chore(sync-service): Fix flakey Publisher tests

[robacourt]

These tests can occasionally fail with:

 Assertion failed, no matching message after 5000ms
     The process mailbox is empty.
     code: assert_receive :publish_finished

The race condition occurs if the TestSubscriber receives the :finish_processing_message message before it receives the publisher call, in which case it swallows the message since handle_info is not implemented on TestSubscriber. A GenServer would log that a message had been received and not handled, but as use GenServer was omitted from the TestSubscriber the message was swallowed silently.

This PR:

• Adds use GenServer to the TestSubscriber as this is best practice (the race condition would have been obvious had it been included in the first place)
• Waits for :message_received from each of the subscribers before sending :finish_processing_message. This eliminates the race condition.
• Renames event to message for consistency

Credit to ChatGPT which worked out what the race condition was. It's solution was terrible, but the hard part was knowing what the race condition was, so that was super helpful.

[codecov]

❌ 1 Tests Failed:

Tests completed	Failed	Passed	Skipped
1785	1	1784	1460

View the full list of 1 ❄️ flaky tests

test/integration.test.ts > HTTP Sync > should correctly select columns for initial sync and updates

Flake rate in main: 22.50% (Passed 31 times, Failed 9 times)

Stack Traces | 5.01s run time

Error: Test timed out in 5000ms.
If this is a long-running test, pass a timeout value as the last argument or configure it globally with "testTimeout".

To view more test analytics, go to the Test Analytics Dashboard
_{📋 Got 3 mins? Take this short survey to help us improve Test Analytics.}

[msfstef]

Very nice, that's what I'm talking about! I wonder if we have any other such "mock" processes that could benefit from implementing GenServer

[robacourt]

Very nice, that's what I'm talking about! I wonder if we have any other such "mock" processes that could benefit from implementing GenServer

Good shout! I just checked and we don't currently. It would be nice if a static analysis tool like credo could ensure things like this. I might look into that at some point.