Eudev has not been maintained for a long time, and guix system still depends on it.
Soon, eudev will probably stop working, and nobody except me in guix community seems to care. If eudev stops working, nothing will work.
Udev can be installed without systemd. It shouldn’t be too difficult for guix maintainers to replace eudev with udev. Are people only going to pay attention after eudev breaks everything else?
why is that? i’m not saying that replacing eudev isn’t necessary (on contrary, it would probably be a good thing), but eudev seems to be a feature-complete project that only has seen maintenance releases once every few years for a while now, so it’s most likely not something that randomly breaks overnight. Would you say systemds udev is the most viable replacement, or are there other options to consider as well?
udev is the most hassle-free substitute for eudev.
Gardenhouse/libudev-garden - Codeberg.org and Gardenhouse/gardendevd - Codeberg.org can substitute for udev in theory, but they are new and unproven. I don’t know whether they will work with gamepads, KDE, GNOME, steam, and so on.
Guix can replace eudev with udev, and contributors can test gardendevd over time.
It will break. When it breaks, people will panic. eudev will not live forever.
udev is constantly changing under the hood. It’s not going to stay compatible with eudev for long.
I see a lot of complacency in software development.
Why will it break and how?
udev “will” introduce breaking changes at some point. It’s just a matter of time.
If eudev doesn’t keep up with udev, it breaks. Eudev is unmaintained.
I thought I didn’t have to explain the nature of change. Change is the only constant in the universe. Every software will introduce breaking changes or die at some point. There is no exception. Even linux kernel will introduce breaking changes at some point although linus torvalds tries very hard to avoid breaking changes. So far, nothing in the physical universe has remained the same over time.
What makes you think udev won’t introduce breaking changes or won’t be replaced with something else?
“Everything” in the physical universe will crumble over time if they are left unmaintained.
I’m not asking people to develop a new software application because eudev is crumbling under the hood.
Replacing eudev with udev should be straightforward for maintainers.
Replacing eudev with udev now consumes a bit of time, but it will cost far less time than discussing any alternative to eudev and pretending udev doesn’t exist when eudev breaks.
The question is why people avoid udev when there is no real proven alternative? It seems they don’t have a strong reason to avoid udev, yet they seem to avoid it for some vague hatred against systemd. We can already use udev without systemd.
A computer that works at all is better than one that doesn’t. Why don’t we just use udev now? People can always develop udev alternatives later.
So
More like philosophical differences and dislike of the project lead.
I find it unlikely that a breaking change will come from the kernel in the near future.
What kind of change in udev do you see breaking eudev?
Why do you think that eudev won’t get an update or maintenance will pick up again if needed?
I hate the ideology and the people behind systemd, udev, and so on, but avoiding udev is a wrong priority.
Content creators who shape human minds with contents are practical with software and often use centralized OSes like windows and mac os. They want to champion open source software, but they see that open source developers don’t care about things like usability and practicality and UX.
Software is merely a tool for getting things done.
linux-libre, eudev, and so on don’t really help content creators influence the grand scheme of things. We can already install udev without systemd. That’s more than enough for now.
Linux-libre, eudev, and so on only prevent content creators from ever considering guix or parabola in the first place. These backward software choices only make things increasingly more impossible for everyone.
Open-source developers should help content creators shape human mind rather than bicker with systemd, linux firmware blobs, and whatnot and make things difficult for everyone.
Content creators who use mac os and iPhone have far more impact than open-source developers who don’t influence human mind with contents. The right priority for software developers should be to help independent content creators for humanity. Linux-libre, eudev, libreboot, and so on don’t really help independent content creators at all. Most independent content creators choose mac os because mac os has lots of high-quality content creation software, and mac os is easy to use for users. Don’t make things unnecessarily difficult.
Even linus torvalds would avoid guix because he would say “Where is my ethernet driver? I have to install linux from nonguix channel? I have to find a way to replace eudev with udev? This is too much. I have work to do. I will go back to fedora.”
Where do you think systemd, udev, and so on come from? They are merely symptoms that come from the perverted centralization ideology pervasive on earth. The ideology led to red hat. You have to stop the ideology in order to stop things like systemd. The root cause of things like systemd is the centralization ideology. To address the root cause, you have to produce contents to influence human mind. You can’t defeat the centralization ideology by trying to avoid things like systemd and udev at all costs. You defeat it through contents. Influencing human minds with contents can take a very long time, but trying to defeat the centralization ideology with linux-libre and eudev will take forever. A very long time is a lot shorter than forever.
There hasn’t been an update for 2 years.
I only make decisions based on the present conditions. I don’t base my decisions on uncertain future scenarios that are just fantasies in human mind. The past is gone. The future is a fantasy. I don’t take future predictions seriously. People who confidently predict the future are scammers.
After 2 yeras, I don’t want to wait more. Most open source projects die permanently.
this is a very philosophical take, but it’s still valid.
I do understand most of what you’ve said here, even if some don’t.
There is nothing wrong with wanting to avoid systemd and udev because even I want to, but when avoiding symptoms of the centralization ideology, such as systemd and udev, at all costs becomes so important that you neglect producing contents to defeat the centralization ideology, you put cart before the horse.
The vast majority of open-source developers didn’t even stop to consider the causal factors behind things like systemd and udev. They never considered content creation as the solution to the engine that endlessly spawns things like systemd, udev, and so on.
As long as you don’t stop things at the source, you will forever be overrun by creatures that spawned from the source.
The centralization ideology is the idea that small groups of people should control the rest of humanity and violate their rights. People who subscribe to the centralization ideology may fight with each other, but they also co-operate to create the global hierarchy on earth because they believe in enforcing the one global hierarchy. They just want to be the ones at the top of the global hierarchy, but all of them want to maintain the global hierarchy. It’s a singularity of control and power. The gravity is so strong that nothing on earth escapes it.
If open-source developers want to permanently stop things like systemd and udev as soon as possible, they must help independent content creators by improving usability and UX and by creating high-quality content creation software applications. Open-source developers can’t stop the endless waves of centralized software without content creators.
Gentoo maintains a systemd agnostic approach to udev - you can have systemd if you want, or use openrc without systemd. Maybe this could provide a vector for guix going forward?
I don’t know how practical this would be, as I am not familiar with the mechanics of the software in question.
It’s not difficult for guix maintainers to replace eudev with udev. They just need to use the build recipe from gentoo linux.
There is also a udev rule engine implement in Guile:
I find it interesting.
goodev doesn’t have libudev implementation.
We need something we can use now.