I've heard the GNU Coreutils in Ubuntu 25.10 will be replaced with a Rust implementation. Should I be concerned about this?
- Vous devez vous identifier ou créer un compte pour écrire des commentaires
I've been made aware that the GNU Coreutils will be replaced with a Rust version in the next Ubuntu release. I absolutely disagree with Ubuntu's decision. Rust uses the MIT license instead of the GPL. Because Trisquel is currently based on Ubuntu, I assume that Trisquel 13 will be based on Ubuntu 26.04 LTS. And if the decision carries on into Ubuntu 26.04 LTS, that would be concerning to me. Therefore, how this will affect future Trisquel releases after Trisquel 12 ecne?
Thanks.
One of the biggest concerns I see is with people thinking GNU is just some low-level utilities rather than an operating system, they might even start referring to so-called Ubuntu as "Ubuntu Rust/Linux" as a means of mockery, similarly to how they use Alpine as an argument against GNU due to it (allegedly) not including GNU code, which is only true if things like Busybox and musl were a full replacement for most things GNU-related, even though a clear search in Alpine's repositories shows that this is not true since Alpine still has massive GNU packages in its repositories (even if one assumes they aren't installed by default), such as GNU Emacs, and even the much smaller (but extremely culturally signficant) GNU Nano editor, and Alpine even offers GNOME which for a long time was originally part of the GNU Project before leaving. Also, from "apt show busybox":
> The utilities in BusyBox generally have fewer options than
their full-featured GNU cousins; however, the options that are included provide the expected functionality and behave very much like their GNU counterparts.
This shows that Busybox is extremely inspired by GNU's design and work, if it wasn't for GNU, Busybox as it is today wouldn't exist at all.
So next time some mean person sends you that horrible copypasta about Alpine arguing that GNU is no longer relevant (
https://rl.bloat.cat/r/confession/comments/1hn5ort/i_use_linux_as_my_operating_system_i_state_proudly/ ), you know they are either ignorant or acting out of genuine malice.
> One of the biggest concerns I see is with people thinking GNU is just some low-level utilities rather than an operating system, they might even start referring to so-called Ubuntu as "Ubuntu Rust/Linux" as a means of mockery
There you go, literally only one day after I posted this and r/"linux"memes is already at it: https://rl.bloat.cat/r/linuxmemes/comments/1jc4tpy/rms_learning_the_news_about_ubuntu/
> Wait so Ubuntu is no longer a GNU+Linux distro?
> > It's a Rust+Linux distro
Rust would be much more of a concern if you agree with Hyperbola's views that Rust is not FSDG-compliant, but Hyperbola is probably the only FSDG-compliant system to see Rust as non-free. But Hyperbola behaves basically fully independently from the FSF, they generally do not care much about what other FSF-endorsed FSDG-compliant distros think and Hyperbola's government has a high level of toxicity with them e.g. labelling thoughts they don't like under the buzzword "conspiracy theories" (as seen in https://wiki.hyperbola.info/doku.php?id=en:project:teamwork ) and pretending to be pro-free-speech even though their e.g. communication channels are extremely strongly moderated, probably much more moderated than any other FSDG-compliant distro. I don't try to pretend I support "free speech", I've stopped advocating for free speech ever since I learned the beautiful things that community-led censorship allows through e.g. the FSDG's censorship against proprietary software. So it would be cool if, like me, Hyperbola stopped pretending to be pro-free-speech when they are basically the most anti-free-speech system out of all FSDG-compliant distros.
You can see on Hyperbola's Wiki ( https://wiki.hyperbola.info/ ) how they argue that they are the only truly freedom-oriented system and how they see all other FSF-approved distros as not being truly concerned with freedom but rather convenience due to the inclusion of things like Rust, systemd, GNOME and/or dbus. Unless you agree with the traits I've shown about Hyperbola, you probably shouldn't worry about their views regarding Rust.
EDIT: Hyperbola was the first FSDG-compliant system for me to have ever used long-term, I have many things to appreciate about Hyperbola, primarily their FSDG-compliance. As a FSDG-compliant system it's still better than 90%+ of distros. But I feel very hurt. I've done very problematic things while I was a part of Hyperbola's community, but the arguments I've put in this reply still stand even when disregarding my somewhat dark past.
I know atleast one of Hyperbola's developers reads the Trisquel Forums, so if such person is reading right now: Leave me alone. Don't try to contact me with a crappy non-apology. I don't want to hear about your thoughts about me. I made this reply as a means to answer the poster's question about Rust, not expecting any pity out of you at Hyperbola. I wish you and everyone else happiness but I don't want to hear you talking to me again. Please leave me alone.
I don't know much about Hyperbola, but it looks like a Parabola fork (correct me if I'm wrong). Didn't the Hyperbola project say they were going to switch development to BSD and no longer be a Linux-libre distro a few years ago? I rarely hear about them. I know Trisquel, PureOS, and Parabola best.
Anyways, interesting note.
> Anyways, interesting note.
Thank you so much, but Hyperbola is not a Parabola fork, it is a GNU/Linux distro independent from the others (Hyperbola GNU/Linux-libre) transitioning to an independent BSD operating system (HyperbolaBSD). During the start of Hyperbola as a FSDG-compliant system they were heavily based on both Arch and Debian (Arch package management + Debian-style stability), in contrast to Parabola which is Arch-based but on Arch's rolling-release model rather than Debian-style stability. Eventually Hyperbola proclaimed independence and behave quite autonomously. Parabola on the other hand doesn't try to be fully autonomous but rather extremely based on Arch development, being a libre version of Arch (and possibly other Arch-based systems) (noted at https://wiki.parabola.nu/Parabola_Social_Contract )
I see. They look so similar, that's what confused me. The website layout looks similar too.
Sounds like someone's living rent free in your head, you should try get over it, or at least I wouldn't hand them ammo so lightly.
That aside, I have nothing but fat respect for the guys from Hyperbola, committing to deliver the first and only freedom respecting BSD distribution (cope) is something worth of respect, and the fact they keep a safe distance from the FSF is good, so long they respect the organization, which they do.
Indeed. They will replace GNU Coreutils with uutils.
"My immediate goal is to make uutils’ coreutils implementation the default in Ubuntu 25.10, and subsequently in our next Long Term Support (LTS) release, Ubuntu 26.04 LTS, if the conditions are right."
(note: this reply is sarcasm)
> if the conditions are right.
That's the same excuse the Chinese government (and generally other Marxist States) uses to promote rapid economic growth under state-capitalist[1] oppression, everyone knows that the material conditions for communism (the final stage of socialism) can only be achieved by promoting wealth inequality, proprietary software, undemocratic government, abusive "social" networks such as TikTok and ignoring any concept about actually implementing socialist democracy (as in, democratic control over both the government and the workplace)
Dengism[2], this is why so-called Ubuntu promotes proprietary software, the only way to end the GNU Coreutils State and transition into uutils anarchocommunist utopia where all software is free is by promoting the same proprietary software contradictions that GNU opposes, this is why the "socialist" government in China adopts an undemocratic governance and promotion of profit motive and imperialism similar to the openly-capitalist governments in NATO. I've got it figured out people, so-called Ubuntu is actually the most free distro there is because through adopting unethical behaviour it has achieved the conditions necessary for the utopia where all software is free, people! Therefore leading to the conditions necessary to software freedom. Trust me people, once so-called Ubuntu achieves worldwide domination they will totally become very free, they are totally not doing this out of selfish interest rather than ideals, so-called Ubuntu is actually so genius by realizing that to achieve freedom you have to go further away from it!
Proof that so-called Ubuntu is just following the wise steps of the Chinese government: https://jing.rocks/2024/01/05/openkylin-attack-on-free-software.html
[1]. Technically all modern forms of capitalism have a State, but in this sense I used state-capitalism to constrast the very high level of State intervention in contrast to liberal-capitalism where the State adopts "free market" ideology during times of stability and mostly intervenes in times of severe crisis
[2]. Technically China under Xi Jingping has an ideology that's something inbetween Mao's and Deng's: Xi Jingping Thought. How amazing of them to name their ideologies after the current dictator, this is surely a sign of democracy, independent thinking required for such democracy without any cult of personality whatsoever and sure a sign that China as a State is sustainable and totally ruled by its people under socialist equality, it's not like there's another example of "socialist" country ("Soviet Union") where defining the ruling ideology simply based on the dictator at the top proved to be disastrous and completely incompatible with democracy. Hooray for the Chinese government!
And yes, I know that wealth inequality in China has technically been rapidly decreasing in the last year or 2, with China's number of billionaires rapidly shrinking, but this doesn't matter if there's a high chance that once dictator Xi dies another person with a different ideology will simply become dictator and crap on Xi's policies by eliminating the policies that led to wealth inequality shrinking in the first place.
To contrast this, an example of actual stability would be the Free Software Foundation, which has maintained mostly the same policy even after Stallman was forcibly overthrown from the presidency, showing that the Free Software Foundation is an actual organization with principles rather than a one-person cult-of-personality-based dictatorship that would crumble the moment Stallman goes missing.
I assume it would be straightforward to fix simply by reverting from uutils to GNU Coreutils in future Trisquel versions.
BTW, I'm not a developer. I'm just a Trisquel user.
> Rust uses the MIT license instead of the GPL.
Rust is licensed under two GPL-Compatible Free Software Licenses:
https://www.gnu.org/licenses/license-list.en.html#Expat
https://www.gnu.org/licenses/license-list.en.html#apache2
Which one is your concern?
The MIT license. At least, what I heard from the YouTube channel "Mental Outlaw," the MIT license allows changes to the program without releasing the modified source while the GPL requires the modified source to be released along with the updated program.
From what I can see, if Ubuntu decides to make an updated version of uutils without the acompanying source code, that would constitute proprietary software, right?
I guess I've been misinformed if that's not the case. I'm not great at understanding licenses.
The title of the YouTube video where I got my information: "Ubuntu Will Replace GNU Core Utilities With Rust" by Mental Outlaw.
Still, it's a free license and Trisquel has always included many packages under the MIT license. In fact, uutils is already there in Trisquel 12 (the unreleased version currently under development). It's just not installed by default. Whatever the Ubuntu/Trisquel developers decide, you'll still have a choice between coreutils and uutils.
Thanks for explanation, i was interested mostly about if it's still choice between still using GNU ocreutils or uutils.
I guess this isn't as big of a deal Mental Outlaw claimed to be. As long as I have a choice to switch back, I think I'll be fine.
That's true.
The package "rust-coreutils" is here: https://archive.trisquel.org/trisquel/pool/main/r/rust-coreutils/
Based on https://directory.fsf.org/wiki?title=Rust&diff=87657&oldid=87651 and on https://www.fsf.org/blogs/directory/fsd-meeting-recap-2022-09-09, I think the software freedom issues that once existed with Rust's trademark policy no longer exist.
https://www.apache.org/licenses/LICENSE-2.0 6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file.
This is not the trademark policy you are looking for.
Mozilla's trademark policy is problematic due to distribution restrictions, such as banning paid redistribution of exact copies and limiting distribution to only the latest version.
The Rust version does have a requirement to rename, which is conditionally okay per the Free Software Definition, "it is acceptable for the license to require that you change the name of the modified version, remove a logo, or identify your modifications as yours. As long as these requirements are not so burdensome that they effectively hamper you from releasing your changes, they are acceptable..."
"A special issue arises when a license requires changing the name by which the program will be invoked from other programs. That effectively hampers you from releasing your changed version so that it can replace the original when invoked by those other programs. This sort of requirement is acceptable only if there's a suitable aliasing facility that allows you to specify the original program's name as an alias for the modified version."
For compatibility, modified versions must still be callable as rustc and cargo, or allow symlinks, so scripts and programs can invoke the expected names. This isn't about trademark confusion but ensuring interoperability, much like invoking gcc as cc. The original (proprietary) UNIX C compiler was invoked as cc. No one's claiming gcc is cc, but it can still be run that way for compatibility with programs that expect the C compiler to be called cc. As long as the Rust Foundation doesn't oppose this usage (and I have no indication they do), it seems fine.
Another thing I hadn't thought to mention at the time is whether the software available via Cargo is all free software. This is separate from the trademark issue but raises the same concerns as other language-specific package managers if it isn't.
Rust's problems and concerns
https://wiki.hyperbola.info/doku.php?id=en:philosophy:rust_issues
We count doku as superseded by this thorough summary of the situation:
https://trisquel.info/en/forum/rust-flash-player-emulator-thing-and-nobody-told-me#comment-174990
The Hyperbola critique appears to conflate multiple issues while also misinterpreting the Free Software Definition. For an example, the argument against Rust's complexity and reliance on Cargo seems to be more about minimalism and simplicity rather than software freedom. For that I say it's not called the Simple Software Foundation: https://jxself.org/complexity.shtml. The claim that the Rust Foundation requires permission for modifications is misleading. The actual restriction is on using the Rust or Cargo names with modified versions. The Hyperbola people are free to have their own standards & viewpoints but remember that the document they're writing is based on those, not things like the Free Software Definition.
Doku clearly states: "The job of Hyperbola, independently of Freedom is, and always is, to develop a stable, minimalistic and privacy-oriented operating-system." So any confusion probably stems from readers of Doku failing to permeate its spirit through and through.
Concerning non-free Rust crates in Cargo, it is indeed the same situation as with other third-party package managers. I am not sure how an FSDG distro may handle the cargo-deny crate in order to keep non-free software out, but there is now a Parabola package:
https://www.parabola.nu/packages/extra/x86_64/cargo-deny
https://deps.rs/repo/github/EmbarkStudios/cargo-deny
Honestly, I am more concerned that they are advocating for using rust as a blank statement for security.
Just because rust exists, doesn't mean everything needs to be built in it.
Some people view rust as an all in one thing to such a degree, where I have heard people refer to people who are big fans of it as being in a cult.
I don't know if this is true, the part about rust lovers being in a cult, but it does feel like rust is a religion where people want to put it everywhere.
Ubuntu replacing coreutils gpl for mit licensed version made in rust, is just... stupid though regardless.
Don't fix what isn't broken and don't be deceptive about your motives is the wise thing.
Ubuntu and other similar projects need to understand this. We already have microsoft, google and apple being this way... we don't need more corporations and/or companies doing this.
To summarize, there are multiple problems beyond just the mit licensing that I can see.
Agree or disagree, this is my viewpoint. I have a bad feeling I am about to see this post thumbed down 5 points due to people's sheer toxicity though.
I ask you, please prove me wrong.
I don't think there is a need to rewrite GNU coreutils in rust as it has proven to be highly reliable, and hasn't had many security issues. Rewriting it for the sake of rewriting could just introduce new bugs.
But as someone who has programmed quite a bit in C and is currently learning Rust, I do believe that we should move away from C at least for more security-critical software going forward. It's just really hard even for an experienced C programmer to avoid memory safety issues, not to mention all of the other undefined behavior.
I do not think Rust developers are a "cult" but it is true that there are some who are a bit fanatical. I would guess that this is just a vocal minority.
I'm not too sure how I feel about the language personally as I've only gone through 5 chapters of the Rust book so far.
I don't think there's anything worth discussing about Ubuntu beyond the plans of Trisquel to move as far away from them as possible and fork off Debian for further releases, Ubuntu is an abysmally bad GNU distribution full of proprietary shitware and with a corporate/anti-freedom agenda, any self-respected free software group should've breached away from them ever since they showed up with those Canonical snap shenanigas, the Amazon adware injection or the command line ads.
Ubuntu is nothing but an embarrassment for the free software community, so it's no surprise they would try to replace long established GNU software with unreliable Rust codebase that has a much weaker licensing, not even mentioning the problems Rust has with branding and the (proprietary) compiler.
The Rust compiler is not proprietary.
GCC is definitely not proprietary, and by extension gccrs, but the reference implementation/"official" compiler surely is proprietary. Good luck bootstrapping that thing without already having a built compiler beforehand.
RustConf 2022 - Bootstrapping: The once and future compiler by Jynn Nelson
https://inv.nadeko.net/watch?v=oUIjG-y4zaA
Rustc is free software.
Okay, but if that's true, why isn't crab language still being worked on?
I would think if rust is free software, more than one compiler could exist for rust yet I think the crab language people stopped due to barriers from trademarks.
Crablang was never a seriously developed fork. It was more of a protest against a draft of the new Rust trademark policy. Pretty much everyone seems to be fine with the new policy so there's no interest in a Crablang.
Technically, languages aren't software so it doesn't make sense to refer to them as free or nonfree. When we refer to Rust being free, we mean the official implementation of Rust.
There is an alternative Rust compiler, it's called gccrs and it is also free.
A lot of people on this forum think Rust is nonfree, or that it is technically bad for various other reasons though I doubt they have ever used it. As far as I can tell these notions mainly stem from a wiki page by Hyperbola, a distribution that has its own definition of "free software" that is quite different from anyone else's.
Huh... nevermind then. I was unaware of this. Perhaps there is a libre implementation of rust now after all.
I think this wiki page might have been in response to this comment: https://wiki.hyperbola.info/doku.php?id=en:philosophy:wording_phrases#there_is_surely_not_an_own_or_different_definition_of_free_software
If one wants to better understand Hyperbola's position on this, reading the wiki is good, but it might also be good to join #hyperbola on Libera.Chat so you can ask questions and get answers in real-time.
The disagreement with Hyperbola could be a difference in interpretation of the definition of free software, but it could be a difference in in interpretation of what Rust allows. For example people I've talked to in #hyperbola think symlinking rust to a modified version would not be allowed (in contrast to jxself's comment [1]). If modifying Rust did also require modifying every program that calls Rust, then I think it would be quite reasonable to say Rust is nonfree software, because modification would be very impractical.
The trademark policy of Rust is just a standard trademark policy that requires changing the name of the software when distributing modified versions (with some limited exceptions). The FSF explicitly states that this does not make software nonfree. jxself is right here.
I do not understand why Hyperbola fans feel the need to come here and tell people that Rust is nonfree. I don't have any problem with them not including Rust in their distro, but I wish that they would stop calling it nonfree, or at least acknowledge that their opinion is not the mainstream one.
Well its not an issue if there is a free implementation I guess.
Since you say there's more than one compiler that exists that seems to suggest you might be right.
Since there is a GCC compiler for rust now.
Although the ideal of some rust fans still makes me wince.
Just because a new programming language comes out, doesn't mean you need to rewrite everything in that language. Some rust fanatics do this and it makes me flinch at their cultishness.
I hope that's a minority though.
I know you refuse to accept it, but the official Rust compiler is also free software.
Generally speaking, online discussions are not representative of real life. There are many people who use Rust professionally, only a minority of these actively discuss it online, let alone call for everything to be rewritten in Rust.
Well, so far it looks like you might be right. But if at some point other implementations that are libre die due to not asking for permission, I will re-evaluate what I just posted right now.
Perhaps its just the rust developers who want to write current programs in rust when they don't need to just to kill the old idea.
Meh... w/e
> if at some point other implementations that are libre die due to not asking for permission
I don't know where you got this misconception from, that anyone has to ask for permission to develop a Rust compiler.
Programming languages are not themselves copyrightable, only their implementations are. The Rust Foundation could not prevent anyone from writing an alternative Rust compiler even if they wanted to (they don't).
This is how GNU Octave exists, for example. Octave developers didn't have to get permission from MATLAB, even though it is proprietary, because they wrote their own implementation.
Oh wasn't aware of this. My knowledge must be somewhat dim on this
Go tell that to the FSF: https://directory.fsf.org/wiki/Rust
Hyperbola welcomes you then. I agree, I think the compiler is non-free. If someone makes another compiler and it goes unchallenged by the rust foundation long enough, then I will say otherwise.
Although you replied to the post where I almost only gave a link to a video, you obviously have not watched it: there are other Rust compilers, in particular gccrs, and the Rust community welcomes them, especially because they are useful to guarantee the chain of trust (Ken Thompson's famous "Trusting Trust" attack).