May 2024

1 commit
Fix segmentation fault in MergeInheritedAttribute()
3 May 2024Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #292 of 3 reviewers saw the final version

The tablecmds.c NULL guard and compression regression-test changes match #29 exactly after whitespace normalization.

  • Robert Haassaw the final versionLLM analysis, 100% sureIn #30, he checked the compression semantics against previous releases and judged #29 committable.
  • Alexander Lakhinsaw an earlier versionLLM analysis, 87% sureHe reported the committed refactor's crash with a reproducer in #24 and confirmed its regression boundary in #26, before the corrective logic in #29.
  • Peter Eisentrautsaw the final versioncommittercommitted the patch

Jan 2024

5 commits
Split some code out from MergeAttributes()
26 Jan 2024Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% suremodified before commitmoderate changesfrom patchset #221 of 1 reviewer saw the final version

Combines the two helper-extraction patches in #22, but keeps name matching and attribute-map updates in MergeAttributes(), changes helper signatures, and leaves NOT NULL deduplication to a separate commit. Peter explains the caller-side matching decision in #23.

  • Peter Eisentrautsaw the final versioncommittercommitted the patch
MergeAttributes code deduplication
26 Jan 2024Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surepart of a posted patchsetmoderate changesfrom patchset #221 of 1 reviewer saw the final version

Extracts the NOT NULL deduplication embedded in the final patch of #22, moving shared constraint handling below the inheritance branches in tablecmds.c while retaining the existing savedef variable.

  • Peter Eisentrautsaw the final versioncommittercommitted the patch
Add a const decoration
25 Jan 2024Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 99% surewritten from the discussion, no matching patch

Adds const to GetColumnDefCollation() in parse_type.c and parse_type.h, a change absent from all attachments. It supports the const ColumnDef helper parameters used when committing the extraction proposed in #22.

MergeAttributes: convert pg_attribute back to ColumnDef before comparing
25 Jan 2024Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #221 of 1 reviewer saw the final version

Applying the first two patches of #22 to the commit's parent reproduces tablecmds.c byte-for-byte. Ashutosh explicitly identifies the first patch as Peter's #21 and supplies the NOT NULL correction in the second.

  • Ashutosh Bapatsaw the final versiontrailerIn #22, he reviewed #21 and supplied the correction to mark prevdef, which was incorporated into the committed version.
Refactor ATExecAddColumn() to use BuildDescForRelation()
12 Jan 2024Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #151 of 1 reviewer saw the final version

Matches the ATExecAddColumn() patch in #15, replacing manual pg_attribute construction with BuildDescForRelation() and fixing the returned attribute number.

  • Alvaro Herrerasaw the final versiontrailerAfter questioning the single-column tuple descriptor in #3, he explicitly withdrew that objection and supported the final patch in #16.

Sep - Oct 2023

7 commits
Push attidentity and attgenerated handling into BuildDescForRelation()
5 Oct 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #131 of 1 reviewer saw the final version

The four changed lines match #13: attidentity and attgenerated assignments move from DefineRelation() into BuildDescForRelation().

  • Alvaro Herrerasaw the final versiontrailerHe called the identity/generated handling change reasonable in #3, and its later differences reflect series reordering rather than a different implementation.
Move BuildDescForRelation() from tupdesc.c to tablecmds.c
5 Oct 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #13

Matches #13, moving BuildDescForRelation() from tupdesc.c to tablecmds.c and relocating its declaration to commands/tablecmds.h.

Push attcompression and attstorage handling into BuildDescForRelation()
5 Oct 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #13

Matches the patch in #13 that moves compression and storage initialization from DefineRelation() into BuildDescForRelation().

Add TupleDescGetDefault()
27 Sep 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #131 of 1 reviewer saw the final version

Matches the TupleDescGetDefault() patch in #13 across tupdesc.c, tablecmds.c, parse_utilcmd.c, rewriteHandler.c, and tupdesc.h.

  • Alvaro Herrerasaw the final versionLLM analysis, 92% sureIn #7, he reviewed the helper's error-reporting rationale and suggested reporting relation OIDs, while the committed implementation preserved the reviewed code.
MergeAttributes() and related variable renaming
26 Sep 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #131 of 1 reviewer saw the final version

Matches the variable-renaming patch in #13, including schema-to-columns renaming in tablecmds.c and the tuple-descriptor interfaces.

  • Alvaro Herrerasaw the final versionLLM analysis, 98% sureIn #3, he said the renaming made sense while identifying conflicts with his NOT NULL work; the eventual patch retained the same changes.
Clean up MergeCheckConstraint()
26 Sep 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #131 of 1 reviewer saw the final version

Matches #13, making MergeCheckConstraint() append new constraints itself and incorporating the clarified matching/conflict comment.

  • Alvaro Herrerasaw the final versionLLM analysis, 100% sureHe reviewed the interface change in #3 and identified ambiguous comment wording in #6, which #13 corrected.
Clean up MergeAttributesIntoExisting()
26 Sep 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #131 of 1 reviewer saw the final version

Matches the MergeAttributesIntoExisting() cleanup in #13, including analogous naming and scope changes in MergeConstraintsIntoExisting() and RemoveInheritance().

  • Alvaro Herrerasaw the final versionLLM analysis, 90% sureHis test-merge review in #3 accepted the cleanup in principle while flagging conflicts with NOT NULL work, followed by rebase adjustments before #13.

Aug 2023

4 commits
Remove useless if condition
29 Aug 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #11 of 1 reviewer saw the final version

Matches the GetAttributeCompression() condition-removal patch in #1 exactly apart from hunk offsets: DefineRelation() in src/backend/commands/tablecmds.c calls the function unconditionally because it already handles NULL.

  • Alvaro Herrerasaw the final versiontrailerHe explicitly endorsed this condition removal in his patch-by-patch review #3, and the committed code is unchanged.
Remove useless if condition
29 Aug 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #11 of 1 reviewer saw the final version

Matches the patch in #1 that removes redundant conditions around attidentity and attgenerated assignments in DefineRelation().

  • Alvaro Herrerasaw the final versiontrailerHe endorsed removal of these redundant conditions in his patch-by-patch review #3.
Make more use of makeColumnDef()
29 Aug 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 99% surelanded as postedfrom patchset #11 of 1 reviewer saw the final version

Matches the makeColumnDef() conversion in #1, modulo wrapped call sites and rebasing over intervening NOT NULL handling in parse_utilcmd.c. The additional removed false initialization is supplied by makeColumnDef().

  • Alvaro Herrerasaw the final versiontrailerIn #3, he endorsed makeColumnDef() reuse, requested shorter call-site lines, and checked its conflicts with his NOT NULL branch.
Add some const decorations
29 Aug 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #1

Matches the final patch in #1, adding const to GetAttributeCompression()'s compression argument in its declaration and definition.

Jul 2023

3 commits
Remove ancient special case code for dropping oid columns
12 Jul 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #11 of 1 reviewer saw the final version

Matches #1, removing the negative-attribute special case from RemoveAttributeById() in heap.c.

  • Alvaro Herrerasaw the final versiontrailerHe commented favorably on this dead-code removal in #3, explicitly noting that he had not verified it.
Remove ancient special case code for adding oid columns
12 Jul 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #11 of 1 reviewer saw the final version

Matches #1, removing obsolete OID-related attribute-number conditions from ATExecAddColumn() and updating the MergeAttributes() comment.

  • Alvaro Herrerasaw the final versiontrailerHe explicitly approved this dead-code removal in #3.
Remove obsolete comment about OID support
12 Jul 2023Peter Eisentrautmaster17beta1REL_19_STABLE17beta1REL_18_STABLE17beta1REL_17_STABLE17beta1
LLM analysis, 100% surelanded as postedfrom patchset #11 of 1 reviewer saw the final version

Matches the first patch in #1, removing the obsolete OID exception from the AddNewAttributeTuples() comment in heap.c.

  • Alvaro Herrerasaw the final versiontrailerHe called this comment cleanup obvious and trivial and endorsed it in #3.

Credits across 20 commits

Author
Peter EisentrautLLM16
Ashutosh Bapat3
Reviewer
Alvaro Herrera8
Peter Eisentrautcommitter3
Ashutosh Bapat1
Reported by
Alexander Lakhin1
Committer
Peter Eisentraut20

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