model-writtenrun 6083 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Peter Eisentraut proposed five Unicode-generator improvements: prevent duplicate parallel execution of UCS_to_most.pl, sort encoding progress output, remove spaces inside generated array bounds, simplify generation using trailing commas, and replace two-space indentation with tabs; he omitted the large regenerated-output diff for the indentation change #1.
- Kyotaro Horiguchi endorsed the changes but could not test them because unicode.org downloads timed out, while Heikki Linnakangas requested an explanation of the Makefile workaround and suggested generating maps separately instead #2 #3.
- Heikki also questioned extra newline escapes and suggested either consistent trailing commas or the existing radix-table technique for avoiding them; Kyotaro supported that alternative, and Peter agreed to revisit the Makefile and newline placement #3 #4 #5.
- On July 19, the encoding-order component of #1 landed unchanged as 4d56115f72, and the generated-whitespace component landed with the questioned newline additions removed as 344dedfd1c.
- Peter then replaced the original Makefile workaround with a reusable rule generating one encoding pair at a time, explaining that the specialized generators also suffered duplicate parallel execution and that UCS_to_most.pl already supported selecting one encoding #6.
- After a two-month gap, Peter supplied a correction to the ALL_TEXTS variable expansion that had prevented the refactored Makefile from working, together with a comment typo fix #7.
- The corrected Makefile from #7 landed unchanged as e752727195, and Peter announced the commit and closure of the CommitFest entry #8.
- The thread ends with the Makefile fix and two independent cleanups landed; the supplied discussion does not establish a landing or explicit withdrawal of the trailing-comma simplification and indentation changes originally proposed in #1, despite the entry being closed in #8.