Changelog

Version 1.1.0

Improved regular-expression compilation by using dense compiler storage, reusing known capacities, and moving generated program segments instead of repeatedly copying them.
Improved matching performance with cached category masks for ASCII characters, reduced counter comparisons, and a more efficient priority-preserving work stack.
Reduced pattern parsing overhead by validating UTF input and enforcing the configured character limit during decoding instead of performing a separate pre-scan.
Reduced hot-path overhead for internal checks and bytecode reads while retaining safety validation and debug diagnostics.
Added focused regression coverage for parser boundaries, category lookup, counter metadata, execution-state comparison, work priority and atomic pruning, program appending, and compiler traversal.
Preserved the public API, regular-expression syntax, capture behavior, generated bytecode semantics, and existing resource limits.