Fix: Numeric Comparison Precision Loss in JsonPathFilterEvaluationEngine [From: Deepanjan-Fork Repo]

[DeepanjanKowar]

Problem

Numeric filter comparisons (less than, greater than, etc.) were using floatValue() for comparisons, which caused precision loss for large integers. This resulted in incorrect filter evaluation results.

Example of the Bug

// Before fix (using floatValue):
26011303 < 26011304  // Returns FALSE (incorrect!)
// Both values convert to same float: 2.6011304E7
     
// After fix (using doubleValue):
26011303 < 26011304  // Returns TRUE (correct!)
// Values preserved: 2.6011303E7 vs 2.6011304E7

Root Cause

Float has only ~7 significant decimal digits of precision, while double has ~15-17 significant digits.

Solution

Changed all numeric comparison methods in JsonPathFilterEvaluationEngine from using floatValue() to doubleValue():
- lessThan()
- lessThanEquals()
- greaterThan()
- greaterThanEquals()
- between()

Testing

Added comprehensive test coverage for large integer comparisons:
- testLessThanFilterWithLargeIntegers - Tests < operator with large integers
- testLessEqualFilterWithLargeIntegers - Tests <= operator with large integers
- testGreaterThanFilterWithLargeIntegers - Tests > operator with large integers
- testGreaterEqualFilterWithLargeIntegers - Tests >= operator with large integers
- testBetweenFilterWithLargeIntegers - Tests between filter with large integers

All existing tests (90 total) continue to pass, ensuring no regressions.

[DeepanjanKowar]

Duplicate of:
#9