model-writtenrun 7726 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Peter Eisentraut opened #1 by explaining that SQL cursor sensitivity concerns changes within the cursor's own transaction and that PostgreSQL's documentation incorrectly equated FOR UPDATE with sensitive cursors.
- David G. Johnston described similar confusion from a PL/pgSQL user in #2, and Peter clarified in #3 that the default behavior is implementation-dependent and PostgreSQL's MVCC naturally supports insensitive cursors.
- Peter posted #4 to correct the documentation and feature claims, add ASENSITIVE as an explicit selection of the default behavior, and test insensitivity using an independent INSERT; implementing sensitive cursors or changing WHERE CURRENT OF visibility remained outside the patch.
- David challenged the addition of ASENSITIVE and proposed a different documentation approach in #5; Peter defended the terminology correction and compatibility keyword in #6, the final thread message, without posting a revised patch.
- Peter subsequently committed #4 as dd13ad9d39 with light flag-allocation edits, completing the submitted work; David's unadopted design suggestions did not establish qualifying patch review under the requested criteria.