Improve doctor.mjs Playwright browser check with distinct failure modes

[Copilot]

checkPlaywright() treated "playwright package missing" and "chromium binary missing" as the same failure with the same fix instruction, which is misleading.

• Separate failure detection: Split the single try/catch into two stages — package import vs binary existence — with targeted fix instructions (npm install vs npx playwright install chromium)
• Show binary path on success: Include executablePath() in the pass label for debuggability

✓ Playwright Chromium browser installed (/home/runner/.cache/ms-playwright/chromium-1217/chrome-linux64/chrome)

✗ Playwright Chromium browser not installed
  → Run: npx playwright install chromium

✗ Playwright package not available
  → Run: npm install

[Copilot]

Pull request overview

This PR refines doctor.mjs’s Playwright prerequisite check so it distinguishes between a missing Playwright package and a missing Chromium browser binary, and improves debuggability by printing the resolved Chromium executable path on success.

Changes:

• Split checkPlaywright() into two stages: Playwright package import vs Chromium binary presence.
• Add targeted fix instructions for each failure mode (npm install vs npx playwright install chromium).
• Include chromium.executablePath() in the success label.

[Copilot]

The import stage uses a bare catch and always reports "Playwright package not available" with the generic npm install fix. This will misdiagnose other import-time failures (e.g., corrupted install, unsupported platform/architecture, or other runtime errors). Capture the thrown error and only treat it as "package not available" when it’s a module-not-found case (e.g., err.code === 'ERR_MODULE_NOT_FOUND' / message indicates missing playwright); otherwise surface a distinct failure label (ideally including err.message) so users aren’t sent to the wrong fix.

Suggested change

	} catch {
	return {
	pass: false,
	label: 'Playwright package not available',
	fix: 'Run: npm install',
	} catch (err) {
	const errCode = err && typeof err === 'object' ? err.code : undefined;
	const errMessage = err instanceof Error ? err.message : String(err);
	const isModuleNotFound =
	errCode === 'ERR_MODULE_NOT_FOUND' ||
	/cannot find package ['"]playwright['"]/i.test(errMessage) ||
	/cannot find module ['"]playwright['"]/i.test(errMessage) ||
	/module ['"]playwright['"] was not found/i.test(errMessage);
	if (isModuleNotFound) {
	return {
	pass: false,
	label: 'Playwright package not available',
	fix: 'Run: npm install',
	};
	}
	return {
	pass: false,
	label: `Playwright import failed: ${errMessage}`,

[Copilot]

The Chromium check’s catch treats any exception as "browser not installed". If chromium.executablePath() throws for reasons other than a missing browser (e.g., Playwright internal error), the output will point to npx playwright install chromium even though that may not resolve it. Capture the error and report a separate failure mode (or include err.message) so this stage only claims "not installed" when the binary is actually absent.

[RajjjAryan]

Closing — superseded by #31 which was merged to main and includes a more comprehensive fix (Playwright browser launch test + YAML validation for profile.yml and portals.yml).

Thank you for the contribution! 🙏