code relies on type inference rules which are likely to change

[Hossein-Noroozpour]

probably msg_send macro is using type inference rules which will be removed in the next coming rustc.
my current rustc version is rustc 1.19.0-nightly (d015610db 2017-06-05)

complete warning is:

• note: #[warn(resolve_trait_on_defaulted_unit)] on by default
• warning: this was previously accepted by the compiler but is being phased out; it will become a hard error in a future release!
• note: for more information, see issue #39216 Tracking issue for future-incompatibility lint resolve_trait_on_defaulted_unit rust-lang/rust#39216
• note: this error originates in a macro outside of the current crate

[SSheldon]

@Hossein-Noroozpour, can you share a snippet of your code that's triggering this warning?

Is it possible that you have some code that looks like this, with no annotated return type?

msg_send![obj, retain];

If so, I believe it should be fixed by adding a type annotation, like:

let _: () = msg_send![obj, retain];

Let me know if the msg_send! usage that causes this error is in the objc crate, but otherwise I think this is something that users of the crate will need to fix themselves.

[Hossein-Noroozpour]

Thanks, yes, this () was the problem. I think maybe you like to see this: https://stackoverflow.com/questions/44597919/simd-encoding-for-objective-c-runtime/44610154#44610154
the SIMD types have a weird problem in encoding, and unfortunately they are important part of the today graphic system of Mac OS.

[SSheldon]

Great, glad that fixed your problem!

Hmm, yeah, that's really interesting that the runtime just entirely skips simd arguments in encodings... definitely not what I would expect. I'll keep that in mind as I work on verify_message! (And now I'm curious to look into simd more, I haven't investigated it much...) Thanks for letting me know!