Feb 2021

2 commits
Fix tuple routing to initialize batching only for inserts
17 Feb 2021Tomas Vondramaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #1092 of 2 reviewers saw the final version

The four-file diff matches #109 exactly after normalizing patch metadata, including the CMD_INSERT guard, fmstate assertion, and regression test. Tomas explicitly confirmed choosing this version in #113, after deferring #111.

  • Tomas Vondrasaw the final versiontrailerRequested regression coverage in #100, examined the initialization mismatch in #110, and selected and committed the smaller fix in #113.
  • Takayuki Tsunakawasaw the final versiontrailerReviewed the fmstate initialization in #102, with only equivalent expression, explanation, and rebase changes between that patch and #109.
Fix pointer type in ExecForeignBatchInsert SGML docs
17 Feb 2021Tomas Vondramaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #971 of 1 reviewer saw the final version

The single-line planSlots pointer correction in doc/src/sgml/fdwhandler.sgml exactly matches Ian Lawrence Barwick's attached patch in #97, despite the commit credit using Reported-by.

  • Tomas Vondrasaw the final versioncommittercommitted the patch

Jan 2021

2 commits
Fix initialization of FDW batching in ExecInitModifyTable
21 Jan 2021Tomas Vondramaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #921 of 4 reviewers saw the final version

Matches the initialization loop in #92, with a clarified comment and removal of the redundant inner CMD_INSERT condition requested in #94.

  • Zhihong Yusaw an earlier versionLLM analysis, 90% sureCompared the posted relocation fix with his alternative in #80 and argued for retaining late initialization, before the loop and operation guard were added.
  • Tom Lanesaw an earlier versionLLM analysis, 98% sureExplained in #83 that every result relation needed initialization, prompting the loop in #86 before the later INSERT-only restriction.
  • Amit Langotesaw an earlier versionLLM analysis, 96% sureChecked the RETURNING initialization dependency in #84 and proposed the outer CMD_INSERT guard in #90, which subsequently changed initialization behavior.
  • Tsunakawa, Takayukisaw the final versionLLM analysis, 100% sureIdentified the redundant inner CMD_INSERT test in #94, and the committed patch removes it with only comment polishing.
Implement support for bulk inserts in postgres_fdw
20 Jan 2021Tomas Vondramaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #672 of 2 reviewers saw the final version

Applying #67 to the commit's parent reproduces the committed files except for two comment edits in contrib/postgres_fdw/deparse.c and postgres_fdw.c, consistent with the polishing reported in #73.

  • Tomas Vondrasaw the final versiontrailerReviewed API design, statement preparation, configuration, and EXPLAIN behavior, then checked the final patch in #71 and committed it.
  • Amit Langotesaw the final versiontrailerFound the trigger crash in #49, reviewed callback and partition initialization in #53 and #58, and examined the final initialization placement in #68.

Credits across 4 commits

Author
Amit Langote1
Ian Lawrence BarwickLLM1
Tomas VondraLLM1
Tsunakawa, Takayuki1
Reviewer
Tomas Vondracommitter3
Amit Langote1
Tsunakawa, Takayuki1
Reported by
Ian Barwick1
Committer
Tomas Vondra4

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