Nov 2019

1 commit
Remove debugging aid
23 Nov 2019Alvaro Herreramaster13beta1REL_19_STABLE13beta1REL_18_STABLE13beta1REL_17_STABLE13beta1REL_16_STABLE13beta1REL_15_STABLE13beta1REL_14_STABLE13beta1REL_13_STABLE13beta1
LLM analysis, 100% surewritten from the discussion, no matching patch

The commit removes only the unintended Assert(false) from errbacktrace() in src/backend/utils/error/elog.c, exactly the correction described and tested by Tom Lane in #26 and acknowledged by Álvaro in #27. No attached patchset contains this follow-up removal.

Nov 2019

1 commit
Add backtrace support for error reporting
8 Nov 2019Alvaro Herreramaster13beta1REL_19_STABLE13beta1REL_18_STABLE13beta1REL_17_STABLE13beta1REL_16_STABLE13beta1REL_15_STABLE13beta1REL_14_STABLE13beta1REL_13_STABLE13beta1Also in: PATCH: backtraces for error messages
LLM analysis, 100% suremodified before commitmoderate changesfrom patchset #211 of 5 reviewers saw the final version

The commit matches #21, with sizeof(buf) corrected to lengthof(buf) in assert.c and elog.c, tab handling corrected in guc.c, diagnostic fprintf removed, and comments expanded; its Authors credit identifies Peter Eisentraut and Álvaro Herrera. This September 2019 attachment is later than the June 2018 patch listed for thread #38893, and supplies the committed GUC parsing implementation.

  • Jaime Casanovasaw an earlier versionLLM analysis, 98% sureJaime compiled #1 with libunwind and reported unresolved-symbol failures in #4, before libunwind was removed and function-list handling was revised.
  • Thomas Munrosaw an earlier versionLLM analysis, 100% sureThomas tested macOS and FreeBSD, identified the missing execinfo linkage in #5, and confirmed the revised patch worked on FreeBSD in #13, before the later function-list changes.
  • Dmitry Dolgovsaw an earlier versionLLM analysis, 96% sureDmitry reproduced the libunwind problem and demonstrated successful backtrace output after defining UNW_LOCAL_ONLY in #6, reviewing an earlier implementation whose libunwind backend was subsequently dropped.
  • Tom Lanesaw an earlier versionLLM analysis, 100% sureTom examined assertion-output buffering in #11 and #12, then identified the process-wide strtok-state hazard in #19, which the subsequent GUC parsing revision removed.
  • Alvaro Herrerasaw the final versioncommittercommitted the patch

Credits across 2 commits

Author
Peter EisentrautLLM1
Reviewer
Alvaro Herreracommitter1
Committer
Alvaro Herrera2

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