model-writtenrun 7216 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Daniel Gustafsson extracted a standalone event-trigger disabling GUC from stalled login-trigger work in #1, aiming to let administrators repair broken triggers without single-user mode.
- Daniel fixed a sample-configuration omission in #2 and rebased in #4, after which Mikhail Gribkov reported successful tests and manual experiments and requested support for disabling individual event types in #6.
- Daniel added individual event-type selection in #7, Mikhail approved that implementation in #8, and Justin Pryzby's wording and error-message corrections in #10 were incorporated into #11.
- Robert Haas argued for positive naming and questioned the need for selective disabling in #14 and #18, Michael Paquier advocated starting with a simple switch in #16, and Tom Lane agreed the design was excessive in #19.
- Daniel accepted the simpler direction and deferred the work to the next development cycle in #20, returning five months later with a boolean debug_event_trigger implementation in #22 that abandoned per-event selection.
- Robert rejected the debug prefix in #23 and reiterated his preference for positive naming in #25, leading Daniel through disable_event_trigger in #24 to the default-on event_triggers setting in #26.
- Michael identified reversed documentation, questioned configuration-file restrictions and requested table-rewrite coverage in #27; Robert supported removing the restrictions in #28, and Daniel removed them and emphasized the GUC over single-user mode in #29, without adding the requested table-rewrite test.
- Michael found that removing GUC_NOT_IN_SAMPLE exposed a TAP failure and supplied the missing postgresql.conf.sample entry alongside the full patch in #30, also requesting clearer documentation wording.
- Daniel folded the sample-file fix and wording correction into #31, and Michael explicitly accepted that final version in #32.
- Daniel committed #31 unchanged as 7750fefdb2 and confirmed closure of the CommitFest entry in #34, completing the standalone boolean GUC while leaving login-trigger support outside this thread.