miri VS chalk

>While we are many missing language features away from this being the case, the noalias case is also magic descended upon box itself, with no user code ever having access to it.

I'm not sure why the author thinks there's magic behind Box. Box is not a special case of `noalias`. Run this snippet with miri and you'll see the same issue: https://play.rust-lang.org/?version=stable&mode=debug&editio...

`Box` _does_ have an expectation that its inner pointer is not aliased to another Box (even if used for readonly operations). See: https://github.com/rust-lang/miri/issues/1800#issuecomment-8...)

Miri [0] is an interpreter for the mid-level intermediate representation (MIR) generated by the Rust compiler. MIR is input for more processing steps of the compiler. However miri also runs MIR directly. This means miri is a VM. Of course it's not a bytecode VM, because MIR is not a bytecode AFAIK. I still think that miri is a interesting example.

And why does miri exist?

It is a lot slower. However it can check for some undefined behavior.

[0]: https://github.com/rust-lang/miri

Provenance is a dynamic property of pointer values. The actual underlying rules that a program must follow, even when using raw pointers and `unsafe`, are written in terms of provenance. Miri (https://github.com/rust-lang/miri) represents provenance as an actual value stored alongside each pointer's address, so it can check for violations of these rules.

Lifetimes are a static approximation of provenance. They are erased after being validated by the borrow checker, and do not exist in Miri or have any impact on what transformations the optimizer may perform. In other words, the provenance rules allow a superset of what the borrow checker allows.

There has been discussion of doing this with MIRI, which would be easier than all of rustc.

This issue has been fixed: https://github.com/rust-lang/miri/issues/2964

Actually, I've done more advanced tests with MIRI (see https://github.com/rust-lang/miri/issues/2920 for example) which allowed me to fix some issues. I've also made the code compatible with loom, but I didn't found the time yet to write and execute loom tests. That's on the TODO-list, and I need to track it with an issue too.

He is one of the big brains behind Miri, which is a interpreter that runs on the MIR (compiler representation between human code and asm/machine code) and detects undefined behavior. Super useful tool for language safety, pretty interesting on its own.

I would also run your tests in Miri (https://github.com/rust-lang/miri) to try to cover more bases.

The Rust compiler uses a Prolog-like query language internally for type checking generic requirements and traits: https://github.com/rust-lang/chalk

rust-analyzer already uses chalk (https://github.com/rust-lang/chalk) which should replace the current trait resolver.

Well it's quite easy to come to that conclusion: The code compiles with rustc, which is currently the reference implementation. If rust-analyzer does not match rustc's behavior it's an issue in their implementation. That written it's not that easy to fix as it's related to how rust-analyzer resolves types/traits. rust-analyzer uses chalk for this, which is known to be incomplete/diverging from the RFC'ed behavior. Now one could argue that we can simplify diesel to the point where it works will with rust-analyzer/chalk, but that would result in basically removing core diesel features that exist way longer than rust-analyzer.

> Not all of those things work well in the real world. E.g. logic programming (prolog) is cool but ultimately never really caught on.

It does have its niches though. For example, there is a trait solver for Rust called Chalk that uses a Prolog-inspired language because trait bounds basically define a logic:

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

Maybe something in the prolog/datalog direction could be useful? Notably Rust has Chalk to help with trait resolution ("Chalk is a library that implements the Rust trait system, based on Prolog-ish logic rules.")

There has been work to implement part of the Rust typing logic in the Chalk Engine which uses a prolog-ish syntax to describe its rules.

That's something Chalk is trying to tackle.

We probably will see all of them at some point -- polonius is a current effort to make the borrow checker accept more valid programs, in a way that also simplifies the logic and is probably a bit faster than the current NLL system, chalk is an attempt to do a similar thing for the trait system, and cranelift is a project that seeks to replace the LLVM codegen backend. But obviously, these are very large and complex projects that are gonna take some time.

Do Rust developers realize that? Oh, yes, absolutely, that's why we have this:

For me an attempt to write where HashMap: FromIterator and then use new and insert was totally bizzare because currently rustc is pretty primitive and doesn't do super-complex machinery needed to do what you want. Chalk may fix that one day, but it's nowhere near to being ready for inclusion into rustc thus I wouldn't even attempt to do what you tried to do… but that's not something you are supposed to know before reading this book!