[RCmerci]

Journal repeats graph-operation cleanup across six reducer transitions, carries unused graph metadata through its readiness notification, and parses/maps attachment selections more than once. This change removes those duplicate obligations while preserving existing behavior.

• Extract the same nineteen constant resets into one private pure helper. Keep each caller's generation handling, pending-effect filtering, cancellation/detach/publication order, and the distinct fail policy.
• Make internal Graph_ready payload-free. Preserve the actual V2_graph_info request, correlation/failure handling, task-catalog startup and separate admission inspection; worker wire formats are unchanged.
• Reuse the decoded attachment event and existing to_import mapping. Preserve public attachment entry points, typed/UUID validation, request identity, errors, batch order and replace_reference=None.

No permanent tests, dependencies, Dune, protected specs or Swift changes.

Validation

The focused build and these seven existing test commands passed both before and after:

• test/journal_routes_test.exe
• test/journal_platform_test.exe
• test/journal_graph_runtime_locality_test.exe — 152 cases
• test/application_view_test.exe — 144 cases, including admission/unlock
• test/logseq_db_worker_application_integration_test.exe — 15 cases
• logseq_sync/test/test_sync.exe test 'pure core' — 87 cases
• logseq_sync/test/test_sync.exe test 'sync recovery reproductions' — 21 cases

Temporary external harnesses compared 1,972 reducer input/public-state/context/ordered-effect records, 5,906 runtime input/request/notification records, and 74 attachment cases byte-for-byte. Baseline replay was deterministic. Runtime comparison normalizes only the intentionally removed Graph_ready argument; scenarios include success, duplicate, failure, abandonment, reset and old completions. Attachment cases cover valid/invalid inputs, exact errors, request IDs and ordered batches. No temporary harness or logs are committed.

Native Swift startup/failure boundary checks also passed (21 scenarios). Independent review of exact head 362796e found no actionable regression. git diff --check and decision-document validation pass.

Limits

Whole-file formatting checks fail on the same four ML files before and after; the MLI passes both. No broad formatting changes were made. Validation uses synthetic inputs and public observations, not opaque reducer internals or a formal equivalence proof. No full App build/full test alias, GUI, Simulator/device, benchmark or real-graph acceptance was run.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-22T11:17:43.326305Z	362796e	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.