Not for the faint of heart

Mirko Brombin, the creator of Bottles, has got Microsoft 365 running on Linux. His posts on X showcase an installation that completes on a Windows 10 or 11 Wine prefix, signs in with his existing license, and launches at least one editor.

He achieved this using two things, an unreleased version of Soda and Bottles 67.4. If you remember, Soda is a Wine-based runner that facilitates easy running of Windows games and applications.

On first glance, what he posted might look like a developer-focused log, but these changes are the foundation of what full Microsoft 365 support via Bottles will one day hopefully look like.

The results of his work

Mirko was able to login into the MS 365 app using his credentials and was able to launch MS Word and use it. Of course that's not the only thing he got working; since two-factor authentication passed, the suite detected his license, and OneDrive integration was functional.

The memory footprint of the office suite was less than 500MB, so this should be light on your Linux system.

The build he is currently testing is 2608 (20326.20144), which he says is the last build the installer will offer for Windows 10 and 11. And before you get all hyped up, wanting to try this out on your computer, know that you will have to wait.

Soda 11.0-11 (error 404 for now) has no release yet, and the component catalog that ships with Bottles 67.4 stops at Soda 11.0-10. Even the long-standing community request for a Microsoft Office installer in Bottles is still open.

Do we need Microsoft 365 on Linux?

Till date, there's not been an official Linux desktop build of Microsoft 365, and seeing how Microsoft currently operates, I don't expect one to appear anytime soon.

What's been achieved here is remarkable. Once fully implemented, this will help draw in people into Linux, as many shudder at the thought of switching to the platform because their preferred office suite doesn't work here.

For context, CodeWeavers sells a paid Wine product, CrossOver, and its database still lists Microsoft Office 365 on Linux as "Installs, Will Not Run," last tested against CrossOver 26.3.0.

Which is why I hope the fix travels further than Bottles. Wine sits beneath every one of these projects, including Soda. So a patch that lands upstream reaches all of them at once, and you would not need a particular runner to get Office working.

Two new apps are coming to Linux Mint later this year: a new calendar and eBook reader.

Plus, the Ubuntu-based Linux distribution has redesigned its (overlooked) document hub, Library, to make it easier to find and manage documents scattered across you device from one place.

Both new apps join the XApp Project, Linux Mint’s collection of home-grown, cross-desktop GTK apps (which now has a new website wearing the popular Claude design aesthetic).

Clockenstien is Mint’s new Calendar app

Clockenstein is Linux Mint’s new calendar app but, lest anyone scream in horror at the terrible name, the user-facing label is just Calendar. It replaces GNOME Calendar in Linux Mint’s default software lineup (about time – it was an older, pre GTK4/libadwaita build).

Now a calendar is a calendar. It needs to let you see and browse dates, add events, create calendars for different tasks – things that Linux Mint’s clockenstein can do from the off.

But integration with online services is important since most of us move between devices during the day. So calendars can be entirely offline and kept local, or sync to/from online providers like Google Calendar or through CalDAV via cloud services like Nextcloud.

Clem says the goal was make sure it was always obvious when remote calendars aren’t working and in-sync, so users can tell if something is up. Failing silently isn’t helpful if there’s an important event you assume you’ll get a reminder about, but never do.

“Down in the laboratory, there’s a little more going on”, says Lefebvre.

A background daemon is used to keep remote calendars in sync and check for any upcoming events that have reminders enabled. If there’s an issue, events stay visible but become read-only – you can’t edit or add – until connect is back online.

Reminders are handled via a notification agent. They get shown on the desktop at the appropriate time so you can action or snooze or be reminded later. Backend interfaces allow other desktop environments to interact or replace the notification agent.

Book Reader ePub viewer

The new Book Reader (xepub) app is designed (surprise) for opening ePub files. It looks to offer a clean reading experience, that Linux Mint’s Clement Lefebvre says is ‘similar’ to a Kindle or a Kobo. It also keeps track of where you are in a book so when you re-open it, you can resume reading.

Features wise, it will offer what you’d expect: chapter and page navigation, full-text search and bookmarks, as well as adjustable page zoom and font options. An annotations feature lets you highlight selections of text, with a colour, and add notes.

Finally, an internal look-up system lets you select a word and ferry it to a search engine (wikipedia, dictionary, etc) to find out more about it. Library gets a dust off Linux Mint's Library app showing document filtersThingy – a personal document hub

Linux Mint has also redesigned its Library document manager, which you may know by the package name thingy.

Documents are now grouped by file type, rather than the app that opens them. So, for instance, you can see just your PDF files in one view, rather than finding them scattered amongst all files the the default document viewer XReader handles.

Common file types are listed in the Library sidebar, but you choose what’s shown and in one order. And, more usefully, you can create your own custom groups that contain a mix of different file types, and even specify them from specific folders.

When can I get the new apps?

Linux Mint’s new Book Reader and Calendar apps will make their official debut in Linux Mint 23, expected to arrive towards the end of the year. That release will also ship with the new and improved Library app too.

Plus, other changes the team has revealed and teased to date (and, no doubt, a few still to come).

If you don’t want to wait, the source code is up on GitHub (each linked in their respective paragraphs above). You can check them out (pun klaxon) to test, and file any issues you find or glaring feature requests you may have.

Exciting stuff!

KaOS 2026.09 is now available as the second stable ISO of the independently developed, Qt-focused Linux distribution, fully based on Dinit, and marks another step away from systemd and KDE Plasma.

The new release uses a startup stack built around Dinit, Turnstile, and seatd for system startup, sessions, and services. Systemd components are now used mainly for udev and tmpfiles. Elogind remains only to provide Polkit integration.

For logins, KaOS has switched to greetd with the console-based Tuigreet as the default display manager. A new Noctalia-based graphical greeter is also available in the repositories, although the developers say it is not yet as complete as Tuigreet.

On the desktop side, KaOS continues its move toward a Niri and Noctalia-based environment. The ISO ships with Niri 26.04 and Noctalia 5.1. Noctalia brings a new screenshot region editor with annotation tools, ModemManager support for mobile data, and collapsible groups in the settings interface.

The distribution remains firmly focused on Qt applications and, according to the project, the new ISO is still GTK-free.

One of the biggest changes, however, concerns KDE Plasma. Plasma 6.7 remains fully available in the repositories, but KaOS has confirmed that it will be the final Plasma series offered by the distribution.

Once KDE Plasma moves to version 6.8, Plasma packages will be removed from the KaOS repositories. The project says this decision is tied to the broader move away from systemd and its preference for the Niri/Noctalia setup going forward.

KaOS 2026.09 also includes a substantial refresh of the underlying system. According to the developers, roughly 70% of the distribution was rebuilt during the past two months because several major updates were held back until the Dinit transition was completed.

Among the updated components are Linux kernel 7.1, GCC 15.3, glibc 2.43, Binutils 2.46, Mesa 26.2, PipeWire 1.6.8, GStreamer 1.28, OpenSSH 10.5, Rust 1.98, CMake 4.4, ZFS 2.4.4, MariaDB 12.3.3, and Poppler 26.09.

Another notable change is the bootloader. Limine is now the default, while systemd-boot has been removed as an installation option. Other UEFI bootloader choices remain available through the Calamares installer.

Speaking of Calamares, it has also been updated so that it can now operate fully under a pure Wayland session, including fixes needed to restore text input in QML-based installer modules such as user creation.

Among the newly added packages are noctalia-greeter, umbriel, xdg-desktop-portal-umbriel, wlroots, and openrc-settingsd. Audacity 4 has also returned to the repositories now that it is Qt 6-based, replacing Tenacity.

For storage, automated partitioning in Calamares continues to support XFS, EXT4, Btrfs, and ZFS. XFS remains the default filesystem for the ISO.

Users planning to test KaOS 2026.09 in VirtualBox should note that 3D acceleration and the VMSVGA graphics controller are required due to the distribution’s Wayland-based desktop. The developers warn that the system will not boot in VirtualBox without this configuration.

There are also a few known limitations. RAID installations are currently unavailable, XFS cannot be used for BIOS installs because of a GRUB compatibility issue, and some privilege escalation actions may not work because Polkit has not yet been fully adapted to the Turnstile/seatd stack.

Lastly, KaOS warns that its installation images do not support Ventoy, Rufus, or UNetbootin for creating bootable media, so users should follow the project’s official installation-media instructions instead.

For more details, see the official KaOS 2026.09 release announcement.

What would you do if you bought some capable hardware that you really like, but there’s no Linux distribution that runs well on it?

When Jason Perlow found himself in that predicament, he broke out some tools and built his own Linux distribution to fully support it. He’s calling the distro NCZ-OS, and it currently supports arm64 machines, with a specific focus on the CIX Sky1 platform, targeting Radxa Orion O6/O6N and Minisforum MS-R1 devices. Support for OrangePi systems is currently under validation.

He’s not stopping there. He’s now working on getting it to work on some older hardware that’s been stranded by Microsoft and Apple.

If Perlow’s name seems familiar, there’s a reason for that. For years his byline graced articles on sites such as CNN, Linux Magazine, ZDNet, and others. He’s also spent stints as an employee at the Linux Foundation, Microsoft, and IBM. Some of you might also know him for his foodie videos.

“I think the story here is interesting because I have heavily used AI pair-programming tools (Codex, Claude, OpenCode/Goose/Pi) to put this distro together, along with Yocto as the build system,” he told me in a conversation that took place via email. “This would have been basically impossible for a single person to do five years ago.”

Making Linux Safe for Arm CIX Sky1

In some ways the project harkens back to the 1990s, when coders were working to build GNU/Linux distributions that would work on whatever hardware they had laying around and needed to get it to work on.

“I bought a Sky1-based system,” Perlow said. “It was an attractive piece of hardware for $600 — 64GB RAM, 12 cores, GPU, and NPU. When it arrived, a lot of stuff wasn’t working out of the box. I decided to keep it and not return it, and see if I could get it to do as advertised. I now own two (Minisforum, Radxa) with a third on the way (OrangePi).”

NCZ-OS originally used an Ubuntu base, but now starts with the arm64 port of Debian testing (forky). It uses a customized Debian installer with a bit of built-in failsafe: the base system is installed directly from the image, doing away with issues, like maybe an iffy network connection or unreachable mirror.

“The goal is specific: make the SoC’s accelerators work under a current kernel and a fully Wayland desktop,” he said. “On new Arm silicon the accelerators are usually the last thing to work and often the thing that never does, with the GPU falling back to software and the NPU and video engine present only as device nodes nothing can drive. Closing that gap is the point of the project.”

There have been numerous workarounds necessary to get the OS working to his liking. That’s not surprising. According to Perlow, the CIX Sky1 is not only new technology, it was apparently initially designed for China’s domestic smartphone market, so architecturally it has more in common with a phone-class system-on-chip than a conventional PC. Although it’s supported by some other distros, the support is incomplete.

“NCZ-OS is not the only option and I would not want to suggest otherwise,” he said. “Armbian maintains a CIX board family, Radxa publish Debian and Fedora images for the Orion O6, Ubuntu has a concept image, and CIX publishes an SDK and is upstreaming core SoC support, which matters more in the long run than any downstream image. A community project, Sky1-Linux, assembled kernel and GPU packaging and drew genuine interest, though its repositories have had no activity since March 2026.

“What differs here is the scope of what is made to work on one image, and the practice of measuring each claim on hardware before making it,” he added.

Perlow isn’t stopping with NCZ-OS as a specialized Arm distro. He’s working on a specialized x86 version too.

“The x86 target is deliberate rather than opportunistic, and it is the one I care most about,” he said.

He’s attacking x86 on two fronts, starting with enabling full desktop support for older Macs that were built before Apple said goodbye to Intel and x86 in 2020.

“I have been a committed Mac user since leaving Microsoft in 2017, and none of the Linux desktops I’ve used since then has given me a compelling reason to switch,” he said. “I own an Intel Mac from 2020 which is largely abandonware, with the exception of the T2 kernel being developed [a Linux kernel with patches or drivers for T2-equipped Macs], but there is no distribution that really is tuned for it. They all kind of stop at the kernel and are the same Ubuntu or Arch or whatever distro experience everyone else gets.”

“That’s not a bad thing necessarily,” he added, “but I think you need a compelling reason for end users with older systems to switch to Linux more than ‘my abandoned hardware now works.’ That doesn’t pull millions of systems necessarily out of the e-waste pile.”

Mac branded x86 machines that were manufactured in the years leading up to Apple’s change to its own chip design are still capable today, Perlow pointed out, but they need a specific kernel patch series to make the keyboard, trackpad, audio, and networking work.

“That series exists and is maintained by the t2linux community,” Perlow said. “Carrying such a series on a current base is the same task this project already performs for Sky1, which is why it is a realistic target rather than a wish.” A Wintel 10 Lifejacket Too…

There’s more in Perlow’s sights than specialized Arm and old Apples.

“I also have several Windows 10 laptops that are perfectly good but also abandonware — right now they are servers, I don’t like the Linux desktops that exist today. I want to develop a fully modernized distribution that is open source but also delivers an end user experience that doesn’t compromise and ‘just works.'”

“The Steve Jobs approach,” he added. “That’s a very lofty thing for one person to try to do. I’m no Google — but I’m arrogant and motivated…”

He’s talking, of course, about Windows 10 machines with modern specs that couldn’t make Microsoft’s cut for the upgrade path to Windows 11. Official support for Windows 10 from Redmond ended last October, although in some cases extended support remains available through October 2027.

NCZ-OS’s desktop environment is Singularity, which is not a Windows clone by any stretch, but offers a conventional layout that won’t confuse Windows (or Mac) users. It’s from Italy-based Mirko Brombin, whose tech resume checks multiple boxes, including UI/UX designer, open-source developer, and Linux project founder. He’s probably best known as the creator of the popular Wine front end, Bottles, and the Debian-based Linux distribution Vanilla OS.

For the x86 version, Perlow said he intends to use the same Debian Forky base, installer, and Singularity Wayland desktop, paired with a Yocto-built Linux 7.2 kernel.

“The initial targets include T2-equipped Intel Macs from 2017-2023 and capable Windows 10 systems that cannot officially move to Windows 11,” he said. “The goal is the same on both Arm and x86: support as much hardware acceleration as possible and deliver a modern, cohesive experience rather than merely making abandoned hardware boot Linux. I expect the x86 work to be easier, although hardware has a habit of punishing statements like that.”

Release information and downloads are available on GitHub.

Found a lot of interesting softwares in there, I might make a list of interesting and potentially useful softwares I discovered

I was talking with a friend about Linux, getting him interested and all that. He was taking to it. I told him about Flathub and since we were in front of a computer he just instantly searched it in the URL bar and figured it was a .com address. We were both caught quite off-guard when a different type of site appeared hahaha. Flathub.org is the correct Linux related site but can be easily mistaken for porn

Protest against the Omarchy/Omacom Foundation and the tech companies and executives funding far-right open source.

Note that i made this for myself and just my first time sharing an application i made here

Came up with this application with one aim in mind...To Control daily limit for myself as my doomscrolling had worsened quite a lot even while working. So decided to make this as an alternative to Digital wellbeing on Android.

It supports:

focus mode

per app daily limits

background daemon

so far i was able to make it work only on GNOME mutter and all x11 because that is my system...will consider updating in future if this receives positive reviews

I use fedora 44 workstation and it runs perfectly for me...and also tied it on manjaro and linux mint on my VMs with x11 , worked there too....so please check for yourselves and do tell me how it goes

NOTE: if you are a dev then please feel free to contribute

GitHub Repository link: https://github.com/adityakrishnan005-a11y/ScreenGuard

AI Disclosure:

Implementation: Manual, Codebase: Manual, Debugging: Assisted, Documentation: Assisted, Packaging: Generated (especially for the Arch custom repository for convenience in updates)

basically something like smarttube but on linux

Omarchy should not matter. But sadly it does. As a power grab.

If you don’t know what Omarchy is here’s the short summary. David Heinemeier Hansson, creator of RubyOnRails, owner of a software business and millionaire, decided a few months ago to go all in on Linux and – because that’s just his MO – thought that turning his setup into a project made sense. Omarchy is just that: A pretty standard Arch Linux distribution with a handful of configuration files, that Heinemeier Hansson (or DHH as he is often called) wrote, or vibed or whatever. There is nothing wrong with this, custom, opinionated special purpose distributions have been part of the Linux culture for a long time and provide a lot of value and experimentation that sometimes even flows back to the original project. In a way Ubuntu Linux is just that to the basis Debian.

There’s also nothing wrong in more opinionated software – quite the opposite. I do think that RubyOnRails and similar frameworks gained traction especially because they do have a clear idea of how they see the respective problem domain. There’s even the statement “There should be one – and preferably only one – obvious way to do it” in the Zen of Python. Having a strong point of view is not a bad thing in software or in general. The problem is the kind of points of view DHH has.

I don’t want to go into all the gritty details here, Brennan Kenneth Brown did a great job with their article “Normalized Fascism in Open Source: $12 Million Given to DHH“, but here’s the short version. DHH is a racist and a fascist. He recently wrote articles on how there are [Content Warning: massive racism] too many brown people in London these days. In July he wrote an [Content Warning: Racism, anti-romanism] article comparing Roma communities (using a slur for them) to “wolves” that should be “shot”. His X account is also know to be “edgy” in a similar way. A few days ago he posted sort of a manifesto for his set of config files:

There’s even a longer version on the website. Even without a strong background in antifascist literature or history this reads like a fascist creed. About unity and identity. About “holding the line” (against codes of conduct and the rights of marginalized people as the website explains). About how there always needs to be a strong leader who makes the decisions. Given DHH’s other writing this is not a misguided joke, this is how he thinks about the world.

I began this text writing “Omarchy should not matter” but I have not explained why it does. It’s surely not about user numbers (it’s a clunky distro aimed at people who love to cosplay hackers). But it is about power.

DHH has used his connection to CEO’s of tech corporations and his position as influencer to collect up to now above 18 million dollars of funding for Omarchy. He uses that money not for himself (he has enough of that) but to fund certain pieces of Software he uses in his distribution. Like for example the window manager hyprland whose main developer the Omarchy money now funds. What is “Vaxry” known for outside of a niche window manager? He runs a community based around LGBTQ discrimination and participates in that (see the short summary on Drew DeVault’s “Weird Little Guys of FOSS” list). His bigotry is so consistent that he was banned from participating in the Freedesktop.org community where all the other developers of window managers and the free software desktop stack collaborate. And it was long before DHH funded him.

One could see a world where a CEO using his connections to drum up some funding for free software projects would be a good thing. But that’s not the world we live in I am afraid. DHH got 3 million USD from Digital Ocean (a cloud provider) for Omarchy. Basically at the same time where Digital Ocean cancelled their support of Flathub and GNOME that amounted to a value of 50 USD per month.

DHH is using his position and his influence to accumulate donations that he then can distribute to projects that align with his fascist worldview and ambitions. And in a time where many important projects and infrastructures are strapped for cash running just on the goodwill of a handful of people that is increasingly dangerous. Not because its corporate money – a lot of funding for open source comes from corporations, they are after all who are sucking up all the surplus from people’s labor. It’s because this amount of money gives him power.

The Omarchy foundation will be where projects in desperate need of funding will be pointed. And will a project focused on human rights and social justice get funding from our fascist overlord? What kind of pressure can that exert on projects who have to decide to either die due to lack of funding or kill their code of conduct?

An open fascist is building one of the bigger sources of funding for open source fully under his individual control and sucks up a lot of money that otherwise might have gone to liberatory projects. Community projects. Antifascist projects.

And that is why Omarchy matters. Sadly. Not because of its technical merit (even though those fascists love to use the “merit” argument to fight human rights) but because of its leader, his connections and his ability to accumulate financial power. Without anyone in that ecosystem saying anything about how that leader operates and talks. And you might know the saying: If there’s a Nazi and 10 folks sitting together at a table and nobody says and does anything? There’s 11 Nazis at that table.

Omarchy got a lot of recognition recently. Not only by corporate donors but by Youtubers and other tech influencers who kept praising it and nudged people (who might not know about DHH’s exploits) towards testing it. To become “a hardcore linux user/dev” or whatever other chauvinist argument they made. This makes it dangerous.

Running Omarchy means contributing to this dynamic. It means willingly integrating oneself into a fascist community. Legitimizing it. Supporting it. Omarchy is a fascist project that normalizes fascist thinking for everyone entering that community (even unknowingly). The baseline has to be not to use that system. And as people who believe in human rights and dignity, in community and a tech world that can be better we need to oppose Omarchy. Need to make sure it gets no seat at any table. Support projects who actively refuse to collaborate. Support projects who care about human rights and flourishing. And do not do business with the companies who willingly support fascists.

Siamo tutti antifascisti!

(Share this message about Omarchy [or some other similar post] around in the tech circles you frequent. Tell people about the background of this hyped Linux distribution and its fascist leader. Quit your 1password or Digital Ocean accounts and mention that it is because the support fascists.)

Just switched Zorin for Win11 on my kids computer (yes, it did hurt and still does 😭). His friends play Fortnite and he felt left out. Windows is the only option if he want to play?

Btw downloaded newest Microslop Windows 11, but still during installation there was a gust from the past; I had to use my other computer to download the wifi card driver, which I had to "select from disk" when installing 😂 in 2026!! (tplink ex3000). Worked flawless on Linux.

I switched from 10 years of Openbox to KDE Plasma. This also means from Xorg to wayland.

Plasma is great, it has almost everything. One thing I'm missing (until I'm able to write my own plasmoids/widgets) is what conky provides: a constantly updating overview of my top processes, what's eating resources.

If you ever tried to hack your own conky you must know how fragile it can be. Plus, it's even more fragile on wayland. Most people run it in Xwayland, many unwittingly, but that didn't work for me either.
Let's not dwell on these things (more than a decade of faffing around with conky. It usually runs stable once it's set up, but getting there is beyond painful).

Every time I thought I had it running stably without freezing or disappearing, something new happened... aaaarrrgghh

Until I decided fuck it, I'll run it in the terminal. Much more stable. And with Konsole* & Kwin I can make it look really good.

* I tried it with Alacritty first, but it flickers!

The workflow, if you are interested:

::: spoiler

The conky config

conky.config = {  
	background = false,  

	out_to_x = false,  
	out_to_console = true,  
	out_to_ncurses = false, -- the terminal gets messed up with console_graphs  
	out_to_stderr = false,  
	out_to_wayland = false,  

	-- replace dumb ascii characters, e.g. in console bars, with nice unicode characters  
	lua_load = "~/.config/conky/lua/scripts.lua",  
	-- see how these chars are replaced in the lua functions above:  
	console_bar_unfill = 'a',  
	console_bar_fill = 'b',  
	console_graph_ticks = "a,b,c,d,e,f,g",  
	use_spacer = 'left',  

	-------------------  
	update_interval = 3,  
	--------------------  

	total_run_times = 0,  

	cpu_avg_samples = 2,  
	diskio_avg_samples = 2,  
	net_avg_samples = 2,  
	no_buffers = true,  

-- 	max_text_width = 30, -- does nothing in console/ncurses  
	pad_percents = 3,  
	short_units = true,  
	format_human_readable = true,  
	if_up_strictness = address,  
	top_name_width = 30,  
	top_name_verbose=true,  
	override_utf8_locale = true,  
	template0 = '${lua rst}${lua blk_brt}${lua fill ┄ 45}${lua rst}', -- ┈🭸 a horizontal line 45 chars wide  
	template1 = [[${lua rst}${lua tab}${top name \1} ${if_match "${top cpu \1}" == "  0.00"}      $else${top cpu \1}$endif ${if_match "${top mem \1}" == "  0.00"}      $else${top mem \1}$endif]],  
};  

conky.text = [[  
${lua rst}${lua cls}${lua cyn_brt}${lua cpad 45 ${kernel}}${lua tab}${lua cyn_brt}Name                              CPU%   MEM%  
${template0}${lua tab}${template0}  
${lua cyn_brt}RAM${lua rst} $mem/$memmax $memperc%    ${lua cyn_brt}SWAP${lua rst} $swap/$swapmax $swapperc%${template1 1}  
${lua cyn_brt}CPU${lua rst} ${freq_g 0}GHz${cpu cpu0}% ${lua mag_brt}${lua bar_to_uni ${cpubar cpu0 1,29}}${template1 2}  
               ${lua mag_brt}  1${lua rst} ${freq_g 1}GHz${cpu cpu1}%  ${lua mag_brt}2${lua rst} ${freq_g 2}GHz${cpu cpu2}%${template1 3}  
               ${lua mag_brt}  3${lua rst} ${freq_g 3}GHz${cpu cpu3}%  ${lua mag_brt}4${lua rst} ${freq_g 4}GHz${cpu cpu4}%${template1 4}  
${lua mag_brt}${lua graph_to_uni ${cpugraph cpu0 1,45 -l}}${template1 5}  
${lua cyn_brt}Load  ${lua rst}1min: ${lua red_brt}${lua rpad 6 ${loadavg 1}}${lua rst} 5min: ${lua red_brt}${lua rpad 6 ${loadavg 2}}${lua rst}  15min:${lua red_brt}${lua lpad 6 ${loadavg 3}}${template1 6}  
${template0}${template1 7}  
${lua cyn_brt}FS root: ${lua rst}${fs_used /}/${fs_size /} ${lua cyn_brt}${lua bar_to_uni ${fs_bar 1,24 /}}${template1 8}  
${lua cyn_brt}FS data: ${lua rst}${fs_used /home/ssd256_data}/${fs_size /home/ssd256_data} ${lua cyn_brt}${lua bar_to_uni ${fs_bar 1,24 /home/ssd256_data}}${template1 8}  
${template0}${lua tab}${template0}  
${if_up enp0s20u2}  
${lua cyn_brt}USB ethernet${lua rst}      up:${lua grn_brt}${upspeedf enp0s20u2}K${lua rst}   dn:${lua grn_brt}${downspeedf enp0s20u2}K${lua tab}$else${if_up wlan0}  
${lua cyn_brt}WLAN${lua rst}              up:${lua grn_brt}${upspeedf wlan0}K${lua rst}   dn:${lua grn_brt}${downspeedf wlan0}K$else  

${lua fill S 45}$endif${lua tab}$endif${lua cyn_brt}${if_match "$mpd_status" == "Playing"}${scroll wait 45 1 ${uppercase ${mpd_smart}}}  
]]  

The accompanying scripts.lua

#!/usr/bin/lua  

-- we make our own ncurses :-D  
function conky_rst() -- reset  
	return ''  
end  
function conky_cls() -- clear screen  
	return ''  
end  
function conky_nl() -- newline  
	return string.char(10)  
end  
function conky_blk_nrm()  
	return ''  
end  
function conky_blk_brt()  
	return ''  
end  
function conky_blk_dim()  
	return ''  
end  
function conky_blk_uln()  
	return ''  
end  
function conky_blk_bln()  
	return ''  
end  
function conky_blk_rev()  
	return ''  
end  
function conky_blk_del()  
	return ''  
end  
function conky_red_nrm()  
	return ''  
end  
function conky_red_brt()  
	return ''  
end  
function conky_red_dim()  
	return ''  
end  
function conky_red_uln()  
	return ''  
end  
function conky_red_bln()  
	return ''  
end  
function conky_red_rev()  
	return ''  
end  
function conky_red_del()  
	return ''  
end  
function conky_grn_nrm()  
	return ''  
end  
function conky_grn_brt()  
	return ''  
end  
function conky_grn_dim()  
	return ''  
end  
function conky_grn_uln()  
	return ''  
end  
function conky_grn_bln()  
	return ''  
end  
function conky_grn_rev()  
	return ''  
end  
function conky_grn_del()  
	return ''  
end  
function conky_ylw_nrm()  
	return ''  
end  
function conky_ylw_brt()  
	return ''  
end  
function conky_ylw_dim()  
	return ''  
end  
function conky_ylw_uln()  
	return ''  
end  
function conky_ylw_bln()  
	return ''  
end  
function conky_ylw_rev()  
	return ''  
end  
function conky_ylw_del()  
	return ''  
end  
function conky_blu_nrm()  
	return ''  
end  
function conky_blu_brt()  
	return ''  
end  
function conky_blu_dim()  
	return ''  
end  
function conky_blu_uln()  
	return ''  
end  
function conky_blu_bln()  
	return ''  
end  
function conky_blu_rev()  
	return ''  
end  
function conky_blu_del()  
	return ''  
end  
function conky_mag_nrm()  
	return ''  
end  
function conky_mag_brt()  
	return ''  
end  
function conky_mag_dim()  
	return ''  
end  
function conky_mag_uln()  
	return ''  
end  
function conky_mag_bln()  
	return ''  
end  
function conky_mag_rev()  
	return ''  
end  
function conky_mag_del()  
	return ''  
end  
function conky_cyn_nrm()  
	return ''  
end  
function conky_cyn_brt()  
	return ''  
end  
function conky_cyn_dim()  
	return ''  
end  
function conky_cyn_uln()  
	return ''  
end  
function conky_cyn_bln()  
	return ''  
end  
function conky_cyn_rev()  
	return ''  
end  
function conky_cyn_del()  
	return ''  
end  
function conky_whi_nrm()  
	return ''  
end  
function conky_whi_brt()  
	return ''  
end  
function conky_whi_dim()  
	return ''  
end  
function conky_whi_uln()  
	return ''  
end  
function conky_whi_bln()  
	return ''  
end  
function conky_whi_rev()  
	return ''  
end  
function conky_whi_del()  
	return ''  
end  
function conky_def_nrm()  
	return ''  
end  
function conky_def_brt()  
	return ''  
end  
function conky_def_dim()  
	return ''  
end  
function conky_def_uln()  
	return ''  
end  
function conky_def_bln()  
	return ''  
end  
function conky_def_rev()  
	return ''  
end  
function conky_def_del()  
	return ''  
end  
function conky_tab()  
    -- ┊🮇⋮┊⦙⸽|︳⏐｜┆  
    return '     ┆     ' -- first black_bright, then reset  
end  
function conky_fill(char,num) -- pass S for spaces (no idea how to pass spaces in conky)  
    return string.rep((char == 'S' and ' ' or char),num)  
end  
function conky_bar_to_uni(str)  
    local bar_fill_map = {  
        a = '⠠',  
--         a = '🞍',  
--         a = '🞌',  
--         a = '⬝',  
--         a = '🭺',  
        b = '⠶',  
--         b = '⯀',  
--         b = '■',  
--         b = '█',  
--         b = '▇',  
--         b = '⠿',  
--         b = '🯦',  
    }  
    return string.gsub(conky_parse(str), "%w", bar_fill_map)  
end  

function conky_lpad(num,str)  
    local pad = '%'..num..'s'  
    return string.format(pad,conky_parse(str))  
end  

function conky_rpad(num,str)  
    local pad = '%-'..num..'s'  
    return string.format(pad,conky_parse(str))  
end  

function conky_cpad(num,str)  
    out = conky_parse(str)  
    len = #out  
    pad = num - len  
    left = pad//2 + pad%2 + len  
    lpad = '%'..left..'s'  
    out = string.format(lpad,out)  
    rpad= '%-'..num..'s'  
    return string.format(rpad,out)  
end  

function conky_graph_to_uni(str)  
    local graph_ticks_map = {  
        a = '⣀',  
        b = '⣄',  
        c = '⣤',  
        d = '⣦',  
        e = '⣶',  
        f = '⣷',  
        g = '⣿',  
--         a = ' ',  
-- --         a = '_',  
--         b = '▁',  
--         c = '▂',  
--         d = '▃',  
--         e = '▄',  
--         f = '▅',  
--         g = '▆',  
--         h = '▇',  
--         i = '█',  
    }  
    return string.gsub(conky_parse(str), "%w", graph_ticks_map)  
end  

Yes, it could do with some improvement.

All this is started through a desktop file in ~/.config/autostart:

[Desktop Entry]  
Name=Conky Konsole  
TryExec=konsole  
Exec=konsole --profile conky-tui --qwindowtitle conky_tui_konsole --separate -e /bin/rbash -c 'tput civis; conky -c ~/.config/conky/conky-tui2.conf'  
Icon=conky-logomark-violet  
NoDisplay=true  
StartupNotify=false  
Type=Application  
SingleMainWindow=true  

I also created a konsole profile that must have the exact dimensions for this conky, also the color scheme likely needs to be edited so that Background color is exactly the same as Color 1, because of transparency. You'll understand when you see it.

So far this gives us something like this (previous version):

Now I right-click on the titlebar -> more actions -> special window settings. The window must be uniquely recognizable through its title 'conky_tui_konsole'. I remove borders, make it appear on all desktops, stay below, and put it in the position where I want it. It takes a while to understand the UI. In the end I get this extra entry in ~/.config/kwinrc (I love KDE! The combo of having a gui menu for everything, yet still simple editable config files is just 😙👌):

[uuid-uuid-uuid-uuid]  
Description=Window settings for conky_tui_konsole  
acceptfocusrule=2  
below=true  
belowrule=3  
desktops=\\0  
desktopsrule=3  
fsplevel=4  
fsplevelrule=2  
noborder=true  
noborderrule=3  
position=740,610  
positionrule=3  
skippager=true  
skippagerrule=3  
skipswitcher=true  
skipswitcherrule=3  
skiptaskbar=true  
skiptaskbarrule=3  
title=conky_tui_konsole — Konsole  
titlematch=1  
types=1  
wmclass=konsole org.kde.konsole  
wmclasscomplete=true  
wmclassmatch=1  

Voilá. I'm a happy hacker now.

:::

I edited this post to show the newest version, both screenshot and code.

I want to sound like Louis Griffin

I recently bought

• a PCIe expansion card for FireWire (1394a and 1394b) capture
• a PCIe expansion card for component video, composite video, VGA and DVI-I capture
• an nvidia GT 220 GPU to make room in my chassi for the other two cards, because my ASUS TUF Gaming GeForce RTX 3080 OC Edition 12GB GDDR6X's big butt was in the way.

The goal is to help mom and dad to digitize old VHS, VHS-C, Mini DV and audio cassette tapes. The above expansion cards will arrive sometime next week and meanwhile, I have been doing some research on ffmpeg and v4l2-ctl. It's absolutely amazing how much you can do - or could do - with ffmpeg if you had about two-three lifetimes to learn. 🤣

Anyway, using a cheap HDMI to USB video capture dongle, I have successfully captured some video with v4l2-ctl -d /dev/video0 --stream-mmap --stream-to=video.raw --stream-count=100 and then converted it using ffmpeg -i video.raw video.mkv.

I am really interested in what this .raw file is, but all I could find was this, which seems to pertain to raw image files, rather than what v4l2-ctl outputs.

I tried to do some Ctrl+f searches on Video for Linux API, but this was far beyond what I was able to comprehend. On the other hand, the manual is far too "shallow" and doesn't explain much. Do you have any recommendations on what to read in order to understand a little bit more about the inner workings without having to take a PhD in how Linux does video? Why are these .raw files so big? What is the difference between them and the standard containers (.mkv, .mpeg, .avi etc.)?

Also, a huge bonus if you can find the time and will to point out any rookie mistakes in the digitizing project itself.

I'm trying to get into QubeOS. Coming from Graphene on phone and its compartmentalizations, I expect an easy transtition. But holy shit, theres like a qube for everything: usb qube, network qube, file manager qubem..etc.

I'm still learning but find it difficult, not because of its features (I like to tinker with things) but because my hardware cannot handle the OS. I have an intel i7 12th Gen with 40GB RAM laptop. It runs blazingly fast on other distros but really suffocating in QubeOS. By "suffocating", i meant just playing a video in Firefox can cause problems. Moving the Firefox window around can freeze the system.

Those who managed to switch to it, what was your hardware (laptop, desktop?) and how long did it take you?

I really like QubeOS, its controls are amazing. But the hardware just couldnt handle it.

cross-posted from: https://lemmy.ml/post/52523151

Exciting news after the Framework 12 just got a fingerprint reader.

Ikey Doherty, the creator of Solus and AerynOS, is back with Barney, a Rust tool that builds Linux distros from source. So that you can easily create your own Omarchy.

In a video posted on YouTube, the creator of Solus and AerynOS called the money behind DHH's distro absurd and promised to do it better with nothing. His answer is Barney, a Rust-based tool for building Linux distros from source.

Omarchy has changed its Kitty configuration after a security issue was identified with allow_remote_control yes.

That setting allows Kitty’s remote-control protocol to accept commands from terminal output, meaning untrusted output could potentially execute commands as the current user. Kitty itself defaults this setting to no; the exposure came from Omarchy explicitly enabling it so other parts of the desktop could query the active terminal’s working directory.

The change disables remote control by default and migrates existing explicit yes settings to no. Users affected by the migration need to fully close and reopen Kitty, as existing processes remain exposed until they exit.

There is a small trade-off: Omarchy can no longer reliably determine the active Kitty tab/pane’s working directory via its socket, so new terminal or Files launches may occasionally fall back to the home directory.

The associated migration tests reportedly pass, although no live Kitty test was run in the test environment.

I understand LUKs can be used to encrypt your data. But what would prevent somebody plug in a USB with and just wipe my drive?

On traditional BIOS like Lenovo, HP, Dell and even Framework you can set a supervisor password that locks the boot menu. So nobody can boot from the USB.

Coreboot is different though. I spoke with Starlabs whose computers run Coreboot, and apparently you can have the boot menu password. OTOH, Sys76's Coreboot doesnt allow such things.

I ask because i want to libreboot my T480, but the number 1 thing i worry is unauthorized USB boot.

This one: https://libreboot.org/docs/linux/grub_hardening.html#grub-password

seems to only lock the ability to edit the grub entry freely, aka press "e" to change stuff when grub fails to boot.