Skip to content

Docs demo E2E scenario catalog

Internal scenario catalog for 21 canonical demos, with one document per catalog ID. Read each demo together with shared-shell.md. The primary workflows are implemented in the docs E2E suite. Observed issues records resolved findings and their verification evidence.

Scope and review baseline

Reviewed on 2026-09-08 against docs HEAD b0c0bd9f851772dfcd30549d6d0a71ab84be12f6 and demos submodule HEAD 666376ae6d046e6f585165f364d5516bc5605311, served at http://127.0.0.1:4173. The checkout was live and dirty: theme changes existed before review, and additional source-panel/test/submodule changes appeared during it. These revisions identify the base, not an immutable build. This task writes only this scenario directory; unrelated changes are preserved. Revalidate findings on the eventual target revision.

Three parallel source reviewers owned Core/Pro grids, planning/Gantt/Scheduler, and Kanban/Pivot. Browser access was unavailable to their agent sessions, so the coordinating reviewer opened every route and supplied screenshots and exact interaction observations. Reviewers cross-checked the documents against mounted Vue components, fixtures, relevant implementation and existing tests. Standalone-only demos and embedded guide examples are excluded. React/Angular/Vanilla source tabs are source examples; the docs live preview remains Vue.

Evidence labels: “Observed” applies only to explicitly recorded actions/results. “Partial” means some steps were exercised; it never means the whole scenario passed. “Source-derived” means the cited mounted configuration/fixture or implementation supplies the expectation, but the action remains to be run. “Blocked” names the prerequisite failure. Unknown editor affordances or plugin outcomes are called out for first-run confirmation, never invented as verified controls. Visual acceptance criteria are intended regression requirements; findings identify where current appearance violates them.

Index and automation order

All entries have source review and partial browser verification. Audit History's ledger is an expected P0 workflow; the earlier missing-ledger observation has been withdrawn after correction from the product owner. The browser column summarizes the exercised subset; individual documents give its precise limits.

ID / documentDemoTierRouteCasesBrowser review subset
grid-at-scalePerformanceCore/demo/grid-at-scale6Partial: Dataset size, save/reset
ai-prompt-libraryAI promptsCore/demo/ai-prompts4Partial: Search empty state, Engineering filter
project-portfolioRow GroupingCore/demo/project-portfolio5Partial: Collapse six departments, expand
project-trackerProject trackerPro/demo/color7Partial: Select Mobile checkout, Mark ready
tree-dataTree dataPro/demo/tree-data5Partial: Collapse/expand, sticky toggle
filteringFilteringPro/demo/filtering5Partial: 217 initial, clear to 10,000, example to 335
infinity-scrollInfinity ScrollPro/demo/infinity-scroll5Partial: Remote chunk loading, pinned rows
column-collapseColumn collapsePro/demo/column-collapse5Partial: Personal and Contact expansion
context-menuContext menuPro/demo/context-menu6Partial: Ada menu open, Escape
row-masterMaster detailPro/demo/row-master5Partial: Warehouse detail content, close
audit-historyAudit historyPro/demo/audit-history5Partial: Invoice edit; ledger workflow required
excelCollaborationPro/demo/excel7Partial: Owner edit, Undo, presence
planningProject workspaceAdvanced/demo/10Partial: All five views, search, source preservation
ganttGanttAdvanced/demo/gantt7Partial: Test Plan completion, hierarchy collapse
gantt-big-data10K GanttAdvanced/demo/gantt-big-data5Partial: Scroll to Task 225–246
gantt-horizontal-big-data20Y GanttAdvanced/demo/gantt-horizontal-big-data5Partial: Horizontal scroll to 2033
event-schedulerSchedulerAdvanced/demo/event-scheduler9Partial: All workspaces, periods, preset selection
kanbanKanbanAdvanced/demo/kanban6Partial: Lane collapse, keyboard move, counts
kanban-performance50K KanbanAdvanced/demo/kanban-performance7Partial: Negative story points rejected; 13 saved
kanban-server-loading100K Remote KanbanAdvanced/demo/kanban-server-loading5Partial: 100 to 2,200 loaded, stable global counts
pivotPivot tableAdvanced/demo/pivot10Partial: Three preset totals, filter, fields, expand

Aliases /demo/planning and /demo/hr are covered by SHELL-02, without duplicate documents. They mount planning and grid-at-scale respectively; they are not required to redirect the URL.

Recommended implementation order:

  1. P0 shell and load gates: SHELL-01/02, then every demo's initial case. Await real content, not only a grid element. Require the Audit ledger as part of the Audit History load gate.
  2. P0 primary workflows: Core search/grouping, Pro selection/edit/filter/collapse/detail, Planning synchronization across all five views, Scheduler workspaces, Kanban moves/editor validation, Pivot numerical totals, large-data scroll/loading.
  3. P1 recovery and combinations: cancel/clear/undo, persistence and isolation, repeated navigation, readonly protection, large-data recycling, export and Audit History restore workflows.
  4. P1 visual coverage: shared shell at all four viewport sizes, then each demo's overlays, selection, scrolling and theme checks. Test each exposed workspace, not just the initial Grid tab.

Viewport and theme evidence

Viewport / themeRoutes inspectedExtent
1280×720 lightAll canonical routes; planning also reviewed in darkInitial appearance and interaction subsets recorded per document. Planning's detailed five-view/search/source actions were in dark.
1440×900 darkAll 21 canonical routesInitial screenshots; additional 10K scrolling retest. Not a full interaction rerun.
1440×900 lightPlanningWorkspace layout baseline.
960×800 lightPlanning, alias /demo/planning, alias /demo/hrInitial appearance, alias identity/canonical metadata; unused sidebar gutter observed. Other demos at this size remain to be checked.
390×844 lightAll 21 canonical routesInitial narrow screenshots. Planning Examples/source drawer also exercised. Deep grids, offscreen columns, all mobile editors and noninitial workspace tabs remain to be verified.

These supplemental observations apply to every demo document even where its earlier local evidence paragraph mentions only the desktop interaction session. No mobile or dark interaction case becomes fully passed from an initial screenshot. Narrow Gantt required waiting beyond the first grid cell before bars settled; remote Kanban required its loaded status rather than skeleton cells. At 390 px, Kanban stacks and long timeline names are clipped within scrollable workspaces: future tests must scroll to inspect complete content and operate controls. Screenshots were inspected in the browser-tool conversation; no persisted screenshot attachments or golden baselines were created.

Reset and automation contract

  • Start each scenario in an isolated context and use its stated seed/reset. A route change alone can retain module state. Preserve storage only in scenarios that explicitly test it. Pivot has named persistence keys; Performance Save view also needs a separate persistence test. Never clear unrelated user storage.
  • Fix the clock and timezone for date-dependent scheduler/audit/planning fixtures. The review date was September 8, 2026, in Europe/Lisbon; literal fixture dates remain literal. Do not replace fixture expectations with today's date.
  • Target stable record/task/card IDs and column properties. Virtual row positions are temporary; reacquire after sort/filter/move/scroll. Pair visible assertions with a supported grid/plugin data contract only when the scenario calls for it; don't mutate application state to simulate a user action.
  • Use role/name locators and scoped demo identities first. Treat generated DOM selectors as candidates until confirmed in the target build. Readiness includes visible fixture content, settled filters and completed remote chunks. No arbitrary FPS, duration or timing threshold is specified.
  • Run actions inside the docs host. Its styling hides some standalone controls; hidden toolbars are not required user steps. Copy/download tests need isolated clipboard/download handling. No external message, upload, or trial submission is required.
  • Keep step-by-step expected values and source references with future tests. Never change an expected result to the current failure just to make a test pass. Capture viewport/theme/route/revision plus actual result for failures.

Coverage and validation

The catalog contains 129 demo-specific cases plus nine shared-shell cases. Every canonical entry has an initial P0 case and primary-interaction scenarios; Planning covers Grid, Kanban, Gantt, Scheduler and Calendar. Review included a cross-check of document inventory, unique scenario IDs, local source links and Markdown whitespace. Source references live in each document; numerical Pivot oracles were independently calculated and matched in the browser.

The Core/Pro docs E2E suite passed 28 tests on 2026-09-10, covering the executable Core/Pro flows recorded in the individual documents. Existing docs E2E, configuration, and source mapping unit tests provide complementary shell and registry coverage. Remaining source-derived cases stay explicitly marked until separately executed.

Catalog source of truth: product catalog, source registry, docs host, and individual mounted components linked from each scenario.