May 2019

3 commits
pg_upgrade: Don't use separate installation for test.
21 May 2019Andres Freundmaster12beta2REL_19_STABLE12beta2REL_18_STABLE12beta2REL_17_STABLE12beta2REL_16_STABLE12beta2REL_15_STABLE12beta2REL_14_STABLE12beta2REL_13_STABLE12beta2+1 moreAlso in: Teach pg_upgrade test to honor NO_TEMP_INSTALL (not this thread)
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #171 of 1 reviewer saw the final version

This lands the first patch from #17, with the obsolete $(DOINST) reference removed following #18. This thread's attachment at 19:52 UTC is later than the related 19:19 UTC submission in thread #40365 and matches the committed Makefile and test.sh changes.

  • Tom Lanesaw the final versiontrailerIn #18, he identified the leftover $(DOINST) reference, which Andres acknowledged in #19 and removed before committing.
pg_upgrade: Avoid check target accidentally breaking make's --output-sync.
21 May 2019Andres Freundmaster12beta2REL_19_STABLE12beta2REL_18_STABLE12beta2REL_17_STABLE12beta2REL_16_STABLE12beta2REL_15_STABLE12beta2REL_14_STABLE12beta2REL_13_STABLE12beta2+1 more
LLM analysis, 100% surepart of a posted patchsetmoderate changesfrom patchset #17

This lands the second patch from #17, introducing NOTSUBMAKEMAKE in src/bin/pg_upgrade/Makefile. Besides correcting a comment typo and accommodating the preceding DOINST cleanup, the committed recipe drops $<, removing the script argument to $(SHELL).

Make pg_upgrade's test.sh less chatty.
21 May 2019Tom Lanemaster12beta2REL_19_STABLE12beta2REL_18_STABLE12beta2REL_17_STABLE12beta2REL_16_STABLE12beta2REL_15_STABLE12beta2REL_14_STABLE12beta2REL_13_STABLE12beta2+6 moreAlso in: Broken defenses against dropping a partitioning column (not this thread)
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #111 of 1 reviewer saw the final version

The commit preserves the initdb -A trust and shell-tracing removal hunks from #11, omitting only the createdb regression workaround. The later material in thread #40898 concerns backporting this change and supplies no competing source patch.

  • Andres Freundsaw the final versionLLM analysis, 85% sureIn #12, he reviewed the cleanup and proposed suppressing notices in pg_regress instead of creating the regression database, a workaround omitted from the committed patch.

Credits across 3 commits

Author
Andres Freund1 LLM2
Tom LaneLLM1
Reviewer
Tom Lane1
Committer
Andres Freund2
Tom Lane1

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