- Improved HTML parsing accuracy
- When
nilis supplied for thelanguageparameter, it is now treated as the empty string for rule-matching purposes (previously it would match no rules)
- Optimize memory usage
- All
Srx::Enginemethods except#segmentare now private - ICU regex syntax
\xhhhhis now no longer converted to Ruby regex, as this syntax was not correct; it now must be\x{hhhh} - ICU regex syntax
\0ooois now supported
- Handle HTML void elements correctly
- Initial release