pcbjam/tests/e2e/modal.spec.ts

293 lines
13 KiB
TypeScript
Raw Normal View History

// Modal dialog styling + drag tests — regression coverage for pcbjam #22.
//
// A wxWidgets top-level window (every modal dialog) is rendered by the wx.js
// runtime glue as `<div class="window toplevel">` (the page-positioned frame)
// containing a `<canvas class="window-canvas">` that C++ paints into. The main
// app frame is id=0 (the page's #canvas) and is NOT a `.window`, so a visible
// `#window-container .window.toplevel` is exactly a dialog.
//
// Two bugs, each asserted here:
// 1. No border/shadow — the `.window` CSS had no border or box-shadow, so
// dialogs looked flat. Fixed by a `.window.toplevel` rule in the shell
// template (and the production shells).
// 2. Black background on drag — `setWindowRect` unconditionally reset
// `canvas.width/height` on every move, clearing the canvas to transparent;
// the `.window` div's `background-color:black` then showed through with no
// repaint queued. Fixed by only resizing the canvas on a real size change.
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
import { test, expect, waitForWxApp, clickByLabel, findByType } from './utils/fixtures';
import { stableShot } from './utils/element-tracker';
const DIALOG_APP = '/standalone/dialog/dialog_test.html';
// Selector for the (single) visible modal frame.
const MODAL_SEL = '#window-container .window.toplevel';
// Wait until a visible modal `.window.toplevel` exists, then return its
// page-space rect (the div keeps a real getBoundingClientRect even though it is
// pointer-events:none).
async function waitForModalRect(page: import('@playwright/test').Page) {
await page.waitForFunction(
(sel) =>
Array.from(document.querySelectorAll(sel)).some((w) => {
const el = w as HTMLElement;
return el.style.display !== 'none' && el.getBoundingClientRect().width > 0;
}),
MODAL_SEL,
{ timeout: 5000 }
);
const rect = await page.evaluate((sel) => {
const el = Array.from(document.querySelectorAll(sel)).find((w) => {
const e = w as HTMLElement;
return e.style.display !== 'none' && e.getBoundingClientRect().width > 0;
}) as HTMLElement | undefined;
if (!el) return null;
const r = el.getBoundingClientRect();
return { left: r.left, top: r.top, width: r.width, height: r.height };
}, MODAL_SEL);
return rect;
}
// Sample the modal's own canvas and report what fraction of a client-area band
// is opaque (alpha) and non-black. With the drag bug the canvas is cleared to
// transparent (alpha 0) → the black div shows through → opaqueFrac ≈ 0.
async function sampleModalCanvas(page: import('@playwright/test').Page) {
return page.evaluate((sel) => {
const modal = Array.from(document.querySelectorAll(sel)).find((w) => {
const e = w as HTMLElement;
return e.style.display !== 'none' && e.getBoundingClientRect().width > 0;
}) as HTMLElement | undefined;
if (!modal) return null;
const canvas = modal.querySelector('canvas.window-canvas') as HTMLCanvasElement | null;
if (!canvas || !canvas.width || !canvas.height) return null;
const ctx = canvas.getContext('2d');
if (!ctx) return null;
// Band across the client area, just below the title bar; skip the edges.
const x0 = Math.floor(canvas.width * 0.1);
const y0 = Math.floor(canvas.height * 0.2);
const w = Math.max(1, Math.floor(canvas.width * 0.8));
const h = Math.max(1, Math.floor(canvas.height * 0.25));
const data = ctx.getImageData(x0, y0, w, h).data;
let opaque = 0;
let nonBlack = 0;
const total = data.length / 4;
for (let i = 0; i < data.length; i += 4) {
const a = data[i + 3];
if (a > 128) {
opaque++;
if (data[i] > 24 || data[i + 1] > 24 || data[i + 2] > 24) nonBlack++;
}
}
return {
total,
opaqueFrac: opaque / total,
nonBlackFrac: nonBlack / total,
canvasW: canvas.width,
canvasH: canvas.height,
};
}, MODAL_SEL);
}
test.describe('Modal dialog border + drag (pcbjam #22)', () => {
test('modal has a visible border and shadow', async ({ page, testLogger }) => {
await page.goto(DIALOG_APP);
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await waitForWxApp(page);
await clickByLabel(page, 'Custom Dialog');
const rect = await waitForModalRect(page);
expect(rect, 'modal should be visible').not.toBeNull();
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await stableShot(page, 'modal-01-border.png', { fullPage: true });
const style = await page.evaluate((sel) => {
const el = Array.from(document.querySelectorAll(sel)).find((w) => {
const e = w as HTMLElement;
return e.style.display !== 'none' && e.getBoundingClientRect().width > 0;
}) as HTMLElement | undefined;
if (!el) return null;
const cs = getComputedStyle(el);
return {
borderTopStyle: cs.borderTopStyle,
borderTopWidth: cs.borderTopWidth,
boxShadow: cs.boxShadow,
};
}, MODAL_SEL);
expect(style, 'modal element should exist').not.toBeNull();
testLogger.consoleLogs.push(`[MODAL_BORDER] ${JSON.stringify(style)}`);
const hasBorder = style!.borderTopStyle !== 'none' && parseFloat(style!.borderTopWidth) > 0;
const hasShadow = style!.boxShadow !== 'none' && style!.boxShadow !== '';
expect(hasBorder, `expected a visible border, got ${JSON.stringify(style)}`).toBe(true);
expect(hasShadow, `expected a box-shadow, got ${JSON.stringify(style)}`).toBe(true);
});
test('modal background stays painted (not black) after drag', async ({ page, testLogger }) => {
await page.goto(DIALOG_APP);
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await waitForWxApp(page);
// The bare test shell lays out #window-container BELOW the full-size main
// canvas, so the modal (and its DOM title bar) render off the bottom of the
// viewport — the real app shells overlay it. Overlay it here so the DOM title
// bar is reachable by the pointer (the old canvas bar was reached indirectly
// via #canvas registry coords; the DOM bar must be clicked where it renders).
await page.evaluate(() => {
const wc = document.getElementById('window-container');
if (wc) {
wc.style.position = 'absolute';
wc.style.top = '0';
wc.style.left = '0';
}
});
await clickByLabel(page, 'Custom Dialog');
await waitForModalRect(page); // wait for the .window.toplevel to exist
// The modal's INPUT model and VISUAL position are decoupled: the wasm places
// the dialog at its registry screen coords (mouse events route through the
// full-viewport #canvas), while the `.window` div renders lower because
// #window-container flows below #main-window. So grab the title bar at the
// registry coords, NOT getBoundingClientRect.
const dlgBefore = (await findByType(page, 'wxDialog'))[0];
expect(dlgBefore, 'dialog should be in the registry').toBeTruthy();
// Sanity: the freshly painted modal canvas is opaque.
const beforeStats = await sampleModalCanvas(page);
testLogger.consoleLogs.push(
`[MODAL_DRAG] before=${JSON.stringify(beforeStats)} dlg=${JSON.stringify(dlgBefore)}`
);
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await stableShot(page, 'modal-02-before-drag.png', { fullPage: true });
// The dialog now drags via its real DOM title bar (`.window-titlebar`):
// pointer events on it → wx_window_move → wxWindow::Move. So grab the element
// at its actual on-screen position (getBoundingClientRect), NOT the registry
// coords the old canvas title bar needed. Center is over the title text, clear
// of the close × at the right.
const titlebar = page.locator(`${MODAL_SEL} .window-titlebar`);
const tbox = await titlebar.boundingBox();
expect(tbox, 'modal should have a DOM title bar').not.toBeNull();
const startX = tbox!.x + tbox!.width / 2;
const startY = tbox!.y + tbox!.height / 2;
await page.mouse.move(startX, startY);
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await page.waitForTimeout(350); // eslint-disable-line -- documented interaction dwell (pointer settle before grabbing the title bar)
await page.mouse.down();
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await page.waitForTimeout(150); // eslint-disable-line -- documented interaction dwell (press commit before the drag begins)
// Drag in many small steps, sampling the modal canvas immediately after each
// move. Each move calls setWindowRect, which clears the canvas; the dialog's
// own content is NOT marked for repaint by a move (DoMoveWindow only
// refreshes the parent), so with the bug the canvas stays transparent and
// the black `.window` div shows through. We track the WORST (minimum) opaque
// fraction seen across the drag.
let minOpaque = 1;
let lowFrames = 0;
const STEPS = 16;
for (let i = 1; i <= STEPS; i++) {
await page.mouse.move(startX + i * 4, startY + i * 3);
const s = await sampleModalCanvas(page);
if (s) {
if (s.opaqueFrac < minOpaque) minOpaque = s.opaqueFrac;
if (s.opaqueFrac < 0.5) lowFrames++;
}
}
await page.mouse.up();
const dlgAfter = (await findByType(page, 'wxDialog'))[0];
const afterStats = await sampleModalCanvas(page);
const moved = dlgAfter
? Math.abs(dlgAfter.screenX - dlgBefore.screenX) + Math.abs(dlgAfter.screenY - dlgBefore.screenY)
: 0;
testLogger.consoleLogs.push(
`[MODAL_DRAG] minOpaque=${minOpaque} lowFrames=${lowFrames}/${STEPS} after=${JSON.stringify(afterStats)} moved=${moved} dlgAfter=${JSON.stringify(dlgAfter)}`
);
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await stableShot(page, 'modal-03-after-drag.png', { fullPage: true });
// Sanity: the drag must actually have moved the modal, otherwise the
// black-background assertion below is meaningless (the bug only triggers on
// a real move).
expect(moved, 'drag did not move the modal — title-bar grab failed').toBeGreaterThan(10);
// With the bug: the canvas is cleared on each move and shows black before any
// repaint → minOpaque drops toward 0. After the fix: the canvas is never
// cleared on a position-only move, so it stays painted throughout.
expect(
minOpaque,
`modal canvas went transparent/black during drag (minOpaque=${minOpaque}, lowFrames=${lowFrames})`
).toBeGreaterThan(0.8);
});
test('modal background stays painted (not black) after resize', async ({ page, testLogger }) => {
await page.goto(DIALOG_APP);
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await waitForWxApp(page);
// As above: the bare shell lays out #window-container below the fold, so overlay
// it at the origin to make the modal's DOM resize handles reachable by the pointer.
await page.evaluate(() => {
const wc = document.getElementById('window-container');
if (wc) {
wc.style.position = 'absolute';
wc.style.top = '0';
wc.style.left = '0';
}
});
await clickByLabel(page, 'Custom Dialog');
await waitForModalRect(page);
// The Custom dialog now carries wxRESIZE_BORDER, so it has DOM resize handles.
const handle = page.locator(`${MODAL_SEL} .window-resize-se`);
const hbox = await handle.boundingBox();
expect(hbox, 'resizable modal should have a se resize handle').not.toBeNull();
const beforeStats = await sampleModalCanvas(page);
testLogger.consoleLogs.push(`[MODAL_RESIZE] before=${JSON.stringify(beforeStats)}`);
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await stableShot(page, 'modal-04-before-resize.png', { fullPage: true });
// Grab the bottom-right corner and grow the dialog in small steps, sampling the
// modal canvas immediately after each move. A resize legitimately reassigns
// canvas.width/height (clears it); inside the modal's Asyncify pump the repaint
// that should refill it is deferred until the next input event, so with the bug
// the canvas stays transparent and the black .window div shows through.
const startX = hbox!.x + hbox!.width / 2;
const startY = hbox!.y + hbox!.height / 2;
await page.mouse.move(startX, startY);
await page.mouse.down();
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await page.waitForTimeout(120); // eslint-disable-line -- documented interaction dwell (press commit before the resize drag begins)
let minOpaque = 1;
let lowFrames = 0;
const STEPS = 16;
for (let i = 1; i <= STEPS; i++) {
await page.mouse.move(startX + i * 5, startY + i * 4);
const s = await sampleModalCanvas(page);
if (s) {
if (s.opaqueFrac < minOpaque) minOpaque = s.opaqueFrac;
if (s.opaqueFrac < 0.5) lowFrames++;
}
}
await page.mouse.up();
const afterStats = await sampleModalCanvas(page);
testLogger.consoleLogs.push(
`[MODAL_RESIZE] minOpaque=${minOpaque} lowFrames=${lowFrames}/${STEPS} after=${JSON.stringify(afterStats)}`
);
test(determinism): deterministic waits + stableShot screenshots; drop blind sleeps/ifs/retries Make the Playwright e2e + kicad suites deterministic so screenshot flake stops tracing to timing races. - Blind page.waitForTimeout -> condition waits (expect.poll, web-first assertions, waitUntil) + readiness helpers (waitForWxApp, waitForCanvasApp). Remaining sleeps are documented interaction dwells (annotated). - Defensive "if element exists" branches -> loud asserts; label-fallback chains -> normalized clickMenuItemByText. First-run wizard for/if loops removed by seeding calculator/gerbview/pcbnew HTMLs. - Screenshots: new stableShot(page, name) settles the render in-page (canvas hash over rAF) then writes a raw PNG to test-results/ for the existing offline gate (tools/screenshots vs baseline-screenshots). Replaces toHaveScreenshot, which did inline compare + its own baselines and had decoupled the specs from the real gate. scale:'css' pinned. - retries: 0 in both configs. - Guard: tests/tools/lint-determinism.ts (npm run lint:determinism) bans blind sleeps / toHaveScreenshot / inline retries / swallowed catches in specs; documented exceptions carry a marker. Rules in tests/TESTING.md. Assertions, coverage, and renders unchanged (semantic-equivalence reviewed; captures pixel-identical modulo inherent timer/timestamp/3d-raytrace variance). Both suites green at retries:0 (e2e 340, kicad 92); ~35-61% faster. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVX1pHMvRPYHdp6ZfEawrk
2026-07-07 10:50:24 +02:00
await stableShot(page, 'modal-05-after-resize.png', { fullPage: true });
// Sanity: the resize actually grew the modal canvas (otherwise the assertion
// below is meaningless — the corner grab must have taken effect).
expect(
afterStats && beforeStats && afterStats.canvasW > beforeStats.canvasW,
`resize did not grow the modal canvas (before=${beforeStats?.canvasW}, after=${afterStats?.canvasW})`
).toBe(true);
// With the bug: the canvas is cleared on each resize and shows black until the
// deferred modal repaint flushes (only on the next click) → minOpaque ≈ 0.
// After the fix (wx_window_resize forces a synchronous repaint): it stays painted.
expect(
minOpaque,
`modal canvas went transparent/black during resize (minOpaque=${minOpaque}, lowFrames=${lowFrames})`
).toBeGreaterThan(0.8);
});
});