@@ -0,0 +1,62 @@
+---
+'@object-ui/plugin-grid': minor
+'@object-ui/plugin-list': minor
+'@object-ui/plugin-kanban': minor
+---
+
+A page size with nothing declared is now the one `@objectstack/spec` declares for
+`pagination.pageSize`, on every surface that has a pager; a fetch that has no
+pager keeps its own fetch batch, which does not follow it (objectui#9853).
+
+"Page size" now means one page, and only where there is a pager. The grid,
+the list and the board read the display default from the spec (`PaginationConfigSchema`)
+instead of keeping numbers of their own, so it is declared once, in the protocol.
+`@objectstack/spec` 17.5.0 declares **50**. The views that fetch one window and
+page nothing keep a fetch batch of their own, with the value they had, so no
+view loses records it could reach before.
+
+⚠️ **Visible change where no page size is declared.** Measured on the renderer
+with `@objectstack/spec` 17.5.0. Declare `pagination: { pageSize: N }` (or, on a
+board, `limit`) to keep an old count:
+
+| surface (nothing declared) | before | after |
+|:--|--:|--:|
+| `object-grid` that fetches its own rows: rows per page, and the `$top` of each page fetch | 50 | 50 (unchanged) |
+| `object-grid` over inline `data`: rows per page | 10 | **50** |
+| grouped `object-grid`: groups per page | 10 | **50** |
+| grouped `object-grid` over a data source that groups on the server: rows per group page (`$top`) | 50 | 50 (unchanged) |
+| `list-view`, grid view (paged on the server): rows per page, and `$top` | 100 | **50** |
+| `list-view`, every view it does not page (kanban, calendar, gallery, timeline and the rest, and a grouped grid): the one fetch's `$top` | 100 | 100 (unchanged) |
+| `object-kanban` with no `limit`: the board's `$top` | 100 | 100 (unchanged) |
+
+What changed underneath:
+
+- **`@object-ui/plugin-grid`.** `ObjectGrid` falls back to the spec's display
+  default on every page it shows, with no local number behind it: if a future
+  spec stops declaring the default, the module throws at load instead of
+  inventing one. The old "server window" constant is renamed
+  `DEFAULT_FETCH_BATCH_SIZE` and keeps its value. Since grouping moved to the
+  server, it is read only when the grid groups a window of rows in the browser
+  because the server does not group it (a grouping key the principal may not
+  read), and never as a page size. The grouped view's rows-per-page selector
+  also lists the size in force when it is not one of its fixed steps.
+- **`@object-ui/plugin-list`.** `ListView` splits its undeclared fallback by
+  whether the view pages. The grid view, which it pages on the server, falls
+  back to the spec's display default. Every other view falls back to a fetch
+  batch of 100, the value it had. A declared `pagination.pageSize` sizes the
+  window on every view, as before. With no declared size, switching between the
+  grid view and an unpaged view now changes the window, so the list fetches
+  again: one request, whose `$top` is the fetch batch. That qualifies the
+  objectui#10512 note in this same release, which says a switch into a dataset
+  chart with no user filter set issues no new request: it still holds when a
+  page size is declared. A refused page size still warns, and the warning names
+  the fallback that view actually used.
+- **`@object-ui/plugin-kanban`.** The board's default `limit` is renamed from
+  `DEFAULT_KANBAN_LIMIT` to `DEFAULT_KANBAN_FETCH_BATCH_SIZE` and stays 100. Its
+  diagnostic for a refused `limit` now calls it the board's fetch batch.
+
+The display default follows the `@objectstack/spec` a host installs. A host
+that resolves an older spec gets that spec's default, on every surface alike.
+
+Released as `minor`, not `major`: this repository keeps its major aligned with
+`@objectstack/spec`, so the behaviour change is spelled out here instead.
@@ -367,6 +367,15 @@ see [Row callbacks are component props](#row-callbacks-are-component-props).
 always carries a rows-per-page picker, and `pageSizeOptions` only replaces the
 choices it offers with your own.
 
+With no `pageSize` declared, every page the grid shows — the server-paged table,
+the table over inline `data`, the grouped view's page of groups, and a group's
+own page of rows — uses the default `@objectstack/spec` declares for
+`pagination.pageSize`; the grid reads it from the spec rather than keeping a
+number of its own. Declare `pagination.pageSize` to choose the count yourself.
+A window of rows the grid groups in the browser, because the server does not
+group it, is not a page: undeclared, it is a fixed fetch batch of the grid's
+own, and it does not follow the display default.
+
 ### Grouping is server-side
 
 `grouping` is answered by the **server**, not by bucketing the records the
 
@@ -208,11 +208,13 @@ rendered on each card resolve in priority order:
 Defaulting to `highlightFields` keeps a card's contents consistent with the
 object's other views without every kanban view having to re-declare its fields.
 
-### Row cap (object-driven boards)
+### Fetch batch (object-driven boards)
 
 An object-driven board fetches at most `limit` records, defaulting to **100**. A
 board renders every fetched record into a lane and offers no pagination control,
-so this is the author's window on the object rather than a page size:
+so this is the author's window on the object — one fetch batch — rather than a
+page size. The default does not follow the display page size `@objectstack/spec`
+declares for `pagination.pageSize`:
 
 ```tsx
 import type { ObjectKanbanSchema } from '@object-ui/types'
 
@@ -0,0 +1,258 @@
+/**
+ * ObjectUI
+ * Copyright (c) 2024-present ObjectStack Inc.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+/**
+ * objectui#9853, ruling 5824040487 — the display default is declared ONCE, in
+ * the protocol, and every paged surface reads it; the fetch batches of the
+ * unpaged surfaces are a different quantity and do not follow it (structure B:
+ * "page size" means one page only where there is a pager).
+ *
+ * ## Why this file lives in `app-shell`
+ *
+ * The surfaces sit in three packages — `ObjectGrid` in `plugin-grid`,
+ * `ListView` in `plugin-list`, `ObjectKanban` in `plugin-kanban` — and the
+ * point of the pin is ONE assertion across all of them, so that a surface
+ * keeping a default of its own shows up as the odd one out. `app-shell` is the
+ * package that already depends on all three, so the real components are
+ * mounted here with no new dependency edge anywhere (the same reason
+ * `rowClickModifierPayloadForward-9462.test.tsx` lives here).
+ *
+ * ## Why the spec's default is STOOD IN FOR
+ *
+ * Reading the real default cannot tell "reads the spec" from "hardcodes the
+ * number the spec says today". So the default is replaced with a number the
+ * spec has never shipped, and every expectation reads it back from the spec
+ * module the renderers import. The stand-in records that its factory ran and
+ * what the REAL default is, so the first row proves the substitution is live.
+ * With the stand-in in force the paged surfaces must move with it and the
+ * fetch batches must not — which is exactly the split the ruling makes.
+ */
+import React from 'react';
+import { describe, it, expect, vi, afterEach } from 'vitest';
+import { render, cleanup, waitFor } from '@testing-library/react';
+import '@testing-library/jest-dom';
+
+const { STUB_DISPLAY_DEFAULT, probe } = vi.hoisted(() => ({
+  /** Deliberately nothing the spec has shipped, and neither fetch batch. */
+  STUB_DISPLAY_DEFAULT: 13,
+  probe: { factoryRan: false, realDefault: undefined as unknown },
+}));
+
+vi.mock('@objectstack/spec/ui', async (importOriginal) => {
+  const actual = await importOriginal<typeof import('@objectstack/spec/ui')>();
+  probe.factoryRan = true;
+  probe.realDefault = actual.PaginationConfigSchema.parse({}).pageSize;
+  return {
+    ...actual,
+    PaginationConfigSchema: actual.PaginationConfigSchema.extend({
+      pageSize: actual.PaginationConfigSchema.shape.pageSize.unwrap().default(STUB_DISPLAY_DEFAULT),
+    }),
+  };
+});
+
+import { PaginationConfigSchema } from '@objectstack/spec/ui';
+import { ObjectGrid } from '@object-ui/plugin-grid';
+import { ListView } from '@object-ui/plugin-list';
+import { ObjectKanban } from '@object-ui/plugin-kanban';
+import { ActionProvider, SchemaRendererProvider } from '@object-ui/react';
+
+/** The protocol's answer, read the way the renderers read it. */
+const SPEC_DISPLAY_DEFAULT: number = PaginationConfigSchema.parse({}).pageSize;
+
+/**
+ * The fetch batch of every unpaged list view and of the board. Spelled by
+ * value: the ruling keeps the value, and what it asks is that neither moves
+ * with the display default.
+ */
+const FETCH_BATCH = 100;
+
+const OBJECT = 'duly_task';
+const OBJECT_FIELDS = {
+  id: { type: 'text', label: 'Id' },
+  subject: { type: 'text', label: 'Subject' },
+  unit: { type: 'text', label: 'Unit' },
+  status: { type: 'select', label: 'Status', options: [{ label: 'Todo', value: 'todo' }] },
+};
+
+/** More rows than any number here, every row its own group. */
+const ROW_COUNT = 240;
+interface Row { id: string; subject: string; unit: string; status: string }
+const ROWS: Row[] = Array.from({ length: ROW_COUNT }, (_, i) => ({
+  id: `r-${i}`,
+  subject: `Task ${String(i).padStart(3, '0')}`,
+  unit: `Unit ${String(i).padStart(3, '0')}`,
+  status: 'todo',
+}));
+
+/** The `FilterCondition` subset the compiled group queries use. */
+function matches(row: Record<string, unknown>, cond: unknown): boolean {
+  if (cond === undefined || cond === null) return true;
+  return Object.entries(cond as Record<string, unknown>).every(([key, value]) => {
+    if (key === '$and') return (value as unknown[]).every((c) => matches(row, c));
+    if (value && typeof value === 'object' && !Array.isArray(value)) {
+      const op = value as Record<string, unknown>;
+      if ('$eq' in op) return row[key] === op.$eq;
+      if ('$null' in op) return op.$null ? row[key] == null : row[key] != null;
+      throw new Error(`double: unsupported operator ${JSON.stringify(op)}`);
+    }
+    return row[key] === value;
+  });
+}
+
+/** Pages `find` by `$top` / `$skip` and answers the group header query. */
+const makeDataSource = () => {
+  const calls: Array<Record<string, unknown>> = [];
+  return {
+    calls,
+    queryGroupHeaders: vi.fn(async (_object: string, query: { where?: unknown; groupBy?: string[]; aggregations?: Array<{ function: string; alias: string }> }) => {
+      const groupBy = query.groupBy ?? [];
+      const buckets = new Map<string, Row[]>();
+      for (const row of ROWS.filter((r) => matches(r as any, query.where))) {
+        const key = JSON.stringify(groupBy.map((f) => (row as any)[f] ?? null));
+        if (!buckets.has(key)) buckets.set(key, []);
+        buckets.get(key)!.push(row);
+      }
+      return [...buckets.entries()].map(([key, members]) => {
+        const values = JSON.parse(key) as unknown[];
+        const out: Record<string, unknown> = Object.fromEntries(groupBy.map((f, i) => [f, values[i]]));
+        for (const agg of query.aggregations ?? []) {
+          out[agg.alias] = agg.function === 'count' ? members.length : null;
+        }
+        return out as { count: number };
+      });
+    }),
+    find: vi.fn(async (_object: string, params: Record<string, unknown>) => {
+      calls.push(params);
+      const where = Array.isArray(params.$filter) ? undefined : params.$filter;
+      const matching = ROWS.filter((r) => matches(r as any, where));
+      const skip = (params.$skip as number | undefined) ?? 0;
+      const top = (params.$top as number | undefined) ?? matching.length;
+      return { data: matching.slice(skip, skip + top), total: matching.length };
+    }),
+    findOne: vi.fn(),
+    create: vi.fn(),
+    update: vi.fn(),
+    delete: vi.fn(),
+    getObjectSchema: vi.fn(async () => ({ name: OBJECT, fields: OBJECT_FIELDS })),
+  };
+};
+type Ds = ReturnType<typeof makeDataSource>;
+
+const mount = (node: React.ReactElement, ds?: Ds) =>
+  render(
+    <SchemaRendererProvider dataSource={ds as any}>
+      <ActionProvider>{node}</ActionProvider>
+    </SchemaRendererProvider>,
+  );
+
+const bodyRows = (container: HTMLElement) => Array.from(container.querySelectorAll('tbody tr'));
+const groupRows = () => [...document.querySelectorAll('[data-testid^="group-row-"]')];
+const tops = (ds: Ds) => ds.calls.map((p) => p.$top);
+/** The `$top` of every per-group row query (a row query carries a group `$filter`). */
+const groupRowTops = (ds: Ds) =>
+  [...new Set(ds.calls.filter((p) => p.$filter && !Array.isArray(p.$filter)).map((p) => p.$top))];
+
+afterEach(() => cleanup());
+
+describe('one display default from the spec across grid, list and board; fetch batches apart (objectui#9853)', () => {
+  it('the spec stand-in is live: the renderers\' spec module answers it, and it is not the real default', () => {
+    expect(probe.factoryRan).toBe(true);
+    expect(SPEC_DISPLAY_DEFAULT).toBe(STUB_DISPLAY_DEFAULT);
+    expect(probe.realDefault).not.toBe(STUB_DISPLAY_DEFAULT);
+    expect(FETCH_BATCH).not.toBe(SPEC_DISPLAY_DEFAULT);
+  });
+
+  it('every paged surface shows the spec default when nothing is declared', async () => {
+    const gridSchema = { type: 'object-grid', objectName: OBJECT, columns: ['subject'] };
+
+    // The grid that fetches its own rows: the page on screen and its `$top`.
+    const flatDs = makeDataSource();
+    const flat = mount(<ObjectGrid schema={gridSchema as any} dataSource={flatDs as any} />, flatDs);
+    await waitFor(() => expect(bodyRows(flat.container).length).toBeGreaterThan(0));
+    const gridServerPaged = bodyRows(flat.container).length;
+    const gridServerTop = tops(flatDs).at(-1);
+    cleanup();
+
+    // The grid over inline rows.
+    const inline = mount(
+      <ObjectGrid schema={{ ...gridSchema, data: { provider: 'value', items: ROWS } } as any} />,
+    );
+    await waitFor(() => expect(bodyRows(inline.container).length).toBeGreaterThan(0));
+    const gridInline = bodyRows(inline.container).length;
+    cleanup();
+
+    // The server-grouped grid: a page of groups, and each open group's page.
+    const groupedDs = makeDataSource();
+    mount(
+      <ObjectGrid
+        schema={{ ...gridSchema, grouping: { fields: [{ field: 'unit' }] } } as any}
+        dataSource={groupedDs as any}
+      />,
+      groupedDs,
+    );
+    await waitFor(() => expect(groupRows().length).toBeGreaterThan(0));
+    await waitFor(() => expect(groupRowTops(groupedDs).length).toBeGreaterThan(0));
+    const groupsShown = groupRows().length;
+    const groupRowTop = groupRowTops(groupedDs);
+    cleanup();
+
+    // The paged list: `ListView` pages its grid view on the server and hands
+    // the real grid the window and the pager.
+    const listDs = makeDataSource();
+    const list = mount(
+      <ListView
+        schema={{ type: 'list-view', objectName: OBJECT, viewType: 'grid', columns: ['subject'] } as any}
+        dataSource={listDs as any}
+      />,
+      listDs,
+    );
+    await waitFor(() => expect(bodyRows(list.container).length).toBeGreaterThan(0));
+    const listPaged = bodyRows(list.container).length;
+    const listTop = tops(listDs);
+
+    expect({ gridServerPaged, gridServerTop, gridInline, groupsShown, groupRowTop, listPaged, listTop }).toEqual({
+      gridServerPaged: SPEC_DISPLAY_DEFAULT,
+      gridServerTop: SPEC_DISPLAY_DEFAULT,
+      gridInline: SPEC_DISPLAY_DEFAULT,
+      groupsShown: SPEC_DISPLAY_DEFAULT,
+      groupRowTop: [SPEC_DISPLAY_DEFAULT],
+      listPaged: SPEC_DISPLAY_DEFAULT,
+      listTop: [SPEC_DISPLAY_DEFAULT],
+    });
+  });
+
+  it('the unpaged list and the board keep their fetch batch, which does not move with the page size', async () => {
+    // An unpaged list view: no pager, so its one window is the batch.
+    const listDs = makeDataSource();
+    mount(
+      <ListView
+        schema={{ type: 'list-view', objectName: OBJECT, viewType: 'kanban', columns: ['subject'], kanban: { groupByField: 'status' } } as any}
+        dataSource={listDs as any}
+      />,
+      listDs,
+    );
+    await waitFor(() => expect(listDs.find).toHaveBeenCalled());
+    const listUnpagedTop = tops(listDs);
+    cleanup();
+
+    // The board on its own: `limit` undeclared, its default fetch batch.
+    const boardDs = makeDataSource();
+    mount(
+      <ObjectKanban
+        schema={{ type: 'object-kanban', objectName: OBJECT, groupBy: 'status' } as any}
+        dataSource={boardDs as any}
+      />,
+      boardDs,
+    );
+    await waitFor(() => expect(boardDs.find).toHaveBeenCalled());
+    const boardTop = tops(boardDs);
+
+    expect({ listUnpagedTop, boardTop }).toEqual({ listUnpagedTop: [FETCH_BATCH], boardTop: [FETCH_BATCH] });
+    expect([...listUnpagedTop, ...boardTop]).not.toContain(SPEC_DISPLAY_DEFAULT);
+  });
+});
@@ -104,7 +104,7 @@ describe('object-kanban inspector · the declared controls are on screen (object
     expect(screen.getByLabelText('Limit')).toBeTruthy();
   });
 
-  it('the row cap states DEFAULT_KANBAN_LIMIT while it is empty, and commits a number', () => {
+  it('the fetch batch box states DEFAULT_KANBAN_FETCH_BATCH_SIZE while it is empty, and commits a number', () => {
     const onPatch = renderInspector(pageDraft({ objectName: 'opportunity' }));
     const box = screen.getByLabelText('Limit') as HTMLInputElement;
     // Empty box, and the placeholder says what applies anyway.
 
@@ -389,7 +389,7 @@ describe('placeholders that must NOT be translated (#3979)', () => {
     'record:related_list.limit': { literal: '10', because: 'a row limit' },
     'object-kanban.limit': {
       literal: '100',
-      because: "a row limit — DEFAULT_KANBAN_LIMIT, the board's fetch cap when the box is empty",
+      because: "a row limit — DEFAULT_KANBAN_FETCH_BATCH_SIZE, the board's fetch batch when the box is empty",
     },
     'record:details.sections.columns': { literal: '2', because: 'a column count' },
   };
@@ -403,7 +403,7 @@ describe('placeholders that must NOT be translated (#3979)', () => {
     // PR objectstack#7115 retired `PageHeaderProps.icon`. Then 17 / 7 / 10
     // until objectui#7772 gave `object-kanban` the `limit` control its schema
     // has declared since objectui#7322 — a VALUE placeholder (`100`,
-    // `DEFAULT_KANBAN_LIMIT`), so it lands on the literal side and the keyed
+    // `DEFAULT_KANBAN_FETCH_BATCH_SIZE`), so it lands on the literal side and the keyed
     // half is what stays untouched this time. Then 18 / 7 / 11 until
     // objectui#8280 removed the `ai:input` panel, whose `agentName` box was a
     // KEYED placeholder — the palette never offered that block.