Aug 2020

1 commit
pg_upgrade doc: mention saving postgresql.conf.auto files
31 Aug 2020Bruce Momjianmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1REL_13_STABLE13rc1+5 more
LLM analysis, 100% surelanded as postedfrom patchset #247

Both hunks in doc/src/sgml/ref/pgupgrade.sgml exactly match #247, adding preservation and migration guidance for postgresql.auto.conf and included configuration files. Bruce announced application in #249.

Jul 2020

3 commits
Add hash_mem_multiplier GUC.
29 Jul 2020Peter Geogheganmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1REL_13_STABLE13beta3
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #2402 of 2 reviewers saw the final version

Commits the multiplier component of the two-patch series in #240, with executable changes preserved modulo rebasing. Additional edits adjust comments and extend documentation in postgres-ref.sgml and runtime.sgml; the prerequisite GUC removal landed separately.

  • Álvaro Herrerasaw the final versiontrailerHis review in #237 requested moving the get_hash_mem() declaration to miscadmin.h, which #240 implements with only light subsequent changes.
  • Jeff Davissaw the final versiontrailerHe assessed the multiplier patch in #193, and subsequent changes to that component were light, with the removed planner branches separately reviewed in #243.
HashAgg: use better cardinality estimate for recursive spilling.
29 Jul 2020Jeff Davismaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1REL_13_STABLE13beta3
LLM analysis, 100% surelanded as postedfrom patchset #2301 of 1 reviewer saw the final version

The HyperLogLog additions and replacements in nodeAgg.c and nodeAgg.h match #230; differences are confined to rebased surrounding context. Jeff announced the commit in #245.

  • Peter Geoghegansaw the final versiontrailerHe tested the patch in #233 and #234, confirming smoother partition counts and a reduction from 2,424 batches to 20.
Remove hashagg_avoid_disk_plan GUC.
28 Jul 2020Peter Geogheganmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1REL_13_STABLE13beta3
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #2402 of 2 reviewers saw the final version

Commits only the hashagg_avoid_disk_plan-removal component of #240, preserving its five-file implementation. The only substantive textual differences within that component are two planner.c comments describing partially grouped paths.

  • Jeff Davissaw the final versiontrailerHe inspected the redundant cheapest_total_path NULL check and approved the planner cleanup in #243.
  • Álvaro Herrerasaw the final versiontrailerHe reviewed unconditional path generation and the simplified DISTINCT hashing decision in #241.

Jun 2020

2 commits
Doc fixup for hashagg_avoid_disk_plan GUC.
22 Jun 2020Jeff Davismaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1REL_13_STABLE13beta2
LLM analysis, 99% surewritten from the discussion, no matching patch

Justin Pryzby reported the renamed setting's incorrect alphabetical placement in #18. The commit relocates its unchanged entry in doc/src/sgml/config.sgml, without a matching attached patchset.

Rework HashAgg GUCs.
11 Jun 2020Jeff Davismaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1REL_13_STABLE13beta2
LLM analysis, 99% surewritten from the discussion, no matching patch

Melanie Plageman supplied the statistics-based testing approach in #8 and #10, while Justin Pryzby proposed clearer planner-only naming in #9. Jeff implemented those proposals without an attached patchset and announced the GUC removal, rename, and regression-test changes in #15.

Credits across 6 commits

Author
Peter Geoghegan2
Bruce MomjianLLM1
Jeff DavisLLM1
Reviewer
Alvaro Herrera2
Jeff Davis2
Peter Geoghegan1
Reported by
Alvaro Herrera1
Justin Pryzby1
Committer
Jeff Davis3
Peter Geoghegan2
Bruce Momjian1

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