Predrag Gruevski

Thanks! I think we'll likely proactively always `cargo package` path dependencies in the "current" arm, to be safe. Otherwise it could be hard to figure out if the local not-yet-published...

I haven't looked at this lint in a while, and I'm not sure I'm convinced anymore that this lint is correct as it currently stands. This witness fails to work...

> I guess this needs more work than I expected. Story of my life in cargo-semver-checks lately 😂

As another person who regularly wrangles a bunch of different Rust versions across several operating systems and targets, the proposed solution and workarounds sound entirely reasonable and adequate to me....

> (PR for Idea Verification) Though given the current test_crates for struct_missing this should work. Yet I feel that it should also cover the cases when the struct contains fields?...

Maybe we just avoid overcomplicating it for the witness hint and just do `use path::to::Struct` instead? In principle someone may have added an module by that name (enum/union/trait are impossible...

> There is another lint namely `constructible_struct_adds_field` which I was trying to work on before. Which would require feild types as well (`let witness = |a: , b: | constructible_struct_adds_field::ExhaustivePlainStruct{foo:...

> Did this just in an attempt to make witness more informative. Well, if it's addition doesn't affect much, does the rest of witness look fine for `constructible_struct_adds_field` lint ?...

What are `foo` and `bar` in that example? If they are field names, can we reuse those field names as arguments in the closure and make the witness shorter?

Yeah, looks great!