Review request: unranked Polaris catalog and Ossie-converter evidence

We published an unranked, source-pinned catalog-community evidence set using Polaris 1.7.0 and request maintainer review of configuration and interpretation. It covers stock Spark/Flink/Trino/DuckDB correctness, scoped restart/restore/migration behavior, and a pinned Apache Ossie Polaris-converter TPC-DS loss report. Please correct any deployment-scope or behavioral claims; corrections are retained against the immutable evidence hash.\n\nPolaris packet: https://github.com/querygraph/catalog-bench/blob/56ffdb0/docs/MAINTAINER-REVIEW-PACKETS.md#apache-polaris\nQ3 report/index: https://github.com/querygraph/catalog-bench/blob/56ffdb0/docs/COMMUNITY-REPORT-2026-Q3.md\nConverter report: https://github.com/querygraph/querygraph/blob/f0e4afd/ossie/polaris-tpcds-report.json\n\nThis is a correctness/interchange review, not a performance comparison.

Start with the Apache Polaris section of docs/MAINTAINER-REVIEW-PACKETS.md and the Q3 report/index in docs/COMMUNITY-REPORT-2026-Q3.md. Review the stated Spark, Flink, Trino, DuckDB, restart/restore/migration, and Ossie-converter claims against the pinned evidence, including ossie/polaris-tpcds-report.json. Done means deployment-scope or behavioral corrections are identified and retained against the immutable evidence hash.

Written by the indexing model from the issue text.