Oct 2022

1 commit
Fix tiny memory leaks
1 Oct 2022Peter Eisentrautmaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1
LLM analysis, 100% surelanded as postedfrom patchset #462 of 2 reviewers saw the final version

Both src/backend/commands/variable.c hunks match #46 exactly after normalizing diff metadata, including cleanup on successful and failed guc_strdup() calls.

  • Jacob Championsaw the final versiontrailerJacob explicitly said the patch looked good in #47 and discussed its memory-context implications in #49.
  • Peter Eisentrautsaw the final versioncommittercommitted the patch

Sep 2022

2 commits
Don't reflect unescaped cert data to the logs
13 Sep 2022Peter Eisentrautmaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #411 of 3 reviewers saw the final version

Takes the complete certificate-escaping patch from the two-patch series in #41; its code, certificates, CRLs, and tests match modulo rebased include context, while the helper change lands separately.

  • Andres Freundsaw an earlier versionLLM analysis, 98% sureAndres rejected SSL-local escaping in #31 and criticized the two-pass API in #36, prompting the shared-helper and allocation redesigns.
  • Tom Lanesaw an earlier versionLLM analysis, 96% sureTom supported making certificate and existing name sanitization behave alike in #33, before the shared-helper redesign.
  • Peter Eisentrautsaw the final versioncommittercommitted the patch
pg_clean_ascii(): escape bytes rather than lose them
13 Sep 2022Peter Eisentrautmaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #411 of 3 reviewers saw the final version

Takes exactly the pg_clean_ascii() patch from the two-patch series in #41, including allocation flags and the postmaster/GUC callers; normalized diffs match, with certificate integration committed separately.

  • Andres Freundsaw an earlier versionLLM analysis, 100% sureAndres proposed returning an allocated string and using guc_strdup() in #36, after which allocation and failure-handling logic changed.
  • Tom Lanesaw an earlier versionLLM analysis, 100% sureTom explained in #38 and #40 that throwing ERROR from these GUC hooks could terminate the postmaster, leading to the nonthrowing allocation changes.
  • Peter Eisentrautsaw the final versioncommittercommitted the patch

Jul 2022

1 commit
Log details for client certificate failures
15 Jul 2022Peter Eisentrautmaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #181 of 2 reviewers saw the final version

Combines #15 with its top-up #18; applying both to the commit's parent leaves only a new cert_errdetail comment and blank-line cleanup in be-secure-openssl.c. The commit explicitly credits Jacob Champion despite Peter Eisentraut sending the final attachments.

  • Graham Leggettsaw an earlier versionLLM analysis, 83% sureGraham proposed RFC 4523 formatting in #8 and identified troubleshooting problems with truncation and ad hoc formatting in #14, before the final error-storage rework.
  • Peter Eisentrautsaw the final versioncommittercommitted the patch

Credits across 4 commits

Author
Jacob Champion3
Masahiko Sawada1
Reviewer
Peter Eisentrautcommitter4
Jacob Champion1
Committer
Peter Eisentraut4