Release 3.1.1

Fixed

• Fixed incorrect score calculation for SIMD implementations of Levenshtein and OSA on 32 bit systems

Release 3.1.0

Changed

• split editops_apply/opcodes_apply into *_apply_str and *_apply_vec. This avoids the instantiation of
  std::basic_string for unsupported types.

Release 3.0.5

Fixed

• the editops implementation didn't properly account for some cells in the Levenshtein matrix.
  This could lead both to incorrect results and crashes.

Release 3.0.4

Fixed

• fix tagged version

Release 3.0.3

Fixed

• fix potentially incorrect results of JaroWinkler when using high prefix weights

Release 3.0.2

Fixed

• fix assert leading to compilation failures

Release 3.0.1

Fixed

• fix doxygen warnings

Release 3.0.0

Performance

• add banded implementation of LCS / Indel. This improves the runtime from O((|s1|/64) * |s2|) to O((score_cutoff/64) * |s2|)

Changed

• changed many types in the interface from int64_t to size_t, since they can't be negative.

Fixed

• fix incorrect transposition calculation in simd implementation of Jaro similarity
• use posix_memalign on android

Release 2.2.3

Fixed

• use _mm_malloc/_mm_free on macOS if aligned_alloc is unsupported

Release 2.2.2

Fixed

• fix compilation failure on macOS