ORE Studio Sprint 19 – Release Notes

• Operational pillars: compass db / services / client / sql replace the recreate/setup/run_sql/start/stop/status scripts; callers (CI, dashboard, recipes, manual) all repointed.
• compass bearings shows the environment: preset, DB restore time with drift signalling in the house staleness UX, service and client state with start hints, and a top-commands cheat sheet.
• PR lifecycle: pr create derives traceability from the branch's task; pr merge refuses on unresolved threads or red CI, closes the task and stamps the journal on success, and --force admin-merges past branch protection (required checks that never report on doc-only PRs).
• sprint audit (five checks) and sprint charts (four PNGs from one command) close the loop between the agile record and reality.

• clang-format everywhere: 16 per-group format passes, CMake format targets, nightly enforcement workflow.
• Components regrouped: projects/ores.<group>/ owns api/core/ service/qt children, modeling, and CMake; codegen validated against the new layout.
• rfl complexity: all nine rates instruments plus equity, FX, bond/ credit/commodity, complex, scripted/composite decomposed; shared audit_record and instrument_identity extracted (now in ores.dq.api).
• API/core split: inventory across 17 components found one systemic offender (ores.dq.api linking ores.database) — fixed; change- reason constants homed; ores.eventing split into api (domain, bus, registry) and core (postgres event source).

• Commission: currency reached 6 of 8 tasks — codegen, SQL, CLI, and shell verified post-NATS; currency NATS message reference and manual chapter work in flight (carried with its branch).
• UI responsiveness: destructor waitForFinished() fixes, spinners, pre-populated combos, dedicated I/O thread pool.
• DB-driven badges with colour semantics (gray = inactive only; orange = missing definition) across list and detail dialogs.
• History dialogs: server-side diffs via the new ores.diff component, HistoryDialogBase adopted, shell unified diff view (5 of 7 tasks; remainder carried).
• PnL tracking improvements captured with the Lightyear knowledge doc (FX locking, timeframe-scoped PnL, FIFO/WAVG).

• Agile process restructured into per-concept pages (stories, tasks, planning, execution, closure, definition of done) with agile-domain context before the ORE-specific conventions.
• Sprint health charts documented as literate org tangling the gnuplot sources; runbooks established as the procedure layer above recipes and skills.
• User manual: chapter 3/4 redesign (Tufte-informed provenance and change-reason display), currency chapters, fresh screenshots, PDF pipeline fixes.
• Knowledge: financial identifiers, PostgreSQL architecture updates, cybernetics pages wired into readme org-roam links.

Dual-axis bar chart. PRs (left axis) and commits (right axis) per day. A high commits-to-PR ratio may indicate scope creep.

Two distinct waves: a build-up peaking on 06-01 (35 PRs, ~205 commits) as the clang-format groups and codegen refactors landed, a mid-week dip on 06-03, then the sprint's biggest day on 06-05 — 48 PRs and ~310 commits — driven by the compass PR-lifecycle work and the paperwork syncs it enabled. Commits track PRs closely all sprint: small, frequent PRs rather than scope creep. The 06-06 tail-off is the close-down itself.

Lines added (green) and deleted (red) per day. Building work produces mostly additions; refactoring produces a mix.

One day dominates: 06-05 touched ~217k lines, and deletions (~157k) far outweigh additions — the profile of the format passes, the component regroup fallout, and eleven operational scripts plus the legacy generator being deleted. 06-02 is also deletion-heavy (~35k). This is a refactoring sprint's signature: the codebase ended smaller and more uniform, not larger.

Hours from PR open to merge, one bar per PR. Long bars indicate review bottlenecks.

Most sampled PRs merge in 4–8 hours. The outliers are early: #941 (~27h) and #947 (~23h) predate the compass PR loop; #1038 (~15h) waited on a review round. From #1039 onward the bars flatten to a uniform ~4–7h — the guarded pr merge visibly tightened the cycle.

Line chart tracking stories marked DONE during the sprint. Steady upward slope is healthy; plateauing signals a stall.

A steady ~1.5 stories/day through 06-04 (13 done), then a jump to 23 on 06-05. The jump is not a delivery burst: it is health reviews 3 and 4 syncing the record — closing stories whose work had already merged. The honest reading of the slope is steady mid-size delivery with the paperwork catching up at the end; the audit now keeps that lag from accumulating again.