Sep 2022

2 commits
pgstattuple: Use double consistently for percentages
12 Sep 2022Peter Eisentrautmaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1
LLM analysis, 97% suremodified before commitmoderate changesfrom patchset #21 of 1 reviewer saw the final version

The commit changes scanned_percent from uint64 to double in contrib/pgstattuple/pgstatapprox.c, matching the pgstattuple fix described in #2. Peter explicitly says in #3 that he additionally changed the multiplier from 100 to 100.0 to produce fractional results; the attachment itself is unavailable because extraction failed.

  • Peter Eisentrautsaw the final versiontrailerPeter committed the fix and explained his additional floating-point arithmetic correction in #3.
Use float8 datatype for percentiles in pg_walinspect stat functions
12 Sep 2022Peter Eisentrautmaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1REL_15_STABLE15rc1
LLM analysis, 78% surelanded as postedfrom patchset #21 of 1 reviewer saw the final version

The commit changes pg_walinspect SQL outputs and documentation from float4 to float8 and uses Float8GetDatum(), matching the attached fix described in #2, which Peter reports applied in #3. No edits to this fix are reported, but failed attachment extraction prevents direct verification of patch equality.

  • Peter Eisentrautsaw the final versiontrailerPeter identified the precision mismatch in #1 and committed the submitted fix, confirming completion in #3.

Credits across 2 commits

Author
Bharath Rupireddy2
Reviewer
Peter Eisentraut2
Reported by
Peter Eisentraut1
Committer
Peter Eisentraut2