Sep 2023

1 commit
pg_dump: tests: Correct test condition for invalid databases
25 Sep 2023Andres Freundmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1REL_16_STABLE16.1REL_15_STABLE15.5REL_14_STABLE14.10REL_13_STABLE13.13+2 more
LLM analysis, 100% surewritten from the discussion, no matching patch1 of 1 reviewer saw the final version

Peter identifies the unsupported not_like key and proposes an explicit empty like condition in #13, which Andres agrees to implement in #14. The commit makes exactly that replacement, with an explanatory comment, in src/bin/pg_dump/t/002_pg_dump.pl; no attached patch contains this correction.

  • Peter Eisentrautsaw the final versionLLM analysis, 100% surePeter reported the invalid test condition and proposed the adopted replacement in #13, before the commit.

Jul 2023

1 commit
Handle DROP DATABASE getting interrupted
13 Jul 2023Andres Freundmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1REL_16_STABLE16beta3REL_15_STABLE15.4REL_14_STABLE14.9REL_13_STABLE13.12+2 moreAlso in: "PANIC: could not open critical system index 2662" - twice (not this thread)
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #101 of 2 reviewers saw the final version

The commit takes Andres's second patch in #10, excluding the separate vacuum lock-release and pg_upgrade diagnostic patches, with light changes to catalog documentation, the equivalent database_is_invalid_form() check, formatting, database names and TAP assertions. Contrary to the attribution in thread #136997, its May 9 prototype #986635 is superseded by this matching July 12 submission, making this thread the commit's source.

  • Daniel Gustafssonsaw the final versiontrailerDaniel reports reading and testing #7 in #8, requests constants, hints and logging, then endorses #9 while carrying its main patch unchanged into #10, after which only light edits affect the committed portion.
  • Thomas Munrosaw an earlier versiontrailerThomas tests the externally posted predecessor in #2 and endorses its approach in #3, identifying autovacuum failures before substantial invalid-database handling was added; that predecessor has no patchset entry here.

Credits across 2 commits

Author
Andres Freund1
Reviewer
Daniel Gustafsson1
Thomas Munro1
Reported by
Evgeny Morozov1
Peter Eisentraut1
Committer
Andres Freund2