Microsoft Continues "Demikernel" Development LibOS For Kernel-Bypass I/O

[Chewi]

Do you mean MSVS (Visual Studio) or do you mean Rust support for the MS C/C++ compiler?
I think theire MSVC is the only actual native compiler, Microsoft have.

Integrating Rust code in C/C++ programs will not really be possible. So how would be looking a Microsoft Rust compiler?
And under what license would it be?

Actually Microsoft use their own build of the regular Rust toolchain. When I last used it, it had some Cargo feature for fetching crates in a slightly different way. I don't know what other differences there are, if any.

[theuserbl]

Actually Microsoft use their own build of the regular Rust toolchain. When I last used it, it had some Cargo feature for fetching crates in a slightly different way. I don't know what other differences there are, if any.

You are right. There is really created a Rust fork by Microsoft called "Rust for Windows". It is currently in version 0.60.0 and published on September, 25.

https://github.com/microsoft/windows-rs

The main Rust on the other side is currently in version 1.82.0 and publiched on October, 17.

https://github.com/rust-lang/rust

So the version number of the Rust for Windows have nothing to do with the main Rust.

That reminds me a little bit to Microsofts Fork of the statistical language R called "R Open".

https://github.com/microsoft/microsoft-r-open

It was a fork of the main R.

https://www.r-project.org/

https://github.com/wch/r-source

And R Open have had also its own version number, additional features and so on.
But R Open is currently dead. Last release for over four years ago.

[theuserbl]

... or are Rust for Windows only libraries for Rust, to call the Windows API from that?
Then there is still missing a link to Microsofts own Rust, if it exists.

[Chewi]

You are right. There is really created a Rust fork by Microsoft called "Rust for Windows". It is currently in version 0.60.0 and published on September, 25.

https://github.com/microsoft/windows-rs

I was talking about Microsoft's Rust toolchain for Linux, which has broadly the same version number IIRC. I'd never heard of Rust for Windows.

[Chewi]

Actually, I have heard of Rust for Windows, but I know it as windows-rs. That has nothing to do with the Rust toolchain, it's for using the Windows APIs from Rust. I've never used it.

[NateHubbard]

Few people realize that good support for Linux is quite near in Redox land. With native or virtualized implementation. This is enough for most desktop use cases, only running AA or AAA games don't fit this paradigm.

Maybe, but I see that Redox is 9 (going on 10) years along already, and at that point Linux itself was a fully usable OS (I had been using full time at home and at work for quite some time by then).
Redox is nowhere near that yet. It can take a very long time indeed to implement compatibility with an existing moving target. See ReactOS.
I would hope that they'd get a fully working operating system up and running before attempting to make it compatible with another one.

[Errinwright]

Maybe, but I see that Redox is 9 (going on 10) years along already, and at that point Linux itself was a fully usable OS (I had been using full time at home and at work for quite some time by then).
Redox is nowhere near that yet. It can take a very long time indeed to implement compatibility with an existing moving target. See ReactOS.
I would hope that they'd get a fully working operating system up and running before attempting to make it compatible with another one.

The complexity and pre-requisites have dramatically changed since Linux was programmed. I don't get the nonsensical noise people output without any insight. Could Redox have been accelerated with more funding early in the process? Sure. Was it? Not in our timeline.

[Unklejoe]

Is this similar in nature to libfabric?

[bob l'eponge]

So it's a Demi kernel for chips that follow the Demi Moore law?