Jul 2025

2 commits
Fix new pg_upgrade query not to rely on regnamespace
4 Jul 2025Álvaro Herreramaster19beta1REL_19_STABLE19beta1REL_18_STABLE18beta2REL_17_STABLE17.6REL_16_STABLE16.10REL_15_STABLE15.14REL_14_STABLE14.19REL_13_STABLE13.22
LLM analysis, 99% surelanded as postedfrom patchset #291 of 1 reviewer saw the final version

The commit makes the same seven-line SQL change as #29 in src/bin/pg_upgrade/check.c: replacing regnamespace with a pg_namespace join. Differences are indentation and placement inside executeQueryOrDie() instead of a query assignment, consistent with the branch adaptation described in #29.

  • Tom Lanesaw the final versionLLM analysis, 100% sureTom tested the replacement query against servers back to PostgreSQL 8.4 in #30, before the commit, which also credits his testing.
pg_upgrade: check for inconsistencies in not-null constraints w/inheritance
4 Jul 2025Álvaro Herreramaster19beta1REL_19_STABLE19beta1REL_18_STABLE18beta2REL_17_STABLE17.6REL_16_STABLE16.10REL_15_STABLE15.14REL_14_STABLE14.19REL_13_STABLE13.22
LLM analysis, 99% suremodified before commitmoderate changesfrom patchset #221 of 2 reviewers saw the final version

The commit preserves the simplified query, diagnostics, and check introduced in #22, but its src/bin/pg_upgrade/check.c diff uses an explicit database loop instead of UpgradeTask and adds a call-site comment. Both the submitted patch and commit explicitly credit Ali Akbar as author.

  • Justin Pryzbysaw an earlier versiontrailerJustin corrected wording in #15 and amended Ali's revised patch in #21, but the later query and execution rewrite followed that work; #23 addresses backpatching without demonstrating review of the rewritten code.
  • Álvaro Herrerasaw the final versioncommittercommitted the patch

Aug 2017

1 commit
Apply ALTER ... SET NOT NULL recursively in ALTER ... ADD PRIMARY KEY.
4 Aug 2017Tom Lanemaster10beta3REL_19_STABLE10beta3REL_18_STABLE10beta3REL_17_STABLE10beta3REL_16_STABLE10beta3REL_15_STABLE10beta3REL_14_STABLE10beta3REL_13_STABLE10beta3+3 more
LLM analysis, 100% suremodified before commitlight changesfrom patchset #51 of 1 reviewer saw the final version

The commit retains both changed files and the regression expectation from #5, but rewrites the comment and passes true rather than is_alter_table to AlterTableInternal(). Tom explains in #6 that the removed guard is unnecessary for the reachable cases.

  • Tom Lanesaw the final versioncommittercommitted the patch

Credits across 3 commits

Author
Ali Akbar1
Alvaro HerreraLLM1
Michael PaquierLLM1
Reviewer
Alvaro Herreracommitter1
Justin Pryzby1
Tom Lanecommitter1
Committer
Alvaro Herrera2
Tom Lane1

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