Merge pull request #332 from gijsk/fix-readability-determination-in-generate-testcase

Use jsdom for parsing the document to determine readability (fixes #325), r=evanxd

test/generate-testcase.js

@@ -111,12 +111,25 @@ function runReadability(source, destPath, metadataDestPath) {
   var myReader, result, readerable;
   try {
     myReader = new Readability(uri, doc);
-    readerable = myReader.isProbablyReaderable();
     result = myReader.parse();
   } catch (ex) {
     console.error(ex);
     ex.stack.forEach(console.log.bind(console));
   }
+  // Use jsdom for isProbablyReaderable because it supports querySelectorAll
+  try {
+    var jsdomDoc = jsdom(source, {
+      features: {
+        FetchExternalResources: false,
+        ProcessExternalResources: false
+      }
+    });
+    myReader = new Readability(uri, jsdomDoc);
+    readerable = myReader.isProbablyReaderable();
+  } catch (ex) {
+    console.error(ex);
+    ex.stack.forEach(console.log.bind(console));
+  }
   if (!result) {
     console.error("No content generated by readability, not going to write expected.html!");
     return;

test/test-pages/blogger/expected-metadata.json

@@ -1,6 +1,7 @@
 {
   "title": "Open Verilog flow for Silego GreenPak4 programmable logic devices",
   "byline": null,
+  "dir": "ltr",
   "excerpt": "I've written a couple of posts in the past few months but they were all for",
   "readerable": true
 }