Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

[bingmatv]

[-- Attachment #1: Type: text/plain, Size: 1050 bytes --]

Richard Matthew Stallman tell people to say GNU/Linux to spread free software, but Rust programming language ecosystem is growing fastly, Rust focuses on memory safety. If one day, Rusty alternative of GNU software is published, but they're not developed by GNU project, someone may say "rust-lang/Linux" instead of "GNU/Linux", which is very bad for spreading free software. So I suggest GNU project use Rust to rewrite all GNU software to prevent future people from saying "rust-lang/Linux" instead of "GNU/Linux", because Rust language users may only support memory safety, but not free software movement. In conclusion: At least glibc and bash need to be rewritten in Rust, glibc is the interface of user-space programs and kernel syscall, I think it's better to write librust instead of glibc, preventing future people from saying "rust-lang/Linux" that may not support free software movement, since rust-lang/Linux doesn't mention GNU and only supports memory safety. And finally, Rust focuses on memory safety, it's currently safer than C++.

Re: Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

[Adhemerval Zanella]

> On 31 Mar 2024, at 09:44, bingmatv--- via Libc-help <libc-help@sourceware.org> wrote:
> 
> Richard Matthew Stallman tell people to say GNU/Linux to spread free software, but Rust programming language ecosystem is growing fastly, Rust focuses on memory safety. If one day, Rusty alternative of GNU software is published, but they're not developed by GNU project, someone may say "rust-lang/Linux" instead of "GNU/Linux", which is very bad for spreading free software. So I suggest GNU project use Rust to rewrite all GNU software to prevent future people from saying "rust-lang/Linux" instead of "GNU/Linux", because Rust language users may only support memory safety, but not free software movement. In conclusion: At least glibc and bash need to be rewritten in Rust, glibc is the interface of user-space programs and kernel syscall, I think it's better to write librust instead of glibc, preventing future people from saying "rust-lang/Linux" that may not support free software movement, since rust-lang/Linux doesn't mention GNU and only supports memory safety. And finally, Rust focuses on memory safety, it's currently safer than C++.

Was it generated using a LLM? In any case you do have some libc implemented in rust (the one from redox project for instance), so you might check this out and send patches to implement the missing parts.

Re: Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

[tomas]

[-- Attachment #1: Type: text/plain, Size: 284 bytes --]

> > [strange drivel elided]
> 
> Was it generated using a LLM?

My impression too. Those startups are desperate for feedback and
training data, so why shouldn't they strip-mine public mailing
lists for that?

This is why I tend to not answer to such stuff.

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

[Dennis Clarke]

On 3/31/24 10:50, tomas@tuxteam.de wrote:
>>> [strange drivel elided]
>>
>> Was it generated using a LLM?
> 
> My impression too. Those startups are desperate for feedback and
> training data, so why shouldn't they strip-mine public mailing
> lists for that?
> 
> This is why I tend to not answer to such stuff.

Pardon me for being very very old school but Do Not Feed The Trolls.


--
Dennis Clarke
RISC-V/SPARC/PPC/ARM/CISC
UNIX and Linux spoken

Re: Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

[馬 冰]

[-- Attachment #1: Type: text/plain, Size: 270 bytes --]

Rust already allows main function returning Result<T, E>, e.g. fn main() -> Result<(), std::io::Error>{Ok(())}, so I think the operating system should accept Result<T, E> instead of int, i.e., value returned from int main(){}. How to write patches for LLM? What's LLM?

Re: Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

[Konstantin Kharlamov]

On Sun, 2024-04-07 at 07:27 +0000, 馬 冰 via Libc-help wrote:
> Rust already allows main function returning Result<T, E>, e.g. fn
> main() -> Result<(), std::io::Error>{Ok(())}, so I think the
> operating system should accept Result<T, E> instead of int, i.e.,
> value returned from int main(){}. 

Nothing needs to be modified. AFAIU `main()` in Rust "returns" a Result
simply to facilitate using "question mark operator" inside `main`. But
it's just an adapter from internal API to one operating on `int`s, i.e.
`Ok()` will be converted to `0`, and an error will be converted to
another number. Current API is fine (and as a side note, the "question
mark operator" unfortunately doesn't exist in most languages. Offhand
Haskell have similar feature, Idk of any other langs).

> How to write patches for LLM? What's LLM?

LLM is a "large language model". The author was referring to the email
possibly being written by an artificial neural network. Although Idk
what's the ponit, but then again the OP didn't reply to this point, so
there may be something to that hypothesis.

Re: Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

[馬 冰]

[-- Attachment #1: Type: text/plain, Size: 382 bytes --]

"it's just an adapter from internal API to one operating on `int`s, i.e.
`Ok()` will be converted to `0`, and an error will be converted to
another number", it may cost more memory if using 2 type systems simultaneously. Rust Result can use Box to allocate errors with unknown size on heap, whereas int is fixed-size, they may be incompatible due to unknown size and fixed-size.

Re: Richard M. Stallman says GNU/Linux, but Rust lang ecosystem is growing.

[Konstantin Kharlamov]

On Thu, 2024-04-11 at 12:05 +0000, 馬 冰 wrote:
> 
> "it's just an adapter from internal API to one operating on `int`s,
> i.e.
> `Ok()` will be converted to `0`, and an error will be converted to
> another number", it may cost more memory if using 2 type systems
> simultaneously. Rust Result can use Box to allocate errors with
> unknown size on heap, whereas int is fixed-size, they may be
> incompatible due to unknown size and fixed-size.

Are you seriously motivated enough to start rewriting POSIX and other
standards and convincing people to break backward compatibility just to
save 8 bytes? Well… good luck