Mar - Apr 2022

6 commits
Update some tests in 013_crash_restart.pl.
5 Apr 2022Tom Lanemaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1REL_14_STABLE14.3
LLM analysis, 100% surewritten from the discussion, no matching patch2 of 2 reviewers saw the final version

No attachment changes src/test/recovery/t/013_crash_restart.pl; its four regex changes implement Tom Lane's diagnosis and revised proposal in #55 and #56, discussed with Mark Dilger in #57 and #59. Tom confirmed completion in #60.

  • Mark Dilgersaw the final versionLLM analysis, 100% sureMark supplied failure logs, examined alternative libpq error paths in #57, and endorsed relaxing the test in #59.
  • Thomas Munrosaw the final versionLLM analysis, 84% sureThomas reported a matching crash-restart test failure with CI logs and separated it from the hook changes in #10.
Force NO_LOCALE / UTF8 for test_oat_hooks tests
23 Mar 2022Andrew Dunstanmaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surewritten from the discussion, no matching patch2 of 2 reviewers saw the final version

Adds ENCODING=UTF8 and NO_LOCALE=1 to src/test/modules/test_oat_hooks/Makefile, changes absent from every attachment. This addresses the non-installcheck failures discussed in #39, and Andrew reported the tested fix in #41.

  • Tom Lanesaw the final versionLLM analysis, 99% sureTom reproduced the encoding-dependent failure and explained why disabling installcheck would leave make check failures in #31 and #39.
  • Justin Pryzbysaw the final versionLLM analysis, 94% sureJustin supplied Windows failure logs showing missing namespace-search notices after the preceding test fix in #36.
Temporarily disable installcheck for test_oat_hooks module
23 Mar 2022Andrew Dunstanmaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surewritten from the discussion, no matching patch2 of 2 reviewers saw the final version

Restores NO_INSTALLCHECK=1 in src/test/modules/test_oat_hooks/Makefile as Andrew proposed in #38 and justified in #40. The attachments already retained that setting and therefore do not contain this restoration.

  • Tom Lanesaw the final versionLLM analysis, 99% sureTom reproduced locale-dependent failures, diagnosed namespace-cache timing, and reviewed the limitations of disabling installcheck in #31 and #39.
  • Justin Pryzbysaw the final versionLLM analysis, 93% sureJustin reported continuing Windows failures with concrete regression diffs before the workaround was committed in #36.
Fix new Object Access hooks test
22 Mar 2022Andrew Dunstanmaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surelanded as postedfrom patchset #232 of 3 reviewers saw the final version

All five file diffs match #23: explicit LOAD replaces preloading, the _PG_init restriction disappears, and emit_audit_message suppresses parallel-worker notices. Andrew confirmed pushing it in #28.

  • Tom Lanesaw the final versionLLM analysis, 100% sureTom reproduced both failures, proposed worker-message suppression and removal of the preload check, and explicitly accepted the final loading design in #27.
  • Julien Rouhaudsaw an earlier versionLLM analysis, 88% sureJulien reported a buildfarm failure of the newly committed test in #14, before the loading and parallel-worker logic was revised.
  • Andrew Dunstansaw the final versioncommittercommitted the patch
Add a test module for Object Access hooks
22 Mar 2022Andrew Dunstanmaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #91 of 1 reviewer saw the final version

Contains the test-module constituent of #9, with the SQL and expected output edited to omit redundant LOAD and client_min_messages manipulation. Andrew explicitly described these small adjustments in #13; the hook implementation landed separately.

  • Andrew Dunstansaw the final versioncommittercommitted the patch
Add String object access hooks
22 Mar 2022Andrew Dunstanmaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 98% surepart of a posted patchsetlight changesfrom patchset #91 of 1 reviewer saw the final version

Exactly matches the first constituent patch of the two-patch series in #9, covering objectaccess.c, guc.c, and three headers; the test-module constituent landed separately. The patch credits Joshua Brindle as its original writer and Mark Dilger as its refactorer, consistent with #1.

  • Andrew Dunstansaw the final versioncommittercommitted the patch

Credits across 6 commits

Author
Mark DilgerLLM2
Joshua BrindleLLM1
Reviewer
Andrew Dunstancommitter3
Committer
Andrew Dunstan5
Tom Lane1

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