Changes
- docs: Explain
write_buffer_sizein MiB instead of asking for a multiple of 1024 (in52faa84)
Performance improvements:
- perf(core): Use
set_iid_to_termsonPUSHwhen assuming new (inc29f50d)
Dependency updates:
- Update dependencies (in
3e99e50) - deps(core): Bump
jieba-rsfrom0.10to0.11(in6e2b5a1) - deps(server): Bump
randfrom=0.10.2to=0.10.3(in11b721d) - deps(server): Bump
configfrom=0.15.25to=0.15.26(in26e3494)
Bug Fixes
- fix(server): Make experimental flag
NEWcompatible with content >buffer_size(in1c43813) - fix(core): Use byte offsets in CJK tokenizer (in
638d310)- Fix panic on Chinese text (
start byte index … is not a char boundary): usejieba_rs::Token::byte_start(byte offset) instead ofstart(Unicode offset) inTokenizer::tokenize. Also uselinderaToken::byte_startinstead of the removedtoken_startfield.
- Fix panic on Chinese text (
- fix(core): Fix object counter on big-endian platforms and after batch ingest (in
6902b0e) - fix(core): Clear
IIDIncrfrom cache during aFLUSHO(in0e1a2ce)