impl_arr_new_type makes it slightly awkward to write downstream code with fuzzing feature

[sanket1729]

#[cfg_attr(not(fuzzing), derive(Eq, PartialEq, Hash, Ord, PartialOrd))]

Don't know nice solutions for this other than creating another macro, but wanted to point this out.

[apoelstra]

What is the problem exactly? That when fuzzing is on, these derives conflict with your own?

[sanket1729]

Yep, that is the issue. Alibit a minor one.

[Kixunil]

Where's the macro? Can't find it in the codebase.

[tcharding]

You are refering to the impl_array_newtype macro in secp256k1-sys/src/macros.rs, right?

I don't understand, the macro impl_array_new_type does nothing with those traits. They are on the type declarations and then manual impls for not(fuzzing).

Can you explain the problem further please?

[Kixunil]

Oh, wrong repository! 🤦‍♂️

Hmm, why is that macro even external? It seems completely unrelated to secp256k1!

Anyway we can drop those impls since 1.48 as those can be done with derives now. Does it resolve the issue?