model-writtenrun 5912 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Peter Eisentraut proposed simplifying pgcrypto regression tests by replacing historical hex encode/decode calls with native bytea literals and output, while retaining explicit escape encoding for readable plaintext results #1.
- Daniel Gustafsson supported the cleanup idea without requesting changes #2.
- Tom Lane supported the approach but recommended replacing removed bytea_output settings with explicit hexadecimal settings so installation defaults could not affect the tests #3.
- Peter explained that pg_regress.c already forces hexadecimal bytea output #4, and Tom withdrew his objection in the thread's final message, leaving no outstanding requested changes #5.
- Peter subsequently committed the submission #1 as 814e1d9ff7, with only three additional redundant-comment removals; the proposed cleanup landed in full.