> When you want to embed a type, you need its definition, but you don’t actually need the full definition. You just need the size/alignment.
Aren't there ABI cases where e.g.
would be passed in e.g. fp registers whereasstruct foo { float X, Y; }
would not?struct { char[8]; }Yeah this is correct. You don't want to pass these values around "by value" but, you should be able to "embed them" and pass "pointers to them". It's a middle-ground between a completely opaque type which you would also pass around by address, but, with the added benefit that you allocate your own storage for it.
I sort of mentioned this in the blog but this is good clarification.
> if you want to pass a shared_ptr to Zig, you need to pass a pointer to the shared pointer
For lore, I believe this GitHub thread is where I first learned about the how types of the same size/alignment can still have different ABIs :) https://github.com/microsoft/win32metadata/issues/623#issuec...
Yes. For example consider this function to add two 2D points, which accepts and returns all variables entirely in xmm registers: https://gcc.godbolt.org/z/hPGKrh6W4 (surprisingly, gcc generates some fairly odd assembly code here)
It's fixed if you pass -fno-trapping-math. There could be junk in the upper half of the registers that causes a floating-point exception.
Oh, that explains the seemingly useless movs. Thank you! I simply thought that it was a bug, since clang did not emit them.
This idea about communicating size/alignment is actually something we're doing on the port of RediSearch to Rust [0]. We have an "opaque sized type" which is declared on the Rust-side, and has its size & alignment communicated to the C-side via cbindgen. The C-side has no visibility into the fields, but it can still allocate it on the stack.
It's a bit ugly due to cbindgen not supporting const-generic expressions and macro-expansion being nightly-only. It seems like this will be a generally useful mechanism to be able to use values which are not traditionally FFI-safe across FFI boundaries.
[0]: https://github.com/RediSearch/RediSearch/blob/cfd364fa2a47eb...
It’s just both using c abi right?
Yeah, this isn't quite C++ interop on its own. It's C++ interop via C, which is an incredibly pertinent qualifier. Since we go through C, opaque pointers are needed for everything, we can't stack allocate C++ values, we need to write extern C wrappers for everything we want to do (like calling member fns), and we don't get any compile-time type/safety checking, due to the opaque pointers.
Direct C++ interop is doable, by embedding Clang into Zig and using its AST, but this is significantly more work and it needs to be done in the Zig compiler. As a Zig user, going through C is about as good as you can do, probably.
It's a bit more than your typical "interop via C". With a "sized opaque" type you actually can stack allocate C++ values in Zig (and vice versa stack allocate Zig values in C++), i.e.
fn stackExample() void {
var some_cpp_type: c.SomeCppType = undefined; c.some_cpp_type_ctor(&some_cpp_type); defer c.some_cpp_type_dtor(&some_cpp_type); // ... }
Seems like it. And the sizes are all hard-coded, which means you are probably wedded very tightly to a particular C++ compiler.
This has reawakened the nightmares about Objective-C++
How so? Having written Objective-C and C++ separately but never written Objective-C++ before, I don't understand what's the hate it. Some of my favorite Mac apps from a decade ago were written in Objective-C++. I think Chrome still has some parts in Objective-C++.
I don't think anyone who could help it wrote software in Objective-C++ directly.
It was an unholy welding of the two languages that mostly let Objective-C apps adopt some libraries from the broader C++ ecosystem. Plus the occasional cross-platform C++ codebase used it as a thin shim to provide a Cocoa-based UI...
It is still around, even if documentation is only on archives.
Objective C++ was great.
Rename a file to .mm and start using Objective C APIs. Very good interop