kamthorn opened a new pull request, #16770: URL: https://github.com/apache/lucene/pull/16770
The Thai work in 10.6 (#16717, #16718, #16720, #16722, #16727) is mostly additive, but it also changes what `ThaiAnalyzer`, its default stopword list and `ThaiTokenizer` produce for some text, so an index built with 10.5 can stop matching until it is reindexed. Downstream projects had to find this out from the diffs (see elastic/elasticsearch#160854, and the questions on #16722). This adds a "Migration from Lucene 10.5 to Lucene 10.6" section to `MIGRATE.md` that: - lists the three changes (the `ThaiAnalyzer` chain, the default stopwords, the `ThaiTokenizer` safe-end rule), - shows example output for 10.5 and 10.6, - explains how to rebuild the 10.5 analyzer and tokenizer behavior. The example outputs were produced with a 10.6 build, comparing `new ThaiAnalyzer()` with an analyzer built from the 10.5 components and the 10.5 `stopwords.txt`; the stopword counts (115 and 87) come from the two versions of that file. The code snippet compiles against 10.6. This is a documentation-only change against `main`; the same section applies to `branch_10x`, where 10.6 is released from. 🤖 Generated with [Claude Code](https://claude.com/claude-code) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
