Rewriting the GNU Coreutils in Rust [LWN.net]

Rewriting the GNU Coreutils in Rust

Posted Jun 9, 2021 23:57 UTC (Wed) by pabs (subscriber, #43278)
In reply to: Rewriting the GNU Coreutils in Rust by khim
Parent article: Rewriting the GNU Coreutils in Rust

Google are replacing the Linux kernel too, see Fuschia.

to post comments

Rewriting the GNU Coreutils in Rust

Posted Jun 10, 2021 9:17 UTC (Thu) by Wol (subscriber, #4433) [Link] (2 responses)

I just wish they'd replace POSIX ... :-)

That's the big worry for me about Fuschia, linux quite specifically implements POSIX, and my experience leads me to believe a lot of the things I don't like about linux are posix-fossilized design decisions. If Fuschia breaks away from that, it could leave linux behind (which is a good thing), but it could leave Linus behind, and that most definitely is a BAD thing.

The point behind Open Source is that "proprietary friction" pushes software writers into collaboration, and linux is successful because of that, not because of copyleft.

But the problem with much software is that the community of users, and the community of developers, is such that the Open Source "friction" doesn't exist. Users can't share development because users don't know how to develop. But copyleft doesn't wrok very well in this scenario, either, because it makes it hard for developers to monetise their work ...

Cheers,
Wol

Rewriting the GNU Coreutils in Rust

Posted Jun 10, 2021 10:13 UTC (Thu) by khim (subscriber, #9252) [Link]

> If Fuschia breaks away from that, it could leave linux behind (which is a good thing), but it could leave Linus behind, and that most definitely is a BAD thing.

For better or for worse, but that's not in a cards. Fuschia have to run Android apps well and for that it needs POSIX.

Sure, it may try to create a sandboxed “jail” for these apps, but then the outcome would be like with MacOS: everyone would just write apps for that “jail” and “native Fuschia” would be irrelevant.

> The point behind Open Source is that "proprietary friction" pushes software writers into collaboration, and linux is successful because of that, not because of copyleft.

Copyleft certainly helped. Lots of stuff in Linux was started as someone's proprietary work and then shared and merged. BSDs are losing such contributions every year (when SONY takes FreeBSD for PlayStation and adds some interesting features to it… what does FreeBSD receives back? nothing, obviously). This made big difference.

But that only worked when copyleft was mild enough for the companies to adopt that software in the first place. When GPLv3 “crossed the pain threshold” companies just stopped using such software so there are no advancements to share back.

> But copyleft doesn't wrok very well in this scenario, either, because it makes it hard for developers to monetise their work ...

Indeed. But that's what GPLv3, ultimately, did. GPLv2 was mild enough that companies adopted GPLv2-licensed programs and then developers got some money. This doesn't often happen with GPLv3: companies prefer to adopt technically inferior alternative rather than deal with GPLv3. Developers are observing that and stop developing GPLv3 software. And then there are no copyleft software at all.

Rewriting the GNU Coreutils in Rust

Posted Jun 11, 2021 20:33 UTC (Fri) by notriddle (subscriber, #130608) [Link]

> I just wish they'd replace POSIX ... :-)

Fuchsia isn't POSIX-compliant. <a href="https://fuchsia.dev/fuchsia-src/concepts/system/libc?hl=en">It does not have signals, and it does not support forking</a>.

Rewriting the GNU Coreutils in Rust

Posted Jun 10, 2021 9:57 UTC (Thu) by khim (subscriber, #9252) [Link] (1 responses)

More correct phrasing would be: Google is trying to replace Linux. And failing.

Industry is just not interested in that and even Google itself spends more resources advancing Linux than developing it's supposed replacement.

And I suspect this is because of Linus adamant position “Linux is purely about tit-for-tat, we don't want to turn it into weapon of mass opinion”.

Note how easy was it for the industry to arrange replacements for GPLv3 pieces (GCC was replaced with clang, Samba abandoned entirely and instead phones are using MTP which is awful from the user's perspective, but doesn't need to put GPLv3 software in the image and so on), while GPLv2 projects are stubbornly clinging and refusing to die en masse. Although there are some limited success like when BlueZ was kicked out from Android, but smaller things (like FLAC library, e.g.) are still there.

Basically position of the industry WRT GPLv2 was “we would gladly use an alternative with more permissive license if someone else would develop it — but wouldn't spend resources themselves” but GPLv3 causes entirely different reaction “we would's use that software and would develop an alternative if needed”.

This made “free software” position much, much, MUCH weaker.

Rewriting the GNU Coreutils in Rust

Posted Jun 28, 2021 14:14 UTC (Mon) by immibis (subscriber, #105511) [Link]

Does it really matter which software their software is based on, if you can't modify it? Apple-heavily-modified-coreutils and Apple-written-from-scratch-coreutils are about the same usefulness to me.