"clippy::single_component_path_imports" False positive on swapped namespaces to enable mocking.

Summary

Imagine you have a 3rd part dependency on module module3.
My component depends on an exported member from that namespace, for example, module3::Foo.
I want to test my component's internal logic without instantiating module3::Foo.
To do so, I mock module3::Foo with mockall. I will now use module3 for actual builds, while use mock::module3 as module3 for test builds.
Lint single_component_path_imports is going to complain about my first use.

Lint Name

single_component_path_imports

Reproducer

I tried this code:

#[cfg(not(test))]
use module3;

#[cfg(test)]
use mock::module3 as module3;

fn computation(foo: module3::Foo) -> u32 { .. }

fn run_computation() {
  let foo =  module3::Foo::new();
  let c = computation(foo);
}

#[cfg(test)]
fn run_test_computation(foo: module::Foo) {
  let foo =  module3::Foo::new();
  let c = computation(foo);
  assert_eq!(c, 100);
}

I saw this happen:

warning: this import is redundant
  --> main.rs
   |
   | use module3;
   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ help: remove it entirely
   |
   = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#single_component_path_imports

I do not know if it is reasonable to expect the lint to deduce that the namespaces are swapped. However it definitely produces a warning where I did not expect it.

Version

rustc 1.74.0 (79e9716c9 2023-11-13)
binary: rustc
commit-hash: 79e9716c980570bfd1f666e3b16ac583f0168962
commit-date: 2023-11-13
host: aarch64-apple-darwin
release: 1.74.0
LLVM version: 17.0.4

Additional Labels

No response

Start with the single_component_path_imports lint implementation and reproduce the warning using the Rust snippet in the issue. Investigate how the lint handles cfg-gated imports with swapped namespaces. Done means the reported false positive has an agreed behavior and regression coverage for the reproducer.