May 2021

1 commit
amcheck: Improve some confusing reports about TOAST problems.
3 May 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1551 of 1 reviewer saw the final version

Matches #155 except for the final early-end diagnostic and its argument order in contrib/amcheck/verify_heapam.c, changed as proposed in #157.

  • Mark Dilgersaw the final versionLLM analysis, 100% sureMark tested the alternatives, identified false positives and the boundary error in #154, and reviewed the final wording in #156; the commit explicitly credits his review.

Mar - Apr 2021

15 commits
amcheck: Reword some messages and fix an alignment problem.
14 Apr 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #1331 of 1 reviewer saw the final version

Exactly matches the first patch of #133, including VARATT_EXTERNAL_GET_POINTER, revised TOAST diagnostics, and corresponding test expectations; the additional validation patch was left pending.

  • Robert Haassaw the final versioncommittercommitted the patch
amcheck: fix multiple problems with TOAST pointer validation
7 Apr 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surepart of a posted patchsetmoderate changesfrom patchset #1231 of 1 reviewer saw the final version

Takes the second patch of #123, deferring TOAST scans until after buffer unlock and avoiding prunable tuples, with list_free changed to list_free_deep to prevent a memory leak.

  • Robert Haassaw the final versioncommittercommitted the patch
amcheck: Remove duplicate XID/MXID bounds checks.
7 Apr 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #1231 of 1 reviewer saw the final version

Exactly matches the first patch of #123, removing duplicate XID/MXID bounds checks and updating comments in contrib/amcheck/verify_heapam.c.

  • Robert Haassaw the final versioncommittercommitted the patch
amcheck: Fix verify_heapam's tuple visibility checking rules.
1 Apr 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #1191 of 1 reviewer saw the final version

The diff exactly matches #119; its header and the commit credit Mark Dilger and Robert Haas jointly, following Robert's consolidation of Mark's #109.

  • Robert Haassaw the final versioncommittercommitted the patch
nbtree VACUUM: Cope with buggy opclasses.
23 Mar 2021Peter Geogheganmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surewritten from the discussion, no matching patch

Peter proposes and explains the missing-downlink recovery in #97 and #99, then confirms committing it in #100; no attached patch matches the nbtpage.c change from ERROR to LOG and return false.

Improve pg_amcheck's TAP test 003_check.pl.
23 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #732 of 2 reviewers saw the final version

Combines the third patch of #73 with its first patch's autovacuum setting for 003_check.pl, leaving autovacuum enabled in 005_opclass_damage.pl as explained in #95.

  • Noah Mischsaw the final versionLLM analysis, 99% sureNoah reports 196 successful runs with the autovacuum-disable patch in #74, testing the unchanged portion subsequently committed from #73.
  • Robert Haassaw the final versioncommittercommitted the patch
Fix a confusing amcheck corruption message.
16 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #731 of 1 reviewer saw the final version

Exactly matches the second patch of #73, suppressing the extra chunk-count complaint for entirely missing TOAST values and updating 004_verify_heapam.pl.

  • Robert Haassaw the final versioncommittercommitted the patch
In pg_amcheck tests, don't depend on perl's Q/q pack code.
13 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #622 of 2 reviewers saw the final version

Exactly matches #62, replacing Perl q packing with two 32-bit fields and using identical nonzero halves for endian-independent test data.

  • Tom Lanesaw the final versionLLM analysis, 99% sureTom explains the remaining word-swapping issue and recommends equal upper and lower halves in #61, which #62 adopts.
  • Robert Haassaw the final versioncommittercommitted the patch
pg_amcheck: Keep trying to fix the tests.
13 Mar 2021Tom Lanemaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #411 of 1 reviewer saw the final version

Exactly matches the one-line correction in #41, moving the database argument after the switches in the remaining 003_check.pl invocation.

  • Tom Lanesaw the final versioncommittercommitted the patch
pg_amcheck: Try to fix still more test failures.
13 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #391 of 1 reviewer saw the final version

Combines all option-order changes from #39 with the first patch of #37, removing a platform-specific authentication-message expectation in 002_nonesuch.pl.

  • Robert Haassaw the final versioncommittercommitted the patch
Try to avoid apparent platform-dependency in IPC::Run
13 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #371 of 1 reviewer saw the final version

Exactly matches the second patch of #37, passing -S* as one argument to avoid apparent IPC::Run shell expansion; the authentication-test patch landed separately.

  • Robert Haassaw the final versioncommittercommitted the patch
Fix portability issues in pg_amcheck's 004_verify_heapam.pl.
12 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #351 of 1 reviewer saw the final version

Matches #35 modulo rebased context, repairing endian-sensitive varlena-header corruption and architecture-dependent TOAST expectations in 004_verify_heapam.pl.

  • Robert Haassaw the final versioncommittercommitted the patch
Try to fix compiler warnings.
12 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surewritten from the discussion, no matching patch

Robert directly fixes the warnings reported by Peter in #29, moving a declaration and marking is_btree PG_USED_FOR_ASSERTS_ONLY in pg_amcheck.c; no corresponding fix attachment appears.

Add pg_amcheck, a CLI for contrib/amcheck.
12 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1Also in: new heapcheck contrib module (not this thread)
LLM analysis, 100% suremodified before commitlight changesfrom patchset #261 of 3 reviewers saw the final version

Matches #26 with diagnostic wording, SGML entity naming, formatting, and MSVC integration edits. This March 12 submission is later than every patchset supplied for thread #42262.

  • Peter Geoghegansaw an earlier versionLLM analysis, 99% surePeter's review in #7 addresses portable block arguments and faithful error reporting, adopted in #11 before further database-selection changes.
  • Peter Eisentrautsaw an earlier versionLLM analysis, 88% surePeter requests moving the submitted application into src/bin in #14, implemented in #26 after intervening database-selection revisions.
  • Robert Haassaw the final versioncommittercommitted the patch
Refactor and generalize the ParallelSlot machinery.
11 Mar 2021Robert Haasmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1Also in: new heapcheck contrib module (not this thread)
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #131 of 1 reviewer saw the final version

Exactly matches the first patch of #13, including ParallelSlotArray, lazy connections, and multi-database slot reuse. This source postdates the latest supplied patchset in thread #42262.

  • Robert Haassaw the final versioncommittercommitted the patch

Credits across 16 commits

Author
Mark DilgerLLM13
Robert HaasLLM1
Reviewer
Robert Haascommitter12
Tom Lanecommitter1
Committer
Robert Haas14
Peter Geoghegan1
Tom Lane1

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