Optimize XLIFF import performance - Fix timeout issue (#30)

[CybotTM]

❌ STATUS: TESTED - OPTIMIZATION FAILED

This PR implements database optimizations that made performance WORSE based on real database testing.

Real Performance Results

Comprehensive testing with actual TYPO3 v13 + MySQL database in DDEV:

Small Test (3,000 trans-units)

Version	Time	Throughput	Result
Baseline (main)	15.2 sec	198 units/sec	-
Optimized	12.2 sec	246 units/sec	✅ +24% faster

Medium Test (10,000 trans-units)

Version	Time	Throughput	Result
Baseline (main)	44.1 sec	227 units/sec	-
Optimized	46.6 sec	215 units/sec	❌ -5.7% SLOWER

Large Test (400,000 trans-units, ~100MB)

Version	Time	Throughput	Result
Baseline (main)	30m 1s	222 units/sec	-
Optimized	31m 45s	210 units/sec	❌ -5.8% SLOWER

What Went Wrong

1. TYPO3 Extbase ORM Already Batches - Manual batching added complexity without benefit
2. TYPO3 May Already Cache - Custom cache arrays added memory overhead
3. Transaction Conflicts - May conflict with existing transaction handling
4. Wrong Assumptions - Optimizations based on theory, not profiling

Lessons Learned

• ❌ Never simulate database operations with usleep()
• ❌ Don't claim performance numbers without real testing
• ❌ Don't optimize without profiling the actual bottleneck
• ✅ Always measure first, optimize second
• ✅ Real data beats theory

Recommendation

Close this PR - optimizations provide no meaningful benefit for large files and cause regressions.

Next Steps

1. Revert to main branch (baseline is better)
2. Profile with Xdebug + MySQL slow query log
3. Measure actual query counts during import
4. Find REAL bottleneck with data, not assumptions
5. Optimize ONE thing at a time
6. Test each optimization individually

See RealPerformanceResults.md for complete test data and analysis.

[CybotTM]

Real Testing Complete

I've completed comprehensive real-world testing with actual TYPO3 v13 + MySQL database operations.

Test Environment:

• DDEV local development with MySQL
• TYPO3 v13.4.20 with PHP 8.2
• Real database operations (not simulations)
• Three test scales: 3K, 10K, and 400K trans-units

Results:

• Small files (3K): 24% improvement ✅
• Medium files (10K): 5.7% regression ❌
• Large files (400K): 5.8% regression ❌

The optimizations that seemed beneficial in theory actually hurt performance at scale.

Root Cause:
The TYPO3 Extbase ORM already handles batching and caching internally. Our manual optimizations added overhead without providing benefit, and may have conflicted with existing transaction management.

Recommendation:
Close this PR and revert to the main branch. The baseline code performs better for large files.

Proper Next Steps:

1. Use Xdebug profiling to find the ACTUAL bottleneck
2. Measure real query counts with MySQL slow query log
3. Test optimizations individually with real data
4. Validate each change provides measurable improvement

Complete test data available in: Documentation/TechnicalAnalysis/RealPerformanceResults.md

[CybotTM]

Closing in favor of PR #57 which provides a more comprehensive high-performance solution using DBAL bulk operations.