test(e2e): implement full sitemap testing logic as per klz-2026 standards

apps/web/scripts/check-forms.ts

@@ -3,10 +3,44 @@ import puppeteer from "puppeteer";
 const targetUrl = process.env.TEST_URL || "http://localhost:3000";
 const gatekeeperPassword = process.env.GATEKEEPER_PASSWORD || "secret";
 
+async function fetchSitemapUrls(baseUrl: string): Promise<string[]> {
+  const sitemapUrl = `${baseUrl.replace(/\/$/, "")}/sitemap.xml`;
+  console.log(`📥 Fetching sitemap from ${sitemapUrl}...`);
+  try {
+    const response = await fetch(sitemapUrl);
+    const text = await response.text();
+
+    // Simple regex to extract loc tags
+    const matches = text.matchAll(/<loc>(.*?)<\/loc>/g);
+    let urls = Array.from(matches, (m) => m[1]);
+
+    // Normalize to target URL instance
+    const urlPattern = /https?:\/\/[^\/]+/;
+    urls = [...new Set(urls)]
+      .filter((u) => u.startsWith("http"))
+      .map((u) => u.replace(urlPattern, baseUrl.replace(/\/$/, "")))
+      .sort();
+
+    console.log(`✅ Found ${urls.length} target URLs.`);
+    return urls;
+  } catch (err: any) {
+    console.error(`❌ Failed to fetch sitemap: ${err.message}`);
+    return [];
+  }
+}
+
 async function main() {
-  console.log(`\n🚀 Starting E2E Form Submission Check for: ${targetUrl}`);
+  console.log(`\n🚀 Starting Strict Asset Integrity Check for: ${targetUrl}`);
+
+  let urls = await fetchSitemapUrls(targetUrl);
+
+  if (urls.length === 0) {
+    console.warn(`⚠️ Falling back to just the homepage.`);
+    urls = [targetUrl];
+  }
 
   // Launch browser with KLZ pattern: use system chromium via env
+  console.log(`\n🕷️ Launching Puppeteer Headless Engine...`);
   const browser = await puppeteer.launch({
     headless: true,
     executablePath:
@@ -26,23 +60,74 @@ async function main() {
 
   const page = await browser.newPage();
 
-  // Enable console logging from the page for debugging
+  let hasBrokenAssets = false;
-  page.on("console", (msg) => console.log(`  [PAGE] ${msg.text()}`));
+  let currentScannedUrl = urls[0] || "";
-  page.on("pageerror", (err: Error) =>
+
-    console.error(`  [PAGE ERROR] ${err.message}`),
+  // Listen for console logging from the page for debugging
-  );
+  page.on("console", (msg) => {
-  page.on("requestfailed", (req) =>
+    const type = msg.type();
-    console.error(
+    // Only capture errors and warnings, not info/logs
-      `  [REQUEST FAILED] ${req.url()} - ${req.failure()?.errorText}`,
+    if (type === "error" || type === "warn") {
-    ),
+      const text = msg.text();
-  );
+      // Exclude common noise
+      if (
+        text.includes("google-analytics") ||
+        text.includes("googletagmanager") ||
+        text.includes("Fast Refresh")
+      )
+        return;
+
+      console.log(`  [PAGE ${type.toUpperCase()}] ${text}`);
+    }
+  });
+
+  page.on("pageerror", (err: Error) => {
+    if (currentScannedUrl.includes("showcase")) return;
+    console.error(`  [PAGE EXCEPTION] ${err.message}`);
+  });
+
+  // Listen to ALL network responses to catch broken assets (404/500)
+  page.on("response", (response) => {
+    const status = response.status();
+    // Catch classic 404s and 500s on ANY fetch/image/script
+    if (
+      status >= 400 &&
+      status !== 429 &&
+      status !== 999 &&
+      !response.url().includes("google-analytics") &&
+      !response.url().includes("googletagmanager")
+    ) {
+      const type = response.request().resourceType();
+
+      // We explicitly care about images, scripts, stylesheets, and fetches getting 404/500s.
+      if (
+        ["image", "script", "stylesheet", "fetch", "xhr", "document"].includes(
+          type,
+        )
+      ) {
+        // Exclude showcase routes from strict sub-asset checking since they proxy external content
+        if (
+          (currentScannedUrl.includes("showcase") ||
+            response.url().includes("showcase")) &&
+          type !== "document"
+        ) {
+          return;
+        }
+
+        console.error(
+          `  [REQUEST FAILED] ${response.url()} - Status: ${status} (${type})`,
+        );
+        hasBrokenAssets = true;
+      }
+    }
+  });
 
   try {
     // Authenticate through Gatekeeper
     console.log(`\n🛡️ Authenticating through Gatekeeper...`);
-    console.log(`  Navigating to: ${targetUrl}`);
+    console.log(`  Navigating to: ${urls[0]}`);
 
-    const response = await page.goto(targetUrl, {
+    const response = await page.goto(urls[0], {
       waitUntil: "domcontentloaded",
       timeout: 60000,
     });
@@ -69,33 +154,74 @@ async function main() {
       console.log(`✅ Already authenticated (no Gatekeeper gate detected).`);
     }
 
-    // Basic smoke test
+    // Scan each page
-    console.log(`\n🧪 Testing page load...`);
+    console.log(`\n🧪 Testing all ${urls.length} pages...`);
-    const title = await page.title();
+    for (let i = 0; i < urls.length; i++) {
-    console.log(`✅ Page Title: ${title}`);
+      const u = urls[i];
+      currentScannedUrl = u;
+      console.log(`\n[${i + 1}/${urls.length}] Scanning: ${u}`);
+      try {
+        await page.goto(u, { waitUntil: "domcontentloaded", timeout: 60000 });
 
-    if (title.toLowerCase().includes("mintel")) {
+        // Simulate a scroll to bottom to trigger lazy-loads if necessary
-      console.log(`✅ Basic smoke test passed!`);
+        await page.evaluate(async () => {
-    } else {
+          await new Promise<void>((resolve) => {
-      throw new Error(`Page title mismatch: "${title}"`);
+            let totalHeight = 0;
+            const distance = 500;
+            const timer = setInterval(() => {
+              const scrollHeight = document.body.scrollHeight;
+              window.scrollBy(0, distance);
+              totalHeight += distance;
+              // Stop scrolling if we reached the bottom or scrolled for more than 5 seconds
+              if (totalHeight >= scrollHeight || totalHeight > 10000) {
+                clearInterval(timer);
+                resolve();
+              }
+            }, 100);
+          });
+        });
+
+        // Small delay for final hydration and asynchronous asset loading
+        await new Promise((resolve) => setTimeout(resolve, 1500));
+
+        const title = await page.title();
+        console.log(`  ✅ Page Title: ${title}`);
+
+        if (!title) {
+          throw new Error(`Page title is missing.`);
+        }
+      } catch (err: any) {
+        console.error(
+          `  ❌ Timeout or navigation error on ${u}: ${err.message}`,
+        );
+        hasBrokenAssets = true;
+      }
     }
   } catch (err: any) {
-    console.error(`❌ Test Failed: ${err.message}`);
+    console.error(`\n❌ Fatal Test Error: ${err.message}`);
-    // Take a screenshot for debugging
+    // Take a screenshot for debugging on crash
     try {
       const screenshotPath = "/tmp/e2e-failure.png";
       await page.screenshot({ path: screenshotPath, fullPage: true });
       console.log(`📸 Screenshot saved to ${screenshotPath}`);
     } catch {
-      /* ignore screenshot errors */
+      /* ignore */
     }
-    console.log(`  Current URL: ${page.url()}`);
+    hasBrokenAssets = true;
-    await browser.close();
-    process.exit(1);
   }
 
   await browser.close();
-  console.log(`\n🎉 SUCCESS: E2E smoke test passed!`);
+
+  if (hasBrokenAssets) {
+    console.error(
+      `\n🚨 The CI build will now fail to prevent bad code from reaching production.`,
+    );
+    process.exit(1);
+  }
+
+  console.log(
+    `\n🎉 SUCCESS: All ${urls.length} pages rendered perfectly with 0 broken assets!`,
+  );
   process.exit(0);
 }