Reexports from other crate are listen in docs even when hidden by other types

[DomantasJ]

Given a two crate set up:

• Crate a:

  [package]
  name = "a"
  version = "0.1.0"
  edition = "2018"

  /// I am `a::Foo`
  pub struct Foo;

• Crate b:

  [package]
  name = "b"
  version = "0.1.0"
  edition = "2018"
  
  [dependencies]
  a = { path = "../a" }

  pub use a::*;
  
  /// I am `b::Foo`
  pub struct Foo;

Documentation for crate b lists two Foo types, with the a::Foo one being inaccessible because it is actually hidden by b::Foo:

If you change a to a module instead of a separate crate, then only b::Foo type is listed in docs.

mod a {
    /// I am `b::a::Foo`
    pub struct Foo;
}

pub use self::a::*;

/// I am `b::Foo`
pub struct Foo;

rustc --version --verbose:

rustc 1.42.0 (b8cedc004 2020-03-09)
binary: rustc
commit-hash: b8cedc00407a4c56a3bda1ed605c6fc166655447
commit-date: 2020-03-09
host: x86_64-pc-windows-msvc
release: 1.42.0
LLVM version: 9.0

Real world example of this: diesel::r2d2 lists a struct Error which is defined by diesel, and an enum Error which is glob-reexported from r2d2.

[fmease]

Seems to have been fixed. As for regression tests, we have src/test/rustdoc/issue-99221-multiple-structs-w-same-name.rs (and similar ones like tests/rustdoc/multiple-mods-w-same-name-99734.rs). So it's likely that PR #99344 fixed this issue (#99738 close second).

Closing as fixed.