Generating C++ to Rust binding so that the Wayland C++ objects can publish events through Rust to the clients, but not yet fully connecting them

[mattkae]

What's new?

• Generating src/middleware.rs, which contains Rust structs that ferry data from Rust to C++
• Generating src/ffi.rs, which contains:
  1. FFI definitions for the middleware structs that sanitize the data going across the language boundary (C++ -> Rust)
  2. FFI definitions for the C++ implementation classes (Rust -> C++)
• Adding an associate method to each Wayland interface's C++ class. This class will "associate" the Box<InterfaceExt> struct with the C++ class so that the C++ class can call methods on its underlying Rust implementation
• Introduced include/ffi_fwd.h to get around a circular dependency issue where the generated C++ headers want to include the Rust headers, but the Rust headers want to include the generated C++ headers. This came about because the <Interface>Impl class now hold an <Interface>Ext box.

How to test

1. Run cargo build from src/wayland/wayland_rs
2. See src/middleware.rs
3. See src/ffi.rs
4. See src/dispatch.rs

Checklist

• Tests added and pass
• Adequate documentation added
• (optional) Added Screenshots or videos

[Copilot]

Pull request overview

This PR restructures the wayland_rs Rust/C++ CXX bridge so generated Wayland interface C++ classes can be associated with Rust “extension” structs, enabling a future path for emitting Wayland events from C++ via Rust to clients.

Changes:

• Switch the CXX bridge entrypoint from src/lib.rs to generated src/ffi.rs and update the demo to include ffi.rs.h.
• Add build-script generation for middleware.rs (Rust extension structs) and ffi.rs (CXX bridge definitions), plus an associate() method on each generated C++ interface.
• Refactor protocol parsing and code generation helpers; add ffi_fwd.h generation to avoid circular includes.

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
src/wayland/wayland_rs/src/lib.rs	Replaces inline module declarations with include!() of generated dispatch.rs, protocols.rs, middleware.rs, and ffi.rs.
src/wayland/wayland_rs/example/main.cpp	Updates the demo to include the new generated CXX header ffi.rs.h.
src/wayland/wayland_rs/build_script/protocol_parser.rs	Sorts protocol XML inputs for deterministic generation and simplifies iteration.
src/wayland/wayland_rs/build_script/protocol_middleware_generation.rs	Adds generator for Rust “Ext” wrappers that forward event calls to wayland_server protocol objects.
src/wayland/wayland_rs/build_script/main.rs	Major generation pipeline rework: middleware + ffi generation, ffi_fwd.h, new dispatch storage model (Arc<Mutex<...>>), and CXX bridge moved to src/ffi.rs.
src/wayland/wayland_rs/build_script/helpers.rs	Extracts shared helpers for writing generated Rust/C++ output and naming helpers.
src/wayland/wayland_rs/build_script/ffi_generation.rs	Generates the unified #[cxx::bridge] mod ffi with Rust->C++ and C++->Rust declarations.
src/wayland/wayland_rs/build_script/cpp_builder.rs	Updates C++ signature generation (return-by-value for unique_ptr, introduces rust::Box<T>, sanitizes identifiers) and pins C++ method receivers in Rust bindings.
src/wayland/wayland_rs/CMakeLists.txt	Points CMake at src/ffi.rs as the bridge source and adjusts include directories for the demo target.
src/wayland/wayland_rs/.gitignore	Updates ignored generated Rust files (ffi.rs, middleware.rs) and removes old ffi_cpp.rs.

[robert-ancell]

I've read through this; there's a lot going on and I can only follow parts of it. I can't see any issues but it needs a second opinion / someone more knowledgeable in Rust to have a look too.

[RAOF]

The actual code generation code is fine.

I'm still digesting the code that it's generating; it seems at this point that we probably need more documentation about the high-level theory of what's happening here, because it's very much non-obvious.

[RAOF]

These are unused, right?

Incidentally, if there's much more case manipulation needed, heck might be a good choice.

[mattkae]

dash_to_snake_ident is still used

[mattkae]

The actual code generation code is fine.

I'm still digesting the code that it's generating; it seems at this point that we probably need more documentation about the high-level theory of what's happening here, because it's very much non-obvious.

I was going to leave this as a task for later where I write a big diagram/markdown of what is getting generated and why. I am waiting for "later" because things feel prone to change now, and I don't want to keep rewriting it

[mattkae]

@RAOF I update to use pin_mut, UniquePtr, and with comments to make things easier to follow :)

[github-actions]

TICS Quality Gate

✔️ Passed

mir

Coding Standards: ✔️ Passed

✔️ Condition “No new Coding Standard Violations for level 1, 2, 3 with respect to Previous analysis” passed.

See the results in the TICS Viewer

The following files have been checked for this project

• src/wayland/wayland_rs/build_script/cpp_builder.rs
• src/wayland/wayland_rs/build_script/ffi_generation.rs
• src/wayland/wayland_rs/build_script/helpers.rs
• src/wayland/wayland_rs/build_script/main.rs
• src/wayland/wayland_rs/build_script/protocol_middleware_generation.rs
• src/wayland/wayland_rs/build_script/protocol_parser.rs
• src/wayland/wayland_rs/example/main.cpp
• src/wayland/wayland_rs/src/lib.rs
• src/wayland/wayland_rs/src/wayland_server.rs

TICS / TICS / Run TICS analysis

[RAOF]

I need GitHub to make it more clear when I haven't actually submitted the review!

This is somewhat weird code, but the weirdness is now documented, and once we've got the whole thing implemented we can see if anything can be simplified.