Jun 2022

1 commit
Improve comments for trivial_subqueryscan().
9 Jun 2022Etsuro Fujitamaster15beta2REL_19_STABLE15beta2REL_18_STABLE15beta2REL_17_STABLE15beta2REL_16_STABLE15beta2REL_15_STABLE15beta2
LLM analysis, 100% surelanded as postedfrom patchset #1621 of 1 reviewer saw the final version

The setrefs.c comment changes exactly match #162; #163 confirms the commit.

  • Zhihong Yusaw the final versionLLM analysis, 100% sureSuggested the numbered-clause formatting in #161, which was adopted in #162, and is credited in the commit message.

Apr 2022

1 commit
Disable asynchronous execution if using gating Result nodes.
28 Apr 2022Etsuro Fujitamaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1REL_14_STABLE14.3
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1582 of 2 reviewers saw the final version

Matches the development-branch variant in #158, with only two regression-test comments changing 'Prevent' to 'Disable'; the other attachment is its PostgreSQL 14 counterpart.

  • Justin Pryzbysaw the final versionLLM analysis, 100% sureReported the planner crash in #146 and confirmed in #152 that the attached fix resolved it.
  • Zhihong Yusaw the final versionLLM analysis, 100% sureReviewed placement of the Result checks in #151 and accepted the retained structure in #157, with only light edits following.

Apr 2022

1 commit
Allow asynchronous execution in more cases.
6 Apr 2022Etsuro Fujitamaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1403 of 3 reviewers saw the final version

Relative to #140, the commit renames SubqueryScan.status to scanstatus and clarifies mark_async_capable_plan()'s return value. The commit credits Lepikhov and Fujita, and #135 explicitly derives the work from Lepikhov's #47.

  • Alexander Pyhalovsaw the final versionLLM analysis, 100% sureReported in #138 that the patch looked good and worked as expected, followed only by code organization, naming, comments, and test changes.
  • Zhihong Yusaw the final versionLLM analysis, 100% sureRequested clearer field naming and return-value documentation in #141, both addressed before committing as explained in #142.
  • Etsuro Fujitasaw the final versioncommittercommitted the patch

Jun 2021

2 commits
Doc: Further update documentation for asynchronous execution.
8 Jun 2021Etsuro Fujitamaster14beta2REL_19_STABLE14beta2REL_18_STABLE14beta2REL_17_STABLE14beta2REL_16_STABLE14beta2REL_15_STABLE14beta2REL_14_STABLE14beta2
LLM analysis, 99% surelanded as postedfrom patchset #1321 of 1 reviewer saw the final version

The postgres-fdw.sgml addition matches #132 apart from line wrapping. The commit credits Lepikhov and Fujita, reflecting the proposed explanation in #119 and its subsequent rewrite.

  • Etsuro Fujitasaw the final versioncommittercommitted the patch
Fix rescanning of async-aware Append nodes.
7 Jun 2021Etsuro Fujitamaster14beta2REL_19_STABLE14beta2REL_18_STABLE14beta2REL_17_STABLE14beta2REL_16_STABLE14beta2REL_15_STABLE14beta2REL_14_STABLE14beta2
LLM analysis, 100% surelanded as postedfrom patchset #1271 of 1 reviewer saw the final version

The nodeAppend.c rescan and initialization changes and postgres_fdw regression tests exactly match #127.

  • Kyotaro Horiguchisaw the final versiontrailerQuestioned removal of explicit initialization in #125, prompting only light initialization changes in #127.

May 2021

2 commits
Doc: Update documentation for asynchronous execution.
17 May 2021Etsuro Fujitamaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #1222 of 2 reviewers saw the final version

The documentation exactly matches #122; the commit explicitly attributes the prose to Stephen Frost, whose replacement wording appears in #107, with Fujita's expansion.

  • Andrei Lepikhovsaw the final versionLLM analysis, 85% sureEndorsed the proposed documentation wording and supplied same-server benchmarks in #108, before the final prose revision.
  • Etsuro Fujitasaw the final versioncommittercommitted the patch
Fix EXPLAIN ANALYZE for async-capable nodes.
12 May 2021Etsuro Fujitamaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #1131 of 1 reviewer saw the final version

The asynchronous instrumentation, postgres_fdw tests, and callback documentation match #113 exactly; the commit explicitly says Lepikhov's original patch was not used.

  • Andrey Lepikhovsaw the final versiontrailerReported a LIMIT-triggered instrumentation error in #109 and confirmed the revised fix while examining tuple counts in #114.

Apr 2021

1 commit
Minor code cleanup in asynchronous execution support.
23 Apr 2021Etsuro Fujitamaster14beta1REL_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 cleanup in nodeAppend.c and postgres_fdw.c exactly matches #97, including the event-count simplification and capability-merging explanation.

  • Kyotaro Horiguchisaw the final versionLLM analysis, 96% sureQuestioned the explanation of merged async_capable options in #88, directly prompting the comment improvement credited in this cleanup commit.

Mar - Apr 2021

2 commits
Adjust input value to WaitEventSetWait() in ExecAppendAsyncEventWait().
6 Apr 2021Etsuro Fujitamaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #931 of 2 reviewers saw the final version

The event-count bound in ExecAppendAsyncEventWait() exactly matches #93; #96 confirms it was applied.

  • Tom Lanesaw an earlier versionLLM analysis, 94% sureReproduced the landed implementation's Valgrind failure and supplied diagnostic traces in #91, before the corrective logic was attached.
  • Kyotaro Horiguchisaw the final versionLLM analysis, 100% sureExamined epoll_wait buffer semantics and questioned the diagnosis in #94, leading to clarification before the unchanged patch was committed.
Add support for asynchronous execution.
31 Mar 2021Etsuro Fujitamaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1Also in: Append with naive multiplexing of FDWs (not this thread)Also in: Re: asynchronous execution
LLM analysis, 100% suremodified before commitlight changesfrom patchset #871 of 4 reviewers saw the final version

Matches #87 except for comment edits in postgres_fdw.c, nodeAppend.c, and execnodes.h. Fujita authored the final integration, with the commit acknowledging contributions from Robert Haas, Kyotaro Horiguchi, and Thomas Munro.

  • Konstantin Knizhniksaw an earlier versionLLM analysis, 100% sureReported LIMIT and stale-connection-state failures and supplied corrective deltas in #30 and #31, before substantial executor and connection-handling redesigns.
  • Andrei Lepikhovsaw an earlier versionLLM analysis, 99% sureReviewed costing and connection handling, supplied the mixed-scan correction in #46, and reported TPC-H testing and interface concerns in #56, before later architectural revisions.
  • Movead Lisaw an earlier versionLLM analysis, 99% sureReported successful early testing in #10 and later reproduced a multi-server assertion and scheduling problems in #54, which preceded substantial scheduling changes.
  • Justin Pryzbysaw the final versionLLM analysis, 98% sureIdentified a missing break and typos in #84, incorporated before landing with only light documentation, assertion, test, and comment changes afterward.

Credits across 10 commits

Author
Etsuro Fujita5 LLM7
Andrei LepikhovLLM2
Stephen FrostLLM1
Reviewer
Etsuro Fujitacommitter3
Andrei Lepikhov1
Kyotaro Horiguchi1
Committer
Etsuro Fujita10

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