Dec 2021

1 commit
Ensure casting to typmod -1 generates a RelabelType.
16 Dec 2021Tom Lanemaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1REL_14_STABLE14.2REL_13_STABLE13.6+3 more
LLM analysis, 100% surewritten from the discussion, no matching patch0 of 1 reviewer saw the final version

John Naylor reported broken partition pruning #7, and Tom Lane identified an unwanted bpchar() coercion call #8, without another attached patch. The commit bypasses coercion-function lookup for negative typmods in src/backend/parser/parse_coerce.c and adds bpchar view and index-scan regression tests.

  • John Naylorsaw an earlier versionLLM analysis, 90% sureJohn bisected a partition-pruning regression to the landed version of #2 and supplied a reproducer and execution plans #7, prompting the subsequent coercion-logic correction.

Aug 2021

1 commit
Don't elide casting to typmod -1.
6 Aug 2021Tom Lanemaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1REL_14_STABLE14beta3REL_13_STABLE13.5+4 moreAlso in: So, about that cast-to-typmod-minus-one business
LLM analysis, 100% surelanded as postedfrom patchset #21 of 1 reviewer saw the final version

The commit diff exactly matches #2 after removing index-hash metadata, including src/backend/parser/parse_coerce.c and both expressions regression files.

  • Dean Rasheedsaw the final versionLLM analysis, 87% sureDean endorsed the fix and explained how it would resolve an existing recursive-CTE typmod mismatch for which the suggested cast was ineffective #3.

Credits across 2 commits

Author
Tom LaneLLM1
Committer
Tom Lane2

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