Timeline: shared precision-aware date-label helper (#778) #824

Merged

marcel merged 6 commits from feat/778-timeline-date-label into main 2026-06-13 14:04:56 +02:00

Conversation 8 Commits 6 Files Changed 11 +122 -10

marcel commented 2026-06-13 13:19:16 +02:00

Owner

Copy Link

Summary

Adds a thin timeline façade frontend/src/lib/timeline/dateLabel.ts over the existing shared formatter formatDocumentDate ($lib/shared/utils/documentDate.ts). A timeline chip now renders a date identically to the same date on a document, localized de/en/es — with zero precision logic duplicated outside documentDate.ts.

timelineDateLabel(eventDate, precision, eventDateEnd?):

• delegates to formatDocumentDate(eventDate, precision, eventDateEnd ?? null, null, getLocale()) for any dated, non-UNKNOWN event;
• short-circuits to null (no formatter call) for UNKNOWN precision or null/undefined/'' eventDate;
• always passes raw = null — timeline events have no verbatim spreadsheet cell, so season words derive only from the structured anchor month.

REQ → test mapping

REQ	Requirement	Proven by (dateLabel.spec.ts)
REQ-001	Render dated entry via shared formatter (de/en/es)	renders a DAY date localized in German, renders a SEASON date with the German season word, delegates a same-year RANGE to formatDocumentDate
REQ-002	Non-UNKNOWN + non-empty → shared label, raw=null, getLocale()	renders a DAY date localized in German, renders a DAY date localized in English
REQ-003	UNKNOWN → null	returns null for UNKNOWN precision even with a date, returns null for UNKNOWN precision without a date
REQ-004	null/undefined/'' eventDate → null, no formatter call	returns null for APPROX with a null eventDate, without calling the formatter, returns null for DAY with an empty-string eventDate, treats undefined eventDateEnd identically to null for RANGE
REQ-005	No rendering logic outside documentDate.ts	delegates a same-year RANGE to formatDocumentDate (asserts byte-identical delegation)
REQ-006	timeline in coverage include	src/lib/timeline/** added to vitest coverage include; exercised by all spec cases

Other edits (per SDD spec review)

• Registered the new timeline frontend domain in frontend/eslint.config.js boundaries (allowed to import only shared).
• Added a drift-risk note above the DatePrecision type: it is a hand-maintained mirror of the Java enum and must NOT be migrated to the OpenAPI-generated type.
• Added src/lib/timeline/** to the coverage include in frontend/vite.config.ts.

Verification

• npx vitest run src/lib/timeline/dateLabel.spec.ts → 9 passed (server project, Node).
• npm run lint (prettier + eslint, incl. boundaries) → green.

Closes #778

🤖 Generated with Claude Code

## Summary Adds a thin timeline façade `frontend/src/lib/timeline/dateLabel.ts` over the existing shared formatter `formatDocumentDate` (`$lib/shared/utils/documentDate.ts`). A timeline chip now renders a date identically to the same date on a document, localized de/en/es — with zero precision logic duplicated outside `documentDate.ts`. `timelineDateLabel(eventDate, precision, eventDateEnd?)`: - delegates to `formatDocumentDate(eventDate, precision, eventDateEnd ?? null, null, getLocale())` for any dated, non-UNKNOWN event; - short-circuits to `null` (no formatter call) for `UNKNOWN` precision or `null`/`undefined`/`''` eventDate; - always passes `raw = null` — timeline events have no verbatim spreadsheet cell, so season words derive only from the structured anchor month. ## REQ → test mapping | REQ | Requirement | Proven by (`dateLabel.spec.ts`) | |---|---|---| | REQ-001 | Render dated entry via shared formatter (de/en/es) | `renders a DAY date localized in German`, `renders a SEASON date with the German season word`, `delegates a same-year RANGE to formatDocumentDate` | | REQ-002 | Non-UNKNOWN + non-empty → shared label, `raw=null`, `getLocale()` | `renders a DAY date localized in German`, `renders a DAY date localized in English` | | REQ-003 | `UNKNOWN` → `null` | `returns null for UNKNOWN precision even with a date`, `returns null for UNKNOWN precision without a date` | | REQ-004 | `null`/`undefined`/`''` eventDate → `null`, no formatter call | `returns null for APPROX with a null eventDate, without calling the formatter`, `returns null for DAY with an empty-string eventDate`, `treats undefined eventDateEnd identically to null for RANGE` | | REQ-005 | No rendering logic outside `documentDate.ts` | `delegates a same-year RANGE to formatDocumentDate` (asserts byte-identical delegation) | | REQ-006 | `timeline` in coverage `include` | `src/lib/timeline/**` added to vitest coverage `include`; exercised by all spec cases | ## Other edits (per SDD spec review) - Registered the new `timeline` frontend domain in `frontend/eslint.config.js` boundaries (allowed to import only `shared`). - Added a drift-risk note above the `DatePrecision` type: it is a hand-maintained mirror of the Java enum and must NOT be migrated to the OpenAPI-generated type. - Added `src/lib/timeline/**` to the coverage `include` in `frontend/vite.config.ts`. ## Verification - `npx vitest run src/lib/timeline/dateLabel.spec.ts` → 9 passed (server project, Node). - `npm run lint` (prettier + eslint, incl. boundaries) → green. Closes #778 🤖 Generated with [Claude Code](https://claude.com/claude-code)

marcel commented 2026-06-13 13:30:23 +02:00

Author

Owner

Copy Link

Requirements Engineer — PR Review

Verdict: ✅ Approved

Traceability of every REQ-001…006 from issue #778 against the diff:

REQ	Implemented?	Real test?	RTM row Done?	Note
REQ-001 (delegate via shared formatter, de/en/es)	✅ dateLabel.ts returns formatDocumentDate(...)	✅ DAY-de, SEASON-de, RANGE delegation	✅	locale proven across two languages
REQ-002 (non-UNKNOWN + non-empty → shared label, raw=null, getLocale())	✅ exact signature formatDocumentDate(eventDate, precision, eventDateEnd ?? null, null, getLocale())	✅ DAY-de / DAY-en	✅	matches the EARS response verbatim
REQ-003 (UNKNOWN → null, no chip)	✅ precision === 'UNKNOWN' guard	✅ UNKNOWN-with-date, UNKNOWN-without-date	✅	both branches covered
REQ-004 (null/undefined/'' → null, no formatter call)	✅ !eventDate guard short-circuits	✅ null-APPROX, empty-string-DAY, undefined-eventDateEnd	✅	falsy guard catches all three
REQ-005 (no rendering logic outside documentDate.ts)	✅ façade owns only 2 decisions	✅ RANGE asserts byte-identical toBe(...) delegation	✅	strongest possible delegation proof
REQ-006 (timeline in coverage include, 80% branch)	✅ src/lib/timeline/** in vite.config.ts	n/a (config)	✅	inside the gate

All six REQ are implemented and tested. The RTM rows were added on the feature branch with Status: Done and correct issue (#778), feature, impl-file, and test-name columns — fully in sync with the diff.

Scope-creep check: no behavior present without a backing REQ. The eslint.config.js registration and the documentDate.ts drift-risk comment are both explicit tasks in the issue body (the SDD-review "Additional task" and Decision 8). No orphan code.

Additional task from the issue (timeline domain → eslint boundary, same commit): satisfied — see Architect comment.

Verdict: ✅ Approved — full end-to-end traceability, RTM accurate.

### Requirements Engineer — PR Review **Verdict: ✅ Approved** Traceability of every REQ-001…006 from issue #778 against the diff: | REQ | Implemented? | Real test? | RTM row `Done`? | Note | |---|---|---|---|---| | REQ-001 (delegate via shared formatter, de/en/es) | ✅ `dateLabel.ts` returns `formatDocumentDate(...)` | ✅ DAY-de, SEASON-de, RANGE delegation | ✅ | locale proven across two languages | | REQ-002 (non-UNKNOWN + non-empty → shared label, `raw=null`, `getLocale()`) | ✅ exact signature `formatDocumentDate(eventDate, precision, eventDateEnd ?? null, null, getLocale())` | ✅ DAY-de / DAY-en | ✅ | matches the EARS response verbatim | | REQ-003 (`UNKNOWN` → `null`, no chip) | ✅ `precision === 'UNKNOWN'` guard | ✅ UNKNOWN-with-date, UNKNOWN-without-date | ✅ | both branches covered | | REQ-004 (`null`/`undefined`/`''` → `null`, no formatter call) | ✅ `!eventDate` guard short-circuits | ✅ null-APPROX, empty-string-DAY, undefined-eventDateEnd | ✅ | falsy guard catches all three | | REQ-005 (no rendering logic outside `documentDate.ts`) | ✅ façade owns only 2 decisions | ✅ RANGE asserts byte-identical `toBe(...)` delegation | ✅ | strongest possible delegation proof | | REQ-006 (`timeline` in coverage `include`, 80% branch) | ✅ `src/lib/timeline/**` in `vite.config.ts` | n/a (config) | ✅ | inside the gate | All six REQ are implemented **and** tested. The RTM rows were added on the feature branch with `Status: Done` and correct issue (#778), feature, impl-file, and test-name columns — fully in sync with the diff. **Scope-creep check:** no behavior present without a backing REQ. The `eslint.config.js` registration and the `documentDate.ts` drift-risk comment are both explicit tasks in the issue body (the SDD-review "Additional task" and Decision 8). No orphan code. **Additional task from the issue (timeline domain → eslint boundary, same commit):** satisfied — see Architect comment. Verdict: ✅ Approved — full end-to-end traceability, RTM accurate.

marcel commented 2026-06-13 13:30:34 +02:00

Author

Owner

Copy Link

🛠️ Developer (Felix Brandt) — PR Review

Verdict: ✅ Approved

#	Item	Status	Note
1	Reuses existing service/util, no parallel structure	PASS	Pure façade over formatDocumentDate; zero precision logic re-typed (KISS-over-DRY, constitution §3.2)
2	Layering / boundary respected	PASS	timeline → shared only; registered in eslint.config.js
3	TDD red/green evidence	PASS	Spec is delegation/integration suite; npx vitest run → 9 passed locally (confirmed)
4	Guard clauses, no nesting	PASS	Single `if (precision === 'UNKNOWN'
5	Naming reveals intent	PASS	timelineDateLabel, eventDate, eventDateEnd, precision — no d/obj
6	Type-only import for DatePrecision	PASS	import { formatDocumentDate, type DatePrecision } per issue import note
7	.js suffix on runtime import	PASS	'$lib/paraglide/runtime.js' — matches the mock path in the spec (mismatch would silently skip the mock)
8	generate:api needed?	PASS	Correctly dropped (Decision 2) — no timeline DTO in the OpenAPI spec yet; codegen would be a no-op
9	New ErrorCode?	PASS	None added — no four-site update needed
10	Dead code / commented-out	PASS	None

Comments-explaining-what vs why: the JSDoc and the inline raw=null comment explain why (security contract, REQ references), not what — consistent with the persona's comment discipline. The façade is genuinely a one-liner that benefits from the contract documentation.

Minor (non-blocking): the function is a thin wrapper whose comments exceed its body; this is intentional and appropriate here (the comments encode cross-language drift and security decisions that the code cannot self-document). No change requested.

Verdict: ✅ Approved — clean, minimal, idiomatic.

### 🛠️ Developer (Felix Brandt) — PR Review **Verdict: ✅ Approved** | # | Item | Status | Note | |---|---|---|---| | 1 | Reuses existing service/util, no parallel structure | PASS | Pure façade over `formatDocumentDate`; zero precision logic re-typed (KISS-over-DRY, constitution §3.2) | | 2 | Layering / boundary respected | PASS | `timeline → shared` only; registered in `eslint.config.js` | | 3 | TDD red/green evidence | PASS | Spec is delegation/integration suite; `npx vitest run` → 9 passed locally (confirmed) | | 4 | Guard clauses, no nesting | PASS | Single `if (precision === 'UNKNOWN' || !eventDate) return null;` — happy path is the last line | | 5 | Naming reveals intent | PASS | `timelineDateLabel`, `eventDate`, `eventDateEnd`, `precision` — no `d`/`obj` | | 6 | Type-only import for `DatePrecision` | PASS | `import { formatDocumentDate, type DatePrecision }` per issue import note | | 7 | `.js` suffix on runtime import | PASS | `'$lib/paraglide/runtime.js'` — matches the mock path in the spec (mismatch would silently skip the mock) | | 8 | `generate:api` needed? | PASS | Correctly **dropped** (Decision 2) — no timeline DTO in the OpenAPI spec yet; codegen would be a no-op | | 9 | New `ErrorCode`? | PASS | None added — no four-site update needed | | 10 | Dead code / commented-out | PASS | None | **Comments-explaining-what vs why:** the JSDoc and the inline `raw=null` comment explain *why* (security contract, REQ references), not *what* — consistent with the persona's comment discipline. The façade is genuinely a one-liner that benefits from the contract documentation. **Minor (non-blocking):** the function is a thin wrapper whose comments exceed its body; this is intentional and appropriate here (the comments encode cross-language drift and security decisions that the code cannot self-document). No change requested. Verdict: ✅ Approved — clean, minimal, idiomatic.

marcel commented 2026-06-13 13:30:49 +02:00

Author

Owner

Copy Link

🧪 Tester (Sara Holt) — PR Review

Verdict: ✅ Approved

#	Item	Status	Note
1	Each REQ has a real, behaviour-named test	PASS	9 tests, full-sentence names; mapped 1:1 to REQ-001…004
2	Test level is right	PASS	*.spec.ts (not *.svelte.spec.ts) → Node server project, as the issue specifies; no browser needed for a pure function
3	Red-then-green meaningful	PASS	Tests exercise real branches; requireAssertions: true is on in this config, so no vacuous test
4	Delegation asserted, not duplicated	PASS	RANGE test computes expected via formatDocumentDate(...) directly and toBe(...) — not a brittle literal; if the shared formatter changes, this test tracks it instead of going stale
5	Edge cases / boundaries	PASS	UNKNOWN-with-date, UNKNOWN-without-date, null, '', undefined end-date all covered — both guard sub-conditions
6	Mock at the right boundary	PASS	Partial mock of runtime.js via importOriginal, overriding only getLocale; real m.date_* exports preserved so the shared formatter runs its real logic
7	Locale switch is meaningful	PASS	de→Juli/en→July proves getLocale() actually threads through, not a no-op mock
8	Branch-coverage gate	PASS	Module added to coverage include; both guard branches + the delegation path are hit → comfortably within 80%
9	Determinism / no flakiness	PASS	Pure function, fixed dates, beforeEach resets locale to de; ran locally → 9 passed in 1.17s

Coverage note: I confirmed the run locally (single file, per the no-full-suite rule). Every executable branch in dateLabel.ts is exercised. No missing error/empty-state coverage — there are no async/error states in a pure formatter.

Verdict: ✅ Approved — exemplary thin delegation suite; tests track the shared contract rather than freezing a literal.

### 🧪 Tester (Sara Holt) — PR Review **Verdict: ✅ Approved** | # | Item | Status | Note | |---|---|---|---| | 1 | Each REQ has a real, behaviour-named test | PASS | 9 tests, full-sentence names; mapped 1:1 to REQ-001…004 | | 2 | Test level is right | PASS | `*.spec.ts` (not `*.svelte.spec.ts`) → Node `server` project, as the issue specifies; no browser needed for a pure function | | 3 | Red-then-green meaningful | PASS | Tests exercise real branches; `requireAssertions: true` is on in this config, so no vacuous test | | 4 | Delegation asserted, not duplicated | PASS | RANGE test computes expected via `formatDocumentDate(...)` directly and `toBe(...)` — not a brittle literal; if the shared formatter changes, this test tracks it instead of going stale | | 5 | Edge cases / boundaries | PASS | UNKNOWN-with-date, UNKNOWN-without-date, `null`, `''`, `undefined` end-date all covered — both guard sub-conditions | | 6 | Mock at the right boundary | PASS | Partial mock of `runtime.js` via `importOriginal`, overriding only `getLocale`; real `m.date_*` exports preserved so the shared formatter runs its real logic | | 7 | Locale switch is meaningful | PASS | de→`Juli`/en→`July` proves `getLocale()` actually threads through, not a no-op mock | | 8 | Branch-coverage gate | PASS | Module added to coverage `include`; both guard branches + the delegation path are hit → comfortably within 80% | | 9 | Determinism / no flakiness | PASS | Pure function, fixed dates, `beforeEach` resets locale to `de`; ran locally → 9 passed in 1.17s | **Coverage note:** I confirmed the run locally (single file, per the no-full-suite rule). Every executable branch in `dateLabel.ts` is exercised. No missing error/empty-state coverage — there are no async/error states in a pure formatter. Verdict: ✅ Approved — exemplary thin delegation suite; tests track the shared contract rather than freezing a literal.

marcel commented 2026-06-13 13:30:59 +02:00

Author

Owner

Copy Link

🔐 Security (Nora "NullX") — PR Review

Verdict: ✅ Approved

#	Item	Status	Note
1	Mutating endpoints / authz	PASS (N/A)	No endpoint, no mutation — a pure client-side formatter. No @RequirePermission surface.
2	Audit fields / mass-assignment	PASS (N/A)	No persistence, no createdBy/updatedBy.
3	IDOR surface	PASS (N/A)	No resource fetch.
4	raw=null passed explicitly (CWE-79 / untrusted-text)	PASS	This is the load-bearing one. The 4th arg is hardcoded null. The shared formatter only ever uses raw to derive a SEASON word via a known-German-token allow-list (seasonFromRaw) — and even that path is bypassed here, so the season word comes solely from the structured anchor month. No untrusted free text reaches the label. Matches Decision 6.
5	{@html} / escaping	PASS	Returns a plain string/null; no HTML, no interpolation. Card-level {@html} concern is correctly deferred to #779 (noted in issue).
6	Secrets / PII in logs	PASS	No logging, no secrets, no PII.
7	New dependency	PASS	None added.

Adversarial read: the only data flowing in is eventDate/eventDateEnd (structured ISO from a trusted DTO) and precision (enum). Even if a future caller passed an attacker-influenced eventDate, the value reaches new Intl.DateTimeFormat(...).format(noon(iso)) and iso.slice(...) only — no eval, no HTML sink, no path/URL/query use. The raw=null decision proactively closes the one channel (verbatim cell) that could carry untrusted text into a season word. Correct least-exposure choice.

Verdict: ✅ Approved — raw=null is exactly the right defensive default; no new attack surface.

### 🔐 Security (Nora "NullX") — PR Review **Verdict: ✅ Approved** | # | Item | Status | Note | |---|---|---|---| | 1 | Mutating endpoints / authz | PASS (N/A) | No endpoint, no mutation — a pure client-side formatter. No `@RequirePermission` surface. | | 2 | Audit fields / mass-assignment | PASS (N/A) | No persistence, no `createdBy`/`updatedBy`. | | 3 | IDOR surface | PASS (N/A) | No resource fetch. | | 4 | `raw=null` passed explicitly (CWE-79 / untrusted-text) | PASS | This is the load-bearing one. The 4th arg is hardcoded `null`. The shared formatter only ever uses `raw` to derive a SEASON word via a known-German-token allow-list (`seasonFromRaw`) — and even that path is bypassed here, so the season word comes solely from the structured anchor month. No untrusted free text reaches the label. Matches Decision 6. | | 5 | `{@html}` / escaping | PASS | Returns a plain `string`/`null`; no HTML, no interpolation. Card-level `{@html}` concern is correctly deferred to #779 (noted in issue). | | 6 | Secrets / PII in logs | PASS | No logging, no secrets, no PII. | | 7 | New dependency | PASS | None added. | **Adversarial read:** the only data flowing in is `eventDate`/`eventDateEnd` (structured ISO from a trusted DTO) and `precision` (enum). Even if a future caller passed an attacker-influenced `eventDate`, the value reaches `new Intl.DateTimeFormat(...).format(noon(iso))` and `iso.slice(...)` only — no `eval`, no HTML sink, no path/URL/query use. The `raw=null` decision proactively closes the one channel (verbatim cell) that could carry untrusted text into a season word. Correct least-exposure choice. Verdict: ✅ Approved — `raw=null` is exactly the right defensive default; no new attack surface.

marcel commented 2026-06-13 13:31:10 +02:00

Author

Owner

Copy Link

⚙️ DevOps (Tobias) — PR Review

Verdict: ✅ Approved

#	Item	Status	Note
1	DB migration / rollback	PASS (N/A)	No Flyway migration in this PR
2	New env vars documented	PASS (N/A)	None introduced
3	CI step compatibility (act_runner)	PASS	No workflow changes; upload/download-artifact pin untouched
4	CI guard steps intact	PASS	No banned-pattern grep or self-tested regex removed/weakened
5	Coverage gate config	PASS	vite.config.ts adds src/lib/timeline/** to include — tightens the gate (more code measured), never loosens it; thresholds (80/80/80/80) unchanged
6	Lint passes	PASS	eslint.config.js boundary entry well-formed; PR states npm run lint green
7	New dependency pinned / audit clean	PASS (N/A)	No dependency added
8	Secrets	PASS	None
9	PII-free logging	PASS	No logging added

Deploy risk: none. Frontend-only, pure-function addition; no migration ordering, no rolling-deploy concern, no new service/healthcheck. The coverage-include change is the only CI-visible effect and it strictly increases measured scope.

Verdict: ✅ Approved — zero operational footprint; coverage gate correctly tightened.

### ⚙️ DevOps (Tobias) — PR Review **Verdict: ✅ Approved** | # | Item | Status | Note | |---|---|---|---| | 1 | DB migration / rollback | PASS (N/A) | No Flyway migration in this PR | | 2 | New env vars documented | PASS (N/A) | None introduced | | 3 | CI step compatibility (act_runner) | PASS | No workflow changes; `upload/download-artifact` pin untouched | | 4 | CI guard steps intact | PASS | No banned-pattern grep or self-tested regex removed/weakened | | 5 | Coverage gate config | PASS | `vite.config.ts` adds `src/lib/timeline/**` to `include` — *tightens* the gate (more code measured), never loosens it; thresholds (80/80/80/80) unchanged | | 6 | Lint passes | PASS | `eslint.config.js` boundary entry well-formed; PR states `npm run lint` green | | 7 | New dependency pinned / audit clean | PASS (N/A) | No dependency added | | 8 | Secrets | PASS | None | | 9 | PII-free logging | PASS | No logging added | **Deploy risk:** none. Frontend-only, pure-function addition; no migration ordering, no rolling-deploy concern, no new service/healthcheck. The coverage-include change is the only CI-visible effect and it strictly increases measured scope. Verdict: ✅ Approved — zero operational footprint; coverage gate correctly tightened.

marcel commented 2026-06-13 13:31:21 +02:00

Author

Owner

Copy Link

🎨 UI/UX (Leonie) — PR Review

Verdict: ✅ Approved

#	Item	Status	Note
1	Interaction states	PASS (N/A)	This is a string helper, not a component — no loading/empty/error UI in scope
2	i18n via Paraglide	PASS	Delegates to the shared formatter which uses Paraglide date_* keys (already present in de/en/es); the en/de tests prove localization works; no new strings hard-coded
3	Reuses established patterns	PASS	Reuses the exact shared formatter that DocumentDate.svelte / DocumentTopBarTitle.svelte / DocumentMultiSelect.svelte use — so a timeline chip reads byte-identically to the same date elsewhere (Nielsen #4 consistency)
4	Undated handling	PASS	UNKNOWN/undated → null (no chip); the warmer "Ohne Datum" bucket-header copy is correctly left to YearBand/TimelineView (#779), per Decision 4 — clean item-label vs bucket-label separation
5	a11y <time> / <abbr>	PASS (deferred)	The issue explicitly notes the <time datetime> wrapper and ca.→<abbr> belong in EventCard/LetterCard (#779), not in this string helper. Correctly out of scope here.
6	RANGE rendering consistency	PASS	RANGE uses the shared full/collapsed-date output (Decision 3), so ranges read the same across timeline/detail/multi-select

Biggest UX consideration (not a gap): the helper returns a bare string, so all <time>/<abbr>/alt-text a11y affordances depend on the consuming card in #779 honouring the issue's notes. That is the correct boundary for this façade; I'll verify it at #779 review time.

Verdict: ✅ Approved — maximally consistent with existing date rendering; a11y wrappers correctly deferred to the consuming component.

### 🎨 UI/UX (Leonie) — PR Review **Verdict: ✅ Approved** | # | Item | Status | Note | |---|---|---|---| | 1 | Interaction states | PASS (N/A) | This is a string helper, not a component — no loading/empty/error UI in scope | | 2 | i18n via Paraglide | PASS | Delegates to the shared formatter which uses Paraglide `date_*` keys (already present in de/en/es); the en/de tests prove localization works; **no new strings hard-coded** | | 3 | Reuses established patterns | PASS | Reuses the *exact* shared formatter that `DocumentDate.svelte` / `DocumentTopBarTitle.svelte` / `DocumentMultiSelect.svelte` use — so a timeline chip reads byte-identically to the same date elsewhere (Nielsen #4 consistency) | | 4 | Undated handling | PASS | `UNKNOWN`/undated → `null` (no chip); the warmer "Ohne Datum" bucket-header copy is correctly left to `YearBand`/`TimelineView` (#779), per Decision 4 — clean item-label vs bucket-label separation | | 5 | a11y `<time>` / `<abbr>` | PASS (deferred) | The issue explicitly notes the `<time datetime>` wrapper and `ca.`→`<abbr>` belong in `EventCard`/`LetterCard` (#779), not in this string helper. Correctly out of scope here. | | 6 | RANGE rendering consistency | PASS | RANGE uses the shared full/collapsed-date output (Decision 3), so ranges read the same across timeline/detail/multi-select | **Biggest UX consideration (not a gap):** the helper returns a bare string, so all `<time>`/`<abbr>`/alt-text a11y affordances depend on the consuming card in #779 honouring the issue's notes. That is the correct boundary for this façade; I'll verify it at #779 review time. Verdict: ✅ Approved — maximally consistent with existing date rendering; a11y wrappers correctly deferred to the consuming component.

marcel commented 2026-06-13 13:31:37 +02:00

Author

Owner

Copy Link

🏛️ Architect (Markus) — PR Review

Verdict: ✅ Approved

#	Item	Status	Note
1	Package-by-domain placement	PASS	New src/lib/timeline/ domain folder; helper lives in its own domain, not smeared across shared
2	Frontend boundary rule honoured	PASS	timeline added to boundaries/elements and a { from: timeline, allow: shared } dependency rule — the frontend analogue of "new backend domain → ArchUnit allow-list". timeline → shared only; nothing broader granted
3	routes → timeline edge	PASS	Deliberately not added yet — no route consumes timeline; the first consumer (#779) adds that edge in its own PR (Decision R7). Adding it now would be premature
4	Irreversible decision needs ADR	PASS	None — façade over an existing util, no new domain data model, no new dependency. Decision 5 explicitly defers the ADR to the backend entity issue. Correct
5	Contradicts an Accepted ADR?	PASS	No. Reuses the shared-formatter / DatePrecision-sharing pattern (ADR-039/040 family) rather than forking it
6	Reuses established pattern	PASS	Single source of truth preserved: all precision/rendering stays in documentDate.ts, inside the docs/date-label-fixtures.json cross-language drift-guard (#666). A forked formatter would have silently dropped timeline from the Java-drift contract — correctly avoided
7	Second source of truth introduced?	PASS	No — the whole point of the façade
8	DatePrecision type discipline	PASS	Hand-declared type kept (not migrated to the OpenAPI-generated enum); drift-risk comment added to documentDate.ts per Decision 8, documenting the migration decision point before #779 lands
9	Blast radius bounded	PASS	One new file + one comment + two config one-liners; no edits forced on unrelated domains
10	Person/AppUser separation	PASS (N/A)	Not touched

Do-Not-Touch (constitution §4) audit: no generated artifact edited (generated/api.ts, paraglide/ untouched); no shipped Flyway migration touched; no Accepted ADR edited; actions/(upload\|download)-artifact pin untouched; no CI guard weakened. The three "other edits" are all in-bounds: documentDate.ts = comment-only (not generated), eslint.config.js + vite.config.ts = config (not on the list).

Verdict: ✅ Approved — textbook reuse-over-reinvent; domain boundary registered in the same commit; no ADR owed.

### 🏛️ Architect (Markus) — PR Review **Verdict: ✅ Approved** | # | Item | Status | Note | |---|---|---|---| | 1 | Package-by-domain placement | PASS | New `src/lib/timeline/` domain folder; helper lives in its own domain, not smeared across `shared` | | 2 | Frontend boundary rule honoured | PASS | `timeline` added to `boundaries/elements` **and** a `{ from: timeline, allow: shared }` dependency rule — the frontend analogue of "new backend domain → ArchUnit allow-list". `timeline → shared` only; nothing broader granted | | 3 | `routes → timeline` edge | PASS | Deliberately **not** added yet — no route consumes timeline; the first consumer (#779) adds that edge in its own PR (Decision R7). Adding it now would be premature | | 4 | Irreversible decision needs ADR | PASS | None — façade over an existing util, no new domain data model, no new dependency. Decision 5 explicitly defers the ADR to the backend entity issue. Correct | | 5 | Contradicts an Accepted ADR? | PASS | No. Reuses the shared-formatter / DatePrecision-sharing pattern (ADR-039/040 family) rather than forking it | | 6 | Reuses established pattern | PASS | Single source of truth preserved: all precision/rendering stays in `documentDate.ts`, inside the `docs/date-label-fixtures.json` cross-language drift-guard (#666). A forked formatter would have silently dropped timeline from the Java-drift contract — correctly avoided | | 7 | Second source of truth introduced? | PASS | No — the whole point of the façade | | 8 | `DatePrecision` type discipline | PASS | Hand-declared type kept (not migrated to the OpenAPI-generated enum); drift-risk comment added to `documentDate.ts` per Decision 8, documenting the migration decision point before #779 lands | | 9 | Blast radius bounded | PASS | One new file + one comment + two config one-liners; no edits forced on unrelated domains | | 10 | `Person`/`AppUser` separation | PASS (N/A) | Not touched | **Do-Not-Touch (constitution §4) audit:** no generated artifact edited (`generated/api.ts`, `paraglide/` untouched); no shipped Flyway migration touched; no Accepted ADR edited; `actions/(upload\|download)-artifact` pin untouched; no CI guard weakened. The three "other edits" are all in-bounds: `documentDate.ts` = comment-only (not generated), `eslint.config.js` + `vite.config.ts` = config (not on the list). Verdict: ✅ Approved — textbook reuse-over-reinvent; domain boundary registered in the same commit; no ADR owed.

marcel added 1 commit 2026-06-13 13:38:16 +02:00

docs(adr): renumber SDD adoption ADR 041 -> 042 (collision with renovate ADR) ... b05990fffb

Two ADR-041 files landed on main in parallel (sdd-adoption and
renovate-runner-setup). Renames the SDD one to 042 and repoints its references
(SPEC_DRIVEN_DEVELOPMENT, constitution, .specify/adrs/README, sdd-gate.yml).
The renovate ADR keeps 041; its references are left untouched. Riding this PR
per request.

Refs #778
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

marcel force-pushed feat/778-timeline-date-label from 33c6035199 to b05990fffb 2026-06-13 13:38:16 +02:00 Compare

gitea-actions bot commented 2026-06-13 14:01:20 +02:00

Copy Link

⚠️ Constitution changed — Sync Impact review

.specify/constitution.md was modified in this PR. Per its §6 Sync Impact rule, re-read and reconcile every file below, and confirm the semantic version bump:

• .claude/skills/draft-spec/SKILL.md
• .claude/skills/implement/SKILL.md
• .claude/skills/review-issue/SKILL.md
• .claude/skills/review-pr/SKILL.md
• .gitea/ISSUE_TEMPLATE/feature.md
• .gitea/workflows/sdd-gate.yml
• .specify/AGENTS.md
• .specify/features/_example/adr-001-avatars-reuse-archive-bucket.md
• .specify/features/_example/design.md
• .specify/features/_example/spec.md
• .specify/templates/adr.md
• .specify/templates/feature-spec.md
• CLAUDE.md
• COLLABORATING.md
• SPEC_DRIVEN_DEVELOPMENT.md
• docs/adr/042-sdd-adoption.md

### ⚠️ Constitution changed — Sync Impact review `.specify/constitution.md` was modified in this PR. Per its §6 Sync Impact rule, re-read and reconcile every file below, and confirm the semantic version bump: - `.claude/skills/draft-spec/SKILL.md` - `.claude/skills/implement/SKILL.md` - `.claude/skills/review-issue/SKILL.md` - `.claude/skills/review-pr/SKILL.md` - `.gitea/ISSUE_TEMPLATE/feature.md` - `.gitea/workflows/sdd-gate.yml` - `.specify/AGENTS.md` - `.specify/features/_example/adr-001-avatars-reuse-archive-bucket.md` - `.specify/features/_example/design.md` - `.specify/features/_example/spec.md` - `.specify/templates/adr.md` - `.specify/templates/feature-spec.md` - `CLAUDE.md` - `COLLABORATING.md` - `SPEC_DRIVEN_DEVELOPMENT.md` - `docs/adr/042-sdd-adoption.md`

marcel merged commit b05990fffb into main 2026-06-13 14:04:56 +02:00

marcel deleted branch feat/778-timeline-date-label 2026-06-13 14:04:56 +02:00

Sign in to join this conversation.

Reviewers

No Reviewers

Labels

Clear labels

P0-critical

Blocks a core user journey, causes data loss, or is a security/accessibility barrier. Work on this before P1+.

P1-high

Significant friction on a main user journey; workaround exists but is non-obvious. Next up after P0.

P2-medium

Noticeable improvement; doesn't block anything; schedule opportunistically.

P3-later

Cosmetic or parking-lot work; fine to stay open indefinitely.

audit

Read-only audit / assessment work; produces a report rather than changing code

bug

Something isn't working

cleanup

Removal of dead code, vague comments, naming violations, and scratch artifacts

collaboration

We want to extend the family archive to have more collaboration tools

conversation

descoped

We will do that later

devops

documentation

README, ARCHITECTURE, GLOSSARY, CONTRIBUTING, per-domain docs

epic

Marker for epic-level issues that group multiple child issues

feature

file-upload

legibility

Codebase Legibility Refactor — preparing the codebase for human evaluation and bus-factor reduction

needs-discussion

Has an open decision or design question that must be resolved before implementation can start.

notification

person

phase-1: security

Security hygiene — must be done first

phase-2: container-images

Production-ready multi-stage Docker images

phase-3: prod-compose

Production compose overlay + Caddy reverse proxy

phase-4: spring-prod-profile

Spring Boot production configuration profile

phase-5: backups

Database and object storage backup strategy

phase-6: deployment-docs

.env.example, DEPLOYMENT.md, runbook

phase-7: monitoring

Prometheus, Loki, Grafana, Alertmanager

refactor

Code restructuring without behaviour change

security

test

ui

UI/UX and accessibility issues

user

No Label

Milestone

No items

No Milestone

Projects

Clear projects

No project

Assignees

Clear assignees

Jens renovate_bot marcel

No Assignees

2 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: marcel/familienarchiv#824