Weird error with inference when a function doesn’t compile

fn foo(_: &[String]) -> Vec<std::std::string::String> {
    Vec::new()
}

fn main() {
    let x = foo(&Vec::new());
    foo(&x);
}

Here, the foo function is obviously wrong – std::std::string::String. However, it looks like rustc tried to typecheck the rest of the program and now thinks that x should have the type [String] – because of its use in foo. I guess the error message would be more explicit with x.as_slice() and that it has something to do with the auto deref coercion?

It results in two messages: the first one, correct, about the std::std thing. The second, about [String], which is really misleading.

1. The errors (on stable):

      Compiling playground v0.0.1 (file:///playground)
   error[E0433]: failed to resolve. Could not find `std` in `std`
    --> src/main.rs:1:42
     |
   1 | fn foo(_: &[std::string::String]) -> Vec<std::std::string::String> {
     |                                          ^^^^^^^^^^^^^^^^^^^^^^^^ Could not find `std` in `std`
   
   error[E0277]: the trait bound `[std::string::String]: std::marker::Sized` is not satisfied
    --> src/main.rs:6:9
     |
   6 |     let x = foo(&Vec::new());
     |         ^ `[std::string::String]` does not have a constant size known at compile-time
     |
     = help: the trait `std::marker::Sized` is not implemented for `[std::string::String]`
     = note: all local variables must have a statically known size
   
   error: aborting due to 2 previous errors
   
   error: Could not compile `playground`.
   
   To learn more, run the command again with --verbose.
   

   Dimitri Sabadie at 2017-09-19 11:29:22

2. Also, feel free to edit/rename the issue – I’m bad at naming that kind of issue, sorry for that.

   Dimitri Sabadie at 2017-09-19 11:30:11

3. I don't see any issue with the error...

   Guillaume Gomez at 2017-09-19 16:02:59

4. @GuillaumeGomez it sees x as a [String], while it’s a Vec<String>.

   Dimitri Sabadie at 2017-09-19 18:54:36

5. Indeed. An unneeded coercion on the second error. Quite strange...

   Guillaume Gomez at 2017-09-20 11:20:09

6. The second error that had weird types no longer appears on 1.52.1 stable, 1.53.0-beta7 beta, or 1.54.0 nightly (2021-06-08)

   Playground link

   asquared31415 at 2021-06-09 14:39:12