Files
rea/src/cli/browserAdvancedCommands.ts
T
morluto 0155b89953 fix(browser): name web-bundle budgets and recover web deadlines (#6036)
* fix(browser): name web-bundle budgets and recover web deadlines

Name the existing heap, per-script, and total web-bundle phases, and reject a per-script deadline above the total. Promote capture, CDP command, source-map decode, and source-map fetch deadlines at today's defaults. Name DevTools timeoutMs, and state that discovery, cleanup grace, byte budgets, and Electron hook retention have no runtime override.

Refs #5910

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): apply selected web deadlines to the owning phase

Keep HTTP and browser-socket discovery on the fixed discovery deadline so a
short command_timeout_ms starts at the handshake. Apply scenario
command_timeout_ms to launch startup and the primary connect session, and
keep service-worker restoration on the fixed 5000 ms grace. Omit
command_timeout_ms from recovery text when the operation does not accept it.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): retain scenario commands that finish after their deadline

A timed-out scenario command stays owned so cleanup can close a late
browser or context and restore a mutation that lands after the deadline.
Child storage replies use that deadline, scenario recovery names
command_timeout_ms in the capture JSON, and the session test connection
implements beginCleanupGrace.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): keep child command timeouts as the public failure

A nested command such as Page.enable that is delivered but never answered
was stored as protocol_error, hiding the command name, selected deadline,
and recovery. Store the typed timeout and assert that public error.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): keep runtime cleanup on the fixed grace

Switch runtime restoration to the cleanup grace before the first send so a
short command_timeout_ms cannot fail an otherwise successful observation.
Stop advertising command_timeout_ms on list_browser_targets, which never
bounds a post-discovery command.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* test(browser): prove a short command deadline cannot fail runtime cleanup

Drive restoration through the runtime fixture. A cleanup command that
outlives command_timeout_ms still succeeds, and a stalled restoration
names the fixed 5000 ms grace.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): name the effective source-map decode deadline

codecBudgetLimitation reported a fixed 20-second deadline after a
caller selected decode_timeout_ms. Result and Evidence limitations
now name that effective millisecond deadline.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): separate handshake and target-attach deadlines

A browser-level Target.attachToTarget no longer shares the WebSocket
handshake budget. Blocked service-worker attachment keeps discovery on
its fixed deadline and applies command_timeout_ms to the later handshake
and attach.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): name the web-bundle resource each runtime exhausts

Bun ignores the V8 old-space flag. A Windows Job reports the process
commit limit, and other Bun hosts no longer suggest raising max_heap_mb.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): cite the browser-socket discovery message budget

WebSocket discovery enforces the 64 MiB protocol budget. Timeout
recovery no longer advertises the 8 MiB HTTP document budget.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): bound the blocked service-worker check before discovery

The first registration check has its own fixed 10s deadline. Discovery,
handshake, and Target.attachToTarget stay on their separate deadlines.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(browser): name the operand that stopped web-bundle admission

Host heap pressure no longer claims the fixed result budget is exhausted.
A Windows Bun allocation before the commit limit still names that limit.

Co-authored-by: morluto <morluto@users.noreply.github.com>

---------

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: morluto <morluto@users.noreply.github.com>
2026-10-22 05:47:38 +08:00

314 lines
10 KiB
TypeScript

import { Cli, z } from "incur";
import { browserContext, browserCliError } from "../cliBrowserContext.js";
import {
browserScopeOptions,
selectedCommandTimeout,
} from "../cliObservationOptions.js";
import {
captureWebScreenshot,
compareWebCaptureEvidence,
compareWebScreenshotEvidence,
discoverWebMcpTools,
} from "../application/BrowserObservationService.js";
import { CdpBrowserProvider } from "../browser/CdpBrowserProvider.js";
import { logCliCommand } from "../cliLogging.js";
import {
AnalysisInputError,
type AnalysisInputIssue,
} from "../domain/analysisErrorCore.js";
import { compareBrowserScenariosInputSchema } from "../domain/browserScenarioDiffValues.js";
import { browserScenarioCaptureSchema } from "../domain/browserScenarioCapture.js";
import { compareWebCapturesInputSchema } from "../domain/webCaptureDiffSchemas.js";
import { projectInputIssues } from "../domain/inputIssueProjection.js";
import { discoverWebMcpToolsInputSchema } from "../domain/webMcpDiscovery.js";
import {
captureWebScreenshotInputSchema,
compareWebScreenshotsInputSchema,
} from "../domain/webScreenshot.js";
import type { JsonValue } from "../domain/jsonValue.js";
import { parseCliJsonInput } from "../cliJsonInput.js";
import type { Logger } from "pino";
import { CLI_COMMANDS } from "../cliCommandNames.js";
import { withCommandCancellation } from "./commandCancellation.js";
import { finishAnalysisCommand } from "./finishAnalysisCommand.js";
/** Register WebMCP, capture-diff, and screenshot CLI equivalents. */
export const registerAdvancedBrowserCommands = (
cli: ReturnType<typeof Cli.create>,
logger: Logger,
): void => {
registerWebMcp(cli, logger);
registerCaptureDiff(cli, logger);
registerScreenshot(cli, logger);
registerScreenshotDiff(cli, logger);
};
const registerWebMcp = (
cli: ReturnType<typeof Cli.create>,
logger: Logger,
): void => {
cli.command(CLI_COMMANDS.discoverWebMcpTools, {
description: "Passively discover page-declared WebMCP tools",
args: z.object({
endpoint: z
.string()
.describe("Loopback CDP HTTP endpoint or browser WebSocket URL"),
targetId: z.string().describe("Target ID from list-browser-targets"),
}),
options: z.object({
...browserScopeOptions,
observationMs: z
.number()
.int()
.min(0)
.default(100)
.describe("Observation duration in milliseconds"),
}),
run: ({ args, options }) =>
withCommandCancellation((signal) =>
logCliCommand(logger, "discover-webmcp-tools", async () => {
const context = browserContext();
const input = {
cdp_endpoint: args.endpoint,
allowed_origins: options.allowedOrigins,
target_id: args.targetId,
observation_ms: options.observationMs,
...selectedCommandTimeout(options.commandTimeoutMs),
};
const parsed = discoverWebMcpToolsInputSchema.safeParse(input);
if (!parsed.success)
return inputError(
"discover_webmcp_tools",
projectInputIssues(parsed.error.issues, input),
);
return await finishAnalysisCommand(
() =>
discoverWebMcpTools(context.provider, parsed.data, { signal }),
{
close: async () => {
await context.provider.close?.();
},
},
context.provider.identity().id,
"discover_webmcp_tools",
signal,
);
}),
),
});
};
const registerCaptureDiff = (
cli: ReturnType<typeof Cli.create>,
logger: Logger,
): void => {
cli.command(CLI_COMMANDS.compareWebCaptures, {
description:
"Compare two normalized passive captures or recorded browser scenarios",
args: z.object({
beforeJson: z
.string()
.describe("Earlier normalized web capture JSON or JSON file path"),
afterJson: z
.string()
.describe("Later normalized web capture JSON or JSON file path"),
}),
options: z.object({
normalizationJson: z
.string()
.default('{"rules":[]}')
.describe(
"Recorded literal normalization policy JSON or JSON file path for browser scenarios",
),
}),
run: ({ args, options }) =>
logCliCommand(logger, "compare-web-captures", async () => {
const inputs = await parseJsonArguments("compare_web_captures", {
before: args.beforeJson,
after: args.afterJson,
normalization: options.normalizationJson,
});
if (!inputs.ok) return inputs.error;
const { before, after, normalization } = inputs.value;
// Passive snapshots own the inspection wrapper. Otherwise recognize
// scenario-owned fields even when the scenario manifest is missing.
const scenarioInput = [before, after].some(
(capture) =>
typeof capture === "object" &&
capture !== null &&
!("inspection" in capture) &&
Object.keys(browserScenarioCaptureSchema.shape).some(
(field) => field in capture,
),
);
const input = scenarioInput
? { before_scenario: before, after_scenario: after, normalization }
: { before, after };
const parsed = scenarioInput
? compareBrowserScenariosInputSchema.safeParse(input)
: compareWebCapturesInputSchema.safeParse(input);
if (!parsed.success)
return inputError(
"compare_web_captures",
projectInputIssues(parsed.error.issues, input),
);
const result = await compareWebCaptureEvidence(
new CdpBrowserProvider(),
parsed.data,
);
return result.ok ? result.value : browserCliError(result.error);
}),
});
};
const registerScreenshot = (
cli: ReturnType<typeof Cli.create>,
logger: Logger,
): void => {
cli.command(CLI_COMMANDS.captureWebScreenshot, {
description:
"Capture one visible page viewport within the supplied origin scope",
args: z.object({
endpoint: z
.string()
.describe("Loopback CDP HTTP endpoint or browser WebSocket URL"),
targetId: z.string().describe("Target ID from list-browser-targets"),
}),
options: z.object({
...browserScopeOptions,
}),
run: ({ args, options }) =>
withCommandCancellation((signal) =>
logCliCommand(logger, "capture-web-screenshot", async () => {
const context = browserContext();
const input = {
cdp_endpoint: args.endpoint,
allowed_origins: options.allowedOrigins,
target_id: args.targetId,
...selectedCommandTimeout(options.commandTimeoutMs),
};
const parsed = captureWebScreenshotInputSchema.safeParse(input);
if (!parsed.success)
return inputError(
"capture_web_screenshot",
projectInputIssues(parsed.error.issues, input),
);
return await finishAnalysisCommand(
() =>
captureWebScreenshot(context.provider, parsed.data, { signal }),
{
close: async () => {
await context.provider.close?.();
},
},
context.provider.identity().id,
"capture_web_screenshot",
signal,
);
}),
),
});
};
const registerScreenshotDiff = (
cli: ReturnType<typeof Cli.create>,
logger: Logger,
): void => {
cli.command(CLI_COMMANDS.compareWebScreenshots, {
description: "Compare two self-verifying PNG artifact JSON values",
args: z.object({
beforeJson: z
.string()
.describe("Earlier screenshot artifact JSON or JSON file path"),
afterJson: z
.string()
.describe("Later screenshot artifact JSON or JSON file path"),
}),
options: z.object({
channelThreshold: z
.number()
.int()
.min(0)
.max(255)
.default(0)
.describe("Per-channel difference threshold for changed pixels"),
}),
run: ({ args, options }) =>
logCliCommand(logger, "compare-web-screenshots", async () => {
const operation = "compare_web_screenshots";
const inputs = await parseJsonArguments(operation, {
before: args.beforeJson,
after: args.afterJson,
});
if (!inputs.ok) return inputs.error;
const { before, after } = inputs.value;
const input = {
before,
after,
channel_threshold: options.channelThreshold,
};
const parsed = compareWebScreenshotsInputSchema.safeParse(input);
if (!parsed.success)
return browserCliError(
new AnalysisInputError(
operation,
{ cause: parsed.error },
projectInputIssues(parsed.error.issues, input),
),
);
const result = await compareWebScreenshotEvidence(
new CdpBrowserProvider(),
parsed.data,
);
return result.ok ? result.value : browserCliError(result.error);
}),
});
};
/**
* Parse each named inline JSON argument or local JSON file, like other
* JSON-input commands. Screenshot artifacts and scenario captures embed PNG
* bytes that can exceed the host's command-line length limit.
*/
const parseJsonArguments = async (
operation: string,
fields: Readonly<Record<string, string>>,
): Promise<
| { readonly ok: true; readonly value: Readonly<Record<string, unknown>> }
| { readonly ok: false; readonly error: JsonValue }
> => {
const values: Record<string, unknown> = {};
for (const [field, value] of Object.entries(fields)) {
const parsed = await parseCliJsonInput(value, operation);
if (!parsed.ok)
return isFileInputFailure(parsed.error)
? parsed
: {
ok: false,
error: inputError(operation, [invalidJsonIssue(field)]),
};
values[field] = parsed.value;
}
return { ok: true, value: values };
};
/** File read and decode failures name the selected path; inline JSON does not. */
const isFileInputFailure = (error: JsonValue): boolean =>
typeof error === "object" &&
error !== null &&
!Array.isArray(error) &&
"input_path" in error;
const invalidJsonIssue = (field: string): AnalysisInputIssue => ({
path: [field],
reason: "invalid_format",
expected: "JSON",
});
const inputError = (
operation: string,
issues: readonly AnalysisInputIssue[] = [],
): JsonValue =>
browserCliError(new AnalysisInputError(operation, undefined, issues));