79: Integrate NenjimProcessManager with NenjimHub runtime status

openspec/specs/nenjim-component-registry/spec.md

@@ -138,11 +138,11 @@ Public component contracts SHALL live in `com.r35157.nenjim.component`; the cano
 - **THEN** exactly one canonical `com.r35157.nenjim.process.NenjimProcess` contract exists and no `NenjimApplication` type remains
 
 ### Requirement: Deferred runtime features remain absent
-The Registry SHALL represent one catalogue of already constructed components and SHALL not implement Contexts, artifact-version identity, resolution, classloading, automatic discovery, contributor APIs, public runtime loading or registration, removal, persistence, events, subscribers, user configuration, automatic selection, automatic start-all, ownership tracking, usage counts, permissions, scopes, filtered views, or a common stop lifecycle. The Hub SHALL not add configurable autorun, process-manager integration, stop, restart, upgrade, health monitoring, recovery, or context-bound dependency access.
+The Registry SHALL represent one catalogue of already constructed components and SHALL not implement Contexts, artifact-version identity, resolution, classloading, automatic discovery, contributor APIs, public runtime loading or registration, removal, persistence, events, subscribers, user configuration, automatic selection, automatic start-all, ownership tracking, usage counts, permissions, scopes, filtered views, or a common stop lifecycle. Except for the Process Manager's read-only polling of the existing started-process snapshot, the Hub SHALL not add configurable autorun, lifecycle events, subscriptions, process controls, stop, restart, upgrade, health monitoring, recovery, or context-bound dependency access.
 
 #### Scenario: Consumer inspects the first Registry API
-- **WHEN** a consumer examines the public Registry, Registry-manager, process, and Hub contracts
-- **THEN** only read-only Registry lookup, passive registration, minimal process startup, and Hub startup/snapshot operations are available and no deferred feature is exposed
+- **WHEN** a consumer examines the public Registry, Registry-manager, process, Hub, and Process Manager contracts
+- **THEN** only read-only Registry lookup, passive registration, minimal process startup, Hub startup/snapshots, and the Process Manager role are available and no deferred runtime feature is exposed
 
 ### Requirement: Components and processes have minimal orthogonal contracts
 `NenjimComponent` SHALL remain a pure marker with no component-ID or lifecycle property. The one canonical `com.r35157.nenjim.process.NenjimProcess` SHALL extend `NenjimComponent` and SHALL expose only `start() throws Exception` as its common lifecycle operation. Service and application SHALL remain independent domain or orchestration roles: an implementation in either role SHALL additionally implement `NenjimProcess` only when it has a Hub-managed active startup lifecycle. Passive components SHALL not implement `NenjimProcess`, and no `NenjimApplication` compatibility type SHALL remain.
@@ -186,15 +186,15 @@ The reference Registry construction boundary SHALL create one Registry object ex
 - **THEN** no component ID identifies the manager and no lookup returns it as a component or process
 
 ### Requirement: NenjimHub owns composition and runtime startup
-`NenjimHub` SHALL be the global runtime owner and SHALL expose `start() throws Exception` plus an immutable ordered snapshot of successfully started process IDs. It SHALL be neither a component nor a process and SHALL not register itself. Its only accepted startup attempt SHALL construct the Registry subsystem, start the Registry process first, register it as `nenjim.registry.service`, construct and register the Journal subsystem as `nenjim.journal.service`, start Journal second, and continue to the complete hardcoded business component set from issue #76 only after Journal startup succeeds. The Hub SHALL then resolve only the explicitly enabled business IDs through `NenjimProcess`, start them in dependency-safe order, announce the existing online point, and preserve the existing blocking wait. Registration SHALL never imply startup, and both the Registry and Journal managers SHALL remain distinct passive, unregistered handles retained privately by the Hub.
+`NenjimHub` SHALL be the global runtime owner and SHALL expose `start() throws Exception` plus an immutable ordered snapshot of successfully started process IDs. It SHALL be neither a component nor a process and SHALL not register itself. Its only accepted startup attempt SHALL construct the Registry subsystem, start the Registry process first, register it as `nenjim.registry.service`, construct and register the Journal subsystem as `nenjim.journal.service`, start Journal second, and continue to the complete hardcoded business component set from issue #76 only after Journal startup succeeds. During application composition the Hub SHALL construct the Process Manager with itself and register it as `nenjim.process-manager.default`. The Hub SHALL resolve only the explicitly enabled IDs through `NenjimProcess`, start them in dependency-safe order with the Process Manager last, announce the existing online point, and preserve the existing blocking wait. Registration SHALL never imply startup, and the Registry and Journal managers SHALL remain distinct passive, unregistered handles retained privately by the Hub.
 
 #### Scenario: Registry starts before graph construction
 - **WHEN** the Hub performs its first startup attempt
-- **THEN** it starts, registers and records Registry first, before constructing Journal or components in the business graph
+- **THEN** it starts, registers, and records Registry first before constructing Journal or components in the business graph
 
 #### Scenario: Journal starts before business composition
 - **WHEN** Registry startup and self-registration have succeeded
-- **THEN** the Hub registers, starts and records Journal before constructing components in the business graph
+- **THEN** the Hub registers, starts, and records Journal before constructing components in the business graph
 
 #### Scenario: Registry is process-discoverable
 - **WHEN** the Registry registration has completed
@@ -204,17 +204,21 @@ The reference Registry construction boundary SHALL create one Registry object ex
 - **WHEN** the Journal registration has completed
 - **THEN** the same object is retrievable as both `JournalService` and `NenjimProcess` under `nenjim.journal.service`, while its manager is not registered
 
+#### Scenario: Process Manager is process-discoverable
+- **WHEN** Hub-owned application composition has completed
+- **THEN** the object constructed with the current Hub is retrievable as both `NenjimProcessManager` and `NenjimProcess` under `nenjim.process-manager.default`
+
 #### Scenario: Required hardcoded bindings are registered
 - **WHEN** Hub composition completes
-- **THEN** Journal and all component IDs and implementation choices enumerated by issue #76 are present exactly as configured, including intentionally inactive components and excluding both service managers
+- **THEN** Journal, the Process Manager, and all component IDs and implementation choices enumerated by issue #76 are present exactly as configured, including intentionally inactive components and excluding both service managers
 
 #### Scenario: Explicit startup subset preserves order
-- **WHEN** Hub startup succeeds through the business-process phase
-- **THEN** successful process order is `nenjim.registry.service`, `nenjim.journal.service`, `assetaz.ticker.default`, `evelyn.service.prod`, `evelyn.service.test`, `evelyn.iou-burner.prod`, and `evelyn.mission-control.default`
+- **WHEN** Hub startup succeeds through the Process Manager phase
+- **THEN** successful process order is `nenjim.registry.service`, `nenjim.journal.service`, `assetaz.ticker.default`, `evelyn.service.prod`, `evelyn.service.test`, `evelyn.iou-burner.prod`, `evelyn.mission-control.default`, and `nenjim.process-manager.default`
 
 #### Scenario: Deliberately inactive processes remain inactive
 - **WHEN** Hub startup reaches its online point
-- **THEN** `jupiter-perps-alarm.default` and the dormant Composer, Process Manager, Test Tool, Soda Task Manager, and Suwimo Client paths have not been started
+- **THEN** `jupiter-perps-alarm.default` and the dormant Composer, Test Tool, Soda Task Manager, and Suwimo Client paths have not been started
 
 #### Scenario: Journal failure is fail-fast
 - **WHEN** Journal process startup throws
@@ -229,7 +233,7 @@ The reference Registry construction boundary SHALL create one Registry object ex
 - **THEN** the exception propagates, no later process starts, the Hub neither announces online nor enters its final wait, and no generic rollback or retry is attempted
 
 ### Requirement: Started-process snapshots are ordered immutable and truthful
-`NenjimHub.getStartedProcessIds()` SHALL be safe before, during, and after startup and SHALL return an immutable point-in-time snapshot in successful startup order. A process ID SHALL be appended only after that process's `start()` call returns successfully. Earlier successful entries SHALL remain visible after a later failure, while a failing process SHALL never be included. The snapshot SHALL make no health or continued-liveness guarantee.
+`NenjimHub.getStartedProcessIds()` SHALL be safe before, during, and after startup and SHALL return an immutable point-in-time snapshot in successful startup order. A process ID SHALL be appended only after that process's `start()` call returns successfully, including `nenjim.process-manager.default`. Earlier successful entries SHALL remain visible after a later failure, while a failing process SHALL never be included. The snapshot SHALL make no health or continued-liveness guarantee.
 
 #### Scenario: Snapshot before process startup
 - **WHEN** the snapshot is requested before any process has started successfully
@@ -239,6 +243,10 @@ The reference Registry construction boundary SHALL create one Registry object ex
 - **WHEN** a process's `start()` call returns successfully
 - **THEN** its ID appears after all previously successful process IDs in every later snapshot
 
+#### Scenario: Process Manager uses the common recording rule
+- **WHEN** Process Manager `start()` returns successfully
+- **THEN** the Hub records `nenjim.process-manager.default` through the same post-start rule used for every other process
+
 #### Scenario: Failed process is not recorded
 - **WHEN** a process's `start()` call throws
 - **THEN** its ID is absent while all earlier successful process IDs remain visible

openspec/specs/nenjim-process-manager/spec.md

@@ -0,0 +1,92 @@
+# nenjim-process-manager Specification
+
+## Purpose
+TBD - created by archiving change 79-integrate-process-manager-runtime-status. Update Purpose after archive.
+## Requirements
+### Requirement: Process Manager is a Hub-managed process
+`com.r35157.nenjim.npm.NenjimProcessManager` SHALL extend the canonical `NenjimProcess` and SHALL add no public refresh, polling, event, start, stop, or query operation. `NenjimProcessManagerImpl` SHALL require and retain its owning non-null `NenjimHub` as an explicit constructor dependency and SHALL use that Hub as the sole source of displayed process status.
+
+#### Scenario: Process Manager is used through the common lifecycle
+- **WHEN** the registered Process Manager is resolved through `NenjimProcess`
+- **THEN** its inherited no-argument `start()` is the only public lifecycle operation used by the Hub
+
+#### Scenario: Owning Hub is required
+- **WHEN** reference Process Manager construction receives a null Hub
+- **THEN** construction fails with `NullPointerException` before startup
+
+#### Scenario: Public API remains minimal
+- **WHEN** a consumer examines `NenjimProcessManager`
+- **THEN** it exposes no Process Manager-specific refresh, polling, event, stop, status, or query method
+
+### Requirement: Process Manager polls immediately on one non-blocking worker
+On its accepted startup attempt, the Process Manager SHALL start or reuse the existing JavaFX runtime mechanism, request the existing Process Manager window, and start exactly one dedicated background polling loop. The loop SHALL request `NenjimHub.getStartedProcessIds()` immediately and again after each normal three-second interval. Polling work SHALL not execute on the JavaFX application thread, and the worker SHALL be a daemon that does not prevent JVM termination.
+
+#### Scenario: First snapshot is immediate
+- **WHEN** Process Manager startup schedules its JavaFX window and starts its polling worker
+- **THEN** the worker requests a Hub snapshot before waiting for the first three-second interval
+
+#### Scenario: Polling continues at the required interval
+- **WHEN** one polling attempt completes without interruption
+- **THEN** the same worker waits three seconds and performs the next attempt
+
+#### Scenario: Polling is not JavaFX work
+- **WHEN** the Hub snapshot is queried and mapped
+- **THEN** that work runs on the dedicated background polling thread rather than the JavaFX application thread
+
+#### Scenario: Worker does not own JVM lifetime
+- **WHEN** only the Process Manager polling worker would otherwise keep the JVM alive
+- **THEN** its daemon status permits normal JVM termination
+
+### Requirement: Successful snapshots replace the JavaFX list exactly
+Each successful poll SHALL map every returned `NenjimComponentId` to its exact `value()` string in Hub-provided order and SHALL publish that complete point-in-time list through the existing JavaFX runtime integration. Every JavaFX control mutation SHALL execute on the JavaFX application thread, and publication SHALL replace the complete displayed list rather than infer additions or removals. The existing title, dimensions, and multiple-selection behavior MAY remain unchanged, and no hardcoded process rows SHALL remain.
+
+#### Scenario: Ordered ID values are displayed
+- **WHEN** the Hub returns an ordered snapshot of component IDs
+- **THEN** the JavaFX list is replaced with the corresponding `value()` strings in the same order
+
+#### Scenario: A later snapshot removes an ID
+- **WHEN** a successful later snapshot omits an ID present in the previously displayed snapshot
+- **THEN** complete replacement removes that string without the Process Manager maintaining authoritative lifecycle state
+
+#### Scenario: Process Manager eventually displays itself
+- **WHEN** the Hub records `nenjim.process-manager.default` after its `start()` returns successfully
+- **THEN** a later successful poll displays that ID in the position supplied by the Hub
+
+#### Scenario: JavaFX controls remain thread-confined
+- **WHEN** the window is created or a successful snapshot is published
+- **THEN** all JavaFX control creation and mutation executes through the existing JavaFX runtime on the JavaFX application thread
+
+### Requirement: Polling failures preserve the last successful view
+A polling failure SHALL be logged clearly, SHALL enqueue no invented or partial replacement, SHALL not terminate the JavaFX application thread, and SHALL leave the last successfully displayed snapshot visible. The background loop SHALL retry after the next normal three-second interval. If background waiting is interrupted, the worker SHALL restore its interrupted status and terminate the polling loop cleanly. Synchronous failures from the Process Manager's `start()` SHALL remain ordinary fail-fast Hub process-start failures.
+
+#### Scenario: Hub polling fails
+- **WHEN** obtaining or mapping a Hub snapshot throws a runtime failure
+- **THEN** the failure is logged, the displayed list is unchanged, and the worker retries after the normal interval
+
+#### Scenario: JavaFX publication fails
+- **WHEN** publishing a successfully polled snapshot fails on the JavaFX application thread
+- **THEN** the failure is logged without deliberately terminating that application thread or publishing invented fallback data
+
+#### Scenario: Polling worker is interrupted
+- **WHEN** the background worker is interrupted while waiting for its next poll
+- **THEN** it restores its interrupted status and exits without another polling attempt
+
+#### Scenario: Initial process start fails
+- **WHEN** a synchronous failure prevents `NenjimProcessManager.start()` from completing
+- **THEN** the failure propagates through the Hub, the Process Manager ID is not recorded, and no later Hub process is started
+
+### Requirement: Displayed IDs are startup milestones only
+The Process Manager list SHALL mean exactly the component IDs in the latest successfully published `NenjimHub.getStartedProcessIds()` snapshot. It SHALL make no claim about current health, worker activity, or continued liveness and SHALL add no names, aliases, metadata, timestamps, controls, icons, colours, history, or persistence. Polling SHALL be treated as a temporary internal observation mechanism without specifying Hub events as delivered functionality.
+
+#### Scenario: Consumer interprets a displayed ID
+- **WHEN** a component ID appears in the Process Manager list
+- **THEN** it means that the Hub recorded a successful return from that process's `start()` and does not guarantee current health or liveness
+
+#### Scenario: Presentation remains limited to IDs
+- **WHEN** the Process Manager renders a Hub snapshot
+- **THEN** every row contains only the exact component-ID value with no translated name or additional process metadata
+
+#### Scenario: Event functionality remains absent
+- **WHEN** the delivered Process Manager implementation is inspected
+- **THEN** it observes the Hub by polling and exposes no event type, subscription, or push-update contract
+