Oct 2020

4 commits
Review format of code generated by PerfectHash.pm
21 Oct 2020Michael Paquiermaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #281 of 1 reviewer saw the final version

The PerfectHash.pm diff exactly matches #28, including right padding and removal of the final comma. Michael Paquier revised John's patch from #27, and the commit explicitly credits John as author.

  • Michael Paquiersaw the final versioncommittercommitted the patch
Fix compilation warning in unicode_norm.c
12 Oct 2020Michael Paquiermaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% surelanded as postedfrom patchset #181 of 1 reviewer saw the final version

Both unicode_norm.c hunks exactly match #18: include port/pg_bswap.h and replace htonl(ch) with pg_hton32(ch).

  • Masahiko Sawadasaw the final versionLLM analysis, 100% sureAfter reporting the FreeBSD warning, he tested the attached fix and confirmed that it resolved the problem in #19.
Use perfect hash for NFC and NFKC Unicode Normalization quick check
11 Oct 2020Michael Paquiermaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 100% suremodified before commitlight changesfrom patchset #122 of 2 reviewers saw the final version

Matches the combined quick-check patch in #12, with unicode_norm_info renamed to pg_unicode_norminfo, a typedefs.list entry, and minor comment and spacing changes. The lookup logic and generated hash-table values are unchanged.

  • Mark Dilgersaw the final versiontrailerHe examined the NFC/NFKC implementation and compared compiler output with and without the wrapper structure in #8; only light edits and patch consolidation followed.
  • Michael Paquiersaw the final versiontrailerHe tested performance with MSVC and gcc in #11, prepared the combined patch in #12, and reviewed the final version as committer.
Improve set of candidate multipliers for perfect hash function generation
8 Oct 2020Michael Paquiermaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 98% surepart of a posted patchsetsignificant changesfrom patchset #62 of 2 reviewers saw the final version

Takes only the multiplier prerequisite from the three-patch series in #6, with its two PerfectHash.pm replacements unchanged. The NFC and NFKC implementation patches were retained for the separate submission #12.

  • Mark Dilgersaw the final versiontrailerHe checked generated keyword-table sizes in #4 and compared assembly before and after the multiplier patch in #7.
  • Michael Paquiersaw the final versiontrailerHe assessed the table-size effect and inspected gcc assembly in #11, then reviewed and committed the unchanged multiplier patch.

Credits across 4 commits

Author
John Naylor3
Michael PaquierLLM1
Reviewer
Michael Paquiercommitter3
Mark Dilger2
Reported by
Masahiko Sawada1
Peter Eisentraut1
Committer
Michael Paquier4

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