contrueCT commented on PR #2994: URL: https://github.com/apache/hugegraph/pull/2994#issuecomment-4634424918
Thanks for your patience. The hstore CI failure turned out to expose an existing latent bug in the hstore range-index query path rather than a flaky test: the upper layer assumed range-index scan results were globally ordered and that the returned page state could be reused as a HugeGraph range key, but hstore's multi-node/tablet scan cursor doesn't guarantee global ordering and its page state is an internal storage cursor. That can make range queries with limit/paging return the wrong subset. I cleaned up the temporary diagnostic commits and kept the fix scoped to hstore range indexes: paged or bounded limit/offset queries are sorted/sliced in the index layer, while unbounded range queries continue to use the original streaming batch path. The cleaned commit has been force-pushed as ebc31c83, and the hstore workflow passed on run 27031748295. -- 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]
