Bump mojo-regex to 0.8.0

[msaelices]

Summary

Bump mojo-regex from 0.7.0 to 0.8.0.

Major performance release. Mojo vs Python win rate improved from 53% to 85% (52 wins, 9 losses out of 61 benchmarks).

Highlights

• Nibble-based SIMD matching: Character class matching using two 16-byte nibble tables with native pshufb, 29-350x faster than Python for [a-z]+, \d+, \w+
• is_match O(1) fast path: 2000-19000x faster than Python, 2-6x faster than Rust
• DFA compiler extensions: Non-capturing alternation groups, variable-length branches, deeply nested groups, capturing group flattening
• .* prefix/suffix fast paths: Skip NFA backtracking entirely for .*LITERAL and LITERAL.* patterns
• Parser bug fixes: Escaped characters, dash outside brackets, DFA start state for bounded quantifiers
• Fair benchmark harness: Pre-compiled regex, median timing, auto-calibration across Mojo/Python/Rust

See full changelog: https://github.com/msaelices/mojo-regex/blob/main/CHANGELOG.md