Mar 2019

2 commits
Use FullTransactionId for the transaction stack.
28 Mar 2019Thomas Munromaster12beta1REL_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 #401 of 1 reviewer saw the final version

Commits the transaction-stack component of the two-patch series in #40, including its SerializedTransactionState implementation. Relative to that component, the only code edit replaces the StartTransaction assertion with the equivalent !FullTransactionIdIsValid check in src/backend/access/transam/xact.c.

  • Heikki Linnakangassaw the final versiontrailerHeikki commented on the transaction-stack proposal in #37 and specifically welcomed the clearer serialization implementation while reviewing the final series in #41.
Add basic infrastructure for 64 bit transaction IDs.
28 Mar 2019Thomas Munromaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 99% surepart of a posted patchsetmoderate changesfrom patchset #401 of 3 reviewers saw the final version

Commits the infrastructure component of #40, with comment edits and restoration of the single-process exception in AdvanceNextFullTransactionIdPastXid's assertion in src/backend/access/transam/varsup.c. That exception fixes Heikki's single-user recovery failure reported in #41, and Thomas confirms the fix and push in #42.

  • Andres Freundsaw an earlier versiontrailerAndres proposed the struct wrapper in #12, reviewed conversions and increment code in #17, and challenged epoch inference in #34, after which #35 substantially rewrote the recovery helper.
  • Tom Lanesaw an earlier versiontrailerTom examined wrapper representation and cross-platform ABI behavior in #13, #24 and #26, and naming in #32, before the recovery-helper rewrite in #35.
  • Heikki Linnakangassaw the final versiontrailerHeikki tested the final series and exposed the single-user WAL-replay assertion failure in #41, which was corrected before committing as confirmed in #42.

Credits across 2 commits

Author
Thomas Munro2
Reviewer
Heikki Linnakangas2
Andres Freund1
Tom Lane1
Reported by
Amit Kapila1
Committer
Thomas Munro2