fix(studio): make message provider fail explicitly instead of returning empty stub results

[123123213weqw]

Fixes #684

Problem

MessageProviderStub is wired as the production MessageProvider, but queryMessages returns an empty list and getMessageTrace returns an empty trace. A successful empty response can mislead operators into thinking no message or trace exists, when in fact no real provider is connected.

Fix

Until a real message provider is connected, both operations now fail explicitly with a structured 501 Not Implemented BusinessException (mapped by GlobalExceptionHandler to a clean Result.error(501, ...)), matching the issue's expected behavior. Added MessageProviderStubTest.

Verification

mvn test -Dtest=MessageProviderStubTest on JDK 21: Tests run: 2, Failures: 0, Errors: 0, Skipped: 0.

Diff

2 files changed, +56 / -10.

[lizhimins]

Thanks for the contribution. This is already covered by #685, merged into rocketmq-studio as 02e7bed: MessageProviderStub.queryMessages and getMessageTrace now both throw BusinessException(501, ...) instead of returning empty results, and MessageProviderStubTest asserts the 501 behaviour. #685 was opened first, so it was merged under first-come-first-served review order.

Two details from your version were not carried over, and a small follow-up PR would be welcome if you think they are worth it: per-method messages (Message query ... vs Message trace ...) instead of one shared message, and the class-level Javadoc explaining why empty results are misleading. Please keep the @Slf4j warn logs, though — they are useful when diagnosing a deployment that still runs the stub.