The preview panel worked by proxying a dev server through OpenChamber's own origin and rewriting the HTML that came back. Anything the rewriter did not anticipate broke, and pages that refuse to be embedded never loaded at all. This deletes the proxy (-1604 lines and its tests) and merges the preview and browser panels into one surface backed by a real Chromium view. What the panel is now - A `<webview>` in its own session partition: logins and cookies persist, hot reload works because nothing is rewritten, DevTools are one click away. - Annotation: pick one element, drag a region, or draw freehand, write a note, and it reaches chat with a screenshot of the visible page with the marks on it. - Toolbar: hard reload, page zoom, device sizes, a light/dark switch that applies to the page rather than the app, and cookie/cache clearing scoped to the panel alone. - Several pages at once, each tab showing the page's own favicon, and an address bar that suggests pages already visited in this project. - Dev servers are listed from what is actually listening on the machine, checked against what a project announced, so a server is offered no matter how it was started. One that is still starting is waited for instead of failing. Remote dev servers The desktop app binds a local port and pipes raw bytes to the OpenChamber host over the existing authenticated connection, so the page keeps its own origin at the root of its own host. The reachable set is exactly what discovery reports and is re-checked per connection, so an authenticated client cannot dial arbitrary local services on the host. Links and redirects to another loopback port stay on the machine that served the page. A tunnel that cannot be opened is reported; it is never replaced by the plain loopback URL, which would answer from the user's own machine under a remote address. Agent control Browser actions are a separate `openchamber_web` tool: open, snapshot, click, type, scroll, inspect computed styles, resize between mobile/tablet/desktop, and capture a screenshot into `.openchamber/screenshots/` in the project. The existing `openchamber` tool keeps sessions, worktrees and scheduled tasks. Each has its own setting in the new Settings -> General -> OpenChamber Tools section, and the plugin is not injected at all when both are off. Capability belongs to the connected client, not to configuration: a client declares on its event stream that it can drive a page, which only a Chromium host does. Exactly one client performs each request — it claims the request before acting, and the first claim wins — because deciding by whose result arrives first would be too late for a click that already happened. No client listening is answered immediately with an explanation rather than a timeout. Runtime boundaries Web tabs get a plain iframe that can display a page but not inspect one. The VS Code extension no longer offers the surface at all, since nothing that makes the panel worth having works there. Mobile is unaffected. Native boundary Camera, microphone, location and device-picker requests from panel pages are denied — Electron grants them by default when no handler is set, and the panel loads whatever address the user types. Page capture, appearance emulation and storage clearing verify that their target belongs to the panel's own session instead of trusting a web-contents id from the renderer. Persisted state Stored `preview` tabs migrate to `browser` (v13 -> v14). Context panel tab limits are now per surface, so filling one surface no longer evicts another's tabs. Address history is stored per project and per runtime. Documentation `preview.mdx` and `desktop-browser.mdx` rewritten across all locales, the agent tool settings path corrected, new `DOCUMENTATION.md` for the browser-control broker and the dev tunnel, and the `ui-api-decoupling` skill updated where it still described the deleted proxy.
148 lines
5.5 KiB
TypeScript
148 lines
5.5 KiB
TypeScript
/**
|
|
* Viewport sizing for the browser panel.
|
|
*
|
|
* The page is rendered at a chosen size and scaled down to fit the panel when
|
|
* it does not. Scaling is visual only: the view still lays out at the chosen
|
|
* width, which is the whole point — a 390px layout has to be measured at 390px,
|
|
* not at whatever the panel happens to be.
|
|
*/
|
|
|
|
export type BrowserViewport =
|
|
| { readonly kind: 'fill' }
|
|
| { readonly kind: 'preset'; readonly id: string; readonly width: number; readonly height: number }
|
|
| { readonly kind: 'custom'; readonly width: number; readonly height: number };
|
|
|
|
export const FILL_VIEWPORT: BrowserViewport = { kind: 'fill' };
|
|
|
|
export const MIN_VIEWPORT_SIZE = 240;
|
|
export const MAX_VIEWPORT_SIZE = 3840;
|
|
|
|
export type ViewportPreset = {
|
|
readonly id: string;
|
|
readonly label: string;
|
|
readonly width: number;
|
|
readonly height: number;
|
|
};
|
|
|
|
/**
|
|
* Sizes worth having, not every device ever made. A long list is harder to pick
|
|
* from than it is useful, and anything missing can be typed in directly.
|
|
*/
|
|
export const VIEWPORT_PRESETS: readonly ViewportPreset[] = [
|
|
{ id: 'iphone-se', label: 'iPhone SE', width: 375, height: 667 },
|
|
{ id: 'iphone-14', label: 'iPhone 14', width: 390, height: 844 },
|
|
{ id: 'iphone-14-pro-max', label: 'iPhone 14 Pro Max', width: 430, height: 932 },
|
|
{ id: 'pixel-7', label: 'Pixel 7', width: 412, height: 915 },
|
|
{ id: 'ipad-mini', label: 'iPad mini', width: 768, height: 1024 },
|
|
{ id: 'ipad-pro', label: 'iPad Pro', width: 1024, height: 1366 },
|
|
{ id: 'laptop', label: 'Laptop', width: 1280, height: 800 },
|
|
{ id: 'desktop', label: 'Desktop', width: 1440, height: 900 },
|
|
];
|
|
|
|
export const clampViewportSize = (value: number): number => {
|
|
if (!Number.isFinite(value)) return MIN_VIEWPORT_SIZE;
|
|
return Math.round(Math.min(MAX_VIEWPORT_SIZE, Math.max(MIN_VIEWPORT_SIZE, value)));
|
|
};
|
|
|
|
export const viewportSize = (
|
|
viewport: BrowserViewport,
|
|
): { width: number; height: number } | null => (
|
|
viewport.kind === 'fill' ? null : { width: viewport.width, height: viewport.height }
|
|
);
|
|
|
|
/** Turns a preset id into a viewport, or null when the id is unknown. */
|
|
export const presetViewport = (id: string): BrowserViewport | null => {
|
|
const preset = VIEWPORT_PRESETS.find((entry) => entry.id === id);
|
|
if (!preset) return null;
|
|
return { kind: 'preset', id: preset.id, width: preset.width, height: preset.height };
|
|
};
|
|
|
|
export const rotateViewport = (viewport: BrowserViewport): BrowserViewport => {
|
|
if (viewport.kind === 'fill') return viewport;
|
|
// Rotating a preset stops it being that preset: an iPhone on its side is no
|
|
// longer the entry in the list, and pretending otherwise makes the picker lie.
|
|
return { kind: 'custom', width: viewport.height, height: viewport.width };
|
|
};
|
|
|
|
export type ViewportLayout = {
|
|
/** Size to lay the page out at, in CSS pixels. */
|
|
readonly width: number;
|
|
readonly height: number;
|
|
/** Visual scale, ≤ 1. Applied with a transform; the page never learns of it. */
|
|
readonly scale: number;
|
|
};
|
|
|
|
/**
|
|
* Fits a chosen viewport into the space available.
|
|
*
|
|
* Only ever scales down. Enlarging a small viewport to fill a big panel would
|
|
* misrepresent the very thing the user asked to see.
|
|
*/
|
|
export const fitViewport = (
|
|
viewport: BrowserViewport,
|
|
available: { width: number; height: number },
|
|
): ViewportLayout | null => {
|
|
const size = viewportSize(viewport);
|
|
if (!size) return null;
|
|
|
|
const usableWidth = Math.max(1, available.width);
|
|
const usableHeight = Math.max(1, available.height);
|
|
const scale = Math.min(1, usableWidth / size.width, usableHeight / size.height);
|
|
return { width: size.width, height: size.height, scale };
|
|
};
|
|
|
|
/** Label for the current viewport, for the size control. */
|
|
export const describeViewport = (viewport: BrowserViewport): string => {
|
|
if (viewport.kind === 'fill') return '';
|
|
if (viewport.kind === 'preset') {
|
|
return VIEWPORT_PRESETS.find((entry) => entry.id === viewport.id)?.label ?? '';
|
|
}
|
|
return '';
|
|
};
|
|
|
|
/**
|
|
* The vocabulary the agent gets.
|
|
*
|
|
* Named sizes rather than pixel dimensions: an agent asked to "check the mobile
|
|
* layout" should not have to invent a width, and a number it invented tells the
|
|
* user nothing about what was actually checked.
|
|
*/
|
|
const VIEWPORT_MODES = ['mobile', 'tablet', 'desktop', 'fill'] as const;
|
|
export type BrowserViewportMode = (typeof VIEWPORT_MODES)[number];
|
|
|
|
const MODE_PRESETS: Record<Exclude<BrowserViewportMode, 'fill'>, string> = {
|
|
mobile: 'iphone-14',
|
|
tablet: 'ipad-mini',
|
|
desktop: 'desktop',
|
|
};
|
|
|
|
export const isViewportMode = (value: unknown): value is BrowserViewportMode => (
|
|
typeof value === 'string' && (VIEWPORT_MODES as readonly string[]).includes(value)
|
|
);
|
|
|
|
export const viewportForMode = (mode: BrowserViewportMode): BrowserViewport => (
|
|
mode === 'fill' ? FILL_VIEWPORT : presetViewport(MODE_PRESETS[mode]) ?? FILL_VIEWPORT
|
|
);
|
|
|
|
/**
|
|
* Reports the current viewport in the agent's own vocabulary, so a snapshot
|
|
* states which layout it describes.
|
|
*/
|
|
export const viewportSummary = (viewport: BrowserViewport): {
|
|
mode: BrowserViewportMode | 'custom';
|
|
width: number | null;
|
|
height: number | null;
|
|
} => {
|
|
const size = viewportSize(viewport);
|
|
if (!size) return { mode: 'fill', width: null, height: null };
|
|
|
|
for (const mode of ['mobile', 'tablet', 'desktop'] as const) {
|
|
const preset = viewportForMode(mode);
|
|
const presetSize = viewportSize(preset);
|
|
if (presetSize && presetSize.width === size.width && presetSize.height === size.height) {
|
|
return { mode, width: size.width, height: size.height };
|
|
}
|
|
}
|
|
return { mode: 'custom', width: size.width, height: size.height };
|
|
};
|