Use supertrait as the suggested bound for a blanket impl instead of the bounds from a different blanket impl

Code

trait A: B {
    fn a();
}

trait B {
    fn b();
}

trait C {}

impl<T> B for T
where
    T: C,
{
    fn b() {}
}

impl<T> A for T {
    fn a(){
        Self::b();
    }
}

Current output

error[E0277]: the trait bound `T: C` is not satisfied
  --> src/main.rs:18:15
   |
18 | impl<T> A for T {
   |               ^ the trait `C` is not implemented for `T`
   |
note: required for `T` to implement `B`
  --> src/main.rs:11:9
   |
11 | impl<T> B for T
   |         ^     ^
12 | where
13 |     T: C,
   |        - unsatisfied trait bound introduced here
note: required by a bound in `A`
  --> src/main.rs:1:10
   |
1  | trait A: B {
   |          ^ required by this bound in `A`
help: consider restricting type parameter `T`
   |
18 | impl<T: C> A for T {
   |       +++

For more information about this error, try `rustc --explain E0277`.

Desired output

error[E0277]: the trait bound `T: B` is not satisfied
  --> src/main.rs:18:15
   |
18 | impl<T> A for T {
   |               ^ the trait `B` is not implemented for `T`
   |
note: required by a bound in `A`
  --> src/main.rs:1:10
   |
1  | trait A: B {
   |          ^ required by this bound in `A`
help: consider restricting type parameter `T`
   |
18 | impl<T: B> A for T {
   |       +++

For more information about this error, try `rustc --explain E0277`.

Rationale and extra context

When providing a blanket impl for some trait A with a supertrait B, and B also has a blanket impl bounded on some other trait C, rustc will suggest bounding A's blanket impl on C rather than the supertrait B. This is a little confusing and likely the incorrect bound to suggest, as it would cause types that directly implement B to not satisfy the blanket impl for A, like the following example:

...

impl<T: C> A for T {
    fn a(){
        Self::b();
    }
}

struct D;

impl C for D {}

struct E;

impl B for E {
    fn b() {}
}

pub fn main() {
    D::a();
    E::a(); // error: method cannot be called on `E` due to unsatisfied trait bounds
}

This works as expected if T: B is specified as the bound for the blanket impl on A instead.

playground link

Other cases

No response

Anything else?

No response

Start by reproducing the minimal Rust example or the linked Playground case and compare rustc's current E0277 output with the desired diagnostic. Trace the compiler's handling of this trait-bound suggestion, add a regression test for the blanket-impl and supertrait case, and verify that the suggested bound is B rather than C.