Jul 2024

1 commit
Add information about access method for partitioned relations in \dP+
2 Jul 2024Michael Paquiermaster18beta1REL_19_STABLE18beta1REL_18_STABLE18beta1
LLM analysis, 100% surelanded as postedfrom patchset #671 of 1 reviewer saw the final version

The three describe.c hunks match #67, which forwards Justin's inline implementation from #65; the commit explicitly credits Justin.

  • Michael Paquiersaw the final versioncommittercommitted the patch

Apr 2024

1 commit
Fix dumps of partitioned tables with table AMs
22 Apr 2024Michael Paquiermaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #621 of 1 reviewer saw the final version

Matches the relkind archive fields, partitioned-table ALTER output, and revised tests in #62; remaining differences are comment wording and indentation in pg_backup_archiver.c.

  • Álvaro Herrerasaw the final versiontrailerProposed distinguishing partitioned tables and using separate output routines in #56 and #58, then reviewed #60 and requested test corrections in #61, after which only light changes followed.

Mar - Apr 2024

6 commits
Don't use the pg_am system catalog in new test
2 Apr 2024Alvaro Herreramaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surewritten from the discussion, no matching patch3 of 3 reviewers saw the final version

Replaces pg_am with tableam_parted_a_heap2 in the foreign-partition regression test, implementing the deadlock diagnosis in #48, #49, and #50; no attachment contains this change.

  • Alexander Lakhinsaw the final versionLLM analysis, 100% sureReported the regression-test deadlock and supplied a concurrent reproducer in #48, directly motivating this discussion-derived fix.
  • Justin Pryzbysaw the final versionLLM analysis, 100% sureIdentified the use of pg_am as unsuitable for concurrent regression tests in #49, the issue corrected by this commit.
  • Tom Lanesaw the final versionLLM analysis, 100% sureReported another buildfarm failure and pinpointed the exclusive pg_am lock in #50 before the fix was committed.
ALTER TABLE: rework determination of access method ID
28 Mar 2024Alvaro Herreramaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 99% surewritten from the discussion, no matching patch2 of 2 reviewers saw the final version

The DefineRelation implementation follows Justin's inline replacement in #46, rather than the attached alternative in #45, with a comment adjustment and regression coverage added. Alvaro explicitly confirms adopting Justin's version in #47.

  • Alexander Lakhinsaw the final versionLLM analysis, 98% sureFound the foreign-table assertion regression and supplied the reproducer in #44 that this inline-derived fix addresses.
  • Alvaro Herrerasaw the final versioncommittercommitted the patch
Allow specifying an access method for partitioned tables
25 Mar 2024Alvaro Herreramaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1Also in: alter table set TABLE ACCESS METHOD (not this thread)
LLM analysis, 99% suremodified before commitlight changesfrom patchset #422 of 2 reviewers saw the final version

The implementation matches #42, with light documentation rewrites in doc/src/sgml/catalogs.sgml and doc/src/sgml/ref/alter_table.sgml; the patch header and #43 establish Michael's authorship alongside Justin and Soumyadeep. I disagree with the partial attribution in thread #43767: its source #483777 dates from July 2021, whereas this thread supplies the matching complete patch in March 2024.

  • The authors themselvessaw the final versiontrailerThe collective credit reflects reciprocal review culminating in Justin's repeated-DEFAULT diagnosis and chgAccessMethod correction in #38 and Michael's implementation and test review in #39, followed by light cleanup.
  • Alvaro Herrerasaw the final versioncommittercommitted the patch
Review wording on tablespaces w.r.t. partitioned tables
20 Mar 2024Alvaro Herreramaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1REL_16_STABLE16.3REL_15_STABLE15.7REL_14_STABLE14.12REL_13_STABLE13.15+1 more
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #40

Extracts and rewords the pg_class.reltablespace documentation hunk from #40, adding removal of a redundant tablespace comment in DefineRelation. Git confirms this is the development commit; the separately listed 58efabdc0c is its REL_12_STABLE backport and is deduplicated here.

Add support for DEFAULT in ALTER TABLE .. SET ACCESS METHOD
8 Mar 2024Michael Paquiermaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #371 of 1 reviewer saw the final version

Takes the complete DEFAULT prerequisite from #37, adding psql completion, an ATPrepSetAccessMethod comment, and a multiple-subcommand regression test; the partitioned-table patch remains separate.

  • Justin Pryzbysaw the final versiontrailerSuggested explicit DEFAULT syntax in #30 and reviewed its purpose and semantics in #36, with only light changes to that prerequisite before commitment.
Update comment of AlterTableCmd->name in parsenodes.h
7 Mar 2024Michael Paquiermaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 98% surewritten from the discussion, no matching patch

Adds access methods to the AlterTableCmd.name field comment in src/include/nodes/parsenodes.h, a file absent from every attachment; the commit identifies this as an issue noticed while working on the feature discussed in #37.

Credits across 8 commits

Author
Justin Pryzby3
Michael Paquier3
Alvaro HerreraLLM1
Soumyadeep Chakraborty1
Reviewer
Alvaro Herreracommitter3
Justin Pryzby1
Michael Paquiercommitter1
The authors themselves1
Reported by
Alexander Lakhin1
Committer
Alvaro Herrera4
Michael Paquier4

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