hdoc-tools 0.60.1 → 0.62.0

package/hdoc-install-browser.js

@@ -1,145 +1,145 @@
-// Resilient browser provisioning for hdoc-tools.
-//
-// Replaces the old `puppeteer browsers install ...` one-liner postinstall.
-// Puppeteer's own bundled download (install.mjs) is disabled via
-// .puppeteerrc.cjs (skipDownload), so this script is the single, controlled
-// place Chrome + chrome-headless-shell are fetched.
-//
-// Why: on some Windows Server 2019 build agents the Chrome archive extracts
-// only partially (Defender quarantining binaries mid-extract, or a truncated
-// download behind a proxy). The first failure leaves the version folder on
-// disk, after which Puppeteer refuses to re-extract and every later install
-// reports "folder exists but executable missing" — a permanent dead end.
-//
-// This script makes provisioning idempotent and self-healing:
-//   * skip when a valid executable already exists,
-//   * delete any stale/partial version folder before (re)installing,
-//   * retry the download a few times,
-//   * verify the executable exists afterwards and fail loudly with concrete
-//     remediation if it still does not.
-
-(async () => {
-	const fs = require("node:fs");
-	const path = require("node:path");
-	const {
-		install,
-		computeExecutablePath,
-		detectBrowserPlatform,
-		Browser,
-	} = require("@puppeteer/browsers");
-
-	// Build id + cache dir come from the shared .puppeteerrc.cjs so install and
-	// runtime launch can never drift apart.
-	const puppeteerConfig = require(path.join(__dirname, ".puppeteerrc.cjs"));
-	const CHROME_BUILD = puppeteerConfig.chromeBuild;
-	const MAX_ATTEMPTS = 3;
-
-	const RED = "\x1b[31m";
-	const YELLOW = "\x1b[33m";
-	const GREEN = "\x1b[32m";
-	const RESET = "\x1b[0m";
-
-	const log = (msg) => console.log(`[hdoc-tools] ${msg}`);
-
-	// Resolve the cache directory the same way Puppeteer does at runtime:
-	// PUPPETEER_CACHE_DIR wins, otherwise the fixed cacheDir declared in
-	// .puppeteerrc.cjs. Reading the shared config (rather than recomputing from
-	// os.homedir()) guarantees we install to exactly the path the browser is
-	// later launched from — critical for `sudo npm i -g`, where postinstall runs
-	// as root but `hdoc` runs as a normal user with a different home directory.
-	const cacheDir = process.env.PUPPETEER_CACHE_DIR || puppeteerConfig.cacheDir;
-
-	let platform;
-	try {
-		platform = detectBrowserPlatform();
-	} catch (err) {
-		console.error(
-			`${RED}Unable to detect browser platform: ${err.message}${RESET}`,
-		);
-		process.exit(1);
-	}
-
-	const sleep = (ms) => new Promise((resolve) => setTimeout(resolve, ms));
-
-	const rmrf = (target) => {
-		try {
-			fs.rmSync(target, { recursive: true, force: true });
-		} catch {
-			/* best effort */
-		}
-	};
-
-	// A usable install is one where the executable exists and is non-empty.
-	const isUsable = (exePath) => {
-		try {
-			return fs.statSync(exePath).size > 0;
-		} catch {
-			return false;
-		}
-	};
-
-	const provision = async (browser, label) => {
-		const exePath = computeExecutablePath({
-			browser,
-			buildId: CHROME_BUILD,
-			cacheDir,
-			platform,
-		});
-		// chrome.exe -> chrome-win64 -> win64-<build>; nuke the whole build folder
-		// so a partial extraction can never block a clean re-extract.
-		const versionFolder = path.dirname(path.dirname(exePath));
-
-		if (isUsable(exePath)) {
-			log(`${label} ${CHROME_BUILD} already present, skipping download.`);
-			return true;
-		}
-
-		for (let attempt = 1; attempt <= MAX_ATTEMPTS; attempt++) {
-			rmrf(versionFolder); // clear any stale/partial extract first
-			try {
-				log(
-					`Installing ${label} ${CHROME_BUILD} (attempt ${attempt}/${MAX_ATTEMPTS})...`,
-				);
-				await install({
-					browser,
-					buildId: CHROME_BUILD,
-					cacheDir,
-					platform,
-					unpack: true,
-				});
-			} catch (err) {
-				console.error(
-					`${YELLOW}  attempt ${attempt} failed: ${err.message}${RESET}`,
-				);
-			}
-
-			if (isUsable(exePath)) {
-				log(`${GREEN}${label} ${CHROME_BUILD} ready.${RESET}`);
-				return true;
-			}
-
-			if (attempt < MAX_ATTEMPTS) await sleep(2000);
-		}
-
-		return false;
-	};
-
-	const chromeOk = await provision(Browser.CHROME, "Chrome");
-	const shellOk = await provision(
-		Browser.CHROMEHEADLESSSHELL,
-		"chrome-headless-shell",
-	);
-
-	if (chromeOk && shellOk) process.exit(0);
-
-	console.error(
-		`\n${RED}Failed to provision a complete browser into:${RESET}\n  ${cacheDir}\n\n` +
-			"This is almost always one of:\n" +
-			"  1. Antivirus (e.g. Windows Defender) quarantining Chrome files mid-extract.\n" +
-			`     Fix: Add-MpPreference -ExclusionPath "${cacheDir}"\n` +
-			"  2. A stale/partial cache folder. Fix: delete the chrome / chrome-headless-shell\n" +
-			`     sub-folders under "${cacheDir}" and reinstall.\n` +
-			"  3. A truncated download behind a proxy / TLS inspection. Check npm/HTTPS proxy config.\n",
-	);
-	process.exit(1);
-})();
+// Resilient browser provisioning for hdoc-tools.
+//
+// Replaces the old `puppeteer browsers install ...` one-liner postinstall.
+// Puppeteer's own bundled download (install.mjs) is disabled via
+// .puppeteerrc.cjs (skipDownload), so this script is the single, controlled
+// place Chrome + chrome-headless-shell are fetched.
+//
+// Why: on some Windows Server 2019 build agents the Chrome archive extracts
+// only partially (Defender quarantining binaries mid-extract, or a truncated
+// download behind a proxy). The first failure leaves the version folder on
+// disk, after which Puppeteer refuses to re-extract and every later install
+// reports "folder exists but executable missing" — a permanent dead end.
+//
+// This script makes provisioning idempotent and self-healing:
+//   * skip when a valid executable already exists,
+//   * delete any stale/partial version folder before (re)installing,
+//   * retry the download a few times,
+//   * verify the executable exists afterwards and fail loudly with concrete
+//     remediation if it still does not.
+
+(async () => {
+	const fs = require("node:fs");
+	const path = require("node:path");
+	const {
+		install,
+		computeExecutablePath,
+		detectBrowserPlatform,
+		Browser,
+	} = require("@puppeteer/browsers");
+
+	// Build id + cache dir come from the shared .puppeteerrc.cjs so install and
+	// runtime launch can never drift apart.
+	const puppeteerConfig = require(path.join(__dirname, ".puppeteerrc.cjs"));
+	const CHROME_BUILD = puppeteerConfig.chromeBuild;
+	const MAX_ATTEMPTS = 3;
+
+	const RED = "\x1b[31m";
+	const YELLOW = "\x1b[33m";
+	const GREEN = "\x1b[32m";
+	const RESET = "\x1b[0m";
+
+	const log = (msg) => console.log(`[hdoc-tools] ${msg}`);
+
+	// Resolve the cache directory the same way Puppeteer does at runtime:
+	// PUPPETEER_CACHE_DIR wins, otherwise the fixed cacheDir declared in
+	// .puppeteerrc.cjs. Reading the shared config (rather than recomputing from
+	// os.homedir()) guarantees we install to exactly the path the browser is
+	// later launched from — critical for `sudo npm i -g`, where postinstall runs
+	// as root but `hdoc` runs as a normal user with a different home directory.
+	const cacheDir = process.env.PUPPETEER_CACHE_DIR || puppeteerConfig.cacheDir;
+
+	let platform;
+	try {
+		platform = detectBrowserPlatform();
+	} catch (err) {
+		console.error(
+			`${RED}Unable to detect browser platform: ${err.message}${RESET}`,
+		);
+		process.exit(1);
+	}
+
+	const sleep = (ms) => new Promise((resolve) => setTimeout(resolve, ms));
+
+	const rmrf = (target) => {
+		try {
+			fs.rmSync(target, { recursive: true, force: true });
+		} catch {
+			/* best effort */
+		}
+	};
+
+	// A usable install is one where the executable exists and is non-empty.
+	const isUsable = (exePath) => {
+		try {
+			return fs.statSync(exePath).size > 0;
+		} catch {
+			return false;
+		}
+	};
+
+	const provision = async (browser, label) => {
+		const exePath = computeExecutablePath({
+			browser,
+			buildId: CHROME_BUILD,
+			cacheDir,
+			platform,
+		});
+		// chrome.exe -> chrome-win64 -> win64-<build>; nuke the whole build folder
+		// so a partial extraction can never block a clean re-extract.
+		const versionFolder = path.dirname(path.dirname(exePath));
+
+		if (isUsable(exePath)) {
+			log(`${label} ${CHROME_BUILD} already present, skipping download.`);
+			return true;
+		}
+
+		for (let attempt = 1; attempt <= MAX_ATTEMPTS; attempt++) {
+			rmrf(versionFolder); // clear any stale/partial extract first
+			try {
+				log(
+					`Installing ${label} ${CHROME_BUILD} (attempt ${attempt}/${MAX_ATTEMPTS})...`,
+				);
+				await install({
+					browser,
+					buildId: CHROME_BUILD,
+					cacheDir,
+					platform,
+					unpack: true,
+				});
+			} catch (err) {
+				console.error(
+					`${YELLOW}  attempt ${attempt} failed: ${err.message}${RESET}`,
+				);
+			}
+
+			if (isUsable(exePath)) {
+				log(`${GREEN}${label} ${CHROME_BUILD} ready.${RESET}`);
+				return true;
+			}
+
+			if (attempt < MAX_ATTEMPTS) await sleep(2000);
+		}
+
+		return false;
+	};
+
+	const chromeOk = await provision(Browser.CHROME, "Chrome");
+	const shellOk = await provision(
+		Browser.CHROMEHEADLESSSHELL,
+		"chrome-headless-shell",
+	);
+
+	if (chromeOk && shellOk) process.exit(0);
+
+	console.error(
+		`\n${RED}Failed to provision a complete browser into:${RESET}\n  ${cacheDir}\n\n` +
+			"This is almost always one of:\n" +
+			"  1. Antivirus (e.g. Windows Defender) quarantining Chrome files mid-extract.\n" +
+			`     Fix: Add-MpPreference -ExclusionPath "${cacheDir}"\n` +
+			"  2. A stale/partial cache folder. Fix: delete the chrome / chrome-headless-shell\n" +
+			`     sub-folders under "${cacheDir}" and reinstall.\n` +
+			"  3. A truncated download behind a proxy / TLS inspection. Check npm/HTTPS proxy config.\n",
+	);
+	process.exit(1);
+})();