fix flaky tests because of ZTestLogger

[keisar]

One test was fixed by moving away from ZTestLogger, as ZTestLogger can cause duplicated log lines.
also identified why another test is flaky and added a TODO comment and TestAspect.flaky, until we find a better way to test it

[Copilot]

Pull Request Overview

Purpose: Replace ZTestLogger with a custom TestLogger to avoid duplicated log lines causing test flakiness, and explicitly mark one still-flaky test while adding a TODO.

• Introduces a simple concurrent TestLogger for capturing log messages.
• Replaces ZTestLogger usage in two tests with the new TestLogger.
• Marks a previously flaky test with TestAspect.flaky and adds a clarifying TODO comment.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Pull Request Overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 1 comment.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Pull Request Overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 1 comment.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

apply is declared to return Unit, but the body evaluates to Boolean (ConcurrentLinkedQueue.add returns Boolean), causing a type mismatch in Scala 3. Discard the Boolean result explicitly, e.g.: override def apply(...): Unit = { messages.add(message()); () }.

Suggested change

	messages.add(message())
	{ messages.add(message()); () }

[keisar]

incorrect, check again

[Copilot]

Pull Request Overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 3 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Counting log messages directly from a ConcurrentLinkedQueue immediately after readState can race with late asynchronous log emissions, potentially reintroducing flakiness (only the first test is marked flaky). Consider synchronizing by awaiting completion of pending fibers or introducing a short drain/settle step (e.g., ZIO.sleep(...)) before snapshotting messages, or capturing log entries with a timestamp and filtering only those emitted before readState completes.

Suggested change

	// Allow time for any late log emissions to settle before snapshotting messages
	_ <- ZIO.sleep(100.millis)

[keisar]

this is a very nice point, since I tested with nonFlaky I think we're good for now

[Copilot]

Counting log messages directly from a ConcurrentLinkedQueue immediately after readState can race with late asynchronous log emissions, potentially reintroducing flakiness (only the first test is marked flaky). Consider synchronizing by awaiting completion of pending fibers or introducing a short drain/settle step (e.g., ZIO.sleep(...)) before snapshotting messages, or capturing log entries with a timestamp and filtering only those emitted before readState completes.

[Copilot]

The custom TestLogger discards useful metadata (logLevel, cause, spans, annotations), which can hinder future debugging or more granular assertions. Consider storing a structured case class (e.g., with level and cause) instead of only message(), or at least prefix messages with logLevel to preserve minimal context.

Suggested change

	class TestLogger extends ZLogger[String, Unit] {
	val messages: ConcurrentLinkedQueue[String] = new ConcurrentLinkedQueue()
	override def apply(trace: Trace, fiberId: FiberId, logLevel: LogLevel, message: () => String, cause: Cause[Any], context: FiberRefs, spans: List[LogSpan], annotations: Map[String, String]): Unit =
	messages.add(message())
	def getMessages: List[String] = messages.asScala.toList
	case class TestLogEntry(
	trace: Trace,
	fiberId: FiberId,
	logLevel: LogLevel,
	message: String,
	cause: Cause[Any],
	context: FiberRefs,
	spans: List[LogSpan],
	annotations: Map[String, String]
	)
	class TestLogger extends ZLogger[String, Unit] {
	val entries: ConcurrentLinkedQueue[TestLogEntry] = new ConcurrentLinkedQueue()
	override def apply(
	trace: Trace,
	fiberId: FiberId,
	logLevel: LogLevel,
	message: () => String,
	cause: Cause[Any],
	context: FiberRefs,
	spans: List[LogSpan],
	annotations: Map[String, String]
	): Unit =
	entries.add(
	TestLogEntry(
	trace,
	fiberId,
	logLevel,
	message(),
	cause,
	context,
	spans,
	annotations
	)
	)
	def getEntries: List[TestLogEntry] = entries.asScala.toList
	def getMessages: List[String] = entries.asScala.toList.map(_.message)