Repository navigation
Release 0.6.0 (M10, identity, lifecycle and visibility) - #64
Merged
Merged
Conversation
CHANGELOG's 0.6.0 section is final (dated 2026-10-24) and reads as release notes in 0.5.0's order: an introduction to M10, then upgrading, behaviour changes, new, fixes and performance. It now covers every M10 pull request (#53-#63): text node ids (#56, #58, #60), set_owner (#55), dropped roles (#57), quotas (#59), observability and progress (#61), the 10M-node soak (#62) and rows made as they are returned (#63). The upgrade list names everything sql/pg_warren--0.5.0--0.6.0.sql does: twelve functions created, the quota's column, table and index, and the library's eleven PL/pgSQL functions replaced to report progress; text ids need no catalog change. Performance records the soak with IRIs against bigint ids, the build's parallel edge intake (79 s against 0.5.0's 173 s), 0.5.0's Louvain and Leiden fixes run at 10M nodes, and the lazy rows' peaks. README's Status describes 0.6.0 as released, with its caveats (walk and neighbours still build their rows first; other backends' rows need the preload on 15 and 16); the install checks out v0.6.0; "Not there yet" starts at M11. docs/semantics.md documents text ids, and docs/limits.md that warren.threads also maps a refresh's edge endpoints. DESIGN §13 marks M10 done, and §5.7 steps 4 and 10. No code or SQL changes: this is the commit to tag v0.6.0. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012Rxfe2L3iVjSwSBPdUu1Ke
sql/pg_warren--0.6.0.sql is frozen by sql/released.sha256, and sql/released.commits records the commit to tag v0.6.0, 7d12900, for 0.6.0. The workspace version is 0.7.0 (in development) with its generated sql/pg_warren--0.7.0.sql (the same entities as 0.6.0's, in pgrx's order). sql/pg_warren--0.6.0--0.7.0.sql is a header only, since the schema has not changed; cargo pgrx install installs it and ALTER EXTENSION UPDATE runs it. CHANGELOG has 0.7.0, and README's Status names what it is for (M11, proof at scale and in deployment, DESIGN.md §13). test/upgrade.py, which reads the releases from sql/released.sha256 and sql/released.commits, now checks six release paths: 0.1.0 to 0.5.0 chained through 0.6.0, and 0.6.0 -> 0.7.0, each seeded by that release's own build (0.6.0's from 7d12900). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012Rxfe2L3iVjSwSBPdUu1Ke
The owner's decision: M11's first objective is a benchmark and test suite on a range of real graph datasets, sourced through keepset (https://mearman.github.io/keepset/), whose per-dataset recipes fetch each dataset and convert it to a common shape. The suite lives in bench/corpus, fetches by recipe and pins what it fetched, and covers social graphs, road networks and knowledge graphs, integer and IRI ids, directed and undirected, weighted and unweighted, from small graphs to past 10M nodes. Each dataset runs oracle checks (NetworkX or igraph) where feasible, the §12 matrix and Appendix A's use cases at 10M nodes. The head-to-head with igraph and Neo4j stays, on graphs from the same suite. It replaces the SNAP and Appendix A bullets. Cross-references follow: §12's SNAP sentence points at the suite (its recipes can include those SNAP graphs), §5.7's front-coding candidate is the suite's IRI-keyed knowledge graphs rather than a slice of Wikidata or DBpedia, and README's 0.7.0 Status and "Not there yet" name the suite. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012Rxfe2L3iVjSwSBPdUu1Ke
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012Rxfe2L3iVjSwSBPdUu1Ke
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Release 0.6.0. Merge with a merge commit, so the tag commit keeps its sha.
Tag
v0.6.0on 7d12900 ("release: 0.6.0"). It changes docs only, with the workspace still at 0.6.0, and is recorded insql/released.commits. 0.5.0 was tagged the same way, on 5f4176f.docs/semantics.mdgains a "Text ids" section.sql/pg_warren--0.6.0.sql(released.sha256,released.commits);pg_warren--0.6.0--0.7.0.sql.bench/corpus(DESIGN §12, §13, §5.7, README).Local checks:
schema-drift.sh: clean.run.pyPG16, all three modes: 62/62.upgrade.py: every release from 0.1.0 to 0.7.0, with 0.6.0 seeded from 7d12900.dump.py: ok.Generated by Claude Code