gsd-pi 2.77.0-dev.eaa4973bc → 2.78.0-dev.aeeb2ca00

package/src/resources/extensions/browser-tools/tests/browser-tools-integration.test.mjs

@@ -12,76 +12,25 @@
 import { describe, it, before, after } from "node:test";
 import assert from "node:assert/strict";
 import { chromium } from "playwright";
-import { readFileSync } from "node:fs";
-import { resolve, dirname } from "node:path";
+import { dirname } from "node:path";
 import { fileURLToPath } from "node:url";
 
 const __dirname = dirname(fileURLToPath(import.meta.url));
-const ROOT = resolve(__dirname, "..");
 
 // ---------------------------------------------------------------------------
-// Source extraction — get the IIFE strings we need for injection
+// Source loading — import the IIFE builders directly via jiti.
+// The test-only named exports in tools/intent.ts and tools/forms.ts exist
+// exactly so this test can call the real, in-tree builders. No brace
+// walking, no regex stripping — a refactor of the signatures just updates
+// the import surface, not the test.
 // ---------------------------------------------------------------------------
 
-// 1. EVALUATE_HELPERS_SOURCE — exported constant, extract via jiti
 import { createRequire } from "node:module";
 const require = createRequire(import.meta.url);
 const jiti = require("jiti")(__dirname, { interopDefault: true, debug: false });
 const { EVALUATE_HELPERS_SOURCE } = jiti("../evaluate-helpers.ts");
-
-// 2. Intent scoring — module-private buildIntentScoringScript.
-//    Extract the function from source, wrap it, and eval to get the builder.
-const intentSource = readFileSync(resolve(ROOT, "tools/intent.ts"), "utf-8");
-
-function extractBuildIntentScoringScript() {
-  // Match the function body: starts with "function buildIntentScoringScript"
-  // and returns a template literal string. We extract up to the matching closing brace.
-  const startMarker = "function buildIntentScoringScript(intent: string, scope?: string): string {";
-  const startIdx = intentSource.indexOf(startMarker);
-  if (startIdx === -1) throw new Error("Could not find buildIntentScoringScript in intent.ts");
-
-  // Walk from start, counting braces to find the end
-  let depth = 0;
-  let foundFirst = false;
-  let endIdx = startIdx;
-  for (let i = startIdx; i < intentSource.length; i++) {
-    if (intentSource[i] === "{") { depth++; foundFirst = true; }
-    if (intentSource[i] === "}") depth--;
-    if (foundFirst && depth === 0) { endIdx = i + 1; break; }
-  }
-
-  let fnBody = intentSource.slice(startIdx, endIdx);
-  // Strip TypeScript type annotations
-  fnBody = fnBody.replace(/\(intent:\s*string,\s*scope\?:\s*string\):\s*string/, "(intent, scope)");
-  return new Function("return " + fnBody)();
-}
-
-const buildIntentScoringScript = extractBuildIntentScoringScript();
-
-// 3. Form analysis — module-private buildFormAnalysisScript.
-const formsSource = readFileSync(resolve(ROOT, "tools/forms.ts"), "utf-8");
-
-function extractBuildFormAnalysisScript() {
-  const startMarker = "function buildFormAnalysisScript(selector?: string): string {";
-  const startIdx = formsSource.indexOf(startMarker);
-  if (startIdx === -1) throw new Error("Could not find buildFormAnalysisScript in forms.ts");
-
-  let depth = 0;
-  let foundFirst = false;
-  let endIdx = startIdx;
-  for (let i = startIdx; i < formsSource.length; i++) {
-    if (formsSource[i] === "{") { depth++; foundFirst = true; }
-    if (formsSource[i] === "}") depth--;
-    if (foundFirst && depth === 0) { endIdx = i + 1; break; }
-  }
-
-  let fnBody = formsSource.slice(startIdx, endIdx);
-  // Strip TypeScript type annotation
-  fnBody = fnBody.replace(/\(selector\?:\s*string\):\s*string/, "(selector)");
-  return new Function("return " + fnBody)();
-}
-
-const buildFormAnalysisScript = extractBuildFormAnalysisScript();
+const { buildIntentScoringScript } = jiti("../tools/intent.ts");
+const { buildFormAnalysisScript } = jiti("../tools/forms.ts");
 
 // ---------------------------------------------------------------------------
 // Browser lifecycle

package/src/resources/extensions/browser-tools/tests/browser-tools-unit.test.cjs

@@ -387,32 +387,44 @@ describe("formatArtifactTimestamp", () => {
 // ---------------------------------------------------------------------------
 
 describe("EVALUATE_HELPERS_SOURCE", () => {
-	it("is a parseable string (valid JavaScript)", () => {
-		assert.doesNotThrow(() => {
-			new Function(EVALUATE_HELPERS_SOURCE);
-		});
-	});
+	// Behaviour test: executing the source in a Node vm sandbox must
+	// populate a `window.__pi` namespace with every expected helper.
+	// No source grep — we actually run the code and verify the resulting
+	// object shape.
+	it("executing the source assigns all expected helpers to window.__pi", () => {
+		const vm = require("node:vm");
+		const expectedFunctions = [
+			"cssPath",
+			"simpleHash",
+			"isVisible",
+			"isEnabled",
+			"inferRole",
+			"accessibleName",
+			"isInteractiveEl",
+			"domPath",
+			"selectorHints",
+		];
 
-	const expectedFunctions = [
-		"cssPath",
-		"simpleHash",
-		"isVisible",
-		"isEnabled",
-		"inferRole",
-		"accessibleName",
-		"isInteractiveEl",
-		"domPath",
-		"selectorHints",
-	];
-
-	for (const fnName of expectedFunctions) {
-		it(`contains assignment for pi.${fnName}`, () => {
-			assert.ok(
-				EVALUATE_HELPERS_SOURCE.includes(`pi.${fnName} = function`),
-				`Expected pi.${fnName} = function assignment in source`,
+		// Playwright evaluates the source in a page context where `window`
+		// exists, so the helpers attach to `window.__pi`. Provide a minimal
+		// window stub in a vm context so we avoid polluting the test globals.
+		const sandbox = { window: {} };
+		const script = new vm.Script(EVALUATE_HELPERS_SOURCE);
+		script.runInNewContext(sandbox, { timeout: 1000 });
+
+		assert.ok(
+			sandbox.window.__pi && typeof sandbox.window.__pi === "object",
+			"executing EVALUATE_HELPERS_SOURCE must assign window.__pi",
+		);
+
+		for (const fnName of expectedFunctions) {
+			assert.equal(
+				typeof sandbox.window.__pi[fnName],
+				"function",
+				`window.__pi.${fnName} must be a function after executing the source`,
 			);
-		});
-	}
+		}
+	});
 });
 
 // ---------------------------------------------------------------------------

package/src/resources/extensions/browser-tools/tests/capture-sharp-optional.test.cjs

@@ -1,93 +1,91 @@
 /**
  * Regression tests for the optional sharp dependency in capture.ts.
  *
- * Verifies two things:
- *   1. Static: the lazy-load pattern is structurally correct in the source.
- *   2. Behavioral: constrainScreenshot returns the raw buffer unchanged when
- *      sharp is unavailable, rather than throwing.
+ * Behaviour:
+ *   - constrainScreenshot must fall back to returning the raw buffer
+ *     unchanged when sharp is unavailable, rather than throwing.
+ *   - When sharp IS available, oversized screenshots get resized.
+ *
+ * No source-grep. The test drives the real constrainScreenshot function
+ * after seeding the module-private `_sharp` cache via the test-only
+ * `__setSharpForTesting` export.
  */
 
-const { describe, it } = require("node:test");
+const { describe, it, afterEach } = require("node:test");
 const assert = require("node:assert/strict");
-const { readFileSync } = require("node:fs");
-const { join } = require("node:path");
-
-// ---------------------------------------------------------------------------
-// 1. Static analysis — verify the lazy-load pattern is present in source
-// ---------------------------------------------------------------------------
+const jiti = require("jiti")(__filename, { interopDefault: true, debug: false });
 
-describe("capture.ts — sharp optional lazy-load (static)", () => {
-	const source = readFileSync(
-		join(process.cwd(), "src/resources/extensions/browser-tools/capture.ts"),
-		"utf-8",
-	);
+const { constrainScreenshot, __setSharpForTesting } = jiti("../capture.ts");
 
-	it("does not have a top-level static sharp import", () => {
-		assert.ok(
-			!source.includes('import sharp from "sharp"'),
-			'capture.ts must not contain a top-level `import sharp from "sharp"` — sharp must be loaded lazily',
-		);
+describe("constrainScreenshot — sharp unavailable (null)", () => {
+	afterEach(() => {
+		// Clear the test override so later tests don't inherit a null sharp.
+		__setSharpForTesting(undefined);
 	});
 
-	it("defines a getSharp lazy-loader function", () => {
-		assert.ok(
-			source.includes("async function getSharp()"),
-			"capture.ts must define an async getSharp() lazy-loader",
+	it("returns the raw buffer unchanged when sharp is null", async () => {
+		__setSharpForTesting(null);
+
+		const rawBuffer = Buffer.from([0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a]); // PNG magic bytes
+		const result = await constrainScreenshot(null, rawBuffer, "image/png", 80);
+
+		assert.strictEqual(
+			result,
+			rawBuffer,
+			"constrainScreenshot must return the exact same buffer instance when sharp is null",
 		);
 	});
 
-	it("guards constrainScreenshot with a null-sharp early return", () => {
-		assert.ok(
-			source.includes("if (!sharp) return buffer"),
-			"constrainScreenshot must return the raw buffer early when sharp is null",
-		);
+	it("returns the raw buffer unchanged for JPEG input when sharp is null", async () => {
+		__setSharpForTesting(null);
+
+		const rawBuffer = Buffer.from([0xff, 0xd8, 0xff, 0xe0]); // JPEG magic bytes
+		const result = await constrainScreenshot(null, rawBuffer, "image/jpeg", 80);
+
+		assert.strictEqual(result, rawBuffer);
 	});
 });
 
-// ---------------------------------------------------------------------------
-// 2. Behavioral — constrainScreenshot passes through buffer when sharp is null
-// ---------------------------------------------------------------------------
-
-describe("capture.ts — constrainScreenshot with sharp unavailable", () => {
-	it("returns the raw buffer unchanged when sharp is null", async () => {
-		// Simulate what getSharp() returns on platforms without sharp by
-		// directly calling constrainScreenshot through a module whose _sharp
-		// cache has been pre-seeded to null via the module-level variable reset.
-		//
-		// Because jiti caches modules across the test suite we use a fresh
-		// require-cache trick: load capture.ts source manually and evaluate the
-		// constrainScreenshot function with a stub getSharp that always returns null.
-		const captureSource = readFileSync(
-			join(process.cwd(), "src/resources/extensions/browser-tools/capture.ts"),
-			"utf-8",
-		);
+describe("constrainScreenshot — sharp available", () => {
+	afterEach(() => {
+		__setSharpForTesting(undefined);
+	});
 
-		// Verify the guard line is reachable (structural check already done above).
-		// For the behavioral test we use the actual constrainScreenshot imported
-		// via jiti — but we force getSharp() to return null by calling the function
-		// with a very small buffer where sharp IS available. Separately we test the
-		// null path by crafting a minimal wrapper.
-		//
-		// The simplest verifiable behaviour: if the guard `if (!sharp) return buffer`
-		// is present, passing a Buffer through a version of constrainScreenshot where
-		// _sharp=null must return that exact buffer. We verify this by extracting and
-		// running a minimal inline version of the guard logic.
+	it("passes through a small image unchanged (below cap)", async () => {
+		const sharp = require("sharp");
+		const small = await sharp({
+			create: {
+				width: 400,
+				height: 300,
+				channels: 3,
+				background: { r: 128, g: 128, b: 128 },
+			},
+		})
+			.jpeg({ quality: 80 })
+			.toBuffer();
 
-		const rawBuffer = Buffer.from([0x89, 0x50, 0x4e, 0x47]); // fake PNG header
+		const result = await constrainScreenshot(null, small, "image/jpeg", 80);
+		const meta = await sharp(result).metadata();
+		assert.equal(meta.width, 400, "small images must not be resized");
+		assert.equal(meta.height, 300);
+	});
 
-		// Inline the guard as it appears in capture.ts so the test is coupled to
-		// the actual contract, not an arbitrary helper.
-		async function constrainScreenshotWithNullSharp(buffer) {
-			const sharp = null; // simulates getSharp() returning null
-			if (!sharp) return buffer;
-			// (remainder of constrainScreenshot would run here with a real sharp)
-		}
+	it("resizes an oversized image to within 1568px", async () => {
+		const sharp = require("sharp");
+		const big = await sharp({
+			create: {
+				width: 3000,
+				height: 2000,
+				channels: 3,
+				background: { r: 128, g: 128, b: 128 },
+			},
+		})
+			.jpeg({ quality: 80 })
+			.toBuffer();
 
-		const result = await constrainScreenshotWithNullSharp(rawBuffer);
-		assert.strictEqual(
-			result,
-			rawBuffer,
-			"constrainScreenshot must return the exact same buffer instance when sharp is null",
-		);
+		const result = await constrainScreenshot(null, big, "image/jpeg", 80);
+		const meta = await sharp(result).metadata();
+		assert.ok(meta.width <= 1568, `width ${meta.width} must be <= 1568`);
+		assert.ok(meta.height <= 1568, `height ${meta.height} must be <= 1568`);
 	});
 });

package/src/resources/extensions/browser-tools/tools/forms.ts

@@ -46,7 +46,11 @@ interface FormAnalysisResult {
  * Runs inside page.evaluate(). Finds the target form, inventories all fields
  * with full label resolution, and returns a structured result.
  */
-function buildFormAnalysisScript(selector?: string): string {
+// Exported for tests only (see tests/browser-tools-integration.test.mjs).
+// Keep this function treated as module-private for production call sites —
+// the only legitimate external caller is the Playwright-driven integration
+// suite that needs to evaluate the returned IIFE against real DOM.
+export function buildFormAnalysisScript(selector?: string): string {
 	// We return a string that will be evaluated in the page context.
 	// This avoids serialization issues with passing functions.
 	return `(() => {

package/src/resources/extensions/browser-tools/tools/intent.ts

@@ -37,7 +37,11 @@ type Intent = (typeof INTENTS)[number];
  * Uses window.__pi utilities (injected via addInitScript) for element
  * metadata — no inline redeclarations.
  */
-function buildIntentScoringScript(intent: string, scope?: string): string {
+// Exported for tests only (see tests/browser-tools-integration.test.mjs).
+// Keep this function treated as module-private for production call sites —
+// the only legitimate external caller is the Playwright-driven integration
+// suite that needs to evaluate the returned IIFE against real DOM.
+export function buildIntentScoringScript(intent: string, scope?: string): string {
 	const scopeSelector = JSON.stringify(scope ?? null);
 
 	return `(() => {

package/src/resources/extensions/claude-code-cli/readiness.ts

@@ -5,8 +5,9 @@
  * Results are cached for 30 seconds to avoid shelling out on every
  * model-availability check.
  *
- * Auth verification follows the T3 Code pattern: run `claude auth status`
- * and check the exit code + output for an authenticated session.
+ * Auth verification runs `claude auth status --json` and inspects the
+ * `loggedIn` field, falling back to a text heuristic when the JSON shape
+ * is unavailable (older Claude CLI builds).
  */
 
 import { execFileSync } from "node:child_process";
@@ -27,17 +28,37 @@ const CLAUDE_COMMAND = process.platform === "win32" ? "claude.cmd" : "claude";
  */
 const CLAUDE_COMMAND_CANDIDATES = process.platform === "win32" ? [CLAUDE_COMMAND, "claude.exe", "claude"] : [CLAUDE_COMMAND];
 
-function execClaude(args: string[]): Buffer {
+// Codes treated as "this candidate didn't run — try the next one" rather than
+// fatal failures. ENOENT/EINVAL cover the original Windows .cmd shim cases.
+// ETIMEDOUT and EAGAIN cover slow-spawn cases where cmd.exe wrapping plus
+// the Claude CLI startup path together exceed the per-attempt timeout
+// (Issue #4997 regression on Windows + Node 25).
+const SOFT_FAIL_CODES = new Set(["ENOENT", "EINVAL", "ETIMEDOUT", "EAGAIN"]);
+
+// Keep the version probe snappy — `claude --version` is a quick path.
+const VERSION_TIMEOUT_MS = 5_000;
+// Auth status can be much slower on Windows because the spawn goes through
+// cmd.exe → claude.cmd → node → Claude CLI. 15s leaves headroom on cold spawns
+// without making startup feel hung when the CLI is genuinely missing.
+const AUTH_TIMEOUT_MS = 15_000;
+
+/**
+ * Run the requested Claude CLI command against each supported executable name.
+ * Returns the first successful output buffer and rethrows hard failures.
+ */
+function execClaude(args: string[], timeoutMs: number): Buffer {
 	let lastError: unknown;
 	for (const command of CLAUDE_COMMAND_CANDIDATES) {
 		try {
-			return execFileSync(command, args, { timeout: 5_000, stdio: "pipe" });
+			return execFileSync(command, args, {
+				timeout: timeoutMs,
+				stdio: "pipe",
+				shell: process.platform === "win32",
+			});
 		} catch (error) {
 			lastError = error;
 			const code = (error as NodeJS.ErrnoException | undefined)?.code;
-			// Windows Git Bash can surface `.cmd` spawn failures as EINVAL instead
-			// of ENOENT. Treat both as "try next candidate".
-			if (code === "ENOENT" || code === "EINVAL") {
+			if (code && SOFT_FAIL_CODES.has(code)) {
 				continue;
 			}
 			throw error;
@@ -46,11 +67,44 @@ function execClaude(args: string[]): Buffer {
 	throw lastError ?? new Error(`Claude CLI executable not found (tried: ${CLAUDE_COMMAND_CANDIDATES.join(", ")})`);
 }
 
+/**
+ * Decide auth state from `claude auth status` output.
+ *
+ * Newer Claude CLI builds emit JSON by default with a `loggedIn` boolean.
+ * Older builds emit free-form text. We prefer the structured signal and fall
+ * back to a text heuristic. Note: the text heuristic only covers English
+ * phrasing — the JSON path is the durable signal.
+ */
+function parseAuthStatus(output: string): boolean {
+	const trimmed = output.trim();
+	if (trimmed.startsWith("{")) {
+		try {
+			const parsed = JSON.parse(trimmed) as { loggedIn?: unknown };
+			if (typeof parsed.loggedIn === "boolean") {
+				return parsed.loggedIn;
+			}
+		} catch {
+			// Fall through to text heuristic.
+		}
+	}
+
+	const lower = trimmed.toLowerCase();
+	if (/not logged in|no credentials|unauthenticated|not authenticated/.test(lower)) {
+		return false;
+	}
+	// Exit-0 with non-error output and no negative markers — treat as authed.
+	return true;
+}
+
 let cachedBinaryPresent: boolean | null = null;
 let cachedAuthed: boolean | null = null;
 let lastCheckMs = 0;
 const CHECK_INTERVAL_MS = 30_000;
 
+/**
+ * Refresh the cached binary/auth state when the cache window has expired.
+ * Preserves a known auth state across soft-fail auth probes.
+ */
 function refreshCache(): void {
 	const now = Date.now();
 	if (cachedBinaryPresent !== null && now - lastCheckMs < CHECK_INTERVAL_MS) {
@@ -62,7 +116,7 @@ function refreshCache(): void {
 
 	// Check binary presence
 	try {
-		execClaude(["--version"]);
+		execClaude(["--version"], VERSION_TIMEOUT_MS);
 		cachedBinaryPresent = true;
 	} catch {
 		cachedBinaryPresent = false;
@@ -70,15 +124,20 @@ function refreshCache(): void {
 		return;
 	}
 
-	// Check auth status — exit code 0 with non-error output means authenticated
+	// Request JSON explicitly so older CLI builds that defaulted to text and
+	// newer builds that default to JSON produce a consistent shape.
 	try {
-		const output = execClaude(["auth", "status"])
-			.toString()
-			.toLowerCase();
-		// The CLI outputs "not logged in", "no credentials", or similar when unauthenticated
-		cachedAuthed = !(/not logged in|no credentials|unauthenticated|not authenticated/i.test(output));
-	} catch {
-		// Non-zero exit code means not authenticated
+		const output = execClaude(["auth", "status", "--json"], AUTH_TIMEOUT_MS).toString();
+		cachedAuthed = parseAuthStatus(output);
+	} catch (error) {
+		const code = (error as NodeJS.ErrnoException | undefined)?.code;
+		// Spawn-shape failures (timeout, transient) shouldn't be treated as
+		// "definitely not authed" — leave the previous value if we have one,
+		// otherwise default to false. The version probe already established
+		// the binary works, so a flaky auth probe is more likely transient.
+		if (code && SOFT_FAIL_CODES.has(code) && cachedAuthed !== null) {
+			return;
+		}
 		cachedAuthed = false;
 	}
 }