Anthippi commented on PR #10557:
URL: https://github.com/apache/seatunnel/pull/10557#issuecomment-4031102446
@DanielCarter-stack @yzeng1618 Thank you again for the incredibly detailed
and constructive review! I have addressed and resolved all the issues you
raised in the latest commits.
Just to give a quick summary of the latest updates to prove the stability
and robustness of these fixes:
1. `Unit Tests Refactoring:` I refactored the unit tests to respect the
Single Responsibility Principle, separating them into `RabbitmqSourceTest.java`
and `RabbitmqSourceReaderTest.java`.
2. `Edge-Case Proof:` These tests explicitly use Mockito to prove that all
the edge-cases discussed are handled safely:
- Issues 1 & 3: Thread-safety of `sourceSplits` and graceful consumer
resource cleanup (no duplicate creations or resource leaks).
- Issues 2 & 5: Strict fail-fast validation for mutually exclusive
configs and immediate exceptions for unconfigured queues (no more silent
message dropping).
3. `E2E Validation (Issue 6):` I initially tried your suggestion to use
separate `Assert` sinks per table. However, while digging into the SeaTunnel
source code, I found that `AssertSinkWriter` relies on a `static
LONG_ACCUMULATOR`. Using multiple Assert sinks in the same job causes
unavoidable race conditions on this shared static counter, making the test
flaky. To achieve the exact per-table validation you requested without hitting
this framework limitation, I configured a single `Assert sink` utilizing the
`tables_configs` array natively inside the `rules` block. It works perfectly
and strictly validates schemas and limits per-table!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
