Fix msgSendSuper regression.

[Nirlah]

Regressed in the Zig 0.15 upgrade.

Not sure how the mentioned test passed, can it be specific for macOS 26?
On my machine (macOS 15), calling msgSendSuper fails to compile without this fix.

[Nirlah]

Ok, it fails on the CI environment.
Alongside the header itself, the docs do not mention any change.

Can it be related to something specific to the CI testing environment?

[Nirlah]

I don’t have experience with using nix, but from my understanding it depends on gnustep-libobjc.

Maybe this can help:
https://github.com/gnustep/libobjc2/blob/7a5114deff221c58b19ce2b18245ab13aac33295/objc/runtime.h#L105

[LaDorade]

same issue with native build on my machine 🫡 (Zig version: 0.15.2
but working with nix develop: Zig version 0.15.1
but no clue if this matter

[amitayas]

The correct field is super_class.
The test fails because it is being linked against a very old SDK (MacOSX11.3)

The class field was removed after SDK 12.3, which still has

#if !defined(__cplusplus)  &&  !__OBJC2__
    /* For compatibility with old objc-runtime.h header */
    __unsafe_unretained _Nonnull Class class;
#else
    __unsafe_unretained _Nonnull Class super_class;
#endif

The Objective-C runtime headers(objc/message.h) on my system(macOS Tahoe 26.2) define struct objc_super as

/// Specifies the superclass of an instance. 
struct objc_super {
    /// Specifies an instance of a class.
    __unsafe_unretained _Nonnull id receiver;

    /// Specifies the particular superclass of the instance to message. 
    __unsafe_unretained _Nonnull Class super_class;

    /* super_class is the first class to search */
};
#endif

IMO, changing the field from class to super_class would be the correct, simplest fix.

EDIT:
Seems like __OBJC2__ is not set for 32 bit platforms, so maybe conditionally support both shapes if the goal is to support both architectures?

[mitchellh]

Seems like OBJC2 is not set for 32 bit platforms, so maybe conditionally support both shapes if the goal is to support both architectures?

I don't think there's any 32-bit Apple targets left so not worth it.

I think fixing CI to link newer headers should fix this and we can merge it up. :)