Bugfix: Flaky test in IonValueDeserializerTest

[cperkkk]

What is the purpose of this PR?

This pull request addresses 1 flaky test, in the jackson-dataformats-binary project. The test was failing intermittently due to assumptions about the ordering of fields returned by the code under test.

Test Affected

• IonValueDeserializerTest#shouldBeAbleToSerializeAndDeserializeIonValueDataWithIncludeNonNull

Why the test fails

The test compares the exact string representation which can be returned in any order (when using HashMap in the underlying implementation). I switched the test to use LinkedHashMap instead to guarantee ordering.

String data = mapper.writeValueAsString(source);
        assertEquals("{a:1,b:null}", data);

How to reproduce the test failure

Use the TestingResearchIllinois's NonDex tool and run:

mvn edu.illinois:nondex-maven-plugin:2.1.7:nondex 

Before the fix, the test failed intermittently with errors similar to:

tools.jackson.dataformat.ion.ionvalue.IonValueDeserializerTest:org.opentest4j.AssertionFailedError: expected: <{a:1,b:null}> but was: <{b:null,a:1}>

Expected results

The test should pass consistently, regardless of the order of fields in the returned response.

Actual results

Before the fix, the test failed intermittently due to changes in the order of the returned response.

Validation

Test Results

• Tests run with NonDex: Passed consistently across multiple seeds

mvn edu.illinois:nondex-maven-plugin:2.1.7:nondex 

Additional Notes

• The fix ensures the test is robust to non-deterministic behaviors and compatible with future updates to it or related dependencies.

Let me know if further improvements or clarifications are needed!

[cowtowncoder]

Thank you @cperkkk !