It's just an SDK (for a bunch of different mcus) that has its own build tools, HAL, and drivers.
I still fail to see where does Zig make a difference vs. Rust. What's the usefulness of the project here...
A much better source on Zig vs Rust would be Alex Kladov [0], one of the authors of TigerBeetle [1], which is one of the best Zig code bases I have seen.
I just wonder what are the advantages of using a language that is not memory-safe and it's not even stable yet...
Edit: If there's any technical reason. Of course, being a hobby project, the author is free to pick whatever he feels is most ergonomic/he likes the most.
An expert Rust programer probably wouldn't have the same friction points I experienced.
Two of the main advantages of the Rust borrow checker is preventing use after free and iterator invalidation.
Zig's deferred free helps with the first, and hardware FIFOS, doorbells etc often caused me to have a non significant amount of unsafe code.
For me, the array safety in Zig removes most of the C foot guns, and the Rust projects decision to error on the constrained side of the static analysis dichotomy was getting in the way.
It isn't even a case of one being 'better' for me, the tradeoffs just made Zig better for this use case for me.