Planet News Roundup
This is a roundup of articles from the openSUSE community listed on planet.opensuse.org. This community blog feed aggregator lists the featured highlights below from Sept. 11 - 17.
This week highlights the release of Agama 24 with a two-for-one announcement covering versions 23 and 24, the beta release of Plasma 6.8 along with its first “This Week in Plasma” review, the first bug-fix update of KDE Gear 26.08 plus feature overviews of Kdenlive and Minuet, the Tumbleweed week 37 review and a September ARM update roundup, Intel’s OpenVINO 2026.4.0 release, the openSUSE Kudos report for August, the reveal of the GNOME 51 wallpapers, and a fond farewell to the Linux Hispano comic strip after 1,144 installments.
Here is a summary and links for each post:
Introducing Enhanced Distribution Support
The Open Build Service blog introduces Enhanced Distribution Support, a beta feature that gives structure to distribution information through a vendor → distribution → release hierarchy, letting maintainers record what they ship, which repositories and architectures each release is built from, and dated lifecycle milestones. It’s an early, purely descriptive first iteration, and the OBS team is asking distribution maintainers what else they need to be able to describe.
OpenVINO 2026.4.0: A New Version Announced
Alessandro’s blog reports on Intel’s OpenVINO 2026.4.0 release, which broadens model support across Intel CPUs, GPUs and NPUs while adding speculative decoding with Multi-Token Prediction. The post also covers EAGLE-3 Tree Drafting, ITT tracing extended to the NPU for VTune, an ASR pipeline for Node.js and idle-model management in the OpenVINO Model Server.
Tumbleweed ARM Updates in September
The openSUSE News blog covers September’s Tumbleweed updates for the ARM port, led by a security-heavy kernel 7.2.5, glibc 2.44 with its large AArch64 push and QEMU 11.1.1. Raspberry Pi owners get a firmware fix that clears up a Wi-Fi scan error, Rockchip and other SBCs receive MediaTek Wi-Fi kernel fixes, and GIMP 3.2.6, shotwell 33.0 and kbd 2.10.0 land AArch64-specific improvements.
GNOME 51 Wallpapers
Jakub Steiner reveals the GNOME 51 wallpaper set, calling the collection evolution rather than revolution as it sticks to its geometric roots. The default is a stylistic touch-up of the GNOME 50 hexagons, images now carry an embedded thumbnail that helps the Appearance panel, and attribution and license data are embedded directly into the files.
Download a Random Linux Hispano Comic Strip in the Terminal
Victorhck shares a pair of scripts for revisiting the 1,144 Linux Hispano comic strips that Danigm published between 2004 and 2026 after the series came to an end. A Python scraper archived every strip into a GitHub repository, while a Bash script picks a random strip and displays it with feh along with its name and publication date.
Syslog-ng repo for Amazon Linux 2023 updated to syslog-ng version 4.12
Peter Czanik reports that his syslog-ng repository for Amazon Linux 2023 has been updated to version 4.12 after a GitHub request finally prompted an update. He recounts how the repo, first created two years ago, sat untouched due to a lack of feedback despite healthy download numbers from Copr, with Java support and slog disabled as in his other repositories.
Kudos August 2026
The openSUSE News blog presents the monthly Kudos report, in which 17 users received 24 kudos and 419 badges were awarded to 251 contributors during August. Bernhard Wiedemann tops the list with five kudos, mostly for his work on Slowroll, and four contributors reached the milestone of 100 Tumbleweed Contributions.
What’s New in Minuet from KDE Gear 26.08, “Enjoy Shiny Stuff” Edition
The KDE Blog closes its KDE Gear 26.08 feature series with Minuet, the music education app that trains the ear through interval, chord, scale and rhythm exercises. Minuet debuts a new interface that works on both desktops and phones, with a reworked home page and navigation sidebar, exercises shown as cards and a search field that filters by translated names and descriptions.
Software Freedom Day 2026
The KDE Blog announces the Dia de la Llibertat del Programari (Software Freedom Day) celebration organized by Caliu on September 19 at the Espai Jove La Fontana in Barcelona, an event where the KDE community is usually well represented. The post translates the Catalan agenda, listing three talks on teaching computing with values, bootstrapping a self-managed forge with Haz-CLI and security on GNU/Linux.
Happy Programmer’s Day
Alessandro’s blog celebrates Programmer’s Day, which falls on the 256th day of the year and explains why the number 256, or 2^8, matters in computing. The post thanks developers for turning coffee into software, errors into lessons and ideas into innovation.
Innit
Jakub Steiner shares how he produced the tune “Innit” for his weeklybeats 2026 project on the Elektron Analog Four rather than his usual Dirtywave M8. After years of avoiding them, he finally learned the box’s performance macros, which bundle up to five track parameters onto a single knob and make live performance far less stressful than pre-programming a sequence.
Plasma 6.8 Beta Released – This Week in Plasma
The KDE Blog translates Nate Graham’s weekly Plasma development report covering the 6.8 beta. Kup joins Plasma as its backup system, clipboard and MIME-action settings become far clearer, the on-screen keyboard button now behaves as expected on lock and login screens, and KWin gains support for the Wayland commit_timing protocol.
Linux Saloon 219 | Butterbian and ButterKnife Distribution Exploration
Nathan Wolf posts episode 219 of the Linux Saloon podcast, covering activities around the Vintage Computer Fest Midwest. The episode gathers user experiences with the Butterknife and Butterbian distributions through polls and links out to upcoming events and participant projects.
First Update of KDE Gear 26.08
The KDE Blog presents the first update of KDE Gear 26.08, released almost a month after the feature set, bringing more stability, better translations and small improvements. It lists resolved issues across apps, libraries and widgets, including an arrow-orientation fix in KDE Connect, room-map handling in Kongress and an unexpected crash when saving documents in Okular.
Plasma 6.8 Beta Released
The KDE Blog reports that the beta of Plasma 6.8 has been released and highlights changes drawn from its changelog, from the new Hanabi wallpaper replacing Waterfall in Breeze and Nepali calendar support to GTK4 window decoration fixes, an AppStream Preview backend in Discover and a QML rewrite of DrKonqi’s coredump viewer. It encourages users to install the beta and report bugs to help polish the release.
Tumbleweed – Review of the Week 2026/37
Dominique Leuenberger and Victorhck review the four Tumbleweed snapshots (0904, 0907, 0908 and 0909) of week 2026/37, which delivered Linux kernel 7.2.3 and 7.2.4, LibreOffice 26.8.0.3, QEMU 11.1.1, zypper 1.14.101 and more.
Tiny Wins for Packagers: End-of-Week Update
The Open Build Service blog shares its end-of-week “Tiny Wins” update, which shipped a fix preventing a second click on the repository dropdown after an autocomplete selection and ensured the generated SBOM includes the SPDX V3 createdBy field. It also reports that build.opensuse.org served 22.8 million HTTP requests and produced 2.38 million package builds over the last seven days.
Agama: Releasing Version 24
The Agama blog announces Agama 24 in a two-for-one post covering versions 23 and 24 after a stabilization phase focused on testing and polish. Language, keyboard and time zone now share a single searchable form, network connections can be bound to devices through a sortable device table, and JSON search sections gained much more expressive device matching.
View more blogs or learn to publish your own on planet.opensuse.org.
akarshFixing a UTF-8 Encoding Crash in OBS’s Workflow YAML Downloader
A few weeks ago I picked up my first real bug in the Open Build Service (OBS) codebase, part of the openSUSE project: a production crash that had been showing up on Errbit every time a GitHub workflow config with the “wrong” character in it got synced. Here’s what broke, why it broke, and how the fix closed it out.
The bug
OBS lets you wire up a GitHub repository so pushes trigger a workflow, things like package rebuilds via webhooks. When that webhook fires, Workflows::YAMLDownloader pulls down the workflow's YAML config file from GitHub and writes it to a Tempfile so OBS can parse it.
The crash looked like this in production:
"\xE2" from ASCII-8BIT to UTF-8
Thrown from IO#write, inside Workflows::YAMLDownloader#create_temp_file.
Why it happened
GitHub’s API returns file contents Base64-encoded. OBS decodes that with Base64.decode64, which hands back a plain Ruby string, but that string is tagged with the ASCII-8BIT encoding (Ruby's way of saying "this is just raw bytes, no assumptions about text"). That's correct and expected; Base64.decode64 has no way to know what encoding the original bytes represent.
The problem shows up one step later. Rails’ Tempfile defaults to expecting UTF-8 when you write to it. So when OBS calls .write with an ASCII-8BIT string, Ruby tries to transcode it, converting the bytes from ASCII-8BIT to UTF-8 on the fly. That's fine for plain ASCII content. But the moment the YAML file contains a multi-byte UTF-8 character, an em dash, a smart quote, anything outside the ASCII range, Ruby has no valid transcoding rule to apply, and it raises Encoding::UndefinedConversionError.
In short: a perfectly valid UTF-8 file was crashing because Ruby thought it was starting from ASCII-8BIT and tried to “translate” bytes that didn’t need translating at all.
The fix
The core of it comes down to one call:
decoded_content.force_encoding('UTF-8')added right before the Tempfile write. The full diff landed at three lines changed across two files: the fix itself in yaml_downloader.rb, plus a matching update to the spec covering that code path, so the regression is actually tested going forward and not just patched.
The key distinction with the fix itself is force_encoding versus a real transcode. force_encoding doesn't touch a single byte; it just relabels the string, telling Ruby "trust me, these bytes are already UTF-8." Since the bytes coming out of Base64.decode64 genuinely were valid UTF-8 all along (they just carried the wrong encoding tag), relabeling is exactly the correct fix. No transcoding attempt means no UndefinedConversionError, because there's nothing left to convert. Ruby writes the bytes exactly as they are.
What I’d tell someone hitting this themselves
If you’re decoding Base64 content in Ruby and writing it somewhere that expects text, check what encoding the decoded string is actually tagged with (str.encoding) before you assume it matches what the destination wants. Base64.decode64 will hand you ASCII-8BIT regardless of what the original content was; it's on you to correctly re-tag it if you know better, and force_encoding is the right tool for that when you're confident the bytes are already valid in the target encoding. Reach for a real transcode (.encode) only when the bytes genuinely aren't in the target encoding yet.
Getting the PR in
This was my first contribution to OBS, and I want to call out something small that made a real difference: I asked the maintainer, hennevogel, if it was okay for me to take the issue before opening a PR. His reply was simple: “no need to ask me. Just PR small things :)”, which told me a lot about how this project, and the openSUSE community around it, actually wants newcomers to engage. Low friction, just ship the fix.
The PR (#20236) was small, three lines changed across two files, reviewed, and merged. Small fix, but a good first real look at how OBS’s Rails frontend talks to external Git hosts, and that boundary, it turns out, is where interesting bugs tend to live.
Introducing Enhanced Distribution Support
Tumbleweed ARM Updates in September
There were a few software package updates for openSUSE Tumbleweed ARM this month with snapshots reaching the ARM port.
September’s snapshots of the ARM port carried the Linux kernel to a security-heavy 7.2.5. There was a jump to glibc 2.44 with its large AArch64 push, and QEMU reaching 11.1.1. Raspberry Pi owners get a firmware fix that clears up a Wi-Fi scan error, the Networking Stack received security patches, and GIMP 3.2.6 brings AArch64-specific fixes.
Raspberry Pi
This month’s raspberrypi-firmware update prevents the spurious “-52” error message that could appear during Wi-Fi scans (bsc#1215134), so scanning for networks is quieter on boards where the firmware reported a phantom failure. The Linux kernel 7.2.4 update fixed a memory leak in the brcmfmac driver’s SDIO control path, which reduces memory pressure over long uptimes on the onboard Cypress Wi-Fi chip used across recent Pi models.
Rockchip & Other SBCs
Owners of Rockchip and other single-board computers get kernel fixes for the MediaTek Wi-Fi chipsets that many boards ship with, including headroom fixes for USB/SDIO transfers on mt7925 and EEPROM size validation on the mt7915 and mt7996 drivers, which protects against malformed device data (bsc#1012628). kernel-firmware-qcom adds DSP firmware for the Qualcomm x1e80100, nord and qcs8300 platforms, improving modem, DSP and graphics support. dracut now includes leds-qcom-lpg in the aarch64 DRM module set, keeping RGB indicator LEDs working when booting from an initrd. Finally, wireless-regdb enables 320 MHz channels for Hong Kong, updates South Africa’s rules, and drops the 60 GHz DFS flag for Togo.
i.MX & NXP
Developers working with NXP hardware gain the imx8mp-evk machine type in QEMU 11.1.1, which emulates the i.MX 8M Plus Evaluation Kit and makes it easier to test software for that platform without physical hardware. The kernel 7.2.3 update fixed the mxs-dcp crypto engine’s source scatterlist length access and improved the Data Co-Processor security block on i.MX System on Chips that use it for accelerated encryption. These changes make NXP hardware slightly more accessible to both emulation and crypto workloads.
aarch64 Virtualization, UEFI & Toolchain
QEMU 11.1.1 is the virtualization highlight of the month, advancing the emulator with Universal Flash Storage emulation for Write Booster and Host-Initiated Defragmentation based on the UFS 4.1 specification, plus vhost-host-user support for offloading real-time clock handling through virtio-rtc. The ARM targets gain new architectural CPU features, and the virt board can now specify cache topology. The ARM-specific fixes include a regression test and boundary-case fix for the SVE2 whilewr/whilerw instructions and a fix for arm_gicv3 kconfig selection, and openSUSE disables GCS linker validation in the spec to keep the aarch64 build working.
glibc 2.44 delivers the strongest AArch64 additions in months. On targets with the Guarded Control Stack extension, glibc now locks all GCS operations, including status, write-on-shadow-stack and push-to-shadow-stack, immediately after enabling GCS with an ENFORCED or OVERRIDE policy, closing a window where a process could still mutate its own shadow stack. Special-case paths for log, exp, sin, cas, sinh, cosh, asinh, acosh and atanh are vectorized for both SVE and AdvSIMD, and vector powr variants are added, which speeds up floating-point-heavy code on capable Cortex-X and Neoverse cores.
ovmf advances to edk2-stable202608, refreshing the UEFI firmware that backs the qemu-uefi-aarch64 subpackage. The update adds AArch64 host test coverage across SecurityPkg, CryptoPkg, PrmPkg, DynamicTablesPkg, MdeModulePkg and MdePkg, plus UnitTestFrameworkPkg support for GCC AArch64, which strengthens firmware quality assurance for arm64 virtual machines. The EFI Memory Attributes Protocol workaround is dropped because GRUB2 now supports the protocol, so booting stays compatible with the cleaner firmware behavior.
sdbootutil improved its systemd-boot handling with a new status command, parallel test execution, a --repair parameter for cleanup, and better reporting of entries with missing files. The update prevents duplicate entries on non-snapper systems, keeps the exit status of sdbootutil calls intact, and avoids writing the recovery PIN to the journal, which is a meaningful privacy improvement for systems using PCR-based secure boot.
GTK4 4.22.5 fixes a build failure on 32-bit ARM and repairs 32-bit Vulkan image builds, so the toolkit and gskvulkanimage both compile cleanly on armv7. A Wayland session crash in gdk_wayland_toplevel_remove_from_session() and a memory buffer fix round out the release for desktop users on ARM.
LLVM 23.1.1 arrives as a bugfix release for the 23.1.0 series, remaining API and ABI compatible while dropping the obsolete i586 build patch.
GIMP 3.2.6 carries the first AArch64-specific wins for the image editor. The clipboard brush and pattern maximum size is bumped to 8192 on AArch64 as well, matching the ceiling that x86-64 and PPC64 have had since the 3.2.0 release candidates, so larger brushes and patterns survive clipboard round-trips on ARM desktops.
shotwell 33.0 makes the jump from the 0.32 series with a full port to GTK4 (requiring at least 4.22), a reworked printing flow, and fixes for the fullscreen toolbar auto-hide and disappearing slideshow icons. The release replaces its dedicated authentication helper with a small localhost web server for publishing, adds a “peek password” toggle to the Piwigo flow and fixes setting desktop backgrounds.
kbd 2.10.0 extends the console stack with XKB support: loadkeys can now generate console keymaps from XKB data, import XKB compose tables and handle XKB group switching, modifier handling, virtual console switching and keypad/editing remaps.
Snapshots at a Glance
| Snapshot | Kernel | Headline changes |
|---|---|---|
| 2026-09-08 | 7.2.3 | glibc 2.44 with AArch64 GCS and SVE/AdvSIMD math, QEMU 11.1.1 with imx8mp-evk, Raspberry Pi firmware Wi-Fi scan fix |
| 2026-09-12 | 7.2.4 | edk2-stable202608 with AArch64 host tests, dracut aarch64 LED module, GTK4 32-bit ARM build fix, LLVM 23.1.1 |
| 2026-09-14 | 7.2.5 | kernel security batch (Tegra241 CMDQV, iommufd, nvmet, Bluetooth RFCOMM, PCI, rndis_host), GIMP 3.2.6 with AArch64 fixes, shotwell 33.0, kbd 2.10.0 |
ARM Resources
For ARM-specific discussions and support, subscribe to the openSUSE ARM mailing list and check the openSUSE ARM wiki for device support information, image downloads, and documentation. New and existing ARM users can also join the openSUSE ARM IRC channel for real-time help.
Tw Arm Update
author: Douglas DeMaio date: 2026-09-16 14:00:00+02:00 layout: post image: /wp-content/uploads/2025/07/sbc.png license: CC-BY-SA-3.0 title: Tumbleweed ARM Updates in September categories:
- Announcements
- openSUSE
- Tumbleweed
- Slowroll
- MicroOS
- arm tags:
- openSUSE
- Tumbleweed
- Developers
- sysadmin
- user
- Open Source
- rolling release
- gamers
- superuser
- distrowatch
- Linux
- kernel
- kernel-source
- Mesa
- graphics
- CVE
- Power Users
- Superuser
- ARM
- AArch64
- Raspberry Pi
- Rockchip
- SBC
- Jetson
- QEMU
- glibc
- UEFI
- snapper
- GIMP
- shotwell
- kbd
There were a few software package updates for openSUSE Tumbleweed ARM this month with snapshots reaching the ARM port.
September’s snapshots of the ARM port carried the Linux kernel to a security-heavy 7.2.5. There was a jump to glibc 2.44 with its large AArch64 push, and QEMU reaching 11.1.1. Raspberry Pi owners get a firmware fix that clears up a Wi-Fi scan error, the Networking Stack received security patches, and GIMP 3.2.6 brings AArch64-specific fixes.
Raspberry Pi
This month’s raspberrypi-firmware update prevents the spurious “-52” error message that could appear during Wi-Fi scans (bsc#1215134), so scanning for networks is quieter on boards where the firmware reported a phantom failure. The Linux kernel 7.2.4 update fixed a memory leak in the brcmfmac driver’s SDIO control path, which reduces memory pressure over long uptimes on the onboard Cypress Wi-Fi chip used across recent Pi models.
Rockchip & Other SBCs
Owners of Rockchip and other single-board computers get kernel fixes for the MediaTek Wi-Fi chipsets that many boards ship with, including headroom fixes for USB/SDIO transfers on mt7925 and EEPROM size validation on the mt7915 and mt7996 drivers, which protects against malformed device data (bsc#1012628). kernel-firmware-qcom adds DSP firmware for the Qualcomm x1e80100, nord and qcs8300 platforms, improving modem, DSP and graphics support. dracut now includes leds-qcom-lpg in the aarch64 DRM module set, keeping RGB indicator LEDs working when booting from an initrd. Finally, wireless-regdb enables 320 MHz channels for Hong Kong, updates South Africa’s rules, and drops the 60 GHz DFS flag for Togo.
i.MX & NXP
Developers working with NXP hardware gain the imx8mp-evk machine type in QEMU 11.1.1, which emulates the i.MX 8M Plus Evaluation Kit and makes it easier to test software for that platform without physical hardware. The kernel 7.2.3 update fixed the mxs-dcp crypto engine’s source scatterlist length access and improved the Data Co-Processor security block on i.MX System on Chips that use it for accelerated encryption. These changes make NXP hardware slightly more accessible to both emulation and crypto workloads.
aarch64 Virtualization, UEFI & Toolchain
QEMU 11.1.1 is the virtualization highlight of the month, advancing the emulator with Universal Flash Storage emulation for Write Booster and Host-Initiated Defragmentation based on the UFS 4.1 specification, plus vhost-host-user support for offloading real-time clock handling through virtio-rtc. The ARM targets gain new architectural CPU features, and the virt board can now specify cache topology. The ARM-specific fixes include a regression test and boundary-case fix for the SVE2 whilewr/whilerw instructions and a fix for arm_gicv3 kconfig selection, and openSUSE disables GCS linker validation in the spec to keep the aarch64 build working.
glibc 2.44 delivers the strongest AArch64 additions in months. On targets with the Guarded Control Stack extension, glibc now locks all GCS operations, including status, write-on-shadow-stack and push-to-shadow-stack, immediately after enabling GCS with an ENFORCED or OVERRIDE policy, closing a window where a process could still mutate its own shadow stack. Special-case paths for log, exp, sin, cas, sinh, cosh, asinh, acosh and atanh are vectorized for both SVE and AdvSIMD, and vector powr variants are added, which speeds up floating-point-heavy code on capable Cortex-X and Neoverse cores.
ovmf advances to edk2-stable202608, refreshing the UEFI firmware that backs the qemu-uefi-aarch64 subpackage. The update adds AArch64 host test coverage across SecurityPkg, CryptoPkg, PrmPkg, DynamicTablesPkg, MdeModulePkg and MdePkg, plus UnitTestFrameworkPkg support for GCC AArch64, which strengthens firmware quality assurance for arm64 virtual machines. The EFI Memory Attributes Protocol workaround is dropped because GRUB2 now supports the protocol, so booting stays compatible with the cleaner firmware behavior.
sdbootutil improved its systemd-boot handling with a new status command, parallel test execution, a --repair parameter for cleanup, and better reporting of entries with missing files. The update prevents duplicate entries on non-snapper systems, keeps the exit status of sdbootutil calls intact, and avoids writing the recovery PIN to the journal, which is a meaningful privacy improvement for systems using PCR-based secure boot.
GTK4 4.22.5 fixes a build failure on 32-bit ARM and repairs 32-bit Vulkan image builds, so the toolkit and gskvulkanimage both compile cleanly on armv7. A Wayland session crash in gdk_wayland_toplevel_remove_from_session() and a memory buffer fix round out the release for desktop users on ARM.
LLVM 23.1.1 arrives as a bugfix release for the 23.1.0 series, remaining API and ABI compatible while dropping the obsolete i586 build patch.
GIMP 3.2.6 carries the first AArch64-specific wins for the image editor. The clipboard brush and pattern maximum size is bumped to 8192 on AArch64 as well, matching the ceiling that x86-64 and PPC64 have had since the 3.2.0 release candidates, so larger brushes and patterns survive clipboard round-trips on ARM desktops.
shotwell 33.0 makes the jump from the 0.32 series with a full port to GTK4 (requiring at least 4.22), a reworked printing flow, and fixes for the fullscreen toolbar auto-hide and disappearing slideshow icons. The release replaces its dedicated authentication helper with a small localhost web server for publishing, adds a “peek password” toggle to the Piwigo flow and fixes setting desktop backgrounds.
kbd 2.10.0 extends the console stack with XKB support: loadkeys can now generate console keymaps from XKB data, import XKB compose tables and handle XKB group switching, modifier handling, virtual console switching and keypad/editing remaps.
Snapshots at a Glance
| Snapshot | Kernel | Headline changes |
|---|---|---|
| 2026-09-08 | 7.2.3 | glibc 2.44 with AArch64 GCS and SVE/AdvSIMD math, QEMU 11.1.1 with imx8mp-evk, Raspberry Pi firmware Wi-Fi scan fix |
| 2026-09-12 | 7.2.4 | edk2-stable202608 with AArch64 host tests, dracut aarch64 LED module, GTK4 32-bit ARM build fix, LLVM 23.1.1 |
| 2026-09-14 | 7.2.5 | kernel security batch (Tegra241 CMDQV, iommufd, nvmet, Bluetooth RFCOMM, PCI, rndis_host), GIMP 3.2.6 with AArch64 fixes, shotwell 33.0, kbd 2.10.0 |
ARM Resources
For ARM-specific discussions and support, subscribe to the openSUSE ARM mailing list and check the openSUSE ARM wiki for device support information, image downloads, and documentation. New and existing ARM users can also join the openSUSE ARM IRC channel for real-time help.
MemberGNOME 51 Wallpapers
With GNOME 51 out the door, it's wallpaper reveal season again. This time around it's evolution, not revolution — the set sticks to its geometric roots.
The default is really just a stylistic touch up of the 50 hexagons. The subtle rim highlight received a spotlight and now shines extra bright.
I do keep hoping landscape nature photos eventually join the lineup, but capturing the same scenery under different conditions so the light and dark variants actually make sense is trickier than it sounds. That one's still on the wishlist.
As usual, plenty of concepts didn't make the cut. For every wallpaper that ships, there's a good pile of experiments that never got past the "that's kinda neat" stage.
One thing we keep struggling with is performance in the Appearance panel. The images now include an embedded small thumbnail, so hopefully a faster way to build the initial cache of thumbnails is on the horizon.
We've also started embedding attribution and license straight into the images themselves, so the credits travel with the file instead of living only in the repo. And with Loupe displaying the metadata nicely, you get to see it conveniently.
Syslog-ng repo for Amazon Linux 2023 updated to syslog-ng version 4.12
Two years ago, I created a syslog-ng repo for Amazon Linux 2023. As I received zero feedback, I left it alone until last week. The repo is now updated to the latest version of syslog-ng: 4.12.
According to Fedora Copr download stats, my Amazon Linux 2023 repo has many downloads. However, for over two years, I received zero feedback from the community, no matter how many GitHub, e-mail or social media posts I used to ask users about their experiences and requirements. As any kind of package maintenance needs some effort, I left the repo alone without updates.
Fast forward to last week: someone responded to my post on GitHub, asking for an update. So, once I got some idle time during a longer task, I updated the repo to syslog-ng version 4.12. During another break, I disabled Java support and slog, just as in any other repository we maintain.
As usual, you can find my Amazon Linux 2023 repo at https://copr.fedorainfracloud.org/coprs/czanik/syslog-ng-amazon23/ and share your feedback at https://github.com/syslog-ng/syslog-ng/discussions/4965
syslog-ng logo
Originally published at https://www.syslog-ng.com/community/b/blog/posts/syslog-ng-repo-for-amazon-linux-2023-updated-to-syslog-ng-version-4-12
Kudos August 2026
Welcome to our monthly report from the openSUSE Kudos recognition platform, where we take a moment to put a spotlight on the people who stepped up, helped others, and made this community a little brighter.
During the month of August 2026, 17 users received 24 kudos, and 419 badges were awarded to 251 contributors. You can see stats for this and previous months.
Spotlight
Top of the list in August: Bernhard Wiedemann (@bmwiedemann) with 5 kudos, more than twice what anyone else collected. Most of them (4) landed in Code & Engineering. That alone accounts for 5 of the month’s 24 kudos. Second place was a three-way tie at 2 kudos each: @bigironman, Gertjan Lettink (@knurpht) and Martin Schreiner (@mschreiner). Behind them, thirteen more people each picked up a kudo of their own, and all of them are listed below.
A milestone is worth stopping for: @dirkmueller, @jengelh, @krop and @pluskalm reached 100 Tumbleweed Contributions. That is the kind of steady, long-haul contribution that rarely makes headlines. Congratulations!
Not all of it is code, either. Documentation, translation, artwork, moderation and event work keep openSUSE usable and welcoming for everyone, and this month @aplanas, @bmwiedemann, @jimed-rand and @lkocman earned Wiki Contributor.
Kudos
Here are the thank-you notes themselves: 24 peer-to-peer kudos written by contributors about contributors, grouped by the person being thanked. The order follows the standings above, so August’s most recognized contributor opens the list.
From @levolet in Code & Engineering
“Thanks for all your great work in maintaining Slowroll. Your community spirit in thinking of it and creating it is also much appreciated.”
From @mdhup in Infrastructure Heroes
“Thank you for the work you do. Using slowroll has been a smooth experience. So much so that it ended my disro hopping. If it’s this good in beta, I can’t wait until it’s finished.”
From @PossiblePrinterPalace in Code & Engineering
“Thank you for the wonderful Slowroll. It has been my most pain-free Linux Desktop experience, in now more than 10 years of using different Linux distros.”
From @dnla in Code & Engineering
“Thanks for maintaining slowroll! Keep up the good work”
From @C7NhtpnK in Code & Engineering
“Thank you for offering Slowroll!”
From @lkocman in Support & User Assistance
“Many thanks for your quick help with re-branching the Leap-Images Wolfgang!”
From @lkocman in Support & User Assistance
“Thanks you for the help with the SPICE devel project for Leap 16.X Wolfgang!”
From @lkocman in Support & User Assistance
“Thank you for raising the concern with spamminess of kudos bot. Thanks to our discussion and the spamminess we actually met a new friend in /bar. So the spam from kudos-bot is actually useful. And we also get to reduce the spamminess”
From @C7NhtpnK in Support & User Assistance
“Thank you for help and assistance at forums.o.o!”
From @C7NhtpnK in Code & Engineering
“Thank you for maintaining LibreOffice!”
From @lkocman in Code & Engineering
“Martin, thank you so much for going the extra mile here! It would have been much easier to simply ship the same LibreOffice version from Factory in Leap, but instead you’re putting in the effort to maintain a separate version that better fits what users expect from a long-term stable distribution. That’s a lot of extra work behind the scenes, and I and many other users really appreciate your work to make Leap an even better LTS-style distro. Thank you! <3”
From @lkocman in Support & User Assistance
“Many thanks for your quick help on building an efficient query for OBS requests accepted within n-days Antonello! I’ll be using it for granting people Tumbleweed contributor badge in kudos. This will make sure we don’t keep OBS busy for too long:)”
From @lkocman in Infrastructure Heroes
“Many thanks for your help with whitelisting access from kudos-o-o to wiki and other services Georg!”
From @victorhck in Code & Engineering
“Thanks for your work keeping Tumbleweed rolling. and for your weekly reviews!”
@firstyear received a kudo:
From @lkocman in Support & User Assistance
“Many thanks you for the help with the 2factor recovery William!”
From @victorhck in Code & Engineering
“Thanks Imobach & Ancor for their work developing Agama, and many other tools since many years..”
From @C7NhtpnK in Code & Engineering
“Thank you for os-update!”
From @C7NhtpnK in Code & Engineering
“Thank you for answering and moderating some user questions about maintenance!”
From @lkocman in Community Moderation
“Many thanks for your quick help with the locked SPICE thread on forums-o-o Malcolm. Users will be happy to find a pointer in case they find this thread in the future. Many thanks!”
From @lkocman in Code & Engineering
“Many thanks for your work on the openSUSE-packaging-skill Martin! I’m really happy that we have somebody who’s proactively looking into how AI can actually improve general packaging experience in openSUSE.”
From @C7NhtpnK in Support & User Assistance
“Thank you for help and assistance at forums.o.o!”
From @C7NhtpnK in Code & Engineering
“Thank you for offering Myrlyn!”
From @lkocman in Code & Engineering
“Many thanks for your support with anything kernel related Takashi! I always really appreciate your help! We might make rtl8812au users happy!”
From @lkocman in Code & Engineering
“Many thanks for all your work on broadcom-wl Tobi! I think many broadcom users will highly appreciate it.”
Badges
Here is an overview of the 8 badges earned in August. The full list of badges you can earn is on the badges page.
Wiki Contributor
Recognition for the first day/page documentation contribution on en.opensuse.org wiki.
Awarded this month to @aplanas, @bmwiedemann, @jimed-rand and @lkocman.
100 Tumbleweed Contributions
For submitting 100 Submit Requests to openSUSE:Factory.
Awarded this month to @dirkmueller, @jengelh, @krop and @pluskalm.
10 Tumbleweed Contributions
For submitting 10 Submit Requests to openSUSE:Factory.
35 contributors earned this badge this month, too many to list here, but every one of them counts.
First Tumbleweed Contribution
For submitting 1 Submit Request to openSUSE:Factory.
210 contributors earned this badge this month, too many to list here, but every one of them counts.
Leap 16.1 Contributor
Recognition for submitting at least one pull request to Leap 16.1 or related appliance repositories on src.opensuse.org.
76 contributors earned this badge this month, too many to list here, but every one of them counts.
Leap 16.0 Contributor
Recognition for submitting at least one pull request to Leap 16.0 or related appliance repositories on src.opensuse.org.
75 contributors earned this badge this month, too many to list here, but every one of them counts.
Got First Kudo
Received your first kudos.
Awarded this month to @atartamo, @bigironman, @firstyear, @IGonzalezSosa, @kukuk, @mschreiner, @pluskalm, @Sauerland and @Tobi_Peter.
First Kudos Given
Shared your first kudos.
Awarded this month to @C7NhtpnK, @dnla, @levolet, @mdhup, @PossiblePrinterPalace and @victorhck.
Congrats on your kudos and badges. Keep up the great work everyone, and don’t forget to have a lot of fun!
Generated by kudos-bot-news-o-o from https://kudos.opensuse.org/api/reports/monthly for the period 1-31 August 2026.
Eight Modules, One Summer: My Amazon ML Summer School 2026 Journey
The email came on July 2nd. The program didn’t start till July 11th, so I had a few days to just sit with it. Getting in wasn’t one step - Amazon first ran an SOP and Resume screening round that narrowed about 134,000 applicants down to 30,000, then an online assessment cut that down to the final 3,000. Clearing that assessment is what actually got me in.
Eight modules followed, taught directly by Amazon Applied Scientists. Here’s what stuck.
Module 1: Supervised Learning
Good refresher on the fundamentals regression vs classification, how a model learns from labeled data. The part that actually stuck came from the QnA: knowing when a model is “good enough to ship” is its own skill, separate from knowing how to build one. Chasing the last 1–2% accuracy isn’t always the right call.
Module 2: Deep Neural Networks
This one went deep, fast.
Loss functions MSE, MAE, Huber loss for outliers, cross-entropy for classification. The line that stuck: picking the wrong loss doesn’t just hurt accuracy, it quietly kills training.
Activation functions sigmoid and tanh were the old standard, but their gradients vanish in deep networks. Rule of thumb now: never use them in hidden layers. ReLU, GELU, and Softmax are the default.
Optimizers AdamW over classic Adam. Classic Adam’s weight decay gets tangled up with its adaptive gradient scaling, so some parameters get under-regularized and others over-regularized. AdamW decouples the two. It’s what trains basically every large model in 2026.
Module 3: Dimensionality Reduction
Kernel PCA was the highlight. Regular PCA can’t separate data like two concentric circles — no straight-line projection untangles them. Kernel PCA lifts the data into a higher-dimensional space where it becomes linear, then runs PCA there — without ever explicitly computing that transformation, thanks to the kernel trick.
Module 4: Unsupervised Learning
Two things stuck. Recommendation systems are really matrix completion the user-item ratings matrix is over 99.9999% empty at Amazon’s scale, and the whole approach rests on assuming a few hidden factors explain most ratings. And NMF vs PCA: same faces, same rank, but NMF’s non-negativity constraint gives you actual interpretable face parts instead of PCA’s abstract, harder-to-read components.
Module 5: Generative AI and LLMs
Covered the basics of how LLMs work next-token prediction at massive scale, the transformer architecture, and attention as the mechanism that lets a model weigh which earlier words matter most. RAG came up too, as the fix for a model not knowing anything past its training cutoff.
Module 6: Sequential Learning
This was about handling anything with order text, time series, audio. RNNs process a sequence step by step but lose information over long stretches. LSTMs fix that with gates that decide what to keep and forget, which is why they held up for years before transformers took over.
Module 7: Causal Inferencing
The core idea: correlation isn’t causation, and ML is very good at finding the former while staying blind to the latter. A/B testing is the gold standard for proving causation because randomization breaks the link to hidden confounding variables but real products can’t always run one, so a lot of this module was about the next-best alternatives.
Module 8: Reinforcement Learning
A different paradigm entirely no labeled data, just an agent taking actions, getting rewards, and learning a policy through trial and error. Also came up in the context of RLHF, the technique behind aligning LLMs with human preferences, which is why RL is having a moment again despite being a decades-old idea.
The QnA Sessions
Honestly, this is where the program earned its reputation more than the slides did. A few things I asked or heard discussed: whether the depth-vs-width tradeoff in a neural net changes with less training data, how you know unsupervised clustering is actually good with no ground truth to check against, and whether the low-rank assumption in recommendation systems holds up once a matrix gets that sparse.
What’s Next
The program ended with an invite to apply for Applied Scientist Intern roles at Amazon. I filled out the interest form, and that led to an online assessment round two DSA questions plus a behavioral section. Now it’s back to sharpening DSA and going deeper on whatever felt shakiest, and waiting to see what comes next.
The Swag and the Certificate
Somewhere in between all the sessions and the assessment, the program wrapped up with a completion certificate — and some Amazon swag showed up too, which was a nice surprise on top of everything else.
Four weeks, eight modules, and more than a few things I thought I already understood turned out to have a lot more underneath.
Next up: the Amazon ML Challenge. Let’s see how that goes.
Grateful to Lathish Kumar for running these sessions so smoothly for 3,000 of us, and to the Amazon Scientists Anoop Saladi, Purav Aggarwal, Apoorva Singh, Vikas Raykar, Sandeep Chatterjee, Upasana Ramakrishnan, and Rudra Singh.
Innit
I've been doing weeklybeats 2026 almost entirely on the Dirtywave M8 -- often on a Sunday night, in bed, on an actual hw. This was before Tim went and elevated a hardware abstraction layer into a proper iOS App (which I still very much recommend). This particular tune though, "Innit", is an oddity for me: it lives on the Analog Four, a box I got from a friend ages ago for an absolutely irresistible price. It mostly collected dust. I'd make a few tunes here and there but never really dug into the stuff that makes it unique. Other than having individual channels ready for effects pedals and exposing the Elektron sequencer via CV for your eurorack addiction (stuff that I for sure will never use).
I've been putting off learning about the performance macros forever, and I hate myself for it. Luckily the amazing Ivar Tryti revealed his mastery on Youtube which helped a lot. The concept is simple: you bundle up to five track parameters onto a single macro knob. Ten of them live on the performance page making is far less stressful of a navigation when performing. There's a setup cost, you need to precisely tune the parameter interactions, but once that's locked in, live performance becomes a far less stressful affair. You can also live map any or all of these ten mappings onto a single encoder for otherwise physically impossible twistings. This is easily remappable during the performance.
I'm still a noob at this, but the point is the same as always: there's real joy in performing and building up energy for the drop in real time -- way more satisfying than preprogramming the whole sequence.