model-writtenrun 8264 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Alvaro Herrera opened this portion of the discussion with a consolidated glossary revision that treated cluster largely as a synonym for instance, folded global SQL objects into SQL object, added storage and transaction terminology, and standardized capitalization and ordering #1.
- Justin Pryzby reviewed the definitions and wording #2, and Alvaro incorporated those suggestions, added several small clarifications and TOAST, and reported pushing the revision for beta1 while inviting definitions for additional terms #3.
- Erik Rijkers supplied nine further grammar and typo corrections after reading the committed glossary #4; these landed unchanged as eeba6c7e43, confirmed in #5.
- Jürgen Purtz proposed clearer SQL-object scope wording and changing bloat from current to relevant row versions, while arguing that instance should describe processes and memory and cluster should describe persistent storage #6; Alvaro disputed the synonym distinction initially and explained that database-local objects need not belong to schemas #7.
- Erik supported distinguishing a running instance from a persistent cluster #10, while Jürgen developed that distinction further #12; Laurenz Albe defended existing contextual usage and argued that server should mean software rather than a machine #13.
- Andrew Grillet emphasized consistent terminology across virtualization layers #14, Peter Eisentraut said careful renaming was possible #15, and Jürgen and Laurenz continued debating the ambiguous meanings of server and host without agreeing to wholesale renaming #16 #17.
- Jürgen later proposed concrete definitions that separated database clusters from running instances while retaining contextual aliases #18; Alvaro accepted that distinction and posted a patch that also clarified SQL-object examples and name uniqueness #19.
- Justin identified wording and completeness issues in that patch #20, and Jürgen supplied an incremental ordering and XML fix while identifying an incorrect cluster cross-reference #21.
- Alvaro folded in the reviews and expanded the revision to introduce routines, distinguish functions and procedures, clarify relations and result sets, repair alphabetical ordering and add TPS #22.
- Erik reviewed the expanded patch and corrected several grammar errors #23; Alvaro incorporated them alongside small additional clarifications in 816cbb59e3 and confirmed their inclusion #24.
- The proposed bloat wording change in #6 was absent from the subsequent terminology patches, following Alvaro's uncertainty about whether bloat includes free space #7; the thread records no later resolution of that proposal.
- Jürgen supplied a separate July patch adding Base Backup, process qualifiers and a Logger cleanup #25; Alvaro split it into a0b2d583db for the three small corrections and 606c384598 for the new definition.
- The thread ended with all July changes pushed: the small corrections reached master and PostgreSQL 13, while Base Backup remained master-only because Alvaro declined to add the new term to PostgreSQL 13 #26.