vismaytiwari commented on issue #15769: URL: https://github.com/apache/lucene/issues/15769#issuecomment-4991938756
@hossman thanks for the detailed write-up and the test patch — made this easy to pin down. Opened #16402. I went with reverting to positionIncrement 0 on the tokens after the first (plain synonyms) rather than the PositionLengthAttribute route, since as you found, posLen > 1 sends QueryBuilder down the graph path and testAncestorQuery still fails. The one wrinkle is assertAnalyzesTo can't validate the overlapping-prefix offsets as a graph, so testTokenizerViaAnalyzerOutput passes graphOffsetsAreCorrect=false for those inputs — which lines up with the tokenizer already being @IgnoreRandomChains and checkRandomData running with offsetsAreCorrect=false. Used your ancestor/descendant tests (with exact hit counts) as the regression coverage. One thing I kept out of scope: #12875 made the same change to ReversePathHierarchyTokenizer, so that one likely has the same problem — happy to do it as a follow-up. -- 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]
