Bump mojo-regex from 0.10.0 to 0.11.0

[msaelices]

Bumps mojo-regex from 0.10.0 to 0.11.0.

Changes since 0.10.0

• #140: OnePass NFA engine for $-anchored patterns. phone_validation ^\+?1?[\s.-]?\(?([2-9]\d{2})\)?[\s.-]?([2-9]\d{2})[\s.-]?(\d{4})$ (5x median): 1.76 us → 0.25 us (6.92x faster). Full bench geom 1.09x across 80 benchmarks.
• #139: Single-byte required-literal memchr prefilter for findall. sparse_email_findall ~160x faster (199 us → 1.25 us, was 167x slower than Rust, now within 5%); simple_phone 4.4x; full bench geom 1.32x.
• #138: Two-way unroll of count_consecutive_matches SIMD loops (range_lowercase 7.77x faster, range_alphanumeric 3.1x faster, range_digits 1.4x; via @always_inline helpers)
• #137: --dev and --filter flags for bench_engine, plus CLAUDE.md / BENCHMARKING_GUIDE.md docs (fast dev iteration: full bench ~4 min -> ~8 s)
• #136: Inline match_next dispatchers + pre-allocate MatchList in findall (sub() 8/0 wins 1.28x geom; top: deep_nested_groups 2.95x, complex_group_5_children 1.70x, simple_phone 1.60x)
• #135: Precompute pattern properties, eliminate string ops from match paths (nanpa_search 17.9x faster)
• #133: Use ref instead of copy for ASTNode in NFA match methods (findall 2.6-2.9x faster, free functions 1.6x faster)
• #132: Fix lazy DFA $ anchor bug, apply cache pointer to all free functions
• #127: Eliminate CompiledRegex copy in sub() via cache pointer lookup (33x faster per call)
• #126: Multi-range SIMD scan for [a-zA-Z0-9]+ patterns (9x faster)
• #124: Increase PikeVM MAX_STATES to 512, enabling lazy DFA for NANPA patterns
• #122: Cache DFA dispatch + inline LazyDFA helpers
• #121: Fix UInt8 overflow for patterns with >255 AST nodes
• #119: Skip dead _create_range_matcher indirection
• #118: Fix 3+ way alternation with character classes failing to match

Benchmark summary

• phone_validation was 55x slower than Rust before Update DeciMojo to v0.4.1 #140, now ~15x (OnePass closes the gap on $-anchored optional-group patterns that LazyDFA declined)
• sparse_email_findall was 167x slower than Rust before Regex package #139, now within 5%
• range_lowercase was 33x slower than Rust before Update decimojo to Mojo v25.4 #138, now ~5x
• vs Python: 96% win rate, geo mean ~14x
• vs Rust: 63% win rate, geo mean 1.86x (improved further by Regex package #139 1.32x and Update DeciMojo to v0.4.1 #140 1.09x full-bench geom)

Recipe changes:

• version: 0.10.0 -> 0.11.0
• rev: pinned to dc3e955