Do not choke on PDF specs

src/browserlib/get-title.mjs

@@ -1,11 +1,14 @@
 /**
  * Gets the title of the document
  */
-export default function () {
+export default function (spec) {
   const title = window.document.querySelector('title');
   if (title) {
     return title.textContent.replace(/\s+/g, ' ').trim();
   }
+  else if (spec?.title) {
+    return spec.title;
+  }
   else {
     return '[No title found for ' + window.location.href + ']';
   }

src/lib/util.js

@@ -350,6 +350,19 @@ async function processSpecification(spec, processFunction, args, options) {
                     return;
                 }
 
+                // Puppeteer does not support loading PDF files and we would
+                // not know how to parse them in any case. Let's return an
+                // empty HTML page instead.
+                if (/\.pdf$/i.test(request.url)) {
+                    await cdp.send('Fetch.fulfillRequest', {
+                        requestId,
+                        responseCode: 200,
+                        responseHeaders: [{ name: 'Content-Type', value: 'text/html' }],
+                        body: ''
+                    });
+                    return;
+                }
+
                 // Abort network requests that return a "stream", they won't
                 // play well with Puppeteer's "networkidle0" option, and our
                 // custom "fetch" function does not handle streams in any case

tests/crawl.js

@@ -73,7 +73,7 @@ if (global.describe && describe instanceof Function) {
         forceLocalFetch: true
       });
       const results = require(path.resolve(output, 'index.json'));
-      assert.equal(refResults.title, results.results[0].title);
+      assert.equal(results.results[0].title, 'Accelerometer');
     });
 
     it("matches spec series shortnames", async () => {