Jul - Aug 2022

3 commits
Add more TAP tests with BASE_BACKUP and pg_backup_start/stop
1 Aug 2022Michael Paquiermaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1
LLM analysis, 100% surelanded as postedfrom patchset #170 of 1 reviewer saw the final version

The complete 55-line addition to src/test/recovery/t/001_stream_rep.pl matches #17 exactly after ignoring diff metadata; #18 confirms application.

  • Fujii Masaosaw an earlier versiontrailerIn #14 and #16, he reviewed #13, suggested progress-phase synchronization, identified connection-string and comment issues, and requested broader coverage; synchronization changes and an additional test followed.
Fix assertion failure and segmentation fault in backup code.
20 Jul 2022Fujii Masaomaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1REL_15_STABLE15beta3REL_14_STABLE14.5REL_13_STABLE13.8+3 more
LLM analysis, 94% surepart of a posted patchsetlight changesfrom patchset #144 of 4 reviewers saw the final version

Commits the state-reset component of the split series described in #14; its src/backend/access/transam/xlog.c hunk exactly matches #7 modulo rebase metadata. The split submission's attachments were not extracted, but #16 confirms both fixes were pushed.

  • Kyotaro Horiguchisaw the final versiontrailerIn #2, he examined the callback problem in #1 and proposed an alternative implementation; the committed state-reset hunk remained unchanged.
  • Masahiko Sawadasaw the final versiontrailerIn #6, he reported that #1 looked good and that testing confirmed the fixes; the state-reset component subsequently landed unchanged.
  • Michael Paquiersaw the final versiontrailerIn #13, he reported a regression test that crashed without the fix and passed with it, then accepted the split in #15.
  • Robert Haassaw the final versiontrailerIn #12, he analyzed the state-reset fix separately from the overlapping-backup guard and recommended distinguishing the two bugs.
Prevent BASE_BACKUP in the middle of another backup in the same session.
20 Jul 2022Fujii Masaomaster16beta1REL_19_STABLE16beta1REL_18_STABLE16beta1REL_17_STABLE16beta1REL_16_STABLE16beta1REL_15_STABLE15beta3REL_14_STABLE14.5REL_13_STABLE13.8+3 more
LLM analysis, 94% surepart of a posted patchsetlight changesfrom patchset #142 of 4 reviewers saw the final version

Commits the overlapping-backup guard component of the split series described in #14; its SendBaseBackup() hunk in src/backend/replication/basebackup.c exactly matches #7 modulo diff metadata. The split attachments are unavailable, but #16 confirms the landing.

  • Kyotaro Horiguchisaw an earlier versiontrailerHe examined #1 and incorporated its duplicate-backup guard into his alternative in #4, before the guard moved to SendBaseBackup() to avoid sink initialization.
  • Masahiko Sawadasaw an earlier versiontrailerIn #6, he tested #1 and proposed moving the guard to SendBaseBackup(), a subsequent error-path change implemented in #7.
  • Michael Paquiersaw the final versiontrailerIn #13, he explicitly accepted grouping the two fixes, then accepted their separation in #15, with no subsequent guard changes.
  • Robert Haassaw the final versiontrailerIn #12, he explained how mixed SQL and replication commands permit overlapping backups and recommended identifying this guard as a separate fix.

Credits across 3 commits

Author
Fujii Masao2
Michael PaquierLLM1
Reviewer
Kyotaro Horiguchi2
Masahiko Sawada2
Michael Paquier2
Robert Haas2
Fujii Masao1
Committer
Fujii Masao2
Michael Paquier1

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