Mar 2021

2 commits
Use FeBeWaitSet for walsender.c.
1 Mar 2021Thomas Munromaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #140 of 1 reviewer saw the final version

The walsender.c patch in #14 has identical added and removed lines, including the WalSndWait() helper and explicit postmaster-death exit.

  • Kyotaro Horiguchisaw an earlier versiontrailerIn #7 he reviewed the WAL-sender conversion and requested clearer event positions, before #14 introduced the helper and explicit death handling; the commit also credits an earlier version.
Introduce symbolic names for FeBeWaitSet positions.
1 Mar 2021Thomas Munromaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #141 of 1 reviewer saw the final version

Exactly matches the symbolic-position patch in #14 across be-secure.c, pqcomm.c, miscinit.c and libpq.h, including the PGINVALID_SOCKET substitutions.

  • Kyotaro Horiguchisaw the final versiontrailerIn #7 he identified unexplained numeric event positions and use of -1 instead of PGINVALID_SOCKET, which Thomas addressed with the naming cleanup introduced in #10.

Jul 2020

3 commits
Introduce a WaitEventSet for the stats collector.
30 Jul 2020Thomas Munromaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #101 of 1 reviewer saw the final version

Matches the pgstat.c patch in #10, with only whitespace around the Windows timeout comment differing.

  • Kyotaro Horiguchisaw the final versiontrailerIn #7 he reviewed the statistics-collector patch and discussed the event variable's scope; its code remained unchanged through #10 apart from rebase and formatting.
Use WaitLatch() for condition variables.
30 Jul 2020Thomas Munromaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #101 of 1 reviewer saw the final version

The condition_variable.c changes exactly match #10: remove the private wait set and use WaitLatch() with the appropriate timeout flags.

  • Kyotaro Horiguchisaw the final versiontrailerHe reviewed the condition-variable conversion in #5 and reported regression testing of the retained early patches in #6, with no subsequent substantive change to this patch.
Use a long lived WaitEventSet for WaitLatch().
30 Jul 2020Thomas Munromaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #100 of 1 reviewer saw the final version

Matches the reusable WaitLatch implementation in #10, with an added comment explaining LatchWaitSetLatchPos and minor formatting changes in latch.c.

  • Kyotaro Horiguchisaw an earlier versiontrailerIn #5 he identified the duplicate close, misleading naming and missing WL_LATCH_SET handling, and #6 reported testing, but #10 subsequently changed the latch-disabled case and removed the faulty cleanup hunk.

Mar 2020

1 commit
Don't use EV_CLEAR for kqueue events.
18 Mar 2020Thomas Munromaster13beta1REL_19_STABLE13beta1REL_18_STABLE13beta1REL_17_STABLE13beta1REL_16_STABLE13beta1REL_15_STABLE13beta1REL_14_STABLE13beta1REL_13_STABLE13beta1
LLM analysis, 100% surelanded as postedfrom patchset #61 of 1 reviewer saw the final version

Exactly matches the two EV_CLEAR removals in latch.c carried unchanged in #6. That attachment retains Thomas Munro's authorship from #4, although Kyotaro Horiguchi sent the series.

  • Kyotaro Horiguchisaw the final versiontrailerIn #5 he checked the change against kqueue documentation, then carried it unchanged and reported regression testing in #6.

Credits across 6 commits

Author
Thomas MunroLLM6
Reviewer
Kyotaro Horiguchi6
Committer
Thomas Munro6

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