Nov 2020

2 commits
doc: wire protocol data type for history file content is bytea
12 Nov 2020Bruce MomjianREL_13_STABLE13.2REL_12_STABLE12.6REL_11_STABLE11.11REL_10_STABLE10.16REL9_6_STABLE9.6.21REL9_5_STABLE9.5.25
LLM analysis, 97% surepart of a posted patchsetlight changesfrom patchset #173 of 3 reviewers saw the final version

This takes the complete stable-branch variant from #17: the raw-byte explanation in doc/src/sgml/protocol.sgml and BYTEAOID comment in src/backend/replication/walsender.c, unchanged apart from branch context. The submission's separate development-branch type change is absent.

  • Michael Paquiersaw the final versionLLM analysis, 92% sureIn #4, he challenged the text-type proposal in #3 because recovery target names can contain non-ASCII characters; subsequent changes clarified the raw-byte behavior.
  • Tom Lanesaw the final versionLLM analysis, 99% sureIn #10, #13, and #16, he rejected escaping, requested raw-file documentation, and required accurate branch-specific labels and a narrowly scoped encoding caveat.
  • Brar Pieningsaw the final versionLLM analysis, 91% sureBeyond his original report, #9 identified the decoding problem with the text-label proposal in #6 and requested explicit documentation of the raw bytes.
change wire protocol data type for history file content
12 Nov 2020Bruce Momjianmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 97% surepart of a posted patchsetlight changesfrom patchset #173 of 3 reviewers saw the final version

This takes the complete development-branch variant from #17, with identical documentation additions and BYTEAOID-to-TEXTOID change in SendTimeLineHistory(). The submission's separate stable-branch variant is absent.

  • Michael Paquiersaw the final versionLLM analysis, 92% sureIn #4, he identified non-ASCII recovery target names as a problem for the text-type proposal in #3, informing the final encoding caveat.
  • Tom Lanesaw the final versionLLM analysis, 99% sureIn #13 and #16, he specified a text label with a raw-file explanation and corrected the proposed scope of the encoding warning, which #17 implemented.
  • Brar Pieningsaw the final versionLLM analysis, 91% sureIn #9, he reviewed the text-label proposal in #6, explaining that clients could incorrectly assume client_encoding and requesting explicit raw-byte documentation.

Credits across 2 commits

Author
Bruce MomjianLLM2
Reported by
Brar Piening2
Committer
Bruce Momjian2

N LLM means N of the credits come from LLM analysis, not from commit trailers.