[beta.2.3] Consolidate framework fixtures and consumer qualification

[SisyphusZheng]

ADR-0152 / #1341 scheduling amendment: execute cleanup alongside every Beta.2.1/Beta.2.2 replacement; Beta.2.3 is the residual closure checkpoint, not the start of cleanup. Remove displaced owners and their callers/checkers/docs together with regression evidence.

[beta.2.3] Consolidate framework fixtures and consumer qualification

Part of #1322. Release: beta.2.3. Excludes examples/**.

Vocabulary (use consistently)

• Package conformance — owned by the package.
• Adapter fixture — minimal app proving adapter semantics.
• External consumer — hermetic scratch project proving published/packed package behavior.
• Product dogfood — real first-party product-shaped application, primarily www.

Stop calling every internal fixture "dogfood".

ui-dogfood ruling

packages/adapter-vite/__fixtures__/ui-dogfood is in scope (package fixture, not example). Its semantic coverage (dialog; dropdown; tabs; forms; focus; keyboard; ARIA; open/light/closed roots; DSD; no-JS; claim; node identity; reconnect) is valuable and must survive. But ask whether it needs a permanent full mini-app (routes/components/islands/deno config/lockfile/Vite config/Playwright setup). Preferred end state: packages/ui -> package/browser conformance; adapter-vite fixture -> adapter-specific integration only; www -> product-shaped dogfood. Preserve coverage under correct owners; delete the miniature application shell if it has no unique responsibility. Never delete real coverage to reduce files.

Packed consumer consolidation

Packed UI qualification tests a different boundary ("does the actual tarball work outside the monorepo") and stays. Consolidate scripts around a scenario harness if beneficial: workspace / packed starter / packed UI / published package / deployment. Do not preserve one bespoke orchestration script per scenario forever.

Release

beta.2.3.

Two-core-product refinement (2026-09-07)

Qualification fixtures cover two adoption paths: standalone Element build/pack/plain HTML, and Router Framework application flow. UI is dogfood/reference input. Reuse shared test infrastructure but preserve these distinct consumer boundaries; do not let monorepo aliases stand in for packed-artifact proof.

Accepted refinement — 2026-09-08 (supersedes conflicting earlier scope)

• Four packed consumers: standalone Element; Router Route Mode without Element runtime; Lit Framework Mode without Native compiler/runtime; Native Framework Mode.
• Verify install/declarations/browser graph, SSR/SSG/client continuation, JS-disabled forms, cleanup and documented runtime/browser matrix.
• Record build/bundle/matching-memory baselines and preserve existing fixture consolidation acceptance.

Governing plan (2026-09-08)

Maintainer-approved: ADR-0152, docs/current/VERSION_PLAN.md and #1341. Execution: https://github.com/orgs/open-element/projects/3. Acceptance checkpoints replace the three-day schedule. No automatic beta.2.4. Native/Lit are formal Framework Mode targets, not already-shipped support. Existing release records/versions remain unchanged; no closure without final-SHA evidence. examples/** stays frozen.

WTR-first browser conformance pilot (2026-09-08)

Audit/proposed pilot: docs/architecture/infrastructure-reduction.md, planning PR #1342. Select WTR for browser conformance; qualification is still pending.

• Migrate facade lifecycle/attribute and composed-event/shadow cases using real Chromium/Firefox/WebKit. Compile Native fixtures through the official existing build path, consume ESM; no second TSX compiler.
• Include a minimal equivalent Lit connection/host case, without claiming full Lit Framework Mode support or introducing Lit into standalone Element just for test helpers.
• Verify nonzero failing tests/setup/browser errors; record build glue, install burden, time and source-map/debug experience.
• Retain Playwright full-document SSR/DSD/native-navigation/POST E2E and Deno pure/runtime tests.
• Replace platform conformance currently simulated by facade-dom.ts and app/dom-stubs.ts after parity. Retain only justified minimal counting doubles from test-dom.ts for operation-count assertions.
• Move UI component conformance out of full adapter mini-app shells while preserving adapter/packed/no-JS boundaries.
• Reconsider WTR only for documented correctness/distribution/sustainable-integration blockers with explicit plan reassessment; less glue alone does not override contribution priority. Do not adopt competing runners for identical responsibilities.
  Lit upstream uses WTR (including SSR integration); this helps portable browser reproductions. URLPatternList upstream uses Node tests, so matching contributions follow its existing harness. No guarantee of PR acceptance; do not optimize for PR count.
  First pilot during Beta.2.2; qualified consolidation closes Beta.2.3. Existing feat: implement Beta.2.1 unified Router and standalone Element tooling #1343 blockers stay first.

Accepted upstream-contribution priority — 2026-09-08

WTR is now the SELECTED browser-conformance direction, not a tied candidate with Vitest. Qualify using official compiled ESM, Chromium/Firefox/WebKit and recipient-native Lit/WTR reproductions. Less Vite glue alone is not grounds to switch; document concrete correctness/distribution/sustainable-integration blockers for explicit plan reassessment. Preserve Playwright full-document E2E, Deno tests and necessary counting doubles. Pilot Beta.2.2, qualified deletion/consolidation Beta.2.3; no second compiler or permanent runner abstraction.

Use docs/architecture/infrastructure-reduction.md and VERSION_PLAN as scope. Prepare useful generic fixes for contribution through normal authorization; no invented bugs, cosmetic PR quota, or upstream-response release dependency. No migration is completed by this update.

Approved executable maturation — 2026-09-09

• WTR consumes official compiled ESM, never a test-only TSX transform.
• First cases: open-input/open-button validation/reset/submitter/overrides; dialog/dropdown Escape/light dismiss/focus/reconnect; paired Native/Lit application journeys.
• Preserve full-document Playwright E2E and Deno pure/runtime tests; counting doubles are not browser conformance.
• Three-browser assertions and missing-browser/setup failures propagate nonzero through required CI.
• Delete migrated fake-platform code only after coverage transfer. Pilot Beta.2.2; consolidation Beta.2.3. Prefer recipient-native upstream reproductions, with no PR quota.

Scope: docs/architecture/alpha-maturation.md and VERSION_PLAN (planning PR #1342). Upstream contribution potential can justify bounded extra integration cost; no invented defects, PR quota or upstream-response release dependency. This update changes planned work, not implementation status.