CR014 persistent details: development verification

Input commit 740c35ce234ec26b16c76a0b77463b273f84863c plus the owner’s CR014 and explicit answers on reuse, per-view scope, preservation and configurable capacity. The accepted concept is Round 12: Persistent per-view details. This record covers a development increment, without starting a release.

Implemented and reviewed criteria

Explicit Details now opens an independent per-view dock, beside a wide graph and below a narrow graph. Need-ID reuse preserves insertion age. The author setting traceability_view_details_limit defaults to 10; positive exact integers are validated before publication. FIFO eviction, manual Close, Close all, independent views and late completion identity are implemented. Existing safe content/cache restoration is shared with transient hover. Graph commits refresh trigger associations without clearing boxes. Reset uses current size; pagehide clears without focus restoration, including back-forward cache suspension.

Requirements Constrain and retain detail... (STV_SWR_033), Separate transient hover an... (STV_SWR_050), Confirm local interactive s... (STV_SWR_057), Own a per-view persistent d... (STV_SWR_063), Present and focus persisten... (STV_SWR_064), Validate and publish the pe... (STV_SWR_065) and Keep explicitly opened arti... (STV_SHR_057) pass this criterion review and actual checks. The new cases are Ordered detail collection c... (STV_TC_SW_049), Persistent limit configurat... (STV_TC_SW_050), Persistent details across n... (STV_TC_SY_063) and Persistent loading cancella... (STV_TC_SY_064). Acceptance and runtime verification remain separate. Unrelated deferred occurrence-mode and historical distribution/CI-runtime procedures were not newly executed; their earlier evidence remains historical.

Actual outcomes

All 156 distinct executable cases pass: 72 Python unit/host/package checks, 36 Node component checks and 48 Firefox browser variants. The browser total includes six new persistent variants, two capacity variants and 40 existing regressions; two usability variants passed on their targeted corrected rerun. The focused report also repeats three already-passing configuration builds. The one-box boundary uses the real persistent manager; real Sphinx fixtures serialize default/1/3 and Firefox enforces default 10 and override 3 in file/HTTP.

The rebuilt 30,000-record fixture passes at 200 visible nodes in three runs per protocol, including rejection at the limit and first/cached details. No numeric performance budget or release qualification is inferred. The fresh wheel and source archive pass boundary, clean installed-consumer and rebuild checks. Windows 11, Python 3.13.17, Firefox 158.0.1, geckodriver 0.37.1; browser external requests are blocked. There is no new dependency or version change.

  • Python/package report (retained in the engineering checkout)

  • Node report (retained in the engineering checkout)

  • Persistent/configuration report (retained in the engineering checkout)

  • Existing browser report (retained in the engineering checkout)

  • Usability corrected rerun (retained in the engineering checkout)

  • Capacity report (retained in the engineering checkout) and measurements (retained in the engineering checkout)

  • Source/package identity (retained in the engineering checkout)

  • Commands and phase notes (retained in the engineering checkout)

  • Strict executable mapping (retained in the engineering checkout)

  • Clean installed consumer (retained in the engineering checkout)

  • Source archive inventory (retained in the engineering checkout)

Rendered and failure review

Wide and narrow screenshots were visually inspected; controls and rich content are readable, the dock scrolls independently and no modal graph overlay remains. The revised hover/persistent Draw.io/SVG lifecycles are synchronized with the source. The pending diagram’s pagehide label was re-exported after feedback.

  • Wide dock (retained in the engineering checkout)

  • Narrow readable box (retained in the engineering checkout)

  • Geometry/order evidence (retained in the engineering checkout)

  • Diagram export/link checks (retained in the engineering checkout)

  • Concept build (retained in the engineering checkout) and concept validation (retained in the engineering checkout)

The initial focused collection error was an escaped-string defect in the new harness test, before any procedure ran. Two subsequent focused attempts failed on off-screen pointer placement and a narrow theme column correctly selecting the below-graph dock. The fixture now explicitly exercises wide and narrow containers. The full existing browser run found two toolbar pointer actions outside the screen after the new below-graph interaction. Scrolling the graph back into view resolves those test assumptions; both complete usability procedures pass, with original hover timing/cancellation assertions retained. No product behavior was weakened to make those assertions pass.

  • Collection diagnostic (retained in the engineering checkout)

  • First focused report (retained in the engineering checkout)

  • Theme/placement attempt (retained in the engineering checkout)

  • Corrected layout report (retained in the engineering checkout)

  • Readable screenshot repeat (retained in the engineering checkout)

Canonical executions

  • Exact runner-to-case mapping (retained in the engineering checkout)

Each execution is one complete canonical procedure. Failed fixture attempts are preserved separately. Timestamps use actual enclosing report completion times; subtest finish timestamps are not invented. Unchanged case designs retain their review states; a passing execution does not accept a draft design.

Final documentation checks

The normal warnings-as-errors documentation build and shared evidence validation pass: 57 stakeholder, 10 system and 65 software requirements; 114 canonical cases and 717 actual execution records. Strict mapping covers 75 system/software requirements and 156 executable cases with no unknown or unmapped active IDs. The refreshed persistent lifecycle was visually inspected standalone and in the normal HTML embedding. The normal embedded demo uses the rebuilt browser assets.

  • Normal documentation build (retained in the engineering checkout)

  • Final evidence validation (retained in the engineering checkout)

  • Requirement paths (retained in the engineering checkout)

  • Requirement-to-case paths (retained in the engineering checkout)

  • Final executable mapping (retained in the engineering checkout)

  • Fresh standalone lifecycle (retained in the engineering checkout)

  • Fresh normal-page embedding (retained in the engineering checkout)