inkwell VS rfcs

Secondly, I have learned about LLVM, and I have learned about the Inkwell library on Rust (It's a New Kind of Wrapper for Exposing LLVM (Safely)). Has anyone used this library before? Is this a good practice? Is it suitable for my compiler? Can I write some optimization passes of my own using this library?

I am trying to use the inkwell Rust crate for working with LLVM. In Cargo.toml, I have specified the LLVM version using features = ["llvm15-0"]

inkwell is a great llvm binding for rust and it has an implementation of kaleidoscope

I'm working on a compiler that uses the LLVM wrapper Inkwell for compilation. In order to compile something in inkwell, unless I'm missing something (which I very well might be), you need two structs:

Compiler backends (If building JIT/machine compiled langauges) 1. cranelift 2. inkwell - safe rust wrapper around llvm

Haha small world. I think we go to the same university and I took the same or similar course a few years ago. If somebody hasn't done the work for you, you may have to do some instrumentation yourself depending on what you want to track IMO https://github.com/TheDan64/inkwell is the best LLVM wrapper for Rustland (Python's llvmlite is a bit easier to use though)

I'm currently using the Inkwell bindings for Rust, which I've found actually pretty nice. In terms of generating LLVM IR, the C bindings (which is what Inkwell uses internally) can do anything you want them to (definitely not limited to trivial languages as someone else here said.) I'm even using the LLVM garbage collection infrastructure, with no problems (well, no problems in generating it; the LLVM GC infrastructure works pretty well but is sparsely documented, so actually writing a GC is fairly difficult, but it's doable). The C bindings are actually more stable than the C++ bindings (!), although not quite as stable as the textual IR format; but without the bindings you would have to write code to generate the IR yourself, the compiler would be slower as it must be emitted as text and then reparsed in a different process, and you would have less control over optimization.

safe llvm bindings https://github.com/TheDan64/inkwell

As we reviewed in this article LLVM IR has many use-cases and allows us to analyze and optimize source code through its passes. Knowing IR language itself will help us to write our passes and build projects around it for debugging, testing, optimizing. Currently, LLVM IR doesn't have Rust API. It's mainly used through the C++ library. However, some user-created repos are available on crates.io. There is a Rust binding to LLVM's C API - llvm-sys and two other, more Rusty APIs that are using LLVM: inkwell and llvm-ir. And finally, if you want to learn how to write a LLVM pass you should start here.

RFC: Add large language models to Rust

https://github.com/rust-lang/rfcs/pull/3603

Man, SO and family has really gone downhill. That top answer is absolutely terrible. In fact, if you care, you can literally look at the RFC discussion here to see the actual debate: https://github.com/rust-lang/rfcs/pull/582

Basically, `for x in y` is kind of redundant, already sorta-kinda supported by itertools, and there's also a ton of macros that sorta-kinda do it already. It would just be language bloat at this point.

Literally has nothing to do with memory management.

Congrats!

> Similarly, uv does not yet generate a platform-agnostic lockfile. This matches pip-tools, but differs from Poetry and PDM, making uv a better fit for projects built around the pip and pip-tools workflows.

Do you expect to make the higher level workflow independent of requirements.txt / support a platform-agnostic lockfile? Being attached to Rye makes me think "no".

Without being platform agnostic, to me this is dead-on-arrival and unable to meet the "Cargo for Python" aim.

> uv supports alternate resolution strategies. By default, uv follows the standard Python dependency resolution strategy of preferring the latest compatible version of each package. But by passing --resolution=lowest, library authors can test their packages against the lowest-compatible version of their dependencies. (This is similar to Go's Minimal version selection.)

> uv allows for resolutions against arbitrary target Python versions. While pip and pip-tools always resolve against the currently-installed Python version (generating, e.g., a Python 3.12-compatible resolution when running under Python 3.12), uv accepts a --python-version parameter, enabling you to generate, e.g., Python 3.7-compatible resolutions even when running under newer versions.

This is great to see though!

I can understand it being a flag on these lower level, directly invoked dependency resolution operations.

While you aren't onto the higher level operations yet, I think it'd be useful to see if there is any cross-ecosystem learning we can do for my MSRV RFC: https://github.com/rust-lang/rfcs/pull/3537

How are you handling pre-releases in you resolution? Unsure how much of that is specified in PEPs. Its something that Cargo is weak in today but we're slowly improving.

In the early days of Rust there was a debate about whether to support "green threads" and in doing that require runtime support. It was actually implemented and included for a time but it creates problems when trying to do library or embedded code. At the time Go for example chose to go that route, and it was both nice (goroutines are nice to write and well supported) and expensive (effectively requires GC etc). I don't remember the details but there is a Rust RFC from when they removed green threads:

https://github.com/rust-lang/rfcs/blob/0806be4f282144cfcd55b...

I did some more digging. By RFC 899, I believe Alex Crichton meant PR 899 in this repo:

https://github.com/rust-lang/rfcs/pull/899

Still, no real discussion of why unbuffered stderr.

Rust has had a stable SIMD vector API[1] for a long time. But, it's architecture specific. The portable API[2] isn't stable yet, but you probably can't use the portable API for some of the more exotic uses of SIMD anyway. Indeed, that's true in .NET's case too[3].

Rust does all this SIMD too. It just isn't in the standard library. But the regex crate does it. Indeed, this is where .NET got its SIMD approach for multiple substring search from in the first place[4]. ;-)

You're right that Rust's standard library is conservatively vectorized though[5]. The main thing blocking this isn't the lack of SIMD availability. It's more about how the standard library is internally structured, and the fact that things like substring search are not actually defined in `std` directly, but rather, in `core`. There are plans to fix this[6].

[1]: https://doc.rust-lang.org/std/arch/index.html

[2]: https://doc.rust-lang.org/std/simd/index.html

[3]: https://github.com/dotnet/runtime/blob/72fae0073b35a404f03c3...

[4]: https://github.com/dotnet/runtime/pull/88394#issuecomment-16...

[5]: https://github.com/BurntSushi/memchr#why-is-the-standard-lib...

[6]: https://github.com/rust-lang/rfcs/pull/3469