wp-fractional-scale-v1 to allow clients to submit buffers with a non-integer scale factor matching the output.
This hopefully means Sway and similar will support real fractional scaling for applications, not just the compositor fractional scaling we already have.
But I don’t know much about application support. Qt and Electron might support it; GTK 4 does not, possibly in a future version).
tearing-control to allow clients to opt-in for tearing page-flips.
That’s great for those who need it. Anyone with a modern display should probably just use variable refresh rate (vrr), but even today some devices don’t support it. E.g. there’s been 240Hz laptops without vrr.
I was sure I read that GTK wants to support true fractional scaling in GTK 5, but I can’t find a source to it. So it was probably just speculation. As far as I understand it, it would require big changes to GTK because everything is build with integer scaling in mind.
At least GTK 4 already has support for this fractional scaling protocol.
At least it does not look blurry with fractional scaling enabled, which is the biggest issue IMO. The current hacky way is not ideal I agree but at least it is functional
Mine was Mandrake 6. RedHat 5.2 was my first, and I was surprised how much easier Mandrake was in comparison. But the one that really wowed me was SuSE (before they became OpenSUSE), I was blown away how polished and user-friendly it was. Windows 9x/ME felt like a joke in comparison at time. And some people still claim Linux isn’t user friendly… and I’m like, bruh it’s been user friendly for about three decades now…
As pointed out, they don’t use it. However, there are loose plan for KWin to migrate to wlroots one day, and in fact a hostile fork exists that is exactly that (KWinFT). So a compositor can make use of wlroots to implement Wayland functionality, sway for example does exactly that, unsurprisingly since they’re sister projects by the same author.
It should be noted that libwayland (mentioned in the patch notes) also exist, and wlroot actually depends on it, so I guess libwayland is like the lower level stuff while wlroots saves you some work to integrate libwayland into your compositor; the motto is “Pluggable, composable, unopinionated modules for building a Wayland compositor; or about 60,000 lines of code you were going to write anyway.”
wlroots is a library that can be used to implement a compositor like KWin or mutter (GNOME). In practice wlroots is used in Sway, Hyprland, river, and more.
What wlroots-based compositors, KWin, and mutter share is that they implement a similar set of the display protocol Wayland. E.g. KWin and Sway implement the Wayland extension wlr_layer_shell
I mean, if it was accidental then… Just turn it off and boot back into Linux? You realise you can just turn it off while it’s downloading updates, right? Heck, you can even pause updates long term if you want! 😱 Crazy!
I never said to do it while it’s in the middle of changing stuff, but if you just booted you can turn it off and nothing will happen because worst case it will just be downloading without installing, as someone else mentioned, once updates are installed you can even turn off without applying updates if you want and you can also tell it to only download and not install unless you tell it to or not download at all.
If you're going that way, Windows is not going to suddenly start updating when you simply boot it. You have to willingly click "Update and Shutdown/Restart" instead of "Shutdown/Restart", assuming your computer even finished downloading the update.
I have a Windows 10 partition on a second machine. I have disabled automatic updates in the options and I never click “Update at restart” or anything. Yet, whenever I need to boot into Windows it decides to automatically start updating itself.
I guess that I use it infrequently so there are always updates available, but it shouldn’t force them on me when I’ve specifically disabled them.
Also, when you choose either of the update or restart/shutdown options, it actually tries to restart, (for me) always boots back into linux because that’s my default. When I’d eventually boot back into Windows, it just continues installing the update I’d long forgotten about.
I’m currently using ROC on my laptop and desktop. Latency is low enough to not be noticeable when playing video on my laptop and streaming audio to the desktop. Audio can get a bit choppy if my laptop is on WiFi. But that is most probably because the signal between the repeater on the second floor and my DSL modem on the ground floor is pretty meh.
Latency is low enough to not be noticeable when playing video on my laptop and streaming audio to the desktop.
That’s basically my use case. Want to use my HTPC as the source and some RasPi’s or repurposed thin clients as the sinks - pretty much what I do now with MPD and Snapcast. I absolutely do not want to have to mess with audio offset settings in Emby to keep the dialog in sync. lol
I’ve only skimmed the docs (holidays are a huge time sink haha), but do you know if it can do one-to-many or just one-to-one? Like, can I have one source and multiple receivers? The docs seemed to imply it could do one-to-many, but I didn’t get to dive into them deep enough.
I don’t know if ROC can do multicast on its own. I use the Pipewire source and sink. And I only do the one-to-one setup.
I did some tests in Pipewire:
Configuring multiple sinks is possible on a machine. They simply present as additional output devices. So if you want to switch audio to another source, that should be doable by switching to another output device.
Doing one-to-many: I don’t know if that is possible with ROC alone. You might be able to do something with Pipewire graphs
i’ve been using sonobus lately and it’s been pretty good, I had latency issues when I tested the android app a long time ago, ill have to test it again
Linux 6.1 will be maintained for another 10 years by the CIP. The hardware in question will be almost 40 years old at that point. I don’t have a violin small enough for users losing free support after 40 years from maintainers who most likely don’t even own the same hardware to test on…
On the other hand, they were probably unchanged for decades. Did anything really change, or is this just a case of we need to remove 500k lines of code, what is most useless ? Let’s cut that.
In other words, removed because it’s a KPI to remove lines, and this makes number go up.
So if it’s been unchanged for decades then you can just add it yourself and recompile the kernel. Elsewhere you argue that you can’t just add old drivers to a newer kernel, which implies these drivers require some nontrivial amount of maintaince. Which is it.
Keeping code around isn’t free. Interfaces change, regressions pop up. You have to occasionally put in work just to keep it in a working state. Usually in cases like this there are discussions on the mailing list about who is going to maintain them and nobody volunteers. You can do that if you’re so passionate about keeping these drivers around.
They were fine all this time, what changed suddenly ? I bet it’s the security nerds stirring shit, making it all a liability and easier deleted than fixed.
I doubt any hardware 25+ years can even run a modern vanilla linux kernel, you’d have to compile it yourself with some serious customization for it even work
Seems like you’re annoyed that I pointed out that what you were saying was irrelevant? And so you reply with more irrelevant crap (on a very nerdy, not-fun-at-parties internet forum for Linux discussion)? Let me know if I got that wrong.
If you’re doing it for the memes then you don’t really need to worry about malware. Your machine is probably too old for anything that’s still floating out there to even work on it.
Many people browse 4-5 pages a day, see a few emails, print a few pdfs, and a core2duo, or x4, for 40#/$/Eu a box run flawlessly with linux and xfce/lxde for example.
Even video-conferencing works fine.
Driver code is still there, you can add it back if you want, same with ide drivers and such, support was removed but code still exists, just add it and compile your own kernel, there are alot of tutorials in internet about it
The drivers were removed in 6.3. Debian 12 is still running on 6.1. Debian 12 just came out and still has many years of support ahead of it (at least 5). You can get plenty of use out of these cards before they stop working.
Someone needs to maintain them for them to keep working. Nobody else is willing to do that anymore, but you can still volunteer as a maintainer. If you don’t, it’s as much your fault as anyone elses.
You can delay all other updates with the group policy editor. You can disable preview builds and you and delay quality updates by 30 days and delay feature updates by 365 days. The bugs are always worked out by then.
Feature updates are necessary after a while. There’s SOME important stuff in there. And if you wait a whole year before installing the new one, all the bugs will be fixed by then
linux
Oldest
This magazine is from a federated server and may be incomplete. Browse more on the original instance.