fix: sort file names with leading zeros, prefixes, and numeric chunks correctly

[AlexAndBear]

Description

Improved file and folder name sorting to match Windows Explorer behavior, with proper handling of leading zeros, numeric prefixes, and complex naming patterns.

Problems Fixed (compared to main branch)

Main Branch Issues:

1. Incorrect leading zero handling: Files like 0, 00, 000 were sorted inconsistently instead of by zero-count
2. Missing prefix recognition: report and report final were not sorted as prefix/suffix pairs
3. Poor numeric chunk comparison: img2, img02, img10 didn't sort correctly (should be img2 < img02 < img10)
4. No basename/extension separation: Files were compared as whole strings instead of separating base name from extension
5. Insufficient suffix handling: Files with spaces, underscores, hyphens, or parentheses after a common root weren't grouped properly

Examples of incorrect sorting in main:

• 00, 000, 0 → now correctly: 0, 00, 000
• report final, report → now correctly: report, report final
• img10, img2 → now correctly: img2, img10
• Project Archive, Project → now correctly: Project, Project Archive

What Changed

New Sorting Features

1. Correct Natural Sorting

• Chunk-based comparison: numeric chunks by value, text chunks by collator
• Leading zero handling: 0 < 00 < 000 (longer zeros sort later when values equal)
• Proper transitivity: deterministic sort order regardless of input order
• Implementation: compareStringChunks() (sortInternals.ts:334-404)

2. File Prefix/Suffix Detection

• Split basename from extension for comparison
• Recognize prefix relationships: report comes before report final
• Group files with common roots and different suffixes
• Implementation: compareFileNamesByBasePrefix() (sortInternals.ts:172-229)

3. Performance Optimization for Non-Name Sorting

• Schwartzian Transform: Precompute sort keys once per item (sortInternals.ts:53-77)
• Eliminates redundant regex/string operations from comparison loop
• Significant speedup for size, date, and nested property sorting
• Trade-off: Name sorting is slower due to additional correctness logic

4. Case/Accent Tiebreaking

• Added finalTiebreakerCollator with sensitivity: 'variant' (sortInternals.ts:36-39)
• True case-sensitive final tiebreak: Apfel < Äpfel, File.txt < file.txt

5. Comprehensive Test Coverage

• Full array validation with .toEqual() instead of partial .filter().toMatchObject()
• Now validates all 119 items in correct order for both asc/desc
• Added test for nested property sorting (useSort.spec.ts:634-675)
• Tests cover: leading zeros, prefixes, numeric chunks, umlauts, case sensitivity

File Structure

• New: sortInternals.ts - Core sorting logic (405 lines)

  • sortItemsByField - Main entry point with Schwartzian transform
  • compareStringChunks - Unified chunk-based comparison
  • compareFileNamesByBasePrefix - File-specific prefix/extension logic
  • Helper functions for name/extension splitting, prefix detection

• Updated: useSort.ts - Delegates to sortInternals.sortItemsByField()

Performance Benchmarks

Comparison with main branch (20,000 items):

Sort Type	Main Branch	This PR	Change	Notes
Name	29.19ms	61.68ms	-53% slower ⬇️	More correctness logic (leading zeros, prefixes, transitivity)
Size	52.97ms	12.36ms	+77% faster ⬆️	Schwartzian transform optimization
Nested Property	48.36ms	14.76ms	+69% faster ⬆️	Schwartzian transform optimization
Date (function)	59.06ms	10.96ms	+81% faster ⬆️	Schwartzian transform optimization

Trade-offs:

• ✅ Non-name sorting (size, date, nested properties): +69% to +81% faster
• ⚠️ Name sorting: -53% slower but more correct (leading zeros, prefixes, case-sensitivity)
• 💡 Real-world impact: Most UIs have 100-1000 items where name sorting takes 2-3ms (still very fast)

Testing

Unit Tests - All 5 tests pass:

• ✅ Nested property paths (spaceQuota.total)
• ✅ Leading zeros (0, 00, 000 sorted by count)
• ✅ Prefix/suffix pairs (report, report final)
• ✅ Case/accent handling (Äpfel/Apfel, Straße/Strasse)
• ✅ Full sort order validation (119 items asc/desc)

Quality Checks:

• ✅ pnpm check:types passes
• ✅ pnpm format:check passes
• ✅ pnpm test:unit --run passes

Related Issue

Fixes #2993

Types of changes

• Enhancement (better sorting algorithm, performance optimization)
• Tests (expanded coverage, proper validation)