model-writtenrun 2796 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Bruce Momjian opened the discussion after a private report that pg_dump --no-comments sounded as though it removed SQL line or block comments rather than COMMENT commands, and proposed clarifying the documentation #1.
- Matthias van de Meent suggested referring to database object comments or COMMENT ON commands #2, while Erik Wienhold favored naming the COMMENT command without the unnecessary “SQL” prefix and noted that --verbose already used “object comments” for line comments #3; Daniel Gustafsson supported that clarification #4.
- Bruce submitted the one-line pg_dump documentation patch #5, which Daniel, Erik, Alvaro Herrera, and Matthias endorsed #6 #7 #8 #9; it landed unchanged as b38bac26e2, and Bruce announced its application to master #10.
- Marcos Pegoraro requested backpatching to supported versions #11, but Bruce explained that his documentation guidelines normally limited unclear-language fixes to master, while remaining open to requests for broader application #12.
- Tom Lane identified matching ambiguity in pg_dump's help output, pg_dumpall's documentation and help output, and pg_restore's help output; he requested consistency on master and opposed backpatching help changes because of translation issues, while finding pg_restore's manual wording adequate #13.
- Bruce supplied the four-hunk follow-up patch #14, which landed unchanged as f722dd32de, completing the identified documentation and help-text clarifications on master without backports.
- The final message contained no further patch request: Alvaro commented on the potentially misleading Reported-by credit for Tom and the blank lines between commit trailers #15; both patches had landed and no implementation work remained pending in the thread.