Bye, copyleft. Hello, corporate interests.
https://fosstodon.org/@gnutools/114168573791483805
GCC - GNU Toolchain (@[email protected])

Bye, GNU Coreutils. Hello, Rust Coreutils. Ubuntu 25.10 starts the preference cascade. https://www.phoronix.com/news/Ubuntu-25.10-Rust-Coreutils

Fosstodon

C is terrible and Rust is so much better, but I don’t think memory safety is the only justification for having ‘ls’ in Rust.

Some of those active in this effort worked on replacing GCC with Clang in distros before.

someone at IBM had warned (threatened?) me about corporate efforts to replace all things GNU and copyleft years ago. they wouldn't say more, but it soon became apparent that they were talking about Rust 😞

we have long had GNAT with strong copyleft, great memory safety and lighterweight compilation, which kind of renders the illusionism behind the push for Rust no better than a rich men's hat trick. but there it is, and people are falling for it.

@lxo @civodul I've looked at Ada a few times over the years, and it looks excruciating to program in. I mean, does it even have closures?

Also, it's object oriented, isn't it? I'm never going back to that.

@lxo @civodul good thing is that thanks to copyleft we can just keep using the good versions. much of the software i use i already used in 2005. i'm also going back to using IRC and other simple protocols i can easily hack things for - turns out i have fun again!
@lxo @civodul well, sadly most of GNU has been mostly in maintenance mode at best for a couple of decades now. We lost many brilliant free software advocates that burnt out trying to turn the nose dive around and also failed to inspire and recruit.

@janneke @lxo @civodul I suspect Janneke is right and that many soft elements like ease of collaboration and response to issues will be at play (as well as "new shiny Rust). As we know being a distro and working with upstreams can be complex.

If it was a central decision for strategic/business reasons the announcement would have come from a Product leader or from Mark. Not that it isn't a big decision, but is most likely _is_ coming from the engineering side.

@lxo @civodul The problem with GNAT is that it still can't be bootstrapped entirely from source (which is why Guix doesn't package it). Rust can, even though the bootstrap chain is really long.