Mar 2023

3 commits
Use nanosleep() to implement pg_usleep().
15 Mar 2023Thomas Munromaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1Also in: Use nanosleep() for pg_usleep() on Unix/Linux?Also in: Use nanosleep(2) in pg_usleep, if available?
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #221 of 1 reviewer saw the final version

Takes the nanosleep() component of the three-patch series in #22, retaining its executable changes in src/port/pgsleep.c. The caution comment was rewritten following Nathan Bossart's feedback in #23 and #25 to explain signal races and latch responsiveness.

  • Nathan Bossartsaw the final versiontrailerEndorsed nanosleep() and challenged the interruption explanation in #23, then requested clearer terminology in #25; only wording changed afterward.
Update obsolete comment about pg_usleep() accuracy.
15 Mar 2023Thomas Munromaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #221 of 1 reviewer saw the final version

Takes the accuracy-comment component of #22, replacing its proposed text with Nathan Bossart's inline wording from #25, modulo wrapping. The commit explicitly credits Nathan as author, and changes only src/port/pgsleep.c commentary.

  • Thomas Munrosaw the final versioncommittercommitted the patch
Fix fractional vacuum_cost_delay.
15 Mar 2023Thomas Munromaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1REL_15_STABLE15.3REL_14_STABLE14.8
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #221 of 2 reviewers saw the final version

Takes exactly the vacuum-fix component of #22, whose two src/backend/commands/vacuum.c hunks are unchanged from #19. It restores pg_usleep(), wait-event reporting, and an explicit IsUnderPostmaster/PostmasterIsAlive() check; the other two series components landed separately.

  • Tom Lanesaw the final versionLLM analysis, 99% sureAccepted the minimal backpatchable fix in #20, and its executable diff remained unchanged in #22 and the commit.
  • Melanie Plagemansaw an earlier versionLLM analysis, 90% sureQuestioned residual-delay behavior after a GUC change in #10, reviewing the accumulator version before it was replaced by substantially different implementations.

Credits across 3 commits

Author
Thomas MunroLLM2
Nathan Bossart1
Reviewer
Nathan Bossart1
Thomas Munrocommitter1
Reported by
Melanie Plageman1
Committer
Thomas Munro3

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