feat: Performance improvements

[rogervila]

Depends on #60

Refactored Array Comparison Logic

• The array comparison logic was refactored for better handling of empty arrays, reducing unnecessary iterations.

• Optimized array key checks to more efficiently identify differences.

Improved Float Comparison

• Logic for comparing float values was refactored to use epsilon handling, avoiding false positives or negatives when values are mathematically very close but not exactly equal.

• Additional edge case tests were added, especially around comparisons involving PHP_FLOAT_MAX and infinity.

Overall Performance Gains

All but 2 benchmarks show performance improvements, with most being substantial:

Improvements (>10% faster):

• benchMixedDataTypes: -27.75% (best improvement)
• benchWideArrayComparison: -18.60%
• benchMediumArrayStrictComparison: -18.79%
• benchMediumArrayLooseComparison: -16.86%
• benchSmallArrayStrictComparison: -14.30%
• benchLargeArrayStrictComparison: -14.22%
• benchLargeArrayLooseComparison: -14.74%
• benchRealisticUserDataset: -14.05%
• benchIdenticalArrays: -14.11%
• benchSmallArrayLooseComparison: -11.38%
• benchArrayWithNumericKeys: -13.37%
• benchComplexNestedConfiguration: -12.73%

Minor Regressions (slower):

• benchCompletelyDifferentArrays: +15.21% slower
• benchEmptyArrayComparison: +22.41% slower

However, these two regressions are on edge cases (completely different arrays and empty arrays), which are less common in real-world usage.

Key Observations:

Real-world scenarios improved significantly: The realistic user dataset (-14.05%) and complex nested configuration (-12.73%) benchmarks show strong improvements in practical use cases.

Consistent improvements across sizes: Small (-14.30%), medium (-18.79%), and large (-14.22%) arrays all showed substantial improvements.

Wide arrays saw major gains: The -18.60% improvement on wide arrays (100 keys) is particularly impressive.

Variance improvements: Many benchmarks also show reduced standard deviation (rstdev), indicating more consistent performance:

• benchComplexNestedConfiguration: -51.63% rstdev
• benchRealisticUserDataset: -31.61% rstdev
• benchMixedDataTypes: -30.59% rstdev

The optimizations are clearly effective across most scenarios, yielding speed improvements of 10-28% in typical use cases. The minor regressions in edge cases are a worthwhile trade-off for the substantial gains elsewhere.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions]

📊 Benchmark Results

+-----------------------+----------------------------------+-----+------+-----+---------------+------------------+-----------------+
| benchmark             | subject                          | set | revs | its | mem_peak      | mode             | rstdev          |
+-----------------------+----------------------------------+-----+------+-----+---------------+------------------+-----------------+
| ArrayCompareBenchmark | benchSmallArrayStrictComparison  |     | 1000 | 10  | 1.982mb 0.00% | 0.476μs -14.30%  | ±2.73% -1.26%   |
| ArrayCompareBenchmark | benchSmallArrayLooseComparison   |     | 1000 | 10  | 1.982mb 0.00% | 0.540μs -11.38%  | ±2.74% +109.71% |
| ArrayCompareBenchmark | benchMediumArrayStrictComparison |     | 1000 | 10  | 1.982mb 0.00% | 2.404μs -18.79%  | ±1.91% +68.71%  |
| ArrayCompareBenchmark | benchMediumArrayLooseComparison  |     | 1000 | 10  | 1.982mb 0.00% | 2.740μs -16.86%  | ±2.29% +11.01%  |
| ArrayCompareBenchmark | benchLargeArrayStrictComparison  |     | 500  | 10  | 1.982mb 0.00% | 5.671μs -14.22%  | ±1.86% -13.54%  |
| ArrayCompareBenchmark | benchLargeArrayLooseComparison   |     | 500  | 10  | 1.982mb 0.00% | 6.078μs -14.74%  | ±1.50% -24.55%  |
| ArrayCompareBenchmark | benchDeeplyNestedArrayComparison |     | 500  | 10  | 1.982mb 0.00% | 2.447μs -9.49%   | ±1.06% -25.56%  |
| ArrayCompareBenchmark | benchWideArrayComparison         |     | 200  | 10  | 1.982mb 0.00% | 77.356μs -18.60% | ±1.30% +45.19%  |
| ArrayCompareBenchmark | benchIdenticalArrays             |     | 1000 | 10  | 1.982mb 0.00% | 0.390μs -14.11%  | ±1.58% -33.49%  |
| ArrayCompareBenchmark | benchCompletelyDifferentArrays   |     | 1000 | 10  | 1.982mb 0.00% | 0.368μs +15.21%  | ±1.21% -12.28%  |
| ArrayCompareBenchmark | benchMixedDataTypes              |     | 500  | 10  | 1.982mb 0.00% | 1.877μs -27.75%  | ±0.85% -30.59%  |
| ArrayCompareBenchmark | benchRealisticUserDataset        |     | 200  | 10  | 1.982mb 0.00% | 90.397μs -14.05% | ±0.82% -31.61%  |
| ArrayCompareBenchmark | benchComplexNestedConfiguration  |     | 100  | 10  | 1.982mb 0.00% | 13.614μs -12.73% | ±1.39% -51.63%  |
| ArrayCompareBenchmark | benchEmptyArrayComparison        |     | 1000 | 10  | 1.982mb 0.00% | 0.333μs +22.41%  | ±2.40% -5.13%   |
| ArrayCompareBenchmark | benchArrayWithNumericKeys        |     | 500  | 10  | 1.982mb 0.00% | 1.622μs -13.37%  | ±1.78% +13.15%  |
+-----------------------+----------------------------------+-----+------+-----+---------------+------------------+-----------------+

Note: Benchmarks compare PR against master branch.
Performance regression threshold: ±2%