Feb 2024

2 commits
Use heap_inplace_update() to unset pg_database.dathasloginevt
11 Feb 2024Alexander Korotkovmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1841 of 1 reviewer saw the final version

The event_trigger.c implementation matches #184; the commit expands the comment explaining why concurrent normal updates losing a flag reset are tolerable.

  • Alexander Lakhinsaw the final versionLLM analysis, 100% sureRetested #184, confirmed that it fixed the reported assertion, and raised concurrent-update concerns in #185.
Fix some typos in event trigger docs
11 Feb 2024Alexander Korotkovmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #1871 of 1 reviewer saw the final version

All three doc/src/sgml/event-trigger.sgml hunks match #187, including the missing fmgr.h include and login event-list entry.

  • Alexander Korotkovsaw the final versioncommittercommitted the patch

Jan 2024

1 commit
Remove the flaky check in event_trigger_login regression test
17 Jan 2024Alexander Korotkovmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surewritten from the discussion, no matching patch2 of 2 reviewers saw the final version

No attached patch matches removal of the event_trigger_login flag-clear assertion; #177 reports the failure, #180 explains its unavoidable race, and #181 announces the removal.

  • Alexander Lakhinsaw the final versionLLM analysis, 100% sureReproduced the autovacuum-dependent failure and supplied the failing assertion and repeatable test procedure in #177.
  • Daniel Gustafssonsaw the final versionLLM analysis, 100% sureExplained in #180 that nonblocking flag cleanup cannot guarantee the asserted result, directly supporting the removal.

Oct 2023

2 commits
Fix instable 006_login_trigger.pl test
29 Oct 2023Alexander Korotkovmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1751 of 1 reviewer saw the final version

The removal of two fatal-login checks from src/test/authentication/t/006_login_trigger.pl matches #175; only a name's spelling in the new comment differs.

  • Tom Lanesaw the final versionLLM analysis, 100% sureDiagnosed the IPC::Run race in #174 and approved #175 while identifying its comment typo in #176.
Add support event triggers on authenticated login
16 Oct 2023Alexander Korotkovmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1682 of 16 reviewers saw the final version

The complete feature diff matches #168 except for the added src/include/catalog/catversion.h bump. Both patch and commit credit Konstantin Knizhnik and Mikhail Gribkov, whose takeover appears in #124, rather than sender Alexander Korotkov.

  • Pavel Stehulesaw an earlier versiontrailerReviewed semantics, performance, permissions and tests, supplied corrections, and accepted the added test coverage in #136, before the final flag and locking redesign.
  • Takayuki Tsunakawano review foundtrailerExpressed support for the feature in #3, but the thread contains no identifiable patch inspection, testing or technical review from him.
  • Greg Nancarrowsaw an earlier versiontrailerSupplied naming and documentation corrections, tested recovery in #13, found tuple leaks in #58, and challenged the restored bypass in #77.
  • Ivan Panchenkosaw an earlier versiontrailerReported successful testing and supplied examples in #56, proposed recovery changes in #75, and contributed adopted standby guidance and tests in #112.
  • Daniel Gustafssonsaw an earlier versiontrailerReviewed naming, recovery and catalog logic, revised the series, and measured the startup overhead in #150 that prompted the final flag redesign.
  • Teodor Sigaevsaw an earlier versiontrailerDemonstrated a lockout scenario and questioned unsuccessful-login logging and error semantics in #71.
  • Robert Haassaw the final versiontrailerTested cancellation in #89 and examined the final locking behavior through #163 and #165, after which only light changes followed.
  • Andres Freundsaw an earlier versiontrailerIdentified standby, invalidation and locking problems and reviewed #158 in #160, before the substantive setter-lock correction.
  • Tom Lanesaw an earlier versiontrailerReviewed configuration-based recovery and audit-bypass risks in #93 and #95, and discussed standby risks in #111.
  • Andrey Sokolovsaw an earlier versiontrailerReported Windows failures with a proposed fix in #102, corrected standby guidance in #106, and identified a typo in #115.
  • Zhihong Yusaw an earlier versiontrailerReviewed cleanup and long strings in #39 and questioned lock placement in #118.
  • Sergey Shinderuksaw an earlier versiontrailerReproduced flag races, concurrent-update failures, template-copy omissions and reenabling bugs in #119 and #121.
  • Gregory Starksaw an earlier versiontrailerReported the pg_dump test failure in #142, leading to the correction in #143.
  • Nikita Malakhovsaw an earlier versiontrailerIdentified the example's timezone-dependent bypass in #126, which was corrected in #127.
  • Ted Yusaw an earlier versiontrailerReviewed wording and duplicated command-tag handling in #128 and accepted the implementation explanation in #130.
  • Alexander Korotkovsaw the final versioncommittercommitted the patch

Sep 2023

1 commit
Add GUC for temporarily disabling event triggers
25 Sep 2023Daniel Gustafssonmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1Also in: GUC for temporarily disabling event triggers (landed as posted)

Credits across 6 commits

Author
Alexander KorotkovLLM2
Bharath Rupireddy1
Daniel GustafssonLLM1
Konstantin Knizhnik1
Mikhail Gribkov1
Reviewer
Alexander Korotkovcommitter2
Robert Haas2
Ted Yu2
Andres Freund1
Andrey Sokolov1
Daniel Gustafsson1
Greg Nancarrow1
Greg Stark1
Ivan Panchenko1
Justin Pryzby1
Michael Paquier1
Mikhail Gribkov1
Nikita Malakhov1
Pavel Stehule1
Sergey Shinderuk1
Teodor Sigaev1
Tom Lane1
Tsunakawa, Takayuki1
Zhihong Yu1
Reported by
Alexander Lakhin2
Tom Lane1
Committer
Alexander Korotkov5
Daniel Gustafsson1

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