[fix][test] Fixed Nondeterministic Assertions in KafkaAbstractSinkTest #24877
+68
−20
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #24876
Motivation
The original test performed six assertions using the
expectThrows
method but was based on two incorrect assumptions. First, it implicitly assumed that field validation within IOConfigUtils.java would occur in a consistent order. However, IOConfigUtils iterates over configuration fields using the loop(Field field : getAllFields(clazz))
, andgetAllFields()
calls Class.getDeclaredFields, which returns an array of fields that are not sorted and not in any particular order. Second, the test initialized multiple required fields as null simultaneously, assuming the validation would fail in a specific order. In reality, the nondeterministic field iteration order can cause different missing fields to be detected first, leading to inconsistent exception messages and nondeterministic test failures due to different environments producing the contents in different orders despite the logical contents being the same.Modifications
Instead of reusing and mutating the same map, each validation case now creates it's own copy of the valid base config data with ``validConfig``` prior to mutating it and later asserting. By isolating the mutations for each test, this ensures the behavior will be as expected. To make the code easier to read, a tiny table of cases was added (
record Case(...)
) which is made up of the basic components of the original test cases: how the data should be mutated, the expected exception type, and the expected message. The final sanity check which ensures the valid config passes was preserved. These changes ensure the test passes consistently, even when the fields are checked in a different order.In essence, these changes keep the spirit of the original tests while eliminating failures caused solely by allowed (but previously unexpected) reordering.
Verifying this change
This change is already covered by existing tests, such as
pulsar-io/kafka/src/test/java/org/apache/pulsar/io/kafka/sink/KafkaAbstractSinkTest.java
Does this pull request potentially affect one of the following parts:
If the box was checked, please highlight the changes
Documentation
doc
doc-required
doc-not-needed
doc-complete
Matching PR in forked repository
PR in forked repository: LucasEby#7