shake VS Rust-for-Linux

Another interesting implementation is Shake: https://shakebuild.com/

It is technically a Haskell DSL, but supports Ninja files, time estimates and has tools for linting and profiling.

They explicitly refer to Shake build system and Build Systems a la Carte paper.

The one paper that gave me hope about build systems was Build systems à la carte: Theory and practice, by Andrey Mokhov, Neil Mitchell, and Simon Peyton Jones. Among other things, it describes the theoretical underpinnings of the Shake build system. To be honest I believe any build system that ignores the maths described in this paper can safely be ignored. (You may however ignore the paper itself if the maths checks out. See Daniel J. Bernstein's redo, which matches Shake very closely.)

Yeah, they don't have to be terrible. I haven't used it, but people in my circles tend to really like Shake, which uses a Haskell embedded DSL to describe builds.

You could try Shake. It's a sane build system written by a former co-worker of mine. https://shakebuild.com/

Shake does require compilation as it's essentially just a Haskell library providing a DSL and it works just fine, I guess in gradle's case it's a thing about Java-typical overengineering and complete blindness to resource usage. Shake's underlying engine can actually go head-to-head with ninja itself when building ninja files.

Did you look at shake: https://shakebuild.com/ ?

For reasons I don't want to get into, I am building my own GHC package without cabal. The documentation is a little sketchy, but I've succeeded in build and installing it in my own user database (I'm on linux x86_64, using GHC 8.6.5). I am using shake to do all of this, and I've been pretty pleased with how it works.

Rust is backwards compatible when you stick to stable features, but the kernel uses unstable features that can and do incur breaking changes.

https://github.com/Rust-for-Linux/linux/issues/2

> How would this work?

Don't know exactly what you're asking.

> And why would it be a better idea?

Poorly written device drivers are a significant attack vector. It's one of the reasons Linux is now exploring using Rust for its own device drivers.[0] You may be asking -- why Rust and not some other language? Rust has many of the performance and interoperability advantages of C and C++, but as noted, makes certain classes of memory safety issues impossible. Rust also has significant mindshare among systems programming communities.

[0]: https://rust-for-linux.com

Ctrl-F "rust"

https://rust-for-linux.com/ links to LWN articles at https://lwn.net/Kernel/Index/#Development_tools-Rust that suggest that only basic modules are yet possible with the rust support in Linux kernels 6.2 and 6.3.

Rust-for-linux links to the Android binder module though:

> Android Binder Driver: This project is an effort to rewrite Android's Binder kernel driver in Rust.

> Motivation: Binder is one of the most security and performance critical components of Android. Android isolates apps from each other and the system by assigning each app a unique user ID (UID). This is called "application sandboxing", and is a fundamental tenet of the Android Platform Security Model.

> The majority of inter-process communication (IPC) on Android goes through Binder. Thus, memory unsafety vulnerabilities are especially critical when they happen in the Binder driver

... "Rust in the Linux kernel" (2021) https://security.googleblog.com/2021/04/rust-in-linux-kernel... :

> [...] We also need designs that allow code in the two languages to interact with each other: we're particularly interested in safe, zero-cost abstractions that allow Rust code to use kernel functionality written in C, and how to implement functionality in idiomatic Rust that can be called seamlessly from the C portions of the kernel.

> Since Rust is a new language for the kernel, we also have the opportunity to enforce best practices in terms of documentation and uniformity. For example, we have specific machine-checked requirements around the usage of unsafe code: for every unsafe function, the developer must document the requirements that need to be satisfied by callers to ensure that its usage is safe; additionally, for every call to unsafe functions (or usage of unsafe constructs like dereferencing a raw pointer), the developer must document the justification for why it is safe to do so.

> We'll now show how such a driver would be implemented in Rust, contrasting it with a C implementation. [...]

This guide with unsafe rust that calls into the C, and then with next gen much safer rust right next to it would be a helpful resource too.

What of the post-docker container support (with userspaces also written in go) should be cloned to rust first?

Rust-for-Linux issue tracker

Yes, I definitely agree that it's a problem that pr_info implicitly wraps its arguments in unsafe {}. I wrote my own Pull Request with a trival fix.

Note that this template won't work with Linux 6.1, which has very minimal Rust support. You'll want the RustForLinux tree, or maybe Linux 6.2.

You can join Rust for Linux zulip chat by requesting invite using the link in https://github.com/Rust-for-Linux/linux 's README.