A new GNOME extension called Simple-taskbar does a pretty good job of giving GNOME 50 a taskbar and start menu very much like the ones from Windows 11. It's so new that it's still not listed on the GNOME Extensions website – at the time of writing, it's still under review. We built it and installed it, and it works. This single extension merges GNOME's built-in panel (across the top of the primary screen) and what it calls the "dash" – the icon bar down the left side – into a single taskbar, styled like Windows 11: icons centered, a Start menu with columns, a search box, and so on. The handy feature here is to have all this in a single, integrated extension. There are already other extensions to do some of this, such as Dash to Panel, whose development is now sponsored by Zorin OS. It doesn't provide a Start menu replacement, though. For that, you need another extension, such as Arc Menu. Like many things in Linux, this sort of tool needs the user to do a little manual config. There is also a risk with this kind of DIY GNOME customization using lots of extensions. Extensions are tied to specific versions of the GNOME desktop – so, for instance, the new Simple-taskbar only supports GNOME 50. If you add lots of extensions and then upgrade the underlying OS to a new version, meaning a new version of GNOME, there's a significant chance that the extensions won't work any more – you may even be unable to log in. Zorin OS takes Ubuntu and does this for you, and the result sells – that's how the company can afford to sponsor extension developers. We rather like Zorin OS as a distro, especially its handy Zorin Appearance accessory, which it showcases on its homepage. If you don't need the customization and just want a Windows 11-like look, last year we looked at AnduinOS, which does this to a stripped-down Ubuntu. At the end of last month, the project released version 2.0, based on Ubuntu 26.04. Another tactic, of course, is to fork the whole of GNOME and change it to make a whole new, more Windows-like environment. Back in 2011, Linux Mint 12 offered MGSE, a set of Mint GNOME Shell Extensions to make GNOME 3 more Windows-like, but by 2013 that became the Cinnamon desktop. Since its launch in 2011, Mint has also offered the MATE Desktop, a fork and continuation of GNOME 2. There are multiple ways to keep GNOME but make it look Windows-like. Even the GNOME ecosystem recognizes the need. Early versions of GNOME 3 offered GNOME Fallback mode, intended for machines that didn't support hardware 3D rendering. GNOME dropped that mode with version 3.8, but two alternatives emerged: GNOME Classic, which uses GNOME Shell extensions, and GNOME Flashback, a continuation of the panel-and-Metacity-based Fallback environment. This is still around and maintained. For instance, Ubuntu 26.04 uses GNOME 50, and includes the packages for GNOME Flashback version 3.58.0. It's installed by default on some distros – in Debian 13 it's a standalone desktop – but on Ubuntu, you must manually install it. You'll probably want to put back the network icon – it's an easy graphical way to manage Wi-Fi connections and so on. All you need is one command: sudo apt install -y gnome-session-flashback network-manager-gnome Log out, or reboot, and on the login screen, you can now use the cogwheel button at the bottom right to choose to log in to the new GNOME Flashback session. By default, Flashback has a GNOME 2 and MATE-style two-panel layout, but if you hold down the Alt key while right-clicking the controls on the panels, you can open a context menu that allows you to move (or remove) them. We moved the Menu Bar control to the beginning of the bottom panel, the Indicator Applet Complete control to the bottom right, and the Show Desktop button to the end, and that's it: a basic classic Windows-style layout. You can place panels on the left or right screen edge. The snag is that the controls on them don't reorient accordingly, which renders vertical panels unusable. That's completely faithful to GNOME 2 at least. There are some small oddities – you need to use GNOME's session applet to shut down or log out, for example. However, it's usable, and it uses no GNOME extensions at all, so in our testing it's unaffected by OS upgrades. It uses the Metacity window manager: this is an X11 environment, and it will pull in X.Org as a dependency. This does mean it works fine in VMs without 3D acceleration, a more relevant benefit today than its original role of supporting very low-end hardware. Flashback's appearance varies between releases – the version in Ubuntu 24.04 had a more GNOME-like, softly rounded look. In the current version of GNOME Flashback, it has an old-fashioned 2D look, which we rather like, but it's also much more customizable: it behaves rather like a drastically cut-down Xfce, and MATE users should find it familiar too. We were surprised but happy to discover the improved customizability in Flashback in GNOME 50. The old GNOME Wiki is coy when it comes to Flashback releases, and doesn't list version 3.58 at all. The banner at the top redirects users to the GNOME handbook instead, but that doesn't contain any mention of Flashback whatsoever. This is still a desktop based on GNOME. Flashback doesn't magically give apps running under it proper title and menu bars or anything like that – but then neither can any combination of GNOME extensions, and that includes environments built around them such as Zorin OS or AnduinOS. If you want to stay close to the GNOME mainstream in most distros, but want a more traditional desktop, we suggest checking out Flashback. Room for improvement? For reconfiguring your desktop quickly to look like existing OS layouts, most distros do have an equivalent to Zorin Appearance – but only if you're using Xfce. This desktop has an optional add-on called Panel Profiles that does the same, although sadly, it doesn't help with plugins such as the Docklike taskbar, Whisker menu, or Xfdashboard. A KDE version of Xfce's Panel Profiles would be a huge win, and something like it for Flashback would really help too. We don't know if GNOME is accepting contributions to the Flashback session; it seems reluctant to mention Flashback at all in its current documentation. The Flashback GitLab does show that it's in active maintenance, though. Development of the MATE desktop is slow. It's two and a half years since the last new version. But then the team has an entire codebase to maintain that was abandoned by its own creators. Porting the functionality of MATE over to the GNOME Panel in Flashback, and replacing a few of GNOME's bundled apps with tradition Gtk ones, such as Linux Mint's XApps, would result in a much more familiar and usable desktop with a smaller maintenance burden. It's a possible future for those who remain disenchanted with GNOME Shell. ®

October 2026 is shaping up to be a busy month for administrators as it heralds the end of several big products, along with the retirement of components that could mean sleepless nights for the ill-prepared. After the demise of Windows 10 in October 2025, sysadmins might have hoped for some respite from Microsoft's ax, but 2026 will see the demise of more beasts. First, there's the end of support for Office LTSC 2021, which means everything from Access 2021 to Word 2021 is for the chop. A move to LTSC 2024 is required to keep the support lights burning, or - and we don't doubt that this is Microsoft's preference - a switch to Microsoft 365. M365 isn't a great option for organizations steering clear of the cloud. The on-prem version is very much a snapshot of where things were at release (although security patches keep flowing). Speaking of on-premises, Windows Server 2022 will drop out of mainstream support on October 13, 2026. Extended support will linger until 2031. Windows 11 24H2 Home and Pro reach the end of the road on the same date. Windows 11 23H2 for enterprise and education will get until November 10 to tidy up their affairs. As for the more minor products? Publisher 2021 will wrap up decades of the Microsoft Publisher brand in October. First appearing as a desktop publisher for Windows 3.0 in 1991, it has had an inexplicably long life, which will come to an end once and for all in a few short months. And unlike the rest of the Office 2021 LTSC suite, there won't be anything to replace it (or its annoyingly proprietary file format). There are also services due for the chop that might cause headaches for those that fail to prepare. Entra ID Sign-In risk policies, formerly known as Identity Protection, is being retired on October 1, 2026. Microsoft wants admins to migrate to Conditional Access, and make sure their policies cover user and sign-in risk. Not done the update? Then risk protection will go away after the retirement date. A Register reader remarked, "Given how many orgs use O365 I can't see how disabling sign-in protection might go wrong." October 2027 looks quieter, with only SQL Server 2017 gasping its last. However, with the coming October loaded with already announced terminations, getting the corporate house in order well ahead of time would be prudent. ®

Swedish VPN vendor Mullvad is facing a backlash from some customers after co-founder Daniel Berntsson personally donated SEK 5 million (£385,000, $519,000) to the controversial populist Örebropartiet, which is running in Sweden's national elections for the first time. Berntsson has this week published a blog post explaining the donation, alongside posts titled "Immigration," "Political individualism," and "Moral individualism." Before that, he had not posted since 2019. The flurry of posts follows weeks of customer criticism over the donation – for example, Finnish blog Moonwriting's post "I'm done with Mullvad." Another blog, Unutterable, is more direct: "Mullvad funding the far-right and the problem with tech company ownership." The Hacker News discussion thread about the donations has reached a remarkable 1,784 comments so far. Mullvad was co-founded and is still jointly owned and run by two CEOs, Berntsson and Fredrik Strömberg. Its name, as reflected in the company's logo, is Swedish for "mole." It also offers the Mullvad Browser, a privacy-enhanced fork of Firefox ESR co-developed with the Tor Project. Swedish socialist newspaper Flamman ("The Flame") revealed the donation in its story "High-profile techie gives millions to the Örebro Party." In a post on X, the company said it was devoted to freedom of speech and tolerance, saying: "It should be obvious that Daniel's private donation to a political party is not part of Mullvad's values or mission, in the same way that someone's opinions on animal rights, taxes or public healthcare policy isn't." The Örebro Party began as a local party and later expanded across Örebro County, but is now contesting Sweden's 2026 parliamentary election. Its English-language Wikipedia page offers a summary for readers outside Sweden. It has a mixture of left-leaning and hard-right policies. It wants to make preschool, public transport, and dental care free of charge, reduce politicians' pay, and calls for a 30-hour working week. However, it also advocates for sharply reduced immigration and what it calls "large-scale remigration" – policies intended to encourage or compel immigrants to leave Sweden. By Swedish standards, Berntsson's donation was unusually large: DonationWatch lists it separately in its Swedish political funding data. Flamman reports that it was the largest single private donation to any Swedish party in 2025, and on its own represented 72 percent of the party's income for the year. The Register has approached Mullvad for comment. ®

WHO, ME? Regrets? We've had a few. And so have many readers of The Register, who we try to help by getting their stories of workplace mistakes off their chests every Friday in a reader-contributed column we call "Who, Me?" This week, meet a reader who asked to be Regomized as "Johnny," who told us he was once hired as a Visual Basic developer, a job he thought was beneath him. He got on with things anyway, until he was given a new project that he decided to write in C# despite that language being new at the time. "To my satisfaction, it handled everything the customer required," Johnny told The Register. But then he hit trouble. A complex string manipulation routine that determined how tasks were routed started to glitch. "About half the time, it worked perfectly. The rest of the time, it failed completely," Johnny wrote. "I stepped through the code repeatedly, checking every transformation, every condition, every variable – but nothing stood out as wrong." He therefore started making small and random changes in the hope it would flush the bug into the open. One of those changes broke everything, so Johnny reverted it – and after that the glitch magically disappeared. "I compared versions, searched for differences, retraced my steps – and found no reason," he told The Register. "To this day, I have no idea what caused it or why it disappeared." Over time, the incident faded from Johnny's memory. Several years later, Johnny was still at the company and had climbed the career ladder to become a senior developer. "One day, the support team escalated an issue to me: an intermittent bug that no one could solve. I agreed to take a look," he said. His first step was to look for comments in the code to understand the intent of whoever wrote this piece of software. "Scrolling through the header, I found the author's note originally written by myself and felt an immediate sense of dread," he told Who, Me? That sense was prescient because the comment read: "Dear future Johnny, I am very, very sorry." Have you written code that came back to haunt you? If so, click here to send us an email to share your story. We promise to handle it carefully so contributing won't haunt your career. ®

Firefox releases will soon get even closer together – but not ESR ones, which remain annual, with the next one due out soon. The change was announced last week on the dev-platform@mozilla.org mailing list by Mozilla’s director of engineering Sylvestre Ledru: “We are planning to move Firefox Desktop and Android from a 4-week release cadence to a 2-week release cadence starting in September 2026. “This will be an experiment… This does not mean that all work needs to ship twice as fast. Work that is not ready should not be rushed, and features can still take the time they need to bake. “The current target is to release Firefox 155 on September 1, 2026, instead of September 15 … “We will closely monitor how this change works in practice and adjust if needed.” The change is already visible in the upcoming Firefox Release Calendar. At the time of writing, it shows the current four-week cadence for Firefox 153 and 154, with Firefox 155 brought forward to September 1. From then on, it lists releases at roughly two-week intervals. This isn’t unprecedented. The last time Mozilla did it was over a decade ago. Google announced a very similar change for Chrome back in March, and The Register reported on it going to six-weekly releases in 2010 and then four-weekly ones in 2021. Brace for the impact of the next ESR In the meantime, Firefox 153 is nearly here: at the time of writing, it’s scheduled for Tuesday, July 21. Right now, the release notes are blank, but the beta release notes are more forthcoming. You can look forward to further improvements to PDF editing. Version 153 will be able to merge multiple PDF documents into one by a simple drag-and-drop in the PDF sidebar, and insert images into PDFs as new pages. It will also improve highlighting text in PDFs. The new version can generate QR codes for sharing web pages offline, such as in printed documents. It can verify and display the EU’s new Qualified Website Authentication Certificates, or QWACs for short. These are part of the new eIDAS legislation, which is a slightly convoluted acronym for “electronic IDentification, Authentication and trust Services”. Sounds cumbersome, but if it helps make the Wild West of the Web a bit safer, we’re all for it. Another security tweak is that extensions will lose local-file access by default. As we use one that relies on it - the excellent Multithreaded Download Manager - that's slightly concerning, although users will be able to grant the permission separately, so we'll see. If a tab is using your location, Firefox 153 will highlight this with a map-pin icon in the address bar. The last release could recognize a few keywords in the address bar to mute all sound: this function will get extended, with typed commands for picking colors from pages, and accessing experimental “Labs” features. Come back, Ubuntu Unity and your HUD, all is forgiven. There’s also experimental JPEG XL support. Pop-up video player controls will work better, and if you use Windows and have a compatible GPU and drivers, 153 will play HDR video. On macOS, Firefox supports the Fn+F keystroke for switching to full-screen mode – which means KeyboardControls.com has to update its description of the function. Firefox 153 will also be a significant release, because it will be the next ESR version: it will get security updates for at least 15 months, meaning until late 2027. (And maybe longer: Firefox 115 is still getting updates, and now will until March 2027.) In turn, that means that Firefox 153 is also the basis of the latest beta of the AI-free Waterfox fork, which recently integrated built-in ad blocking. ®

Spelled out in big colorful letters that even Google can understand, the EU is now requiring the Chocolate Factory to share search data with competitors while enhancing Android AI interoperability for bots other than Gemini. Needless to say, the company would like to find a way to disable this default. The European Commission (EC) announced a pair of specification decisions on Thursday. The first covers AI vendors' ability to integrate into Google's mobile OS, and the other forces Google to give search data to other search engines in order to “rebalance the playing field," as the EC put it. The specification proceeding, the Commission said, isn’t a noncompliance decision and doesn’t attempt to determine whether Google has been flouting its obligations as a Digital Markets Act gatekeeper (i.e., it’s big enough that it controls access to markets for smaller firms). It only wanted to make sure both parties were clear on what the EC actually wants Google to do to lessen its monopolistic hold on search and its control over the operating system on the majority of the world’s smartphones. Gemini gets pried loose from Android “Currently, on Android phones, competitors' AI assistants only have restricted access to key functionalities,” the EC said. “Today's decision will ensure that users can activate their preferred AI assistant.” Specifically, Google will be required to give third-party AI providers extensive access to Android-powered devices, including allowing them to be voice-activated in place of Gemini, and to take actions in apps on user’s behalf. Google cried foul, as it has when faced with practically any EU decision spanking it for unfair competition. “Today's decisions risk undermining vital privacy and security guardrails for millions of Europeans,” Google and Alphabet president of global affairs Kent Walker said in a statement. More specifically to the AI portion of the decision, Walker said the EC’s ruling is unnecessary, as AI assistants "already safely access Android’s capabilities” inasmuch as OEMs let them. “This Android ruling threatens device security by granting external apps sensitive and powerful device permissions without [OEM] safeguards,” Walker said. Never mind the fact that the Commission spelled out in its FAQ for the AI interoperability specifications that Google and OEMs still retain a good deal of control over their implementations. “In view of the sensitive nature of some of the features, Google may put in place objective and non-discriminatory eligibility conditions to limit access to third parties meeting certain privacy, security and integrity standards,” the EC said. An unreasonable search seizure? As for the search half of the specification decision, the EU was concerned Google’s prior attempts at offering to open up were “innefective so far.” With that in mind, the Commish wants other search engines to be able to access Google data in order to improve their own services and compete more effectively. “Subject to anonymisation, Google should share the same data that it collects to optimise its own search services,” the Commission said, noting that that includes giving AI chatbots access to that data to improve their services as well. “The aim of these measures is to allow companies to be able to offer European users a wider and more feature-rich range of options to choose from, both when it comes to their AI services on Android and to search services.” Google, again, is predictably unhappy. “Europeans’ private searches would be exposed to unfamiliar companies, without adequate anonymisation of the data and without user knowledge or consent,” Walker argued in his response. “This would weaken citizens' privacy, risk business trade secrets, and endanger national security.” The Chamber of Progress, which speaks as a voice for the tech industry and is funded in part by Google, similarly decried the risks of the decision to user safety. “The Commission is ignoring well documented privacy risks to impose a vision of the digital economy, rather than working with industry to find solutions that are safe and supported by consumer demand,” said the group's VP for Europe, Kay Jebelli. “The likely result is services will be pulled while legal challenges are mounted.” Neither Google nor the Chamber bothered to get specific in their responses, and likely with good reason: The Commission made it clear that Google has final say over what gets shared, and spells out various elements of how it will require search data to be anonymized in an FAQ for the search section of the decision. “The decision … allows Google to assess, before sharing any data, whether sharing such data with a specific third party poses serious cyber security and data protection risks,” the Commission explained. Per the FAQ, particular technical elements of the search data sharing anonymization include suppressing search records that contain rare items (e.g., usernames and passwords, addresses, bank account info, etc.), generalization of metadata and grouping users into bundles of at least 1,000 people with similar geographic and device data, and the removal of all direct and indirect identifiers. Additionally, search data shared by Google with other providers “will only be made available to eligible beneficiaries with verified investment plans to improve online search services,” and the use of that data is limited to improving search services. Recipients won’t be allowed to link the data to any other datasets, share it with any third party, and must undergo independent audits to verify safeguards and compliance with EU policy before ever getting access, followed by yearly audits afterward. “The measures will also be subject to biennial reviews to take into account practical experience and any new developments,” the Commission added. Cambridge Analytica redux? Google's complaints are nearly identical to those made when the EU first told it to open up search data in April. So we reached out to understand which points were particularly sticky for the search giant. Jebelli from the Chamber of Progress told us he's concerned that the Commission's decision is overly reliant on contractual restrictions to enforce search privacy – "the same kind … that led to the Cambridge Analytica scandal." Recall, that's when Facebook allowed a third-party firm to collect detailed data about users by offering them a "quiz," then attempted to use this data to influence the US presidential election with microtargeted advertising. When we asked Google for comment, Google pointed us to the statement above, and also provided us with additional background counterpoints. The company called attention to the fact that there's no explicit user consent in the EU's specifications, and also noted that the Commission is only asking for pseudonymization of search data instead of full anonymization. Google also expressed concern over handing search data to AI chatbots, saying that companies like OpenAI could slurp up search data en masse to train their models – an argument we note is somewhat rebuffed by the Commission's point that search data won't be freely available under its proposed schema. As for the AI interoperability concerns, Google told us third-party AI assistants are gaining distribution on Android all the time thanks to OEM agreements, calling the EU's decision a side-step on manufacturers' ability to safely vet AI. The company also told us that giving third-party AI the ability to screen-scrape with system-level access is a huge window for threat actors, dismissing the EU decision as a rushed checklist instead of actual law. The EU said that Google has a right of defense regarding the decisions, and independent judicial scrutiny of today’s announcements is still warranted too. If nothing manages to change the Commission’s mind, Google will be forced to comply with the search data sharing specifications beginning in January 2027, while the AI portion will go into effect in July 2027. In all likelihood it'll be tied up in litigation for far longer than that. ®

Forrester warns that customers should brace for bigger software bills next year as software and AI vendors raise prices and pile on usage charges. Working from a survey of more than 2,600 business and technology decision-makers, the tech research company said software budgets were expected to rise "as vendors increase prices or add usage charges to pass their AI costs to customers." In the last six months, Anthropic, OpenAI, and GitHub have shifted some services away from flat-rate subscriptions toward usage-based billing, prompting cost concerns among users. Forrester added Microsoft to the list, citing its recent launch of the premium E7 license, which bolts M365 Copilot, Agent 365, and security tools onto E5. Last year, consultants Bain & Company estimated that the build cost for AI datacenters would hit $2 trillion by 2030. Forrester said that AI would drive increases in data and software spending, with 80 percent of decision-makers expecting those budgets to rise. Sharyn Leaver, chief research officer at Forrester, said: "The organizations that outperform in 2027 won't be those that spend the most on AI. They'll be the ones that invest in the foundations that make AI effective: trusted data, strong governance, organizational readiness, and the ability to continuously adapt as technology and customer behavior evolve." Forrester also found that personnel costs have yet to fall, despite the "AI washing of layoffs" in the tech industry. "While several tech giants, including Oracle, Microsoft, and Meta, have announced significant layoffs in recent months, IT staffing spend has not declined in recent years," the report said. Staffing accounted for 35 percent of IT budgets in 2025. For 2027, 67 percent of tech decision-makers expected to increase their staffing budget, while 23 percent said it would stay flat, and 10 percent expected it to decline. "The AI washing of layoffs will continue as vendors trim for financial and restructuring reasons. Guard against inflated promises that AI can replace employees across the board. Staffing for data/analytics-specific roles is expected to rise, with 68 percent of data technology decision-makers expecting this budget item to increase," the report said. Forrester said that organizations should adapt their FinOps practices to help manage the unpredictable costs associated with AI. "Traditional FinOps wasn't built for token-based, usage-driven AI costs, but that team is certainly best positioned to build these new capabilities and must make this leap in 2027. Fund runtime cost controls such as model routing, semantic caching, and usage guardrails to prevent runaway spend," the report recommended. In July, KPMG research found that nearly a third of corporate leaders reported difficulty understanding and controlling operating costs when implementing business AI at scale. "As usage-based pricing models become more common, many organizations are still building the capabilities required to forecast, monitor, and manage AI spending effectively," the consultancy said. ®

Imagine never having to reach for your mouse to navigate around Windows again. Sounds like a dream, doesn’t it? Well wake up: We have some peripherals to burn. Neverclick, from developer Lazo Velko, was published recently with the promise to allow users to perform mouse actions on every single object on their screen with nothing but keyboard shortcuts. Want to close a window, open an application, or click a particular spot on the screen with character-level precision? It’s capable of doing all that, along with selecting multiple spots to click at once to, say, close a bunch of windows at the same time. However, drag and drop and highlight don't work yet. What’s more, Neverclick is completely free, has no account signup, doesn’t serve any ads or collect user data, and works entirely offline, Velko notes in the GitHub readme (the repo is currently for issue reporting only and doesn’t contain the full Neverclick source code). Velko explains on the Neverclick website that he designed the app when he was dealing with a repetitive strain injury that made using a mouse difficult. “I recovered years ago, and I owe it to this software,” the self-professed C++ hater said of his project. “I still use Neverclick every day and can't imagine using my computer without it.” Foolish El Reg reporter, I can hear you saying. This app is nothing new and will have the same problems as other ones: It simply won’t work with software that wasn’t built with accessibility in mind. Velko has heard those complaints, which as he explains in a comment in the Hacker News thread about his app, he didn’t build it with accessibility APIs. He used a computer vision model to perform raw pixel analysis and identify individual elements on the screen instead. “I've had a poor experience with accessibility apis, they're clunky, slow, and unpredictable,” Velko wrote in the thread. “With computer vision you don't have to worry about that.” Velko told The Register in an email that he built the computer vision system himself without relying on any third-party libraries, and he claims that it works far faster than the UI Automation accessibility system in Windows. “I have users that use Neverclick on 10 year old hardware and they tell me that the [computer vision] runs instantly whereas UI Automation is super laggy for them,” Velko wrote on Hacker News. “I can't believe that accessibility APIs are so poorly optimized that raw pixel analysis is faster.” According to Velko, the computer vision model used by Neverclick is small, with the entire app only needing around 40 MB of hard drive space, and using around 200 MB of memory on a typical 1080p monitor when in use, though that varies by resolution and other factors. Neverclick doesn’t run in a continuous loop either, so shouldn’t eat up any CPU unless it’s called on for use. El Reg US Editor Avram Piltch tried Neverclick out on his Windows PC. As described in the documentation, he was able to activate the software by hitting Ctrl + Enter for left-click or Alt + Enter for Ctrl + Left Click (there's no right click mapped by default but you can configure one). After hitting one of those two key combos, his screen was covered in two-letter key combos that the software calls "hints" (e.g. ek, af, jj, etc). If he typed one of those combos, the cursor clicked in that exact spot, whether it was a button, a menu option, or just the middle of a word. Word hints click in the center of a word by default, but you can hit a number like 0 to get to a particular letter. There are also keyboard shortcuts for switching windows. And you get a choice of having the hints appear only in the active window or all throughout your screen. However, the UI can be overwhelming. It is nearly impossible to see all the menus and icons with the hints covering over them. And to make the overlay go away, you have to click. We also noticed that the default shortcut of Ctrl + Enter overrides the same shortcut in other apps. For example, Ctrl + Enter is supposed to send an email in Gmail, but with Neverclick running, it brings up the overlay instead. Neverclick is a Windows-only app for now. Velko said a macOS version is in the works, but he doesn’t have a timeline, he told us in an email. “I'm quite overworked and still need to release a few features for Windows,” he explained. Those features include window dragging and text highlighting by selecting two different hints, triggering whatever text editor is being used to highlight whatever’s between them. You can install Neverclick from its website and it's available now. Once it comes to macOS, you’ll be able to combine it with the AirPods app we wrote about last week that allows users to scroll their screens by tilting their heads up and down for a truly mouse-free experience. ®

HANDS-ON Dortania's OpenCore Legacy Patcher lets you run newer versions of macOS on unsupported Intel Macs. It's handy, but there are a few things to beware of – including macOS Tahoe. OpenCore Legacy Patcher brings Hackintosh techniques to genuine Mac hardware. It's an inspired hack that helps you to install newer versions of Apple macOS on older Macs for which Apple has dropped support in recent versions of macOS. Its co-author, Mykola Grymalyuk, has a fascinating retrospective on its development. We introduced OCLP a few years ago when it was still quite new. Our screenshot then showed version 1.01, but as the project's GitHub page shows, it's now up to version 2.4.1. It does still have one big drawback, though. At present, the newest macOS it can help with is version 15 "Sequoia," whose release we covered back in 2024. In 2025, The Register warned that the forthcoming macOS 26 "Tahoe" would be the last version to have any support for x86 Macs. As we confirmed in September, it supports a mere four x86 Macs. This limited Intel support makes it a significant challenge for the OCLP development team, and nearly a year later, the latest OCLP still does not support Tahoe. As we've mentioned a few times, the main desktop computer at the Irish Sea wing of Vulture Towers is a 27-inch Retina 5K iMac (late 2015). A couple of years ago, we maxed out this machine with a quad-core i7, 32 GB of RAM, a 1 TB NVMe SSD, and an 8 TB hard disk. It also drives a 27-inch Apple Thunderbolt Display. It's a capable box, but the last supported OS was 2021's macOS 12 "Monterey." As we mentioned last month, this is now too old for the latest Raspberry Pi Imager. It was time. Getting started with OCLP Alongside OCLP's own documentation, there are several how-to guides out there, such as the iFixit Guide and Greg Gant's 10-step guide. We are not setting out to recreate them, just to point out the stumbling blocks we encountered. Ever since OS X 10.9 "Mavericks" in 2013, macOS has been a free download. You can get it from Apple's App Store, and the company even has a downloads page to help you find older versions. When you download a version of macOS, what you get is a macOS application, called "Install macOS [Codename]." Buried inside this is a little command-line app to create a bootable USB key to reinstall a completely blank machine – for instance, if you fit a new empty SSD. You'll need a fairly big, and ideally fast, key. The recommended minimum is 32 GB. Then you boot off the key and install or upgrade macOS. However, OCLP can also patch a system disk to make it bootable on an unsupported Mac. It will even download various releases of macOS and create the installation USB key for you. The basic way that OCLP works is by creating a standard macOS installer and adding a model-specific OpenCore configuration that bypasses Apple's firmware checks. It may also apply post-install root patches for unsupported hardware. Intel Macs use EFI firmware, which means a small hidden partition containing the EFI bootloader, called the EFI System Partition, or ESP for short. OCLP adds a new entry to this, which shows a special "EFI Boot" entry, as shown in the manual. You must use this to start your newly created key. OCLP does not build a generic modified macOS – it creates a model-specific OpenCore configuration for the Mac on which you run it. There is an option to bypass that and create a custom USB key for a different Mac, but watch out: by default, it creates one for the machine you run it on. For an easier life, you might want to keep a small partition with the last officially supported macOS for your machine – and if you're installing on a blank disk, do that first, before continuing with the process. Even after it's created a custom USB, you can't boot this the normal way. You must use the custom "EFI Boot" entry it creates. You will probably do a lot of rebooting. Don't do this process with a wireless keyboard or mouse, even if they're official Apple ones. Plug in plain old wired USB ones. If it's a PC keyboard, make sure that you know which keys are which in Mac terms. If the instructions say "hold down the Option key," that means Alt; if they say "press Cmd plus [some other key]," that means the Windows key. Our normal Das Keyboard is a USB 3 device, which might be why it doesn't always register when the machine boots, so we needed to use an additional USB 2 as well. To pick custom boot options, press the Option key (Alt) immediately after the power-on chime, and hold it down. Wait for all the options to appear, and pick EFI Boot, then in the following boot-device picker, choose the installer for your new version. Don't worry: after it's all installed, you don't need to do this every time – only if you change the Startup Disk to an unpatched, supported version, then need to switch back again. Once you've booted your new, customized macOS installer key, you can either do a clean install or upgrade an existing copy of macOS, just like normal. As a fallback measure, we used Bombich's Carbon Copy Cloner to make a backup copy of our Monterey partition on another drive, just in case. This has been an essential Mac utility for decades, and it has a handy trick up its sleeve. Since 10.14 Mojave, macOS has defaulted to Apple's proprietary APFS, and since 10.15 Catalina, it uses a system of container volumes to keep the core OS semi-immutable and safe. For an example, the Eclectic Light Company explains the Sequoia config. This limits how much a macOS volume can be shrunk, but Carbon Copy Cloner can image a volume onto any destination big enough to hold the data, which is very handy. Post-install gotchas Of all our apps, only one didn't work. A couple of years ago, Broadcom made VMware Fusion freeware, but we couldn't update past Fusion 13. With Sequoia, we could update to Fusion 26 – but it didn't work. All VMs had a blank black display. No biggie: we have VirtualBox and the very capable UTM, and both work fine. Our iMac came with a "Fusion drive" – a tiered storage volume combining a tiny 24 GB SSD and a 1 TB HDD. When we upgraded the drives, to keep life simple, we didn't recombine them: it boots from the SSD, and we have a big HFS+ data volume on the hard disk that holds our home directory. Sequoia seems not to expect home directories on HFS+, and various settings would not save. It asked some of the same first-run questions every boot. To get around this, we had to move it back. You can't move the home directory of a logged-in account, so we were glad of the standby "Administrator" user account we created years ago for emergencies. Logging in as "Administrator," we moved all the big data folders out of our home directory (Documents, Downloads, Pictures, Music, Videos, and so on) elsewhere on the hard disk. Then we moved the home directory into its normal location in the boot volume on the SSD. Apple's Library folder is huge, and this process took a long time – about five hours. We didn't really expect it to work, but we created aliases to the data folders in our newly relocated home directory. (Aliases are the macOS equivalent of symbolic links, and to make them by drag-and-drop, hold down Cmd and Opt – the mouse pointer gains a curved arrow emblem, much like Windows shortcut icons.) This worked perfectly, and all of our apps find the folders in their new location. With the whole home directory on SSD, the Mac runs much faster and apps load much quicker too. (The folders don't have their usual icons, which is just cosmetic; also, Apple Music isn't very happy – but we never use it: Foobar 2000 does all we want.) Another snag is that once Sequoia was running smoothly, it started nagging us to upgrade to Tahoe. Don't do this: OCLP doesn't support macOS 26 yet, and while it might boot, USB doesn't work – and no Mac is very useful without a keyboard and mouse. If you do this, just say no to Tahoe. (If you only need some specific app then an older version of macOS may be all you need. Last year, we Hackintoshed a Dell Latitude to recover some data from a network Time Machine backup, and we found that macOS 11 Big Sur was much quicker than macOS 12 Monterey on that hardware. If an older version will do what you need, it may be easier or lighter-weight, but we decided to just go direct to the latest version on offer.) Back in March, we looked at "Stop Tahoe Upgrade," a tiny macOS policy file that you can install to make the nags go away. This worked perfectly on our M1 MacBook Air, but it didn't on our iMac. Doing some further investigation, we found some additional steps that got it working. You need to manually create two UUIDs and add them to the scripts, then the profile worked on our Intel box. The comments on that article offer an alternative method, by opting into beta updates, but that resulted in us being offered a beta of the update to 15.7.8, which isn't ideal. Why not Linux, or even Windows? We have considered using Boot Camp Assistant to see if we can crowbar Windows 10 LTSC onto this machine, but to be honest, we didn't buy a Mac to run Windows on it. This vulture adopted Windows with 2.01 in 1988. We've done our time, served our sentence, and we don't want to go back. So why not Linux? This is, after all, The Reg FOSS desk. Well, we tried it. The main problem is the iMac's dual displays. Its built-in screen is a massive 5,120 2,880, but the Thunderbolt display is a quarter of the resolution at just 2,560 x 1,440. Running two different dot-pitches currently means using Wayland, and that massively reduces the choice of desktops. First, we tried Pop!_OS 24.04 – we quite like its COSMIC desktop. This normally snappy distro took tens of minutes to boot and was almost totally unresponsive. After some research, we disconnected the Thunderbolt display, and then it ran fine, but we couldn't install it. As we reported back in 2021, the Pop!_OS installer expects an ESP of at least a gigabyte for its systemd-boot setup, while Apple, like Microsoft, typically creates one only a couple of hundred megabytes in size. Next, as a control, we tried Ubuntu 26.04. With the second screen disconnected, Ubuntu worked fine: HiDPI screen, Wi-Fi, power management, everything. We installed it and updated it, and it ran very quickly indeed… until we connected the Thunderbolt screen, when it instantly fell over. It seems that Linux's Thunderbolt support is distinctly lacking. Back in 2022, we reported on the difficulties connecting this display via a USB-C to Thunderbolt adapter. The adapter won't work at all on our company Dell Latitude 5420. Apparently, tenth-generation and later Intel processors can't talk to Thunderbolt 1 anymore. We had more luck on an older Dell, the 2025 FOSS desk testbed XPS 13. That can drive the display through an adapter, but only as a display. The 27-inch Thunderbolt display is also a docking station: it has Firewire, Ethernet, three USB-B 2 ports, a webcam, microphone, and stereo speakers, and can power an Apple laptop over MagSafe too. None of these ports work in Linux. It becomes just a big, hot, external screen. For now, Linux's hardware support just isn't up to making the most of some Apple hardware, even an 11-year-old iMac. After we moved the home directory onto the APFS boot drive, Sequoia runs very well on this machine. Thanks to having the home directory on SSD, it's faster than ever. For a couple of days, it ran hot and slow while Spotlight re-indexed some three terabytes of files, but once that was done, it happily purrs along at 63ºC, the same as before. If you have an older but well-specified Mac, OCLP can give it the gift of life. It doesn't support all models, but it supports most from the last decade and a half. So long as current browsers are available, running a dated version of macOS is no big problem. It's not trivial to install; you want as many backups as you can make, and once running, you must make very sure that it doesn't try to update itself to macOS 26. We hope that the OCLP developers add support for that final x86 release soon. ®

Enterprise customers should use SAP's commitments to the European Commission to reconsider complex ERP migration timelines and gain leverage in negotiations with the German software giant, according to Gartner. Last week, the European Commission ended an investigation into possible anticompetitive practices after SAP agreed to abolish reinstatement fees and reduce back-maintenance fees, among other concessions. The case stems from September last year, when the executive branch of the EU launched a formal investigation into SAP's policies for maintenance and support services in Europe, responding to concerns that SAP restricted competition in this crucial aftermarket. In two papers published late last week, Gartner said the agreement represented an opportunity for users grappling with mammoth ERP migrations. Thousands of large businesses rely on SAP ERP Central Component (ECC) to run core business operations. SAP's mainstream support for ECC ends in December 2027, while customers can opt for extended maintenance until December 2030 by paying an additional two percentage points on their maintenance fees. As well as trying to move customers off legacy software, SAP is pushing them toward the cloud and SaaS through its lift-shift-and-transform migration program RISE with SAP, launched in early 2021 in conjunction with cloud providers and third-party service vendors. Yet Gartner estimates that the majority of SAP's installed base is still using on-premises software. By 2030, more than 10,000 SAP customers will continue to support major parts of their business with solutions based on SAP ECC, Gartner said. "Premature SAP modernization leads to excessive technical debt and spiraling costs. The EU ruling fundamentally endorses third-party software support not just as a cost-cutting measure, but as a vital strategic resource to reduce the risks of ECC or S/4HANA migration. Heads of enterprise applications must urgently leverage this newfound flexibility to buy preparation time and avoid being locked into inflexible cloud commitments before they are ready," a recent Gartner research paper says. In its October response to the Commission's formal investigation, SAP promised to improve "the financial attractiveness for customers who wish to reinstate SAP maintenance and support services." SAP has now agreed to abolish reinstatement fees and reduce back-maintenance fees charged to customers who return to SAP support after a period of absence, the Commission confirmed. SAP also agreed to clarify conditions that allow customers to choose different maintenance and support service providers and different levels of support from SAP. "Previously, SAP customers faced an unavoidable dilemma: rush a highly complex migration to SAP S/4HANA or SAP private services before 2030, or stay on-premises with limited support and compliance guarantees via SAP's Customer-Specific Maintenance," the Gartner paper states. "Because this new ruling legally prevents SAP from restricting customers who want to 'mix and match' M&S services from different suppliers, it shatters the historical barriers to utilizing third-party support services (TPSS), providing essential extra options to best manage risk." Gartner recommends users review their SAP ERP modernization timetable in view of the new rules agreed by SAP and the EU, placing less weight on vendor deadlines. Technical debt, readiness for operating model change, and urgency of new ERP capabilities should instead govern the migration plan. In a second paper responding to the agreement, Gartner says customers can use the commitments to re-evaluate SAP support, optimize costs, and create new leverage in vendor negotiations. "Increased competition gives ECC and S/4HANA on-premises customers greater negotiation power on support contracts and cloud deals. Customers should explicitly reference the ruling when seeking alternatives to standard maintenance increases," the paper says. SAP user groups, however, told The Register they did not expect a wholesale shift to third-party support. Michael Bloch, licensing expert and board member of German-speaking user group DSAG, said: "It will be interesting to see whether customers will take that choice actively: most of them who are using ECC and may be looking for third-party support will not come back to SAP. As S/4HANA will still be supported until 2040, I personally do not believe that customers will cancel their support contracts and come back afterwards to SAP while having support until 2040." Conor Riordan, chair of the UK & Ireland SAP User Group, said SAP users had long called for greater flexibility, transparency and predictability. "Ultimately, this should give customers a clearer framework for responding to changing business conditions while evolving their SAP estates at a pace that suits them. We do not expect a wholesale shift away from SAP maintenance. For most organisations, especially those with complex or business-critical SAP landscapes, extended maintenance remains the lower-risk bridge while they continue their transition planning." ®

Another month, another ambitious idea from Yuri Zaporozhets: a proposal for a standard PC-style BIOS for RISC-V computers. The Harmonic Firmware Initiative, or HFI for short, is simple and appealing in principle, but will be considerably harder to implement. From the project's own description, it aims to provide a BIOS-like experience, familiar from x86 PCs, for RISC-V hardware. When the machine starts, the firmware would identify the system and its hardware, list what peripherals are connected, offer the option to enter a setup program – and then hand over to U-Boot to load an OS. If the name seems familiar, this may be because last month we looked at several of Zaporozhets's previous projects, including his GateMate PC, the System/359 micro-mainframe, his 64-bit RISC-V version of QNX 6 QRV, and most recently, his QSOE RISC-V RTOS, which offers a choice of seL4 or its own in-house kernels. One part of the GateMate RISC-V PC project is a PC-style modular BIOS. This includes a video BIOS for the GateMate PC's bespoke video adapter. The HFI proposal is somewhat different. It's intended to work with an existing, standard RISC-V bootloader, the RISC-V version of the FOSS cross-platform bootloader Das U-Boot. It does include a video BIOS, though, one designed to initialize arbitrary video hardware in a standard way and bring it up to a VGA-like text mode. For now, Zaporozhets is developing HFI on his SiFive HiFive Unmatched. In that motherboard's PCIe slot, he has an old Nvidia GK208 graphics card, and his firmware can initialize that and show a text-based boot-up display, without using any old x86 BIOS code. As well as the homepage, there's also a six-page white paper [PDF] laying out the idea in detail. In this, he does mention what could be seen as a rival – another RISC-V firmware offering, the EDK2 implementation of UEFI. UEFI is large and complex, and the white paper says that most current boards use U-Boot instead. As the last section of the page says, this is an just an invitation with aspirations: "HFI is offered as an initiative, not a product to license. Its reference software is open by construction – HFI BIOS links U-Boot and carries U-Boot's license – and QSOE Systems stewards the whole: the interface specification, a high-quality reference implementation, and the porting to new controllers." It seems like an excellent idea to us. The absence of standard firmware has been a handicap for all Arm-based computers ever since the chips caught on outside Acorn Computers. This is why Arm OSes are hard to transfer from one hardware platform to another, why every phone needs its own ROM images, and it's why the move to Apple Silicon meant the end of the Hackintosh world. It would be very good news for RISC-V if it avoided the same fate. ®

Wayland is dominating the recent news about FOSS GUIs – even dignified elder Xfce’s official support is getting close. However, X11 is very much not dead yet, and new developments keep appearing. Last week, Norwegian FOSS developer Geir Isene announced his all-new server for the venerable X11 display protocol. Its description is in the title of the announcement post: Frame - the first Linux Assembly X server. Isene explains his motivation thus: “On my quest to own my software, one foundational piece kept itching… the X server. The underlying graphics engine, the thing that puts pixels on the screen. X11 is 4 million lines of code, a beast very few can claim they understand. So I did the reasonable thing. I wrote my own, in Assembly.” (This, for clarity, is extremely dry Norwegian humor. Writing your own X11 server in assembly language is the polar opposite of what most Unix developers would consider “reasonable…” But there is a reason behind the decision, and you may already have guessed it.) Isene also has his own window manager, Tile; his own terminal emulator, Glass; and even his own shell, called Bare. He calls the whole stack CHasm — CHange to ASM. All the tools are standalone binaries, implemented in x86-64 assembly language, targeting Linux, with no external dependencies. To go with them, he also has a similarly compact suite of Rust-based tools as well, named Fe₂O₃. It reminds us of the Suckless collection of extremely minimal, statically linked Linux apps, which once included a entire statically-linked Linux distro called Stali. Although Stali has been unmaintained for nearly a decade now, there are others: one current statically linked Linux distro is Oasis.) If you like a hyper-minimalist Linux setup, Isene’s custom stack of tools sounds amazing – perhaps even too good to be true. (That’s a hint, or what we gather fiction writers call foreshadowing.) yserver Remarkably enough, Frame is not the first new X11 server for Linux that has appeared recently. That distinction goes to yserver, which appeared last month. Developer Jos Dehaes describes it as “a modern X11 server written from scratch in Rust.” This also sounds good, especially if you’re an admirer of Rust. We have started using a few Rust-based projects here in the Irish Sea wing of Vulture Towers. They tend to be small, fast, stable, and quite clean of legacy baggage. They offer a significant contrast with Electron-based ones. In terms of tech legacy baggage, they score very well on what sometime Reg contributor Verity Stob measured with the Stob Cruft index. Yserver looks impressively complete: for instance, Dehaes has a list of 13 existing environments that run fine under it, from Enlightenment to Xfce. But – there’s always a “but” – before you get too excited, there’s a catch. Both Frame and yserver were built using LLM bots. (“Vibe coded” feels too harsh for programs of this complexity, but some will apply it anyway.) So, although Isene says of Frame, “I wrote my own in Assembly,” further down the same page, he concedes: “When something breaks or I want a feature, I turn to my buddy Claude and describe the itch.” To be honest, for this staunchly AI-skeptical vulture, this means Isene didn’t really write the CHasm programs himself: he presumably wrote a detailed spec, from which a bot generated code. This also arguably means he doesn’t strictly own the software, but on that point, to his credit, he does not call it open source or free software: instead, he put it in the public domain under the Unlicense. There’s less info about yserver development than Frame, but yserver’s Github repository contains both CLAUDE.md and AGENTS.md, and both OpenAI Codex and Github Copilot have commits in its history. Previously… There is also a third new X11 server, which we already mentioned back in January. That one is called Phoenix, and it’s written in the relatively new Zig language. Phoenix’s original Git repo seem to be down, and although the homepage is there, it hasn’t been updated in a couple of years and doesn’t mention Phoenix at all. Even so, the Github repo is active, with commits just three weeks ago at the time of writing. We have no information as to whether the developer, known only as dec05eba, is using any LLM tools to work on it. As well as all these, there is the XLibre fork of the X.org server, whose announcement we covered a year ago, followed by its gathing vocal endorsement, including Devuan. Multiple distros and OSes include or support XLibre. Development is active: the project has put out 27 releases so far. (XLibre development does seem to be happening significantly faster than the Wayback project, which appeared a month later. Wayback is a sort of “Wayland display server” to enable entire X11 desktop environments to run on a pure-Wayland display stack. As far as we can see, there have only been three Wayback releases so far, and the last was six months ago.) Arcan Alongside whole new X11 servers, whatever they’re written in (or how), there is also an active project that is far more technologically ambitious than all of them put together. The Arcan display server hit a significant anniversary in June, and developer Björn Ståhl marked it with a blog post: Arcan: 10 Years of Online Obscurity. We wrote about the project back in 2022, in Lash#Cat9: A radical new Linux UI for keyboard warriors. That article focused on just part of the project, its shell Cat9. Cat9 is a shell in the sense of the Windows Shell: it’s a user interface. Cat9 happens to be a command-line one, and it also contains quite a lot of what in traditional Unix terms would be a terminal emulator. Arcan is a lot more than that. Arcan covers a display protocol (called A12) and a display server (comparable to X.org). It includes a command-line shell (Cat9), but also a desktop environment called Durden. Durden isn’t the only one: there’s also a Zooming user interface (ZUI) called Pipeworld. So you can run X11 apps, Arcan has an X server called xarcan, and it can also run Wayland client apps over a module called Waybridge – a pun on weighbridge. Arcan is the sort of ambitous rethink of the Unix display stack we wanted to see – and which Wayland’s developers didn’t even think about, let alone attempt. Arcan doesn’t just let programs open windows over the network like X11. Arcan lets networked computers send media streams to each other, and because it understands hardware 3D it enables remote 3D acceleration, for instance for network gaming. It integrates multitasking and 3D and media right into the Unix command line: shell commands can fade into the background while displaying grapical progress meters, or play media streams in the terminal, or multiple updating parallel zones of text output. There are multiple prices to pay for this, and they are formidable: in the learning curve, and the maturity of the tools, and in interoperability with existing solutions. Arcan really deserves wider attention. Its potential is amazing. It’s also all hand-coded by a tiny team. As far as we can tell, there are no coding assistants here. This is the sort of thing the Linux world deserved, not some modest improvement to local 2D desktops that has taken nearly 20 years to reach minimal usability. ®

OpenMandriva has accused a former contributor of using his trusted admin access to trash repositories and push a package that could have broken desktop installations after a community dispute spilled over into the project's infrastructure. The Linux distribution disclosed the incident in a forum post this week, describing what it called an attempted act of "distribution sabotage" allegedly involving Davide Beatrici, a developer known for his work on the Mumble instant messaging app. According to OpenMandriva, Beatrici joined the project some time ago and later offered to migrate its repository infrastructure from GitHub to his privately operated OneDev instance, mirroring several dozen repositories in the process. While some maintainers were uneasy about concentrating so much of the project's infrastructure in one person's hands, the proposal went ahead because, as the project put it, "he was such a well-known figure that we didn't expect anything bad." OpenMandriva says trouble started after two other contributors joined alongside Beatrici. One allegedly engaged in repeated abusive behavior toward users and project members, much of it in private messages. The project says several contributors left before the maintainers finally stepped in, kicking the individual out of the OpenMandriva-Cooker Matrix chat. He wasn't banned from the project, but OpenMandriva says the decision "triggered a cascade of events." Beatrici and another contributor then resigned. When OpenMandriva later decided there was little point continuing to mirror repositories to Beatrici's private infrastructure, it says it began severing those connections. According to the project, that didn't go down well. "This infuriated Davide so much that, abusing of the administrative privileges he still had, he sabotaged the distribution today in the early morning hours," the statement says. OpenMandriva alleges Beatrici deleted parts of its GitHub repositories containing years of development work. It also says he "decided to publish an empty package in the cooker repository, which obsoleted all gnome and cosmic packages, which could have damaged the systems of people using gnome or cosmic." The Cooker repository is OpenMandriva's rolling development branch, not a stable release, so the damage appears to have been confined to bleeding-edge users rather than to everyone running the distro. Even so, having one disgruntled admin yank years of work and potentially break package updates isn't the sort of resilience test most projects volunteer for. The project says it is restoring the deleted repositories and repairing the affected packages. It also says it carried out "a full system audit" and found that "aside from the removed packages, we found no other violations." OpenMandriva adds that it considered legal action, saying the alleged sabotage "constituted a criminal offense," but ultimately decided against it. According to tech publication The Lunduke Journal, Beatrici said that "this was by no means a sabotage. The objective was not to harm the distribution I cared for." He reportedly admitted deleting Cosmic and Gnome repositories and said he did this because someone was "messing with my work." The Register has contacted OpenMandriva to ask whether any stable releases were affected, how many repositories were deleted or modified, and what changes the project plans to make to administrative access. We also reached out to Beatrici but have not heard back. Every project needs contributors, but they don't all need the kind of access that can turn a disagreement into a recovery exercise. ®

The European Commission has ended an investigation into possible anticompetitive practices after SAP agreed to abolish reinstatement fees and reduce back-maintenance fees. The move could reduce barriers for customers considering third-party support for products nearing the end of their vendor support terms, including thousands of large businesses that rely on SAP ERP Central Component (ECC) to run their business operations. SAP's mainstream support for ECC ends in December 2027, while customers can opt for extended maintenance until December 2030 by paying an additional two percentage points on their maintenance fees. The most recent figures from Gartner showed that in Q4 2024 only 39 percent of worldwide ECC customers – from a total of 35,000 – had bought or subscribed to licenses to start their transition to SAP S/4HANA, the replacement ERP product. In September last year, the European Commission launched a formal investigation into SAP's behavior in the aftermarket for maintenance and support services in Europe. It said it was responding to concerns that SAP restricted competition in this crucial aftermarket by making it harder for rivals to compete, leaving European customers with fewer choices and higher costs. In October, SAP published its response. “SAP’s commitments aim at improving the financial attractiveness for customers who wish to reinstate SAP maintenance and support services. Thus, future costs associated with reinstatement will not financially prevent customers from choosing to terminate SAP maintenance and support for a given period of time,” the document said. SAP has now agreed to abolish reinstatement fees and reduce back maintenance fees charged to customers who return to SAP's support after a period of absence, the Commission confirmed. It also agreed to clarify conditions that allow customers to choose different maintenance and support service providers and different levels of support from SAP. The agreement is relevant to customers considering third-party support to extend their use of ECC beyond vendor maintenance. For example, last year, European retailer Kingfisher — owner of well-known UK brands B&Q and Screwfix — told a Gartner conference it had chosen Rimini Street to support ECC 6.0 because it saw insufficient value in migrating to SAP S/4HANA. EC executive VP Teresa Ribera said in a statement, “SAP’s software is critical to businesses across Europe, and indeed globally. Today’s decision gives customers using SAP’s popular on-premises business management software more freedom to choose maintenance and support services without unfair restrictions that raised their costs and stifled competition. The legally binding commitments secured by the Commission set a benchmark for the industry more broadly and should serve as a warning against practices with similar effects in the cloud markets, where customers are increasingly moving.” SAP said in a statement that its maintenance practices were aligned with industry standards and that customers had a broad range of deployment, licensing, and support options across its on-premise and cloud products. “The commitments strengthen customer choice and predictability by making policies more transparent, introducing targeted flexibility for exceptional shelfware situations and reinforcing consistent execution through improved guidance, training and independent oversight," it said. “The decision relates solely to on-premise maintenance policies and does not concern SAP’s cloud offerings. However, the added clarity and flexibility support customers as they modernize toward an AI-enabled autonomous enterprise at their own pace. In closing this matter, SAP is able to move forward with a clear framework for customers, partners and investors.” The commitments offered by SAP will remain in force globally for ten years. ®