Take a journey through key milestones in AI’s evolution and how #opensource shaped its future. Watch Harish Pillay break it down in this insightful session.
Auf YouTube findest du die angesagtesten Videos und Tracks. Außerdem kannst du eigene Inhalte hochladen und mit Freunden oder gleich der ganzen Welt teilen.
If you still need to run an X11 session instead of Wayland (e.g. for accessibility use cases not covered yet, or old school software/WMs) then forget about cursed Xorg forks, *this* is the future-proof way! social.treehouse.systems/users…
nice to see, that someone actually did what I had on my mind for ages. Only I had called it Xenomorph (get it? Alien themed, Wayland Yutani reference).
Sorry but I really don't understand the logic behind recommending something experimental with "a gazillion bugs" over something that already exists and is really stable.
@SuperDicq This solution is built upon mature Wayland components for compositing and display backend that are mature, modern, future proof, and already work better than the Xorg equivalents today.
The bugs in the integration bits, which are comparatively simple are a lot easier to fix than legacy Xorg DDX technical debt.
@SuperDicq Not on all hardware. On Apple hardware (which doesn't implement legacy VBlank IRQs or legacy cursors) Xorg has tearing, flickering cursors, and animations run too fast because there's no VSync.
Xorg only runs well on old drivers that implement all the bloat and legacy workarounds that it needs...
@SuperDicq All modern computers do. At least on Apple hardware that firmware does not have full system access that could potentially snoop on or backdoor your OS. On every current Intel and AMD system it does.
>At least on Apple hardware that firmware does not have full system access that could potentially snoop on or backdoor your OS. I highly doubt that. How do you know for certain?
Also define "modern computers". My GNU Booted computers are still modern, as they can do everything that a computer sold today can do (except enforce hardcuffs to the user obviously, but I do not consider that a desirable feature).
We know because we reverse engineered it. I personally fully reverse engineered the GPU microPPL, the only firmware component with privileged system access on a design technicality (which is only a couple kilobytes of code), and reported a vulnerability that got fixed. There are no backdoors or holes left as far as I can tell. Every other firmware blob is firewalled by an IOMMU.
The bug I found earned me a $150k bounty, so there is lots of incentive for people to find those kinds of bugs that can break security.
BTW, even the decade old ThinkPads that people flash with open firmware and say are fully free have chips running proprietary firmware on the LPC bus with full DMA access. Apple machines do much better than even those.
If you want to reduce it to "there could still be a secret silicon backdoor", that is true of every modern system by definition.
You have a list of firmware blobs and a rough idea of what they do. As far as I can tell you do not have the original source code or rewritten replacement software.
If this becomes available in any point in the future I might switch to an Apple laptop, but for now I will stay on my GNU Booted hardware instead.
@SuperDicq It doesn't matter what the firmware blobs do, the point is what they can't do, which is take over your system or steal your data because they don't have the hardware access to do that. It's no different than the firmware blob running your mouse and keyboard.
It doesn't matter what the firmware blobs do, the point is what they can't do, which is take over your system or steal your data because they don't have the hardware access to do that.
When you run proprietary blobs I agree it is obviously better to run them in such a way that they can not access anything. But that still is not good, because the blobs are still there denying you freedom over your own devices. They should be replaceable with free software.
It's no different than the firmware blob running your mouse and keyboard.
That's not true. I have built my own keyboard and it my own customized fork of QMK Firmware.
But I assume you mean in the case of a regular keyboards (excluding the ones that run proprietary flashable firmware like a lot of "gamer" branded products) this firmware is backed into the device on a read-only chip. In this case is not denying its users any freedom as not being able to customize it is in this case just a hardware limitation, not something specifically put in place by manufacturer to handcuff the user.
You're repeating the arguments from the FSF word for word, so I guess you've fully drunk their kool-aid. I strongly disagree with their posture and I think it is very harmful to the free software ecosystem as a whole, but it's almost impossible to convince anyone in your position in my experience, so we'll just have to agree to disagree.
P.S. a huge majority of USB products have flashable firmware, they just don't advertise it or provide the flashing tools outside of manufacturing. You're still being "denied your freedom", you just don't know it.
@foxido @SuperDicq Wayland is exactly as old as the Xorg modesetting driver (both were introduced in 2008) and Wayland compositors (many of which are much younger) do a much better job of driving display controllers and compositing than the Xorg solution.
You can say whatever you want about the ecosystem and desktop features, but Wayland was designed to work well with modern hardware and rendering pipelines and it has always done this better than Xorg. And that's what this solution does, run Wayland's rendering backend with the Xorg feature set on top.
@ariadne If Xwayland already works for the application one wants to use, then they're not really stuck with X right?
I thought the whole issue was that some applications are not (maybe never?) supported by Xwayland, so no solution that's based on Xwayland would work for those applications?
No, that's for rootless XWayland running under a full Wayland compositor, which is intended to run X11 apps but not full X11 sessions (because it's running under a Wayland session), that's why it can't do everything a full X11 session can.
This is rootful XWayland which can run any X11 window manager and supports everything Xorg supports that isn't directly hardware device related. It's a full native X11 session. XWayland is the same codebase as Xorg minus the hardware drivers.
looking at the code I can't see anything special that enables you to use an x window manager, were those always just supported by xwayland and nobody bothered to just like... do this?
No offense to you, this is cool, but it's also so simple that I'm stunned no one did it before
@lina @Mae Well, and Xwayland needed some QoL improvements to actually properly host an X11 desktop environment, and those were done at the end of 2023. So someone finally getting around to it less than two years later isn't that shabby. 😃
Reduce code duplication and bypass all the bugs and legacy warts of the largely unmaintained DDX drivers (e.g. why Xorg is quite broken on Apple display controllers and this wouldn't be).
Auf YouTube findest du die angesagtesten Videos und Tracks. Außerdem kannst du eigene Inhalte hochladen und mit Freunden oder gleich der ganzen Welt teilen.
Auf YouTube findest du die angesagtesten Videos und Tracks. Außerdem kannst du eigene Inhalte hochladen und mit Freunden oder gleich der ganzen Welt teilen.
@titipat โอ้พอดีตอนนี้ในหัวมี side project เยอะมากอยากทำหมดเลยทีนี้จะ deploy ขึ้นพวก managed service อย่าง fly.io ทั้งหมดก็ไม่ไหวน่าจะหมดตัวเลยว่าจะหาทางง่ายๆดีพลอยทุกตัวใน server เดียวเห็นช่วงก่อน dhh ขาย kamal ไว้เลยว่าจะลองดู 55555
ระบบเลือกตั้งออนไลน์ประเทศเดียวในโลกเค้าทำงานกันยังไง แลกมาด้วยอะไรกันบ้าง fun fact—คนมากกว่าครึ่งเลือกที่จะลงคะแนนเสียงออนไลน์ในการเลือกตั้งระดับประเทศครั้งล่าสุดในปี 2023 Read “เลือกตั้งออนไลน์? เป็นไปได้ที่เอสโตเนีย“ by iNViTiON on Medium: blog.invition.dev/online-elect...
Online voting is something many people desire — maybe expect that everything to be faster, or might be because it will eliminate errors, or perhaps simple because they’re too lazy to travel to a…
I have to tip my hat to Microsoft for having worked so hard to convince the world that the City of Munich failed with their Linux migration.
It's of course just a big fat lie. Munich is since 20+ years growing their Open Source footprint. They are sharing and they are growing their use of open source.
Announced back in February 2024, we're still waiting on a release date for the Orange Pi Neo gaming handheld with Manjaro Linux but now we have prices.
Need a new first-person dungeon-crawling RPG? Perhaps it's time to check out Monomyth. The devs just updated it with better Steam Deck and gamepad support.
Hi #fediverse ! We’re the European Southern Observatory, and we design, build and operate ground-based telescopes.
One of them is our Extremely Large Telescope, currently under construction in #Chile. It will have a 39 m mirror, and its rotating enclosure will weigh 6100 tonnes, or about 700 mastodons!
We’re looking forward to chatting with all of you about #astronomy
And many thanks to @sebinthestars for running our former instance!
Asahi Lina (朝日リナ) // nullptr::live
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •datenwolf
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •spikederailed
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •SuperDicq
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •Asahi Lina (朝日リナ) // nullptr::live
in reply to SuperDicq • • •@SuperDicq This solution is built upon mature Wayland components for compositing and display backend that are mature, modern, future proof, and already work better than the Xorg equivalents today.
The bugs in the integration bits, which are comparatively simple are a lot easier to fix than legacy Xorg DDX technical debt.
SuperDicq
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •Asahi Lina (朝日リナ) // nullptr::live
in reply to SuperDicq • • •@SuperDicq Not on all hardware. On Apple hardware (which doesn't implement legacy VBlank IRQs or legacy cursors) Xorg has tearing, flickering cursors, and animations run too fast because there's no VSync.
Xorg only runs well on old drivers that implement all the bloat and legacy workarounds that it needs...
SuperDicq
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •Asahi Lina (朝日リナ) // nullptr::live
in reply to SuperDicq • • •SuperDicq
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •>At least on Apple hardware that firmware does not have full system access that could potentially snoop on or backdoor your OS.
I highly doubt that. How do you know for certain?
Also define "modern computers". My GNU Booted computers are still modern, as they can do everything that a computer sold today can do (except enforce hardcuffs to the user obviously, but I do not consider that a desirable feature).
Asahi Lina (朝日リナ) // nullptr::live
in reply to SuperDicq • • •asahilinux.org/docs/platform/i…
We know because we reverse engineered it. I personally fully reverse engineered the GPU microPPL, the only firmware component with privileged system access on a design technicality (which is only a couple kilobytes of code), and reported a vulnerability that got fixed. There are no backdoors or holes left as far as I can tell. Every other firmware blob is firewalled by an IOMMU.
The bug I found earned me a $150k bounty, so there is lots of incentive for people to find those kinds of bugs that can break security.
BTW, even the decade old ThinkPads that people flash with open firmware and say are fully free have chips running proprietary firmware on the LPC bus with full DMA access. Apple machines do much better than even those.
If you want to reduce it to "there could still be a secret silicon backdoor", that is true of every modern system by definition.
Introduction to Apple Silicon - Asahi Linux Documentation
asahilinux.orgSuperDicq
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •You have a list of firmware blobs and a rough idea of what they do. As far as I can tell you do not have the original source code or rewritten replacement software.
If this becomes available in any point in the future I might switch to an Apple laptop, but for now I will stay on my GNU Booted hardware instead.
Asahi Lina (朝日リナ) // nullptr::live
in reply to SuperDicq • • •SuperDicq
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •When you run proprietary blobs I agree it is obviously better to run them in such a way that they can not access anything. But that still is not good, because the blobs are still there denying you freedom over your own devices. They should be replaceable with free software.
That's not true. I have built my own keyboard and it my own customized fork of QMK Firmware.
But I assume you mean in the case of a regular keyboards (excluding the ones that run proprietary flashable firmware like a lot of "gamer" branded products) this firmware is backed into the device on a read-only chip. In this case is not denying its users any freedom as not being able to customize it is in this case just a hardware limitation, not something specifically put in place by manufacturer to handcuff the user.
Asahi Lina (朝日リナ) // nullptr::live
in reply to SuperDicq • • •You're repeating the arguments from the FSF word for word, so I guess you've fully drunk their kool-aid. I strongly disagree with their posture and I think it is very harmful to the free software ecosystem as a whole, but it's almost impossible to convince anyone in your position in my experience, so we'll just have to agree to disagree.
P.S. a huge majority of USB products have flashable firmware, they just don't advertise it or provide the flashing tools outside of manufacturing. You're still being "denied your freedom", you just don't know it.
SuperDicq
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •Asahi Lina (朝日リナ) // nullptr::live
in reply to SuperDicq • • •foxido [cutie status: expired]
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •@lina@vt.social @SuperDicq@minidisc.tokyo
Asahi Lina (朝日リナ) // nullptr::live
in reply to foxido [cutie status: expired] • • •@foxido @SuperDicq Wayland is exactly as old as the Xorg modesetting driver (both were introduced in 2008) and Wayland compositors (many of which are much younger) do a much better job of driving display controllers and compositing than the Xorg solution.
You can say whatever you want about the ecosystem and desktop features, but Wayland was designed to work well with modern hardware and rendering pipelines and it has always done this better than Xorg. And that's what this solution does, run Wayland's rendering backend with the Xorg feature set on top.
shironeko
in reply to Asahi Lina (朝日リナ) // nullptr::live • • •@ariadne If Xwayland already works for the application one wants to use, then they're not really stuck with X right?
I thought the whole issue was that some applications are not (maybe never?) supported by Xwayland, so no solution that's based on Xwayland would work for those applications?
Asahi Lina (朝日リナ) // nullptr::live
in reply to shironeko • • •No, that's for rootless XWayland running under a full Wayland compositor, which is intended to run X11 apps but not full X11 sessions (because it's running under a Wayland session), that's why it can't do everything a full X11 session can.
This is rootful XWayland which can run any X11 window manager and supports everything Xorg supports that isn't directly hardware device related. It's a full native X11 session. XWayland is the same codebase as Xorg minus the hardware drivers.