Aria Desires

Delete is looking to be the better option here... A replacement is tantamount to a rewrite...

This issue is partially ill-defined in scope -- when it's time to tackle it in earnest, negotiate what the boundary of it is.

Note that the original semantic values -- (tagged/untagged) union variants -- are not eligible for this. they are *truly* purely semantic.

NB the annoying/big thing to do here is implement "kdl type => c format specifier" translation. Right now we just `%d` everything! Also picking the appropriate assert function (ideally a...

A quick TL;DR of the RFC: The `repr(C)` layout is the *most* obvious lowering to a `struct` containing a c-style `enum` and a `union`. The `enum` says which case of...

For OptionU128 i believe you want `Busted(Check)` which is the case that involves us printing out bytes. I should make the diagnostic better indicate which phase failed.

Hmm yeah not sure what's happening there, it's like the process aborts or something -- you can use `result.run = Skip` to make them not run at all, for the...

EmptyEtruct is definitely another test that's expected to make compilers sad... The raw logs are interesting here; > `cd "/home/runner/work/rustc_codegen_cranelift/rustc_codegen_cranelift/./build/abi-cafe" && CARGO_ENCODED_RUSTDOCFLAGS="" CARGO_ENCODED_RUSTFLAGS="" RUSTC="/home/runner/.rustup/toolchains/nightly-2024-07-09-x86_64-unknown-linux-gnu/bin/rustc" RUSTDOC="/home/runner/.rustup/toolchains/nightly-2024-07-09-x86_64-unknown-linux-gnu/bin/rustdoc" "/home/runner/.rustup/toolchains/nightly-2024-07-09-x86_64-unknown-linux-gnu/bin/cargo" "run" "--manifest-path" "/home/runner/work/rustc_codegen_cranelift/rustc_codegen_cranelift/./build/abi-cafe/Cargo.toml" "--target-dir"...

Assuming this is in fact a linux error code that would suggest it's SIGILL...

I do not, but we do a lot of tokio async and running FFI code in-process so there's a lot of places for weird panic boundaries.