TRE

TRE uniquely supports approximate matching using Levenshtein distance, allowing configurable costs for insertions, deletions, and substitutions, which is rare in free regex libraries and included in the agrep tool.

The algorithm guarantees O(M²N) time complexity for most operations, ensuring consistent matching speed regardless of text length, with modest memory usage (less than 50K during compilation).

Adheres closely to IEEE Std 1003.1-2001, providing correct subpattern matching and standard behavior, which the README notes is uncommon even in non-free implementations.

Fully re-entrant APIs allow safe concurrent use across multiple threads, and it compiles on various platforms including Windows, FreeBSD, and macOS as listed in the README.