Add a base element to the DOM with current URL

This makes the DOM parse with a correct base URI instead of the extension's base URI, which fixes deathau#1

Author: gordonped

background/background.js

@@ -3,13 +3,13 @@ browser.runtime.onMessage.addListener(notify);
 
 // creates the readable article object from Readability
 function createReadableVersion(dom) {
-  var reader = new Readability(dom);
+  var reader = new Readability(dom, { debug: true });
   var article = reader.parse();
   return article;
 }
 
 // convert the article content to markdown using Turndown
-function convertArticleToMarkdown(article, url) {
+function convertArticleToMarkdown(article) {
   var turndownService = new TurndownService()
   var markdown = turndownService.turndown(article.content);
 
@@ -21,9 +21,6 @@ function convertArticleToMarkdown(article, url) {
     markdown = "> " + article.excerpt + "\n\n" + markdown;
   }
 
-  //add url
-  markdown = url + "\n\n" + markdown;
-
   return markdown;
 }
 
@@ -63,14 +60,22 @@ function downloadMarkdown(markdown, title) {
 function notify(message) {
   // message for initial clipping of the dom
   if (message.type == "clip") {
+
+    // parse the dom
     var parser = new DOMParser();
     var dom = parser.parseFromString(message.dom, "text/html");
     if (dom.documentElement.nodeName == "parsererror") {
       console.error("error while parsing");
     }
 
+    // make markdown document from the dom
     var article = createReadableVersion(dom);
-    var markdown = convertArticleToMarkdown(article, message.url);
+    var markdown = convertArticleToMarkdown(article);
+
+    // add url to the top of the markdown
+    markdown = dom.baseURI + "\n\n" + markdown;
+
+    // send a message to display the markdown
     browser.runtime.sendMessage({ type: "display.md", markdown: markdown, article: article });
   }
   // message for triggering download

contentScript/pageScraper.js

@@ -1,7 +1,17 @@
 function notifyExtension() {
-    //var serializer = new XMLSerializer();
-    //var content = serializer.serializeToString(document);
+    // if the document doesn't have a "base" element make one
+    // this allows the DOM parser in future steps to fix relative uris
+    if (document.head.getElementsByTagName('base').length == 0) {
+        let baseEl = document.createElement('base');
+        // use the current uri
+        baseEl.setAttribute('href', window.location.href);
+        document.head.append(baseEl);
+    }
+
+    // get the content of the page as a string
     var content = document.documentElement.outerHTML;
-    browser.runtime.sendMessage({ type: "clip", dom: content, url:window.location.href});
+    
+    // send a message that the content should be clipped
+    browser.runtime.sendMessage({ type: "clip", dom: content});
 }
 notifyExtension();