Mar - Apr 2022

6 commits
Fix aggregate logging of pgbench.
19 Apr 2022Tatsuo Ishiimaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surelanded as postedfrom patchset #2001 of 2 reviewers saw the final version

Applying #200 to the commit's parent reproduces both committed files exactly, removing the aggregate failures column and updating its documentation.

  • Tom Lanesaw an earlier versionLLM analysis, 88% sureIn #196, he inspected the landed aggregate implementation from #194 and proposed removing the redundant total, before the removal patch was submitted.
  • Fabien COELHOsaw the final versionLLM analysis, 100% sureHe reported successful compilation, global and local checks, and documentation generation for #200 in #201.
Doc: reformat descriptions of pgbench output formats.
10 Apr 2022Tom Lanemaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1962 of 2 reviewers saw the final version

Compared with #196, the committed doc/src/sgml/ref/pgbench.sgml adds 'transaction start delay' to the schedule_lag definition and removes the example command's newline, as discussed in #198 and #199.

  • Tatsuo Ishiisaw the final versionLLM analysis, 100% sureHe explicitly said the field-list rewrite in #196 looked good in #197.
  • Fabien COELHOsaw the final versionLLM analysis, 100% sureHe built the HTML documentation and requested the adopted example-line adjustment in #198.
Change aggregated log format of pgbench.
6 Apr 2022Tatsuo Ishiimaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surelanded as postedfrom patchset #1941 of 3 reviewers saw the final version

Applying #194 to the commit's parent exactly reproduces src/bin/pgbench/pgbench.c and doc/src/sgml/ref/pgbench.sgml, including fixed aggregate columns and the revised ordering.

  • Alvaro Herrerasaw an earlier versionLLM analysis, 88% sureHe criticized the aggregate-format documentation in #181 and proposed the adopted fixed-column, zero-fill design in #189 before its implementation.
  • Tom Lanesaw an earlier versionLLM analysis, 92% sureHe identified ambiguity in the landed optional-column format in #182 and endorsed zero-filled fixed columns in #191.
  • Fabien COELHOsaw the final versionLLM analysis, 94% sureHe reviewed #192 and requested the column ordering adopted by #194 in #193.
Second attempt to fix pgbench TAP test.
25 Mar 2022Tatsuo Ishiimaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surelanded as postedfrom patchset #1692 of 2 reviewers saw the final version

Applying #169 to the commit's parent exactly reproduces the TAP test, replacing both unsupported Perl backreferences.

  • Tom Lanesaw the final versiontrailerHe caught the remaining unsupported backreference in #168 and confirmed the corrected patch worked with Perl 5.8.8 in #170.
  • Yugo Nagatasaw the final versiontrailerHe tested the final correction and reported success in #171.
Fix pgbench TAP test.
24 Mar 2022Tatsuo Ishiimaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surelanded as postedfrom patchset #1642 of 2 reviewers saw the final version

Applying Yugo Nagata's #164 to the commit's parent exactly reproduces the capture-expression restoration in src/bin/pgbench/t/001_pgbench_with_server.pl.

  • Tom Lanesaw the final versionLLM analysis, 96% sureHis buildfarm failure report in #163 exposed the broken retry test from #160, followed by the small test-only correction in #164.
  • Tatsuo Ishiisaw the final versioncommittercommitted the patch
Allow pgbench to retry in some cases.
23 Mar 2022Tatsuo Ishiimaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surelanded as postedfrom patchset #1601 of 9 reviewers saw the final version

Applying both patches in #160 to the commit's parent reproduces all six committed files exactly. Sender Yugo Nagata authored the final patches, and the commit also credits Marina Polyakova's original work.

  • Fabien Coelhosaw an earlier versiontrailerHe repeatedly tested and redesigned the feature through #152, but #153 subsequently changed buffer lifetime, pipeline-helper structure, and reporting logic before the final documentation revision.
  • Tatsuo Ishiisaw the final versiontrailerHe investigated timeout and pipeline behavior, reviewed the final revisions, approved #160 in #161, and committed the result.
  • Alvaro Herrerasaw an earlier versiontrailerHe advocated retries, requested the preparatory split and logging changes, and reviewed the series through #114 before the later error-handling and pipeline rework.
  • Kevin Grittnersaw an earlier versiontrailerHe explained whole-transaction retries, successful-throughput accounting, and reevaluation of dependent statements in #5 and #12.
  • Andres Freundsaw an earlier versiontrailerHe recommended PGOPTIONS instead of an isolation-level flag in #4 and test-drove the retry implementation in #38.
  • Arthur Zakirovsaw an earlier versiontrailerHe reviewed logging-function naming and recursion in #88 and #90, before that component became optional and was eventually superseded.
  • Alexander Korotkovsaw an earlier versiontrailerHe discussed isolation-option usability in #19 and distinguished retryable failures from ordinary SQL errors in #39.
  • Teodor Sigaevsaw an earlier versiontrailerHe challenged transaction-boundary complexity in #43 and questioned retry limits and variable restoration in #48.
  • Ildus Kurbangalievsaw an earlier versiontrailerHe reported successful functional testing and requested logging, naming, and time-calculation improvements in #52.

Nov 2018

1 commit
pgbench: introduce a RandomState struct
16 Nov 2018Alvaro Herreramaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #1012 of 2 reviewers saw the final version

Takes only the RandomState component of the four-patch series in #101. Changes in src/bin/pgbench/pgbench.c rename state members, move initRandomState(), and adjust style and base context, consistent with the commit announcement in #112.

  • Fabien Coelhosaw the final versiontrailerHe tested the RandomState patches, shaped independent thread seeds, and confirmed the final submitted component was satisfactory in #102.
  • Alvaro Herrerasaw the final versioncommittercommitted the patch

Credits across 7 commits

Author
Tatsuo Ishii2 LLM3
Yugo Nagata1 LLM2
Marina Polyakova1
Tom LaneLLM1
Reviewer
Alvaro Herreracommitter2
Fabien COELHO2
Tatsuo Ishiicommitter2
Alexander Korotkov1
Andres Freund1
Arthur Zakirov1
Ildus Kurbangaliev1
Kevin Grittner1
Teodor Sigaev1
Tom Lane1
Yugo Nagata1
Reported by
Tom Lane2
Committer
Tatsuo Ishii5
Alvaro Herrera1
Tom Lane1

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