Release build failure with `IggyMessageBus::background_tasks_len`

[jpds]

Bug description

This accessor here:

• iggy/core/message_bus/src/lib.rs

  Lines 611 to 623 in 239d7ee

  	/// Number of background-task handles currently retained by the bus.
  	///
  	/// Test-only accessor: lets integration tests pin the
  	/// reap-on-push invariant in `track_background` without exposing
  	/// the underlying `RefCell<Vec<JoinHandle<()>>>` to production
  	/// callers. A long-running bus is expected to keep this number
  	/// bounded under sustained accept traffic; a leak shows up here as
  	/// monotonic growth proportional to total accepts.
  	#[cfg(any(test, debug_assertions))]
  	#[must_use]
  	pub fn background_tasks_len(&self) -> usize {
  	self.background_tasks.borrow().len()
  	}

...is called by this integration test here:

iggy/core/message_bus/tests/installer_panic_cleanup.rs

Lines 179 to 183 in 239d7ee

	assert_eq!(
	bus.background_tasks_len(),
	0,
	"background_tasks vec must be empty post-shutdown",
	);

When I tried to build a release build of the latest iggy master and I received this compile error:

  error[E0599]: no method named `background_tasks_len` found for struct `Rc<IggyMessageBus>` in the current scope
     --> core/message_bus/tests/installer_panic_cleanup.rs:180:13

Deployment

Not applicable

Versions

master

Hardware / environment

Nix build

Sample code

No response

Logs

No response

Iggy server config

No response

Reproduction

No response

[hubcio]

good catch. we intentionally dont run tests with --release because with build step it'll take more than 30 minutes in github CI. so, the fix will be only ad hoc, no actual regression test for CI to catch this issue in future. (i couldn't figure out any fast and viable solution besides dirty regex gate).

See #3242.