refactor!: contract the redesign's legacy fields

The expand steps of the redesign kept older app versions and API clients working. This removes those compatibility paths: - Entry date text (period/date) is written from the structured dates on every save; an edit to the text alone is overwritten. - application.archived is dropped; archived rows close first. CSV import still reads the flag from older exports. - resume_version.label is dropped; versions are named by kind and name. - rejected is no longer accepted as a stage; remaining rows and history close first. CSV import still maps it to closed. BREAKING CHANGE: older app versions and API clients that read or write archived, version labels, the rejected stage or date text alone no longer work. migrations/20260929063245_contract_redesign_legacy_fields has a rollback.sql that restores the dropped columns.
2026-10-03 18:23:47 +10:00 · 2026-09-29 08:34:27 +02:00
parent 29ab0326b2
commit cbc76b03b1
32 changed files with 7412 additions and 395 deletions
@@ -235,7 +235,13 @@ Today `cover_letter` has name, recipient (rich-text HTML), content (HTML), `styl
 | M8 | `application.closed_reason`, `cover_letter_id`, `sent_*`, `requirements` | `rejected`/`archived` → `closed` | Reverse script |
 | M9 | `cover_letter.sender_linked`, `design_linked`, `recipient_name`, `recipient_company`, `letter_date`, `layout`; `cover_letter_version` | Existing letters → freeform, unlinked | Additive |

-Contract steps, not done without your approval: stop dual-writing `period`/`date`; drop `application.archived`; drop `resume_version.label`; drop `rejected` from the stage enum.
+Contract steps (approved on 29 Sep 2026, migration `20260929063245_contract_redesign_legacy_fields`, with `rollback.sql`):
+
+- **`period`/`date` are written from `dates` only.** The text stays in the data as printed output, rewritten from the dates on every save; an edit to the text alone is overwritten. Entries without dates (imports, older data) still get dates from their text.
+- **`application.archived` is dropped.** Rows still archived are closed first. The list, bulk update, MCP tools and CSV export lose the flag; CSV import still reads it from older exports as closed.
+- **`resume_version.label` is dropped.** Versions are named by `kind` and `name`.
+- **`rejected` is no longer accepted as a stage.** Remaining rows and history close first; CSV import still reads it from older exports.
+- Older app versions and API clients that relied on these stop working against this schema.

 ---

@@ -7,7 +7,6 @@ const application: Application = {
 	company: 'Müller, "Partners"',
 	role: "Engineer",
 	status: "interview",
-	archived: false,
 	closedReason: null,
 	coverLetterId: null,
 	sentResumeVersionId: null,
@@ -75,7 +74,6 @@ describe("application CSV export", () => {
 			"Stage History": "Applied (2026-08-03) → Interview (2026-08-12)",
 			Timeline: "2026-08-03: Applied\n2026-08-08: Called recruiter\n2026-08-12: Interview",
 			Contacts: "Ada (Recruiter, Referral)",
-			Archived: "false",
 		});
 		expect(application.activity[0]?.id).toBe("interview");
 	});
@@ -138,10 +136,10 @@ describe("application CSV export", () => {

 describe("application export selection", () => {
 	const early = { ...application, id: "early", appliedAt: new Date("2026-08-02T23:59:59Z") };
-	const late = { ...application, id: "late", archived: true, appliedAt: new Date("2026-08-03T23:59:59Z") };
+	const late = { ...application, id: "late", status: "closed" as const, appliedAt: new Date("2026-08-03T23:59:59Z") };
 	const all = [early, application, late];

-	it("exports exactly current filtered rows, or all rows including archived", () => {
+	it("exports exactly current filtered rows, or all rows including closed ones", () => {
 		expect(selectApplicationsForExport(all, [application], { scope: "filtered" })).toEqual([application]);
 		expect(selectApplicationsForExport(all, [application], { scope: "all" })).toEqual(all);
 	});
@@ -1,7 +1,7 @@
 import type { ApplicationStatus, Contact } from "@reactive-resume/schema/applications/data";
 import type { Application } from "./types";
 import {
-	applicationStatusInputSchema,
+	applicationStatusSchema,
 	contactSchema,
 	INTERVIEW_KINDS,
 	STAGES,
@@ -207,7 +207,8 @@ export function mapCsvToApplications(table: string[][], mapping?: readonly (CsvF
 	// The confirmed match, else the automatic one.
 	const fieldFor = mapping ?? autoMapHeaders(headers);
 	const recognized = [...new Set(fieldFor.filter((f): f is CsvField => !!f))];
-	const isReactiveResumeExport = ["Stage History", "Timeline", "Archived", "Created At", "Updated At"].every((header) =>
+	// Older exports also carry an Archived column.
+	const isReactiveResumeExport = ["Stage History", "Timeline", "Created At", "Updated At"].every((header) =>
 		headers.includes(header),
 	);

@@ -224,7 +225,9 @@ export function mapCsvToApplications(table: string[][], mapping?: readonly (CsvF
 			if (!value) return;
 			if (field === "tags") record.tags = parseTags(value);
 			else if (field === "status") {
-				const parsed = applicationStatusInputSchema.safeParse(value.toLowerCase());
+				// Older exports may still say `rejected`, the stage `closed` replaced.
+				const stage = value.toLowerCase();
+				const parsed = applicationStatusSchema.safeParse(stage === "rejected" ? "closed" : stage);
 				if (parsed.success) record.status = parsed.data;
 			} else if (field === "stageEnteredAt") {
 				record.stageEnteredAt = dateOnly(value);
@@ -297,7 +300,6 @@ export function exportApplicationsCsv(applications: readonly Application[]): str
 		"Closed Reason",
 		"Stage History",
 		"Timeline",
-		"Archived",
 		"Created At",
 		"Updated At",
 	];
@@ -334,7 +336,6 @@ export function exportApplicationsCsv(applications: readonly Application[]): str
 			application.closedReason ?? "",
 			stages.map((entry) => `${stageLabel(entry.stage)} (${dateOnly(entry.at)})`).join(" → "),
 			timeline.map((entry) => `${dateOnly(entry.at)}: ${timelineText(entry)}`).join("\n"),
-			String(application.archived),
 			new Date(application.createdAt).toISOString(),
 			new Date(application.updatedAt).toISOString(),
 		];
@@ -1,9 +1,8 @@
 import { orpc } from "@/libs/orpc/client";

 // A single source of truth for the applications list query so the query key stays identical
-// across the board (optimistic drag), the detail panel, and the route. We always fetch archived
-// rows too and filter them per-view client-side, so "unarchive" has something to act on.
-const LIST_INPUT = { includeArchived: true } as const;
+// across the board (optimistic drag), the detail panel, and the route.
+const LIST_INPUT = {} as const;

 export const applicationsListQueryOptions = () => orpc.applications.list.queryOptions({ input: LIST_INPUT });
 export const applicationsListQueryKey = () => orpc.applications.list.queryKey({ input: LIST_INPUT });
@@ -110,8 +110,7 @@ describe("ResumesCommandGroup", () => {
 	])("opens the %s list page from the root palette with Enter", async (page, keys, createLabel) => {
 		mockUseQueryData((entity) => {
 			if (entity === "resumes") return [{ id: "resume-1", name: "Evil Apricot Pike", slug: "apricot" }];
-			if (entity === "applications")
-				return [{ id: "application-1", company: "Umbrella", role: "Staff Engineer", archived: false }];
+			if (entity === "applications") return [{ id: "application-1", company: "Umbrella", role: "Staff Engineer" }];
 			if (entity === "threads")
 				return [{ id: "thread-1", title: "Cover letter rewrite", resumeName: "Product Resume" }];
 			return [];
@@ -133,8 +132,7 @@ describe("ResumesCommandGroup", () => {
 	])("keeps arrow-key navigation active on the %s list page", async (_page, keys, createLabel, itemLabel) => {
 		mockUseQueryData((entity) => {
 			if (entity === "resumes") return [{ id: "resume-1", name: "Evil Apricot Pike", slug: "apricot" }];
-			if (entity === "applications")
-				return [{ id: "application-1", company: "Umbrella", role: "Staff Engineer", archived: false }];
+			if (entity === "applications") return [{ id: "application-1", company: "Umbrella", role: "Staff Engineer" }];
 			if (entity === "threads")
 				return [{ id: "thread-1", title: "Cover letter rewrite", resumeName: "Product Resume" }];
 			return [];
@@ -201,9 +199,7 @@ describe("ResumesCommandGroup", () => {
 	it("loads applications on the applications page", () => {
 		useCommandPaletteStore.setState({ pages: ["applications"] });
 		mockUseQueryData((entity) =>
-			entity === "applications"
-				? [{ id: "application-1", company: "Umbrella", role: "Staff Engineer", archived: false }]
-				: [],
+			entity === "applications" ? [{ id: "application-1", company: "Umbrella", role: "Staff Engineer" }] : [],
 		);

 		renderGroup();
@@ -217,8 +213,8 @@ describe("ResumesCommandGroup", () => {
 		mockUseQueryData((entity) =>
 			entity === "applications"
 				? [
-						{ id: "application-1", company: "Umbrella", role: "Staff Engineer", archived: false },
-						{ id: "application-2", company: "Wayne Enterprises", role: "Product Engineer", archived: false },
+						{ id: "application-1", company: "Umbrella", role: "Staff Engineer" },
+						{ id: "application-2", company: "Wayne Enterprises", role: "Product Engineer" },
 					]
 				: [],
 		);
@@ -232,9 +228,7 @@ describe("ResumesCommandGroup", () => {
 	it("opens the selected application by id", () => {
 		useCommandPaletteStore.setState({ pages: ["applications"] });
 		mockUseQueryData((entity) =>
-			entity === "applications"
-				? [{ id: "application-1", company: "Umbrella", role: "Staff Engineer", archived: false }]
-				: [],
+			entity === "applications" ? [{ id: "application-1", company: "Umbrella", role: "Staff Engineer" }] : [],
 		);

 		renderGroup();
@@ -565,9 +565,9 @@ export const useResumeStore = create<ResumeStore>()(
 				}

 				fn(state.resume.data as WritableDraft<ResumeData>);
-				// The server keeps dates and their legacy text in step on save; doing the same here keeps its
-				// echo identical to the draft, so an autosave never reads as an outside change.
-				syncResumeDates(state.resume.data as ResumeData, currentResume.data);
+				// The server writes the date text from the dates on save; doing the same here keeps its echo
+				// identical to the draft, so an autosave never reads as an outside change.
+				syncResumeDates(state.resume.data as ResumeData);
 				state.saveStatus = "saving";
 				state.canUndo = state.undoStack.length > 0;
 				state.canRedo = state.redoStack.length > 0;
@@ -559,7 +559,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 									},
 									"period": {
 										"type": "string",
-										"description": "The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Kept in sync from `dates`; prefer writing `dates`."
+										"description": "The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Written from `dates` on every save; write `dates` instead."
 									},
 									"dates": {
 										"type": "object",
@@ -659,7 +659,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 												},
 												"period": {
 													"type": "string",
-													"description": "The period of time this role was held, as text. Kept in sync from `dates`; prefer writing `dates`."
+													"description": "The period of time this role was held, as text. Written from `dates` on every save; write `dates` instead."
 												},
 												"dates": {
 													"type": "object",
@@ -822,7 +822,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 									},
 									"period": {
 										"type": "string",
-										"description": "The period of time the education was obtained over, as text. Kept in sync from `dates`; prefer writing `dates`."
+										"description": "The period of time the education was obtained over, as text. Written from `dates` on every save; write `dates` instead."
 									},
 									"dates": {
 										"type": "object",
@@ -990,7 +990,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 									},
 									"period": {
 										"type": "string",
-										"description": "The period of time the project was worked on, as text. Kept in sync from `dates`; prefer writing `dates`."
+										"description": "The period of time the project was worked on, as text. Written from `dates` on every save; write `dates` instead."
 									},
 									"dates": {
 										"type": "object",
@@ -1470,7 +1470,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 									},
 									"date": {
 										"type": "string",
-										"description": "The date when the award was received, as text. Kept in sync from `dates`; prefer writing `dates`."
+										"description": "The date when the award was received, as text. Written from `dates` on every save; write `dates` instead."
 									},
 									"dates": {
 										"type": "object",
@@ -1639,7 +1639,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 									},
 									"date": {
 										"type": "string",
-										"description": "The date when the certification was received, as text. Kept in sync from `dates`; prefer writing `dates`."
+										"description": "The date when the certification was received, as text. Written from `dates` on every save; write `dates` instead."
 									},
 									"dates": {
 										"type": "object",
@@ -1808,7 +1808,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 									},
 									"date": {
 										"type": "string",
-										"description": "The date when the publication was published, as text. Kept in sync from `dates`; prefer writing `dates`."
+										"description": "The date when the publication was published, as text. Written from `dates` on every save; write `dates` instead."
 									},
 									"dates": {
 										"type": "object",
@@ -1977,7 +1977,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 									},
 									"period": {
 										"type": "string",
-										"description": "The period of time the author was volunteered at the organization or company, as text. Kept in sync from `dates`; prefer writing `dates`."
+										"description": "The period of time the author was volunteered at the organization or company, as text. Written from `dates` on every save; write `dates` instead."
 									},
 									"dates": {
 										"type": "object",
@@ -2617,7 +2617,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 										},
 										"period": {
 											"type": "string",
-											"description": "The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Kept in sync from `dates`; prefer writing `dates`."
+											"description": "The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Written from `dates` on every save; write `dates` instead."
 										},
 										"dates": {
 											"type": "object",
@@ -2717,7 +2717,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 													},
 													"period": {
 														"type": "string",
-														"description": "The period of time this role was held, as text. Kept in sync from `dates`; prefer writing `dates`."
+														"description": "The period of time this role was held, as text. Written from `dates` on every save; write `dates` instead."
 													},
 													"dates": {
 														"type": "object",
@@ -2892,7 +2892,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 										},
 										"period": {
 											"type": "string",
-											"description": "The period of time the education was obtained over, as text. Kept in sync from `dates`; prefer writing `dates`."
+											"description": "The period of time the education was obtained over, as text. Written from `dates` on every save; write `dates` instead."
 										},
 										"dates": {
 											"type": "object",
@@ -3072,7 +3072,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 										},
 										"period": {
 											"type": "string",
-											"description": "The period of time the project was worked on, as text. Kept in sync from `dates`; prefer writing `dates`."
+											"description": "The period of time the project was worked on, as text. Written from `dates` on every save; write `dates` instead."
 										},
 										"dates": {
 											"type": "object",
@@ -3590,7 +3590,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 										},
 										"date": {
 											"type": "string",
-											"description": "The date when the award was received, as text. Kept in sync from `dates`; prefer writing `dates`."
+											"description": "The date when the award was received, as text. Written from `dates` on every save; write `dates` instead."
 										},
 										"dates": {
 											"type": "object",
@@ -3771,7 +3771,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 										},
 										"date": {
 											"type": "string",
-											"description": "The date when the certification was received, as text. Kept in sync from `dates`; prefer writing `dates`."
+											"description": "The date when the certification was received, as text. Written from `dates` on every save; write `dates` instead."
 										},
 										"dates": {
 											"type": "object",
@@ -3952,7 +3952,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 										},
 										"date": {
 											"type": "string",
-											"description": "The date when the publication was published, as text. Kept in sync from `dates`; prefer writing `dates`."
+											"description": "The date when the publication was published, as text. Written from `dates` on every save; write `dates` instead."
 										},
 										"dates": {
 											"type": "object",
@@ -4133,7 +4133,7 @@ The full JSON Schema for Reactive Resume follows. You can also fetch the latest
 										},
 										"period": {
 											"type": "string",
-											"description": "The period of time the author was volunteered at the organization or company, as text. Kept in sync from `dates`; prefer writing `dates`."
+											"description": "The period of time the author was volunteered at the organization or company, as text. Written from `dates` on every save; write `dates` instead."
 										},
 										"dates": {
 											"type": "object",
@@ -0,0 +1,18 @@
+-- Contract steps for the redesign (§3.10): the closed stage replaced `archived` and `rejected`, and versions are
+-- labelled by `kind`. Older app versions stop working against this schema. rollback.sql restores the columns.
+-- Anything still marked archived or rejected closes first, so no application reopens when the flag goes.
+UPDATE "application" SET "status" = 'closed', "closed_reason" = 'not-selected' WHERE "status" = 'rejected';--> statement-breakpoint
+UPDATE "application" SET "status" = 'closed' WHERE "archived" = true AND "status" <> 'closed';--> statement-breakpoint
+UPDATE "application" SET "activity" = (
+	SELECT jsonb_agg(
+		CASE WHEN entry->>'type' = 'stage' AND entry->>'stage' = 'rejected'
+			THEN jsonb_set(entry, '{stage}', '"closed"')
+			ELSE entry
+		END
+		ORDER BY position
+	)
+	FROM jsonb_array_elements("activity") WITH ORDINALITY AS items(entry, position)
+)
+WHERE "activity" @> '[{"type": "stage", "stage": "rejected"}]';--> statement-breakpoint
+ALTER TABLE "application" DROP COLUMN "archived";--> statement-breakpoint
+ALTER TABLE "resume_version" DROP COLUMN "label";
@@ -0,0 +1,15 @@
+-- Restores the columns the contract step dropped, for older app versions. Closed applications without a reason
+-- were archived ones; version labels come back from their kind.
+ALTER TABLE "application" ADD COLUMN "archived" boolean DEFAULT false NOT NULL;
+UPDATE "application" SET "archived" = true WHERE "status" = 'closed' AND "closed_reason" IS NULL;
+ALTER TABLE "resume_version" ADD COLUMN "label" text DEFAULT '' NOT NULL;
+UPDATE "resume_version" SET "label" = CASE "kind"
+	WHEN 'created' THEN 'Created'
+	WHEN 'import' THEN 'Imported'
+	WHEN 'auto' THEN 'Manual save'
+	WHEN 'before-restore' THEN 'Before restore'
+	WHEN 'restored' THEN 'Restored version'
+	WHEN 'ai' THEN 'AI edit'
+	WHEN 'sent' THEN 'Sent'
+	ELSE coalesce("name", '')
+END;
@@ -4,7 +4,6 @@ import * as schema from "@reactive-resume/db/schema";
 import {
 	aiMetadataSchema,
 	applicationClosedReasonSchema,
-	applicationStatusInputSchema,
 	applicationStatusSchema,
 	applicationTimelineEntrySchema,
 	contactSchema,
@@ -40,7 +39,6 @@ const applicationSchema = createSelectSchema(schema.application, {
 	closedReason: applicationClosedReasonSchema
 		.nullable()
 		.describe("Why a closed application ended: not-selected, withdrew, accepted-other or no-response."),
-	archived: z.boolean().describe("Deprecated: the closed stage replaced it. Kept in step for older clients."),
 	resumeId: z.string().nullable().describe("The linked Reactive Resume, if any."),
 	coverLetterId: z.string().nullable().describe("The linked saved cover letter, if any."),
 	sentResumeVersionId: z
@@ -109,7 +107,7 @@ const editableSchema = applicationSchema.pick({
 const createInputSchema = editableSchema.partial().extend({
 	company: applicationSchema.shape.company,
 	role: applicationSchema.shape.role,
-	status: applicationStatusInputSchema.optional(),
+	status: applicationStatusSchema.optional(),
 	closedReason: applicationClosedReasonSchema.nullable().optional(),
 	stageEnteredAt: timelineDateSchema.optional(),
 });
@@ -120,10 +118,9 @@ export const applicationDto = {
 			.object({
 				status: applicationStatusSchema.optional(),
 				tags: z.array(z.string()).optional(),
-				includeArchived: z.boolean().optional().default(false),
 			})
 			.optional()
-			.default({ includeArchived: false }),
+			.default({}),
 		output: z.array(applicationSchema.omit({ userId: true })),
 	},

@@ -146,12 +143,11 @@ export const applicationDto = {
 	update: {
 		input: editableSchema.partial().extend({
 			id: z.string(),
-			status: applicationStatusInputSchema.optional(),
+			status: applicationStatusSchema.optional(),
 			closedReason: applicationClosedReasonSchema
 				.nullable()
 				.optional()
 				.describe("Why the application closed; kept when status stays closed, cleared by any other stage."),
-			archived: z.boolean().optional(),
 		}),
 		output: applicationSchema.omit({ userId: true }),
 	},
@@ -225,13 +221,12 @@ export const applicationDto = {
 		output: z.void(),
 	},

-	// Table bulk actions: move stage, archive/unarchive, add tags across a selection.
+	// Table bulk actions: move stage, close, add tags across a selection.
 	bulkUpdate: {
 		input: z.object({
 			ids: z.array(z.string()).min(1).max(200, "Too many items in a single bulk operation"),
-			status: applicationStatusInputSchema.optional(),
+			status: applicationStatusSchema.optional(),
 			closedReason: applicationClosedReasonSchema.nullable().optional(),
-			archived: z.boolean().optional(),
 			addTags: z.array(z.string()).optional(),
 		}),
 		output: z.object({ updated: z.number() }),
@@ -31,7 +31,6 @@ const versionSchema = z.object({
 			"What made the version: created, import, auto (an editing session's autosave), named, before-restore, restored, ai (an AI or API edit) or sent.",
 		),
 	name: z.string().nullable().describe("The user's name for a named version."),
-	label: z.string().describe("A short English description of the version."),
 	createdAt: z.date().describe("When this state was saved."),
 });

@@ -12,7 +12,7 @@ export const crudRouter = {
 			operationId: "listApplications",
 			summary: "List job applications",
 			description:
-				"Returns all job applications belonging to the authenticated user, most recently updated first. Archived applications are excluded unless includeArchived is set. Optionally filter by pipeline stage. Requires authentication.",
+				"Returns all job applications belonging to the authenticated user, most recently updated first. Optionally filter by pipeline stage. Requires authentication.",
 			successDescription: "A list of the user's job applications.",
 		})
 		.input(applicationDto.list.input)
@@ -22,7 +22,6 @@ export const crudRouter = {
 				userId: context.user.id,
 				...(input.status ? { status: input.status } : {}),
 				...(input.tags ? { tags: input.tags } : {}),
-				includeArchived: input.includeArchived,
 			}),
 		),

@@ -178,7 +178,7 @@ describe("applicationService.update", () => {
 			[Record<string, unknown>],
 		];
 		expect(closing).toMatchObject({ status: "closed", closedReason: "withdrew" });
-		expect(reopening).toMatchObject({ status: "applied", closedReason: null, archived: false });
+		expect(reopening).toMatchObject({ status: "applied", closedReason: null });
 	});
 });

@@ -187,11 +187,11 @@ async function recordSentResume(row: ApplicationRow): Promise<ApplicationRow> {
 	return updated ?? row;
 }

-/** Closing keeps (or takes) a reason; any other stage clears it, and reopens an application that was archived. */
+/** Closing keeps (or takes) a reason; any other stage clears it. */
 function stageFields(status: ApplicationStatus | undefined, closedReason: ApplicationClosedReason | null | undefined) {
 	if (status === undefined) return closedReason !== undefined ? { closedReason } : {};
 	if (status === "closed") return closedReason !== undefined ? { closedReason } : {};
-	return { closedReason: null, archived: false };
+	return { closedReason: null };
 }

 async function assertOwnedResumes(userId: string, resumeIds: (string | null | undefined)[]) {
@@ -270,7 +270,7 @@ const stripUserId = <T extends { userId: string; activity?: ApplicationTimelineE
 };

 export const applicationService = {
-	list: async (input: { userId: string; status?: ApplicationStatus; tags?: string[]; includeArchived?: boolean }) => {
+	list: async (input: { userId: string; status?: ApplicationStatus; tags?: string[] }) => {
 		const rows = await db
 			.select()
 			.from(schema.application)
@@ -283,7 +283,7 @@ export const applicationService = {
 			)
 			.orderBy(desc(schema.application.updatedAt));

-		return rows.filter((row) => input.includeArchived || !row.archived).map(stripUserId);
+		return rows.map(stripUserId);
 	},

 	getById: async (input: { id: string; userId: string }) => {
@@ -364,12 +364,11 @@ export const applicationService = {
 			userId: string;
 			status?: ApplicationStatus | undefined;
 			closedReason?: ApplicationClosedReason | null | undefined;
-			archived?: boolean | undefined;
 		},
 	) => {
 		await requireOwned(input.id, input.userId);

-		const { id, userId, status, closedReason, archived, ...fields } = input;
+		const { id, userId, status, closedReason, ...fields } = input;
 		await assertOwnedResume(userId, fields.resumeId);
 		await assertOwnedCoverLetter(userId, fields.coverLetterId);

@@ -394,7 +393,6 @@ export const applicationService = {
 				...fields,
 				...(status !== undefined ? { status } : {}),
 				...(appliedAtExpr ? { appliedAt: appliedAtExpr } : {}),
-				...(archived !== undefined ? { archived } : {}),
 				...stageFields(status, closedReason),
 				...(activityExpr ? { activity: activityExpr } : {}),
 			})
@@ -675,7 +673,6 @@ export const applicationService = {
 		ids: string[];
 		status?: ApplicationStatus | undefined;
 		closedReason?: ApplicationClosedReason | null | undefined;
-		archived?: boolean | undefined;
 		addTags?: string[] | undefined;
 	}) => {
 		const scope = and(inArray(schema.application.id, input.ids), eq(schema.application.userId, input.userId));
@@ -713,7 +710,6 @@ export const applicationService = {
 				...(input.status !== undefined ? { status: input.status } : {}),
 				...(appliedAtExpr ? { appliedAt: appliedAtExpr } : {}),
 				...(activityExpr ? { activity: activityExpr } : {}),
-				...(input.archived !== undefined ? { archived: input.archived } : {}),
 				...stageFields(input.status, input.closedReason),
 				...(tagsExpr ? { tags: tagsExpr } : {}),
 			})
@@ -742,7 +738,7 @@ export const applicationService = {

 	// Raw counts for Insights; funnel/sankey/tiles are derived client-side from these.
 	stats: async (input: { userId: string }) => {
-		const scope = and(eq(schema.application.userId, input.userId), eq(schema.application.archived, false));
+		const scope = eq(schema.application.userId, input.userId);

 		const byStage = await db
 			.select({ status: schema.application.status, count: sql<number>`count(*)::int` })
@@ -22,15 +22,11 @@ function parseApiResumeData(data: unknown, code: "BAD_REQUEST" | "INTERNAL_SERVE
 	}
 }

-/**
- * Validates data before it's saved, then keeps structured dates and their legacy text in step. `previous` is
- * the stored data being replaced, which tells an edit to the text alone (from a client that doesn't know
- * structured dates) from an edit to the dates.
- */
-export const parseWritableResumeData = (data: unknown, previous?: ResumeData) => {
+/** Validates data before it's saved, then writes each entry's date text from its structured dates. */
+export const parseWritableResumeData = (data: unknown) => {
 	const parsed = parseApiResumeData(data, "BAD_REQUEST", "Resume data does not match the canonical schema.");
 	upgradeResumeDates(parsed);
-	syncResumeDates(parsed, previous);
+	syncResumeDates(parsed);
 	return parsed;
 };

@@ -54,7 +54,6 @@ vi.mock("@reactive-resume/db/schema", () => ({
 		resumeId: "resume_id",
 		userId: "user_id",
 		data: "data",
-		label: "label",
 		createdAt: "created_at",
 	},
 	user: { id: "id", username: "username" },
@@ -77,7 +77,6 @@ async function applyResumePatchTx(
 		userId: string;
 		operations: JsonPatchOperation[];
 		expectedUpdatedAt?: Date;
-		versionLabel?: string;
 	},
 ) {
 	const [existing] = await client
@@ -118,7 +117,7 @@ async function applyResumePatchTx(
 		});
 	}

-	patchedData = parseWritableResumeData(patchedData, storedData);
+	patchedData = parseWritableResumeData(patchedData);
 	// The version guard is the ms-precision JS check above, under the SELECT ... FOR UPDATE lock.
 	// Never compare expectedUpdatedAt in SQL: rows stamped by Postgres now() (defaultNow() on
 	// insert) carry microseconds, while JS Dates are ms-truncated — SQL equality then matches
@@ -154,7 +153,6 @@ async function applyResumePatchTx(
 		userId: input.userId,
 		data: resume.data,
 		kind: "ai",
-		...(input.versionLabel ? { label: input.versionLabel } : {}),
 	});

 	return resume;
@@ -609,9 +607,7 @@ export const resumeService = {
 					});
 				}

-				const normalizedData = input.data
-					? parseWritableResumeData(input.data, parseStoredResumeData(existing.data))
-					: undefined;
+				const normalizedData = input.data ? parseWritableResumeData(input.data) : undefined;
 				// A blank resume is named after its headline until someone names it by hand.
 				const followedName =
 					existing.autoName && input.name === undefined && normalizedData
@@ -16,7 +16,7 @@ const selectChain = (rows: unknown[]) => ({

 const insertValues = () => {
 	const values = vi.fn((_input: Record<string, unknown>) => ({
-		returning: async () => [{ id: "v1", kind: _input.kind, name: _input.name, label: _input.label }],
+		returning: async () => [{ id: "v1", kind: _input.kind, name: _input.name }],
 	}));
 	dbMock.insert.mockReturnValue({ values });
 	return values;
@@ -40,26 +40,27 @@ beforeEach(() => {
 });

 describe("writeVersion", () => {
-	it("stores normalized data with its kind and an English label, then applies retention", async () => {
+	it("stores normalized data with its kind, then applies retention", async () => {
 		const values = insertValues();
 		const input = { ...data(), basics: { ...data().basics, name: "Ada" } };

 		await writeVersion(dbMock as never, { resumeId: "r1", userId: "u1", data: input, kind: "before-restore" });

 		expect(values).toHaveBeenCalledWith(
-			expect.objectContaining({ kind: "before-restore", label: "Before restore", name: null, sessionId: null }),
+			expect.objectContaining({ kind: "before-restore", name: null, sessionId: null }),
 		);
 		expect(values.mock.calls[0]?.[0]).toHaveProperty("data.basics.name", "Ada");
 		// Expired versions, then autosaves beyond the cap.
 		expect(dbMock.delete).toHaveBeenCalledTimes(2);
 	});

-	it("labels a named version with its name", async () => {
+	it("keeps a named version's name", async () => {
 		const values = insertValues();

 		await writeVersion(dbMock as never, { resumeId: "r1", userId: "u1", data: data(), kind: "named", name: "Sent" });

-		expect(values).toHaveBeenCalledWith(expect.objectContaining({ kind: "named", name: "Sent", label: "Sent" }));
+		expect(values).toHaveBeenCalledWith(expect.objectContaining({ kind: "named", name: "Sent" }));
+		expect(values.mock.calls[0]?.[0]).not.toHaveProperty("label");
 	});
 });

@@ -19,22 +19,10 @@ const MAX_AUTOSAVES = 500;
 // History shows at most this many versions, newest first.
 const LIST_LIMIT = 100;

-// `label` keeps an English description for API clients and older rows; the app labels versions by kind.
-const LABELS: Record<Exclude<ResumeVersionKind, "named">, string> = {
-	created: "Created",
-	import: "Imported",
-	auto: "Manual save",
-	"before-restore": "Before restore",
-	restored: "Restored version",
-	ai: "AI edit",
-	sent: "Sent",
-};
-
 const summary = {
 	id: schema.resumeVersion.id,
 	kind: schema.resumeVersion.kind,
 	name: schema.resumeVersion.name,
-	label: schema.resumeVersion.label,
 	createdAt: schema.resumeVersion.createdAt,
 };

@@ -44,7 +32,6 @@ type VersionInput = {
 	data: ResumeData;
 	kind: ResumeVersionKind;
 	name?: string;
-	label?: string;
 	sessionId?: string;
 };

@@ -58,7 +45,6 @@ export async function writeVersion(client: DbOrTx, input: VersionInput) {
 			kind: input.kind,
 			name: input.name ?? null,
 			sessionId: input.sessionId ?? null,
-			label: input.label ?? (input.kind === "named" ? (input.name ?? "") : LABELS[input.kind]),
 		})
 		.returning(summary);
 	if (!version) throw new ORPCError("INTERNAL_SERVER_ERROR", { message: "Failed to save the version." });
@@ -180,7 +166,7 @@ export async function getVersion(input: { resumeId: string; userId: string; vers
 export async function renameVersion(input: { resumeId: string; userId: string; versionId: string; name: string }) {
 	const [version] = await db
 		.update(schema.resumeVersion)
-		.set({ name: input.name, label: input.name })
+		.set({ name: input.name })
 		.where(and(ownedVersion(input), eq(schema.resumeVersion.kind, "named")))
 		.returning(summary);

@@ -34,8 +34,6 @@ export const application = pg.pgTable(
 		status: pg.text("status").$type<ApplicationStatus>().notNull().default("saved"),
 		// Why a closed application ended (not selected, withdrew, accepted another offer, no response). Insights use it.
 		closedReason: pg.text("closed_reason").$type<ApplicationClosedReason>(),
-		// Superseded by the closed stage, which absorbed it; kept while older app versions may still read it.
-		archived: pg.boolean("archived").notNull().default(false),
 		// Live link to one of the user's resumes. Kept on resume delete (set null) so the
 		// application history survives.
 		resumeId: pg.text("resume_id").references(() => resume.id, { onDelete: "set null" }),
@@ -88,8 +88,6 @@ export const resumeVersion = pg.pgTable(
 			.references(() => user.id, { onDelete: "cascade" }),
 		// Snapshot of the resume data. A session's autosave is refreshed while the session lasts; the rest are immutable.
 		data: pg.jsonb("data").notNull().$type<ResumeData>(),
-		// English description kept for API clients and rows written before `kind`; the UI labels versions by `kind`.
-		label: pg.text("label").notNull(),
 		kind: pg.text("kind", { enum: RESUME_VERSION_KINDS }).notNull().default("auto"),
 		// The user's name for a `named` version.
 		name: pg.text("name"),
@@ -102,12 +102,12 @@ describe("buildMcpServerCard", () => {
 		expect(props.apiKey).toBeDefined();
 	});

-	it("matches the create/update application archived contract", () => {
+	it("closes applications through their stage: no archived flag", () => {
 		const create = TOOL_META[MCP_TOOL_NAME.createApplication].inputSchema;
 		const update = TOOL_META[MCP_TOOL_NAME.updateApplication].inputSchema;

 		expect(create.safeParse({ company: "Acme", role: "Engineer", archived: true }).success).toBe(false);
-		expect(update.safeParse({ id: "app-1", archived: true }).success).toBe(true);
+		expect(update.parse({ id: "app-1", archived: true })).not.toHaveProperty("archived");
 	});

 	it.each([{ content: "Updated" }, { recipient: "Dear Hiring Manager" }, { template: "onyx" }])(
@@ -312,7 +312,7 @@ export const TOOL_META = {
 			"                                            year-months ('2022' or '2022-03')",
 			"",
 			"Dates: write `dates`; the text in `period` (or `date` for awards, certifications and",
-			"publications) is rewritten from it in the resume's locale. Editing only the text still works.",
+			"publications) is rewritten from it in the resume's locale, so an edit to the text alone is lost.",
 			"Important: HTML content fields (description, summary.content) must use valid HTML.",
 			"New items must include a valid UUID as `id` and `hidden: false`.",
 			`Locked resumes cannot be patched; use \`${T.unlockResume}\` first.`,
@@ -542,7 +542,6 @@ export const TOOL_META = {
 		inputSchema: z.object({
 			status: applicationStatusSchema.optional(),
 			tags: z.array(z.string()).optional().default([]),
-			includeArchived: z.boolean().optional().default(false),
 		}),
 		annotations: READ_IDEMPOTENT,
 	},
@@ -577,7 +576,6 @@ export const TOOL_META = {
 		inputSchema: z.object({
 			id: applicationIdSchema,
 			...applicationMutableFieldsSchema,
-			archived: z.boolean().optional().describe("Deprecated: close the application instead. Kept for older clients."),
 		}),
 		annotations: WRITE_DESTRUCTIVE,
 	},
@@ -652,7 +650,6 @@ export const TOOL_META = {
 			ids: z.array(z.string()).min(1),
 			status: applicationStatusSchema.optional(),
 			closedReason: applicationClosedReasonSchema.nullable().optional(),
-			archived: z.boolean().optional().describe("Deprecated: close the applications instead."),
 			addTags: z.array(z.string()).optional(),
 		}),
 		annotations: WRITE_DESTRUCTIVE,
@@ -262,9 +262,9 @@ describe("registerTools", () => {
 		registerTools(server as never, clientMock as never, new Headers());

 		const tool = registered.find((item) => item.name === "list_applications")!;
-		const result = await tool.handler({ includeArchived: true, tags: ["remote"] });
+		const result = await tool.handler({ tags: ["remote"] });

-		expect(clientMock.applications.list).toHaveBeenCalledWith({ includeArchived: true, tags: ["remote"] });
+		expect(clientMock.applications.list).toHaveBeenCalledWith({ tags: ["remote"] });
 		expect(JSON.parse(result.content[0]!.text)).toEqual([{ id: "app-1", company: "Acme", role: "Engineer" }]);
 	});

@@ -8,12 +8,6 @@ export const applicationStatusSchema = z.enum(APPLICATION_STATUSES);

 export type ApplicationStatus = z.infer<typeof applicationStatusSchema>;

-/** A stage as clients may send it: the retired `rejected` stage still arrives from older clients and means closed. */
-export const applicationStatusInputSchema = z.preprocess(
-	(value) => (value === "rejected" ? "closed" : value),
-	applicationStatusSchema,
-);
-
 export const applicationClosedReasonSchema = z.enum(["not-selected", "withdrew", "accepted-other", "no-response"]);

 export type ApplicationClosedReason = z.infer<typeof applicationClosedReasonSchema>;
@@ -79,8 +73,7 @@ export type InterviewDetails = z.infer<typeof interviewDetailsSchema>;
 export const applicationTimelineEntrySchema = z.discriminatedUnion("type", [
 	timelineBaseSchema.extend({
 		type: z.literal("stage"),
-		// History written before the closed stage may still say `rejected`.
-		stage: applicationStatusInputSchema,
+		stage: applicationStatusSchema,
 	}),
 	timelineBaseSchema.extend({
 		type: z.literal("note"),
@@ -137,7 +137,9 @@ export const awardItemSchema = baseItemSchema.extend({
 	awarder: z.string().describe("The awarder of the award."),
 	date: z
 		.string()
-		.describe("The date when the award was received, as text. Kept in sync from `dates`; prefer writing `dates`."),
+		.describe(
+			"The date when the award was received, as text. Written from `dates` on every save; write `dates` instead.",
+		),
 	dates: resumeDatesSchema.optional(),
 	website: itemWebsiteSchema.describe("The website of the award, if any."),
 	description: z.string().describe("The description of the award. This should be a HTML-formatted string."),
@@ -149,7 +151,7 @@ export const certificationItemSchema = baseItemSchema.extend({
 	date: z
 		.string()
 		.describe(
-			"The date when the certification was received, as text. Kept in sync from `dates`; prefer writing `dates`.",
+			"The date when the certification was received, as text. Written from `dates` on every save; write `dates` instead.",
 		),
 	dates: resumeDatesSchema.optional(),
 	website: itemWebsiteSchema.describe("The website of the certification, if any."),
@@ -167,7 +169,7 @@ export const educationItemSchema = baseItemSchema.extend({
 	period: z
 		.string()
 		.describe(
-			"The period of time the education was obtained over, as text. Kept in sync from `dates`; prefer writing `dates`.",
+			"The period of time the education was obtained over, as text. Written from `dates` on every save; write `dates` instead.",
 		),
 	dates: resumeDatesSchema.optional(),
 	website: itemWebsiteSchema.describe("The website of the school or institution, if any."),
@@ -179,7 +181,9 @@ const roleItemSchema = z.object({
 	position: z.string().describe("The position or job title for this role."),
 	period: z
 		.string()
-		.describe("The period of time this role was held, as text. Kept in sync from `dates`; prefer writing `dates`."),
+		.describe(
+			"The period of time this role was held, as text. Written from `dates` on every save; write `dates` instead.",
+		),
 	dates: resumeDatesSchema.optional(),
 	description: z.string().describe("The description of this specific role. This should be a HTML-formatted string."),
 });
@@ -199,7 +203,7 @@ export const experienceItemSchema = baseItemSchema.extend({
 	period: z
 		.string()
 		.describe(
-			"The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Kept in sync from `dates`; prefer writing `dates`.",
+			"The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Written from `dates` on every save; write `dates` instead.",
 		),
 	dates: resumeDatesSchema.optional(),
 	website: itemWebsiteSchema.describe("The website of the company or organization, if any."),
@@ -254,7 +258,7 @@ export const projectItemSchema = baseItemSchema.extend({
 	period: z
 		.string()
 		.describe(
-			"The period of time the project was worked on, as text. Kept in sync from `dates`; prefer writing `dates`.",
+			"The period of time the project was worked on, as text. Written from `dates` on every save; write `dates` instead.",
 		),
 	dates: resumeDatesSchema.optional(),
 	website: itemWebsiteSchema.describe("The link to the project, if any."),
@@ -267,7 +271,7 @@ export const publicationItemSchema = baseItemSchema.extend({
 	date: z
 		.string()
 		.describe(
-			"The date when the publication was published, as text. Kept in sync from `dates`; prefer writing `dates`.",
+			"The date when the publication was published, as text. Written from `dates` on every save; write `dates` instead.",
 		),
 	dates: resumeDatesSchema.optional(),
 	website: itemWebsiteSchema.describe("The link to the publication, if any."),
@@ -321,7 +325,7 @@ export const volunteerItemSchema = baseItemSchema.extend({
 	period: z
 		.string()
 		.describe(
-			"The period of time the author was volunteered at the organization or company, as text. Kept in sync from `dates`; prefer writing `dates`.",
+			"The period of time the author was volunteered at the organization or company, as text. Written from `dates` on every save; write `dates` instead.",
 		),
 	dates: resumeDatesSchema.optional(),
 	website: itemWebsiteSchema.describe("The link to the organization or company, if any."),
@@ -240,12 +240,12 @@ describe("syncResumeDates", () => {
 		if (!entry) throw new Error("Missing entry.");
 		entry.dates = { start: "2019-04", end: null, present: true };

-		syncResumeDates(next, previous);
+		syncResumeDates(next);

 		expect(job(next)?.period).toBe("Apr 2019 – Present");
 	});

-	it("reads the text again when only the text changed, as clients unaware of dates do", () => {
+	it("overwrites an edit to the text alone: the dates are the only source", () => {
 		const previous = parseResumeData(resume("2020 - 2022"));
 		syncResumeDates(previous);
 		const next = structuredClone(previous);
@@ -253,10 +253,10 @@ describe("syncResumeDates", () => {
 		if (!entry) throw new Error("Missing entry.");
 		entry.period = "Jan 2018 - Present";

-		syncResumeDates(next, previous);
+		syncResumeDates(next);

-		expect(job(next)?.dates).toEqual({ start: "2018-01", end: null, present: true });
-		expect(job(next)?.period).toBe("Jan 2018 – Present");
+		expect(job(next)?.dates).toEqual({ start: "2020", end: "2022", present: false });
+		expect(job(next)?.period).toBe("2020 – 2022");
 	});

 	it("reformats every entry when the locale or format changes, without reading the text again", () => {
@@ -266,7 +266,7 @@ describe("syncResumeDates", () => {
 		next.metadata.page.locale = "ja-JP";
 		next.metadata.page.dateFormat = "long";

-		syncResumeDates(next, previous);
+		syncResumeDates(next);

 		expect(job(next)?.dates).toEqual({ start: "2022-03", end: null, present: true });
 		expect(job(next)?.period).toBe(
@@ -278,7 +278,7 @@ describe("syncResumeDates", () => {
 		const data = resume("Summer 2016 - 2018");
 		syncResumeDates(data);
 		const once = structuredClone(data);
-		syncResumeDates(data, once);
+		syncResumeDates(data);

 		expect(job(data)?.period).toBe("Summer 2016 - 2018");
 		expect(job(data)?.dates?.raw).toBe("Summer 2016 - 2018");
@@ -42,10 +42,6 @@ export const DEFAULT_DATE_FORMAT: DateFormat = "short";

 export const EMPTY_RESUME_DATES: ResumeDates = { start: null, end: null, present: false };

-/** Dates of the same entry are the same when their fields are; `raw` counts, since it changes what prints. */
-export const areResumeDatesEqual = (a: ResumeDates | undefined, b: ResumeDates | undefined) =>
-	a?.start === b?.start && a?.end === b?.end && a?.present === b?.present && a?.raw === b?.raw;
-
 const labels: Record<string, string> = presentLabels;

 /**
@@ -241,44 +237,20 @@ type PageDateOptions = { locale: string; dateFormat?: DateFormat | undefined };
 export const formatEntryDates = (dates: ResumeDates | undefined, text: string, page: PageDateOptions) =>
 	dates ? formatResumeDates(dates, { locale: page.locale, format: page.dateFormat }) : text;

-type Snapshot = { text: string; dates: ResumeDates | undefined };
-
-const entryKey = (entry: DatedEntry) => (entry as { id?: string }).id;
-
 /**
- * Keeps structured dates and the legacy text (`period` or `date`) in step before resume data is saved, so
- * older app versions and API clients that read the text keep working (the dual write).
- *
- * Dates are the source of truth: the text is rewritten from them in the resume's locale and date format.
- * The one exception is an edit that changed only the text since `previous` (a client that doesn't know
- * structured dates); then the text is read into new dates. Entries without dates get them from their text.
+ * Rewrites each entry's text (`period` or `date`) from its structured dates, in the resume's locale and date
+ * format. Dates are the only source: an edit to the text alone is overwritten. Entries without dates (new
+ * items from imports or older data) get them from their text first.
 *
 * Mutates `data` in place and writes only what changed, so it's safe inside an immer draft.
 */
-export function syncResumeDates(data: ResumeData, previous?: ResumeData) {
+export function syncResumeDates(data: ResumeData) {
 	const { locale, dateFormat } = data.metadata.page;
 	const options: DateFormatOptions = { locale, format: dateFormat, presentLabel: getPresentLabel(locale) };

-	const before = new Map<string, Snapshot>();
-	if (previous) {
-		forEachDatedEntry(previous, (entry, field) => {
-			const key = entryKey(entry);
-			if (key) before.set(key, { text: getLegacyDateText(entry, field), dates: entry.dates });
-		});
-	}
-
 	forEachDatedEntry(data, (entry, field) => {
 		const text = getLegacyDateText(entry, field);
-		const read = () => readLegacyDates(text, { locale, single: field === "date" });
-
-		if (!entry.dates) {
-			entry.dates = read();
-		} else if (text !== formatResumeDates(entry.dates, options)) {
-			const key = entryKey(entry);
-			const snapshot = key ? before.get(key) : undefined;
-			const onlyTextChanged = snapshot && snapshot.text !== text && areResumeDatesEqual(snapshot.dates, entry.dates);
-			if (onlyTextChanged) entry.dates = read();
-		}
+		entry.dates ??= readLegacyDates(text, { locale, single: field === "date" });

 		const formatted = formatResumeDates(entry.dates, options);
 		if (formatted === text) return;
@@ -169,7 +169,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `sections.experience.items[].company` | `string` | yes | — | The name of the company or organization. Empty while the entry is a draft, which isn't printed. |
 | `sections.experience.items[].position` | `string` | yes | — | The position held at the company or organization. Used when there is only a single role. If multiple roles are provided in the 'roles' field, this serves as a summary title or can be left blank. |
 | `sections.experience.items[].location` | `string` | yes | — | The location of the company or organization. |
-| `sections.experience.items[].period` | `string` | yes | — | The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Kept in sync from `dates`; prefer writing `dates`. |
+| `sections.experience.items[].period` | `string` | yes | — | The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Written from `dates` on every save; write `dates` instead. |
 | `sections.experience.items[].dates` | `object` | no | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `sections.experience.items[].dates.start` | `string or null` | yes | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `sections.experience.items[].dates.start` | `string` | yes (variant 1 at sections.experience.items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -188,7 +188,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `sections.experience.items[].roles[]` | `object` | — | — | — |
 | `sections.experience.items[].roles[].id` | `string` | yes | — | The unique identifier for the role. Usually generated as a UUID. |
 | `sections.experience.items[].roles[].position` | `string` | yes | — | The position or job title for this role. |
-| `sections.experience.items[].roles[].period` | `string` | yes | — | The period of time this role was held, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `sections.experience.items[].roles[].period` | `string` | yes | — | The period of time this role was held, as text. Written from `dates` on every save; write `dates` instead. |
 | `sections.experience.items[].roles[].dates` | `object` | no | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `sections.experience.items[].roles[].dates.start` | `string or null` | yes | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `sections.experience.items[].roles[].dates.start` | `string` | yes (variant 1 at sections.experience.items[].roles[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -216,7 +216,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `sections.education.items[].area` | `string` | yes | — | The area of study or specialization. |
 | `sections.education.items[].grade` | `string` | yes | — | The grade or score achieved. |
 | `sections.education.items[].location` | `string` | yes | — | The location of the school or institution. |
-| `sections.education.items[].period` | `string` | yes | — | The period of time the education was obtained over, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `sections.education.items[].period` | `string` | yes | — | The period of time the education was obtained over, as text. Written from `dates` on every save; write `dates` instead. |
 | `sections.education.items[].dates` | `object` | no | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `sections.education.items[].dates.start` | `string or null` | yes | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `sections.education.items[].dates.start` | `string` | yes (variant 1 at sections.education.items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -244,7 +244,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `sections.projects.items[].id` | `string` | yes | — | The unique identifier for the item. Usually generated as a UUID. |
 | `sections.projects.items[].hidden` | `boolean` | yes | — | Whether to hide the item from the resume. |
 | `sections.projects.items[].name` | `string` | yes | — | The name of the project. Empty while the entry is a draft, which isn't printed. |
-| `sections.projects.items[].period` | `string` | yes | — | The period of time the project was worked on, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `sections.projects.items[].period` | `string` | yes | — | The period of time the project was worked on, as text. Written from `dates` on every save; write `dates` instead. |
 | `sections.projects.items[].dates` | `object` | no | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `sections.projects.items[].dates.start` | `string or null` | yes | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `sections.projects.items[].dates.start` | `string` | yes (variant 1 at sections.projects.items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -326,7 +326,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `sections.awards.items[].hidden` | `boolean` | yes | — | Whether to hide the item from the resume. |
 | `sections.awards.items[].title` | `string` | yes | — | The title of the award. Empty while the entry is a draft, which isn't printed. |
 | `sections.awards.items[].awarder` | `string` | yes | — | The awarder of the award. |
-| `sections.awards.items[].date` | `string` | yes | — | The date when the award was received, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `sections.awards.items[].date` | `string` | yes | — | The date when the award was received, as text. Written from `dates` on every save; write `dates` instead. |
 | `sections.awards.items[].dates` | `object` | no | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `sections.awards.items[].dates.start` | `string or null` | yes | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `sections.awards.items[].dates.start` | `string` | yes (variant 1 at sections.awards.items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -355,7 +355,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `sections.certifications.items[].hidden` | `boolean` | yes | — | Whether to hide the item from the resume. |
 | `sections.certifications.items[].title` | `string` | yes | — | The title of the certification. Empty while the entry is a draft, which isn't printed. |
 | `sections.certifications.items[].issuer` | `string` | yes | — | The issuer of the certification. |
-| `sections.certifications.items[].date` | `string` | yes | — | The date when the certification was received, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `sections.certifications.items[].date` | `string` | yes | — | The date when the certification was received, as text. Written from `dates` on every save; write `dates` instead. |
 | `sections.certifications.items[].dates` | `object` | no | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `sections.certifications.items[].dates.start` | `string or null` | yes | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `sections.certifications.items[].dates.start` | `string` | yes (variant 1 at sections.certifications.items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -384,7 +384,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `sections.publications.items[].hidden` | `boolean` | yes | — | Whether to hide the item from the resume. |
 | `sections.publications.items[].title` | `string` | yes | — | The title of the publication. Empty while the entry is a draft, which isn't printed. |
 | `sections.publications.items[].publisher` | `string` | yes | — | The publisher of the publication. |
-| `sections.publications.items[].date` | `string` | yes | — | The date when the publication was published, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `sections.publications.items[].date` | `string` | yes | — | The date when the publication was published, as text. Written from `dates` on every save; write `dates` instead. |
 | `sections.publications.items[].dates` | `object` | no | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `sections.publications.items[].dates.start` | `string or null` | yes | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `sections.publications.items[].dates.start` | `string` | yes (variant 1 at sections.publications.items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -413,7 +413,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `sections.volunteer.items[].hidden` | `boolean` | yes | — | Whether to hide the item from the resume. |
 | `sections.volunteer.items[].organization` | `string` | yes | — | The name of the organization or company. Empty while the entry is a draft, which isn't printed. |
 | `sections.volunteer.items[].location` | `string` | yes | — | The location of the organization or company. |
-| `sections.volunteer.items[].period` | `string` | yes | — | The period of time the author was volunteered at the organization or company, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `sections.volunteer.items[].period` | `string` | yes | — | The period of time the author was volunteered at the organization or company, as text. Written from `dates` on every save; write `dates` instead. |
 | `sections.volunteer.items[].dates` | `object` | no | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `sections.volunteer.items[].dates.start` | `string or null` | yes | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `sections.volunteer.items[].dates.start` | `string` | yes (variant 1 at sections.volunteer.items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -507,7 +507,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `customSections[].items[].company` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The name of the company or organization. Empty while the entry is a draft, which isn't printed. |
 | `customSections[].items[].position` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The position held at the company or organization. Used when there is only a single role. If multiple roles are provided in the 'roles' field, this serves as a summary title or can be left blank. |
 | `customSections[].items[].location` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The location of the company or organization. |
-| `customSections[].items[].period` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Kept in sync from `dates`; prefer writing `dates`. |
+| `customSections[].items[].period` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The overall period of time at the company, as text. When multiple roles are used, this should reflect the total tenure. Written from `dates` on every save; write `dates` instead. |
 | `customSections[].items[].dates` | `object` | no (type experience, schema experienceItemSchema at customSections[]) | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `customSections[].items[].dates.start` | `string or null` | yes (type experience, schema experienceItemSchema at customSections[]) | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `customSections[].items[].dates.start` | `string` | yes (type experience, schema experienceItemSchema at customSections[]; variant 1 at customSections[].items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -526,7 +526,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `customSections[].items[].roles[]` | `object` | — (type experience, schema experienceItemSchema at customSections[]) | — | — |
 | `customSections[].items[].roles[].id` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The unique identifier for the role. Usually generated as a UUID. |
 | `customSections[].items[].roles[].position` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The position or job title for this role. |
-| `customSections[].items[].roles[].period` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The period of time this role was held, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `customSections[].items[].roles[].period` | `string` | yes (type experience, schema experienceItemSchema at customSections[]) | — | The period of time this role was held, as text. Written from `dates` on every save; write `dates` instead. |
 | `customSections[].items[].roles[].dates` | `object` | no (type experience, schema experienceItemSchema at customSections[]) | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `customSections[].items[].roles[].dates.start` | `string or null` | yes (type experience, schema experienceItemSchema at customSections[]) | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `customSections[].items[].roles[].dates.start` | `string` | yes (type experience, schema experienceItemSchema at customSections[]; variant 1 at customSections[].items[].roles[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -557,7 +557,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `customSections[].items[].area` | `string` | yes (type education, schema educationItemSchema at customSections[]) | — | The area of study or specialization. |
 | `customSections[].items[].grade` | `string` | yes (type education, schema educationItemSchema at customSections[]) | — | The grade or score achieved. |
 | `customSections[].items[].location` | `string` | yes (type education, schema educationItemSchema at customSections[]) | — | The location of the school or institution. |
-| `customSections[].items[].period` | `string` | yes (type education, schema educationItemSchema at customSections[]) | — | The period of time the education was obtained over, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `customSections[].items[].period` | `string` | yes (type education, schema educationItemSchema at customSections[]) | — | The period of time the education was obtained over, as text. Written from `dates` on every save; write `dates` instead. |
 | `customSections[].items[].dates` | `object` | no (type education, schema educationItemSchema at customSections[]) | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `customSections[].items[].dates.start` | `string or null` | yes (type education, schema educationItemSchema at customSections[]) | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `customSections[].items[].dates.start` | `string` | yes (type education, schema educationItemSchema at customSections[]; variant 1 at customSections[].items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -588,7 +588,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `customSections[].items[].id` | `string` | yes (type projects, schema projectItemSchema at customSections[]) | — | The unique identifier for the item. Usually generated as a UUID. |
 | `customSections[].items[].hidden` | `boolean` | yes (type projects, schema projectItemSchema at customSections[]) | — | Whether to hide the item from the resume. |
 | `customSections[].items[].name` | `string` | yes (type projects, schema projectItemSchema at customSections[]) | — | The name of the project. Empty while the entry is a draft, which isn't printed. |
-| `customSections[].items[].period` | `string` | yes (type projects, schema projectItemSchema at customSections[]) | — | The period of time the project was worked on, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `customSections[].items[].period` | `string` | yes (type projects, schema projectItemSchema at customSections[]) | — | The period of time the project was worked on, as text. Written from `dates` on every save; write `dates` instead. |
 | `customSections[].items[].dates` | `object` | no (type projects, schema projectItemSchema at customSections[]) | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `customSections[].items[].dates.start` | `string or null` | yes (type projects, schema projectItemSchema at customSections[]) | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `customSections[].items[].dates.start` | `string` | yes (type projects, schema projectItemSchema at customSections[]; variant 1 at customSections[].items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -680,7 +680,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `customSections[].items[].hidden` | `boolean` | yes (type awards, schema awardItemSchema at customSections[]) | — | Whether to hide the item from the resume. |
 | `customSections[].items[].title` | `string` | yes (type awards, schema awardItemSchema at customSections[]) | — | The title of the award. Empty while the entry is a draft, which isn't printed. |
 | `customSections[].items[].awarder` | `string` | yes (type awards, schema awardItemSchema at customSections[]) | — | The awarder of the award. |
-| `customSections[].items[].date` | `string` | yes (type awards, schema awardItemSchema at customSections[]) | — | The date when the award was received, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `customSections[].items[].date` | `string` | yes (type awards, schema awardItemSchema at customSections[]) | — | The date when the award was received, as text. Written from `dates` on every save; write `dates` instead. |
 | `customSections[].items[].dates` | `object` | no (type awards, schema awardItemSchema at customSections[]) | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `customSections[].items[].dates.start` | `string or null` | yes (type awards, schema awardItemSchema at customSections[]) | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `customSections[].items[].dates.start` | `string` | yes (type awards, schema awardItemSchema at customSections[]; variant 1 at customSections[].items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -712,7 +712,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `customSections[].items[].hidden` | `boolean` | yes (type certifications, schema certificationItemSchema at customSections[]) | — | Whether to hide the item from the resume. |
 | `customSections[].items[].title` | `string` | yes (type certifications, schema certificationItemSchema at customSections[]) | — | The title of the certification. Empty while the entry is a draft, which isn't printed. |
 | `customSections[].items[].issuer` | `string` | yes (type certifications, schema certificationItemSchema at customSections[]) | — | The issuer of the certification. |
-| `customSections[].items[].date` | `string` | yes (type certifications, schema certificationItemSchema at customSections[]) | — | The date when the certification was received, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `customSections[].items[].date` | `string` | yes (type certifications, schema certificationItemSchema at customSections[]) | — | The date when the certification was received, as text. Written from `dates` on every save; write `dates` instead. |
 | `customSections[].items[].dates` | `object` | no (type certifications, schema certificationItemSchema at customSections[]) | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `customSections[].items[].dates.start` | `string or null` | yes (type certifications, schema certificationItemSchema at customSections[]) | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `customSections[].items[].dates.start` | `string` | yes (type certifications, schema certificationItemSchema at customSections[]; variant 1 at customSections[].items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -744,7 +744,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `customSections[].items[].hidden` | `boolean` | yes (type publications, schema publicationItemSchema at customSections[]) | — | Whether to hide the item from the resume. |
 | `customSections[].items[].title` | `string` | yes (type publications, schema publicationItemSchema at customSections[]) | — | The title of the publication. Empty while the entry is a draft, which isn't printed. |
 | `customSections[].items[].publisher` | `string` | yes (type publications, schema publicationItemSchema at customSections[]) | — | The publisher of the publication. |
-| `customSections[].items[].date` | `string` | yes (type publications, schema publicationItemSchema at customSections[]) | — | The date when the publication was published, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `customSections[].items[].date` | `string` | yes (type publications, schema publicationItemSchema at customSections[]) | — | The date when the publication was published, as text. Written from `dates` on every save; write `dates` instead. |
 | `customSections[].items[].dates` | `object` | no (type publications, schema publicationItemSchema at customSections[]) | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `customSections[].items[].dates.start` | `string or null` | yes (type publications, schema publicationItemSchema at customSections[]) | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `customSections[].items[].dates.start` | `string` | yes (type publications, schema publicationItemSchema at customSections[]; variant 1 at customSections[].items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -776,7 +776,7 @@ Choose one coherent shape for each union value. Required fields are local to tha
 | `customSections[].items[].hidden` | `boolean` | yes (type volunteer, schema volunteerItemSchema at customSections[]) | — | Whether to hide the item from the resume. |
 | `customSections[].items[].organization` | `string` | yes (type volunteer, schema volunteerItemSchema at customSections[]) | — | The name of the organization or company. Empty while the entry is a draft, which isn't printed. |
 | `customSections[].items[].location` | `string` | yes (type volunteer, schema volunteerItemSchema at customSections[]) | — | The location of the organization or company. |
-| `customSections[].items[].period` | `string` | yes (type volunteer, schema volunteerItemSchema at customSections[]) | — | The period of time the author was volunteered at the organization or company, as text. Kept in sync from `dates`; prefer writing `dates`. |
+| `customSections[].items[].period` | `string` | yes (type volunteer, schema volunteerItemSchema at customSections[]) | — | The period of time the author was volunteered at the organization or company, as text. Written from `dates` on every save; write `dates` instead. |
 | `customSections[].items[].dates` | `object` | no (type volunteer, schema volunteerItemSchema at customSections[]) | — | Structured dates. Write these rather than the legacy text field (`period` or `date`), which is kept in sync from them. |
 | `customSections[].items[].dates.start` | `string or null` | yes (type volunteer, schema volunteerItemSchema at customSections[]) | default: null | When it started, or the date itself for single-date entries (awards, certifications, publications). |
 | `customSections[].items[].dates.start` | `string` | yes (type volunteer, schema volunteerItemSchema at customSections[]; variant 1 at customSections[].items[].dates.start) | — | A year ("2022") or a year and month ("2022-03"). |
@@ -22,20 +22,20 @@ test("exports filtered or all owned applications with an inclusive date range",
 			String(testInfo.project.use.baseURL),
 		);
 		await otherContext.close();
-		for (const [email, company, date, archived] of [
-			[account.email, "Alpha Company", "2026-08-03T00:00:00Z", false],
-			[account.email, "Beta Archived", "2026-08-03T23:59:59Z", true],
-			[account.email, "Gamma Later", "2026-08-04T00:00:00Z", false],
-			[foreign.email, "Foreign Secret", "2026-08-03T12:00:00Z", false],
+		for (const [email, company, date, status] of [
+			[account.email, "Alpha Company", "2026-08-03T00:00:00Z", "applied"],
+			[account.email, "Beta Closed", "2026-08-03T23:59:59Z", "closed"],
+			[account.email, "Gamma Later", "2026-08-04T00:00:00Z", "applied"],
+			[foreign.email, "Foreign Secret", "2026-08-03T12:00:00Z", "applied"],
 		] as const) {
 			await pool.query(
-				`INSERT INTO application (id,user_id,company,role,status,applied_at,archived,notes,activity)
-				 SELECT $1,id,$2,'Engineer','applied',$3,$4,$5,$6 FROM "user" WHERE email=$7`,
+				`INSERT INTO application (id,user_id,company,role,status,applied_at,notes,activity)
+				 SELECT $1,id,$2,'Engineer',$4,$3,$5,$6 FROM "user" WHERE email=$7`,
 				[
 					randomUUID(),
 					company,
 					date,
-					archived,
+					status,
 					'Quoted "note"\nSecond line',
 					JSON.stringify([{ id: randomUUID(), type: "stage", stage: "applied", at: date }]),
 					email,
@@ -57,7 +57,7 @@ test("exports filtered or all owned applications with an inclusive date range",
 		if (!path) throw new Error("CSV download was not saved");
 		let csv = await readFile(path, "utf8");
 		expect(csv).toContain('"Alpha Company"');
-		expect(csv).not.toMatch(/Beta Archived|Gamma Later|Foreign Secret/);
+		expect(csv).not.toMatch(/Beta Closed|Gamma Later|Foreign Secret/);
 		expect(csv).toContain('"Applied (2026-08-03)"');
 		expect(csv).toContain('"Quoted ""note""\nSecond line"');

@@ -76,7 +76,7 @@ test("exports filtered or all owned applications with an inclusive date range",
 		if (!path) throw new Error("CSV download was not saved");
 		csv = await readFile(path, "utf8");
 		expect(csv).toContain('"Alpha Company"');
-		expect(csv).toContain('"Beta Archived"');
+		expect(csv).toContain('"Beta Closed"');
 		expect(csv).not.toMatch(/Gamma Later|Foreign Secret/);
 		await page.setViewportSize({ width: 390, height: 844 });
 		await page.reload();