there's simply no ignoring the flood of mostly AI apps that take up resources at Garmin and elsewhere to support. Heck, my inbox is flooded with these apps daily requesting reviews. It sounds like when Garmin shut off access to the API last spring, they were seeing inbound API requests at a daily scale that they'd previously only seen over a period of weeks. Simply put, there's no good answer. Ray Maker at DC Rainmaker The excesses of the easy availability of AI" code generation impacts the entire industry. At least one good thing has come out of this: hobbyists can now access the Connect API developer program much easier than before.

If you've ever written C (or c++) on a mac, you'll know that the so-called gcc is just clang in disguise. Most people (fairly) assume that this is just an alias, but this is not true! It's a binary! Now, what is this binary? I naively assumed that it would just be shelling out to clang but it turns out the answer is more interesting than that. juliethefoxcoon It's all smoke and mirrors.

Unlike traditional OSs that separate user applications from the kernel, a unikernel bundles everything into a single, isolated image that runs directly on a hypervisor (like KVM, BHyve, or OpenVMM). Because it removes unnecessary drivers and background services, a unikernel is extremely small, fast to boot, and efficient with resources. Here, we will learn how to develop and deploy unikernels built using the OCaml language and the Solo5 unikernel. These are the two fundamental components that will then enable us to develop our own applications in the form of unikernels, which we can subsequently deploy using Albatross. Unikernels in OCaml Some light reading for the weekend.

Graybeards like me remember trying to painstakingly type source code from a magazine. This is a very accurate account of what a young computer user in 1986 needed to do to get a new game for the Commodore 64, complete with an immersive emulation. It's time to install an app! But here in 1986, we still call them programs," not apps. So, you'll want to set aside some time to get started installing a program.

Early 2025, I wrote about how I think macOS' UNIX certification was basically a lie. It required so many convoluted and cumbersome changes to macOS that literally nobody applies to their system that it the certification had kind of become a farce. At the end of the article I wondered for how much longer Apple would be playing this little charade to hammer macOS through a meaningless UNIX-shaped hole, and it turns out the answer was about 18 months". When you go to the Open Group's website and the list of certified UNIX operating systems, macOS is no longer listed. For the past few weeks, it was still listed if you drilled down specifically to the UNIX 03 certification, but that, too, has now been removed. Assuming macOS 27 doesn't return in the near future with a brand new certification, we can safely say that Apple got tired of paying a few employees to get its desktop operating system certified. With one more operating system gone from the list, there's now only one UNIX operating systems left: AIX. While the registry also lists HP-UX, z/OS, and UnixWare/OpenServer, those either only conform to really old UNIX certifications (from the '90s), or simply aren't developed anymore (HP sunset HP-UX at the end of last year). Only one operating system version conforms to the latest version of the certification, UNIX 07, and that's AIX 7.2 TL5 and later", which possibly includes AIX 7.3, the most recent version of AIX. Not that any of this really matters.

Hot on the heels of Wine being ported to OpenBSD, we now have the release of OpenBSD 8.0. It adds hibernate support for arm64, switches the sparc64 toolchain to LLVM, and updates the DRM graphics drivers to Linux 6.18.53, which also adds support for Intel Arc A-Series (DG2/Alchemist) graphics cards. There's a whole bunch of improvements, fixes, and changes for RISC-V and ARM in general, of course, as well as to the various Wi-Fi drivers. On top of that, it also comes with new versions of OpenSSH and LibreSSL, as well as a massive list of other improvements and fixes. If you've never installed OpenBSD before, be sure to read this detailed guide to know what to do once everything's installed, while listening to the official release song, of course.

Apparently, macOS 27 comes with a small mystery people are trying to figure out. Recently, while troubleshooting a PATH issue, I noticed an entry for /pkg/env/global/bin. At first, I worried that something had gone horribly wrong. Turns out, this is part of macOS 27.0... though no one is sure what it's for yet. In a post about how macOS's boot volume structure has been updated, Howard Oakley explains that the pkg directory was added to the Data volume in macOS 26.4, but not added to the root of the Signed System Volume until 27.0. He also notes that this is sort of a big deal. The weird thing is, as far as I can tell, no one seems to know what this new directory is for! It's odd that Apple would update their filesystem structure, add it to everyone's PATH, and there not be some explicable reason. Gavin Anderegg I guess we'll find out eventually, but this does highlight something that I find extremely unsettling about this. With proprietary software like macOS, you literally are at the mercy of whatever Apple decides to do to your" operating system. You install an update, and aside from all the documented changes, there's going to be changes you simply have no idea what they're for, or why they're there. Don't get me wrong, there's obviously not something nefarious going on in this particular case, but it still feels wild to me that people just willingly accept this. In the open source world, stuff like this simply does not happen. You're not going to install a bunch of Fedora updates one day and find a mysterious, undocumented folder in your PATH. Such a change will be documented, but even in the unlikely event it's not, you can just go through the commit history and find out exactly why the new folder is there, and what its intended purpose is or is going to be. Considering the state of the world and the people corporations like Apple really answer to, I would not be comfortable with letting them dictate even something as simple as my PATH.

In what is quite a major achievement, Wine has been ported and made to work on OpenBSD. We have since tested several dozens of 64-bit Windows software, mostly successfully. See screenshots, livestream recordings for example. Even the Remaster of The Witcher 3 which released just last month runs, and runs well (requires adding in vkd3d-proton; see README; livestream which looks less fluent than it is due to the on-system recording with ffmpeg). Thomas Frohwein The biggest limitation seems to be the lack of support for WoW64, so 32bit Windows software won't work. Still, this is a massive milestone for OpenBSD, and makes especially gaming a somewhat viable option on OpenBSD.

What if a regular, normal NetBSD installation is too plebeian for you? This is a quick overview which attempts to synthesize DW Armstrong's tutorial on NetBSD encryption with the cgdroot page from the NetBSD wiki in a single MBR partition, divided into multiple disklabel slices to preserve MBR entries. indigoparadox Web Zone What if you need this?

Significant changes in this release include some significant binfmt_misc extensions, some group-scheduling work aimed at improving performance on large systems, the new failfs filesystem that fails all operations, the fchroot() system call, global per-CPU data for BPF programs, the completion of the sched_ext subscheduler feature, working-set tracking for virtual machines, alternative data-stream support in the ntfs3 filesystem, a long list of SMB server improvements, and much more. corbet" at LWN.net It'll make it way to your distribution of choice soon enough.

In 1980, Intel introduced the 8087, a chip that made floating-point operations much faster in the IBM PC and other systems. Although the 8087 supports advanced operations such as tangents, square roots, and logarithms, this article focuses on basic arithmetic operations: addition and subtraction. The 8087 provided a spectacular speed improvement over the 8086 processor, speeding up addition by a factor of almost 100, and complicated operations by even more. In this article, I examine the 8087's microcode and circuitry to explain exactly how floating-point addition and subtraction are implemented, including numerous special cases for infinities and other special values. Ken Shirriff Shirriff took an actual chisel to the 8087 coprocessor to gain access to the insides for highly detailed photographs, taken with a microscope. They look amazing.

Only a few day ago I linked to a massive collection of original Palm OS development resources, and today we can take a look at what it was like to put all of those resources into practice: the original source code for the Palm OS port of Lemmings. With these resources combined, there's no excuse to not building any Palm OS applications. I'd really like a modern email client, a Fediverse client, and a slightly more modern browser surely wouldn't hurt. Get on it!

It's been two years since my initial post about making GNOME OS an end-user general purpose OS. Today, we're excited to announce that we've hit a major milestone! TL;DR: We're launching a new edition called GNOME OS Preview", which is built from the latest stable GNOME release. This means that GNOME OS is no longer just a nightly OS for GNOME developers and testers! Aaron Ardiri A number of changes and improvements have made the move from nightly status to preview status possible. They've made GNOME OS handle environments with limited amounts of RAM much better - not a luxury, these days - including the ability to dynamically increase the swap space if needed. In addition, the work to make it much easier to install and test daily Flatpak builds of applications is now part of GNOME OS, as is the effort to make command-line utilities manageable in the same way as Flatpaks. As the blog post notes, there's still a ton of work left to do before GNOME OS is ready for everyone. They're going to bet heavily on systemd-homed to enable, among other things, per-user disk encryption, and there's a new installer coming, as well. They also want to ship GNOME's preinstalled applications as Flatpaks, integrate delta updates, and much more. It all looks incredibly appealing, even if I am not the target user for something like GNOME OS.

This repository contains approximately 1 GB of historical Palm OS development material, collected over many years of working with the platform. The archive includes a mixture of publicly distributed development resources, personal development material, technical references, and historical artifacts. Aaron Ardiri An incredible resource for anyone wishing to dive into Palm OS programming and development in this day and age.

OpenVMS isn't exactly the easiest platform to get into, with even the first major hurdle - licensing! - being a bit of a complex issue. To address this and other issues people new to OpenVMS might run into, VSI, the company currently developing OpenVMS, launched its new VSI How-to Portal. For now, it covers things like licensing, installing and updating OpenVMS for x86, TCP/IP configuration, and setting up SSH. The company plans to expand the portal with more articles based on customer and user feedback, so hopefully more information will appear soon. Most of us are probably not particularly not particularly experienced with OpenVMS, so having a portal like this can really help get you started.

Microsoft's been claiming for a while now that they're working on improving performance and responsiveness in Windows, with a particular focus on making more efficient use of memory. This is probably the only positive thing to come out of the RAM crisis Microsoft helped create and is currently still fueling. The company's now claiming that Windows 11 is up to 20% more responsive" thanks to these efforts. The result is a Windows experience that's up to 20% more responsive, with improvements you can feel as soon as you boot into your PC and extending to Start, File Explorer, Search, Snipping Tool, Microsoft 365, Edge and more. A big part of that work is making smarter use of memory and system resources. We've improved memory management and efficiency across Windows to help keep more of your PC's resources focused on what you're actively doing. That work can make an especially meaningful difference on PCs with 8GB of memory. Windows can now keep up to 2GB more of your active data in memory instead of moving it to slower storage, helping your PC stay responsive as you open more apps or browser tabs. Windows is also making additional memory-saving improvements that reduce background memory use. For example, on PCs with 8GB of memory or less, Widgets load only when you open them, making more than 200 MB of memory available for your apps and other work.It's one example of the broader work underway to reduce the Windows memory footprint and deliver a fast, responsive experience across the PCs people use every day, an area we've made an explicit focus for Windows. Navjot Virk on the Windows Blogs If, for a moment, we take these claims at face value, they're simply good news for every Windows 11 user, and legitimately the only good thing to come out of the RAM crisis. These kind of claims are always hard to measure, especially considering just how varied the Windows software and hardware ecosystem is, but I'm sure people good at benchmarking are going to try and find some more definitive answers here. In the same blog post, Microsoft also highlights several other recent improvements it's made to Windows 11, like the new search experience we talked about recently, a brand new Photos application, some new whiteboard-like application, and several more things. Considering how complex Windows' versioning scheme and various branches have become, I have no idea when these features and improvements actually make their way to your machine. On a slightly related note, Microsoft also unveiled that it's bringing Microsoft Defender to Linux, with a particular focus on desktop Linux. Even Microsoft has to face the reality of the rising popularity of Linux on the desktop.

Graphical user interfaces for DOS were a dime-a-dozen back in the '90s, but of course faded into irrelevance as time went on. But what if you still run DOS today, and want a maintained graphical shell instead of the rather limited DOS command line? Costa is a small and lightweight application which runs on top of MS-DOS or compatible operating systems. It replaces the DOS-prompt with a clean, user-friendly interface - a so-called GUI (Graphical User Interface). Costa comes packed with a bunch of useful accessories and can be used with or without a mouse. Developed by Jacob Palm, a Danish IT-professional and hobby developer, the first release of Costa came way back in 2004. Over the years - and through several releases - Costa has gradually evolved into what it is today, with new releases still coming out regularly as of 2025. Costa website Costa comes bundled with a number of useful utilities, a few games, and useful features like a desktop with custom icons, virtual desktop support, a file selector, and much more. The latest release, version 1.8.0, was released in late 2024, and should run on just about any IBM PC-compatible, but a 386 is recommended. It's been tested with a variety of DOS flavours, including FreeDOS ad DOSBox, so it should run pretty much anywhere.

Ubuntu 26.10 has landed. Ubuntu 26.10 brings significant progress for security, and strengthens the platform with the full adoption of Rust-based coreutils, post-quantum-ready package signing, and expanded TPM-backed full-disk encryption support. Enterprise authentication gains deeper integration with Microsoft Entra ID, including password and multi-factor authentication. Linux kernel 7.3 extends support for the latest hardware, alongside new RISC-V desktop images and optimized amd64v3 installation images. Ubuntu 26.10 release announcement Of course, it comes with the latest version of GNOME, and the whole family of Ubuntu variants focusing on other desktops have also been updated. While I doubt many people here run Ubuntu, it's still quite popular, and serves as the base for countless other distributions.

KDE turns 30 today, and they're celebrating it with a brand new KDE Plasma release: version 6.8 is here. On the accessibility front, 6.8 brings a new dwell clicker function (meaning a click will register if the pointer is held still for a short while) and support for the Bikram Sambat calendar and the Amharic language. I always love it when I see support for new languages and scripts, especially when, as is the case here, it's a language spoken as a first language by 25 million people. The Plasma remote desktop has also been improved with bi-directional copy/paste, performance and quality improvements, and more, and you can now record microphone and audio while screen recording with Spectacle. In System Monitor, you can now assign processes and tasks to specific processors/cores, which is a great feature for people running intensive tasks but who would still like to, you know, do other things while those tasks are running. KDE 6.8 drops support for X11, focusing entirely on Wayland. Of course, X11 applications will remain usable thanks to Xwayland, and KDE's applications will continue to support X11 in case you're using them outside of KDE. This release also further expands the Union theming system, now adding support for styling Qt Widgets. Union is still in beta, so you'll have to enable it to test it. There's a ton more, too, but these are some of the highlights. KDE Plasma 6.8 will find its way to your distribution of choice soon enough.

You know the drill by now - a new month, a new Haiku activity report. This past month doesn't seem to have any big ticket items, but instead contains a long list of smaller but still welcome fixes. Various filesystem drivers saw some work, a few bugs in virtio drivers were fixed, and NVMM execution is now easily interruptible by signals which should improve performance in various emulators like QEMU. Be sure to comb through the complete list to see if there's anything in there that impacts you.

The ESA - not, not the cool one - a lobby group made up of various videogame and entertainment companies, has asked the US Trade Representative (USTR) to add Anbernic to its list of companies that are supposedly notorious for engaging in and facilitating piracy. If you're unaware, Anbernic is one of the bigger names in Android and Linux-based handheld gaming devices targeted at the retrocomputing market. While many online sellers do include an SD card full of ROMs alongside emulation hardware of all types, Anbernic itself does not appear to do the same when selling directly from its website. Regardless, the ESA argues that these ROM-laden Anbernic devices selling through normal retail channels" help to normalize hardware-based piracy by delivering illicit catalogues through a polished consumer product." This, the ESA says, meets the threshold for a superfake," an ill-defined designation that applies to high-quality counterfeit trademark goods ... on a commercial scale." Kyle Orland at Ars Technica There's a lot of murkiness surrounding whether or not Anbernic sells their devices with preloaded ROMs or not. The number of reports from people who say they did get a ton of ROMs on their Anbernic devices, bought from Anbernic's own store, is large enough that it's clearly something they do. However, there's an equal number of people who say their Anbernic devices were completely free of ROMs, and I'm one of them - I've bought like three or so devices from their store over the years, and they never came with any ROMs. My most recent purchase was an RG 35XXPro Linux-based model, and there were no ROMs on it (I replaced the stock operating system with Knulli, which has been a great experience). If Anbernic does indeed ship at least some of their devices with ROMs preinstalled, then it's their own damn fault if that lands them in legal trouble. It's not like downloading ROMs is particularly difficult (just do an internet search for your preferred system and something like ROM pack" and you'll get the console's complete library in a single download), so they really have no need to do so. There's one extra layer here though. Third-party resellers like Amazon offer loads of Anbernic devices preloaded with SD cards full of ROMs, all from third parties not affiliated with Anbernic. And wouldn't you know it - Amazon is a member of the ESA lobby group. In other words, the problem also lies with lax retailers not properly vetting what they're offering, but instead of going after places like Amazon, they're going after Anbernic instead. Amazon is a haven for illegal and counterfeit products, but they're so wealthy and powerful they just get away with it.