model-writtenrun 8653 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Peter Eisentraut proposed enabling color by default for PostgreSQL 13, using TERM heuristics with PG_COLOR and NO_COLOR overrides, and supplied a documentation appendix in #1; Juan José Santamaría Flecha identified how the pending Windows support could integrate with this logic in #2.
- The default change drew conflicting reactions: Tom Lane questioned readability in #3, Andreas Joseph Krogh favored highlighted errors in #6, and Jeff Janes reported that trying colored PostgreSQL command output showed little benefit in #12; Alvaro Herrera also found the appendix unclear about rule precedence in #7.
- After the separate Windows work landed, Juan José Santamaría Flecha proposed salvaging terminal checks for PG_COLOR=auto and the documentation independently in #15; Michael Paquier illustrated the difficulty of discovering PG_COLORS syntax and advocated centralized documentation in #16 and #23.
- Peter abandoned default coloring because there was no consensus and posted a documentation-only revision covering existing PG_COLOR and PG_COLORS behavior in #24.
- Michael reviewed the revised appendix in #26, requesting descriptions of the three color fields, an SGR reference, and links from utility documentation instead of deleting its color-support mentions.
- Peter committed #24 as e1ff780485, adding descriptions of the color keys and default values; his announcement in #27 closed the commitfest item while explaining that man-page links were omitted because of complications and that a comprehensive SGR value list would depend on the terminal.
- Terminal detection remained unresolved: Juan wanted the checks applied to PG_COLOR=auto in #29, Michael warned that embedded TERM knowledge could become stale in #30, and Peter requested concrete examples of terminals needing detection before refining it in #32.
- Michael continued advocating appendix links and an SGR reference in #31, and Peter invited concrete wording in #33; Michael then supplied the follow-up patch in #34, adding an ECMA-48 link and consistent PG_COLOR/PG_COLORS references across utility documentation.
- At the end of the thread, the appendix had landed through e1ff780485, default coloring had been abandoned in #24, terminal detection still needed justification in #32, and Michael's documentation follow-up in #34 awaited review.