model-writtenrun 8647 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
Oct - Nov 2019 · 2 entries
- Corey Huinker proposed a glossary appendix to help newcomers and non-native English speakers understand database terminology, search for help, and cite definitions, while explicitly requesting technical and editorial review of the initial draft #1.
- Fabien COELHO supported the idea but questioned duplication and maintenance effort, requested links between definitions and into the manual, and identified markup problems #2; after a pause, Corey returned with a draft revised with technical writer Roger Harkavy #4.
Feb - May 2020 · 19 entries
- Corey merged Jürgen Purtz's complementary administration and internals definitions and added stable identifiers for individual terms #5, after which Jürgen supplied a correction patch that also raised questions about schema naming and transaction semantics #7.
- Corey and Jürgen discussed exceptions to schema-name uniqueness and how much transaction detail belonged in concise definitions #9 #11, and Corey incorporated the resulting edits into a consolidated draft #12.
- Alvaro Herrera took up the patch, reformatted it, replaced explicit link titles with standard xref markup, and requested ACID definitions #13; his next delta split aggregate terminology, reduced repeated markup, and temporarily hid questionable entries #16.
- Justin Pryzby supplied a detailed review covering process names, foreign keys, storage terminology, locks, schema objects, and many wording issues #19 #22, while Jürgen emphasized the need to distinguish historically inconsistent uses of server, host, instance, and cluster #24.
- Peter Eisentraut reinforced the use of standard xref links #26, and Robert Haas clarified that WAL buffers and shared_buffers are separate allocations within shared memory #30.
- Corey consolidated the review into a draft checked by Roger, adding the four ACID properties, Tablespace, a separate Vacuum definition, and WAL Archiver, while leaving disputed process suffixes and host/server terminology unresolved #33.
- Jürgen supplied another wording delta #34, and Justin reviewed both that delta and the full draft, questioning the separate Aggregating entry and requesting clearer definitions and consistent terminology #35 #36.
- Alvaro then rewrote the WAL-related definitions, introduced DocBook glosssee and glossseealso markup, removed identifiers from redirect-only entries, and added process labels and a Window frame definition #41.
- Justin proposed additional query-related terminology #42; Alvaro added Query and Query Planner, revised the ACID definitions, and redirected Master to Primary, while explicitly deferring Normalized because of its scope #43 #46.
- Corey added missing glossary links and verified that the documentation built #48, then Alvaro merged that work while replacing Unique Index with Unique Constraint, correcting global-object terminology, and revising Client, Connection, Session, and window-function definitions #49.
- Justin challenged the remaining host/server ambiguity #50, and Alvaro committed the glossary as 347d2b07fc from #49 with further definition changes, additional entries, and capitalization cleanup.
- Alvaro announced that landing as a starting point and explained that Serializable had an inaccurate definition, Sequence was too detailed, and exclusion and NOT NULL constraints still needed coverage #51.
- Erik Rijkers immediately supplied wording and capitalization corrections #54, which Alvaro turned into a follow-up patch #55; Justin requested consistent SQL articles across the glossary #56, and Erik proposed an ACID definition #57.
- Laurenz Albe proposed clearer SQL Object and Null definitions, a Standby alias, and reconsideration of server, instance, role, and function terminology #58, while Jürgen argued for distinguishing processes and memory from the stored database cluster #60.
- Fabien and Corey discussed linking the rest of the manual to glossary entries, automatic hover definitions, acronym improvements, and possible backpatching #59 #61 #62, but Alvaro declined to undertake the stylesheet work and expressed little enthusiasm for backpatching #65.
- Jürgen submitted an independent patch to restore alphabetical ordering and fix two WAL writer links #63, which Alvaro accepted for a future push #64.
- Jürgen subsequently requested that sequences be explicitly listed as relations #66 #68, and Peter questioned capitalization of glossary terms #69.
- Alvaro committed the follow-up work as 756abe2bc7, combining #63 with selected corrections from #55 and discussion contributions, while adding ACID and storage terminology, restoring a concise Sequence definition, and consolidating Cluster and global-object explanations into other entries.
- By the final message, both the glossary and the expanded cleanup had landed, but Peter clarified that his objection concerned initial-letter capitalization, leaving that question unresolved #71; broader manual linking and several omitted definitions remained follow-up work identified in #51 and #61.