Add guid link inference option

Author: danmactough

README.md

@@ -121,6 +121,10 @@ You can also check out this nice [working implementation](https://github.com/scr
   If you only need the metadata available when each article streams, you can
   use `item.meta` as usual.
 
+- `guidlink` - Set to `false` to override Feedparser's default behavior, which
+  is to use an RSS item's `guid` as the item `link` when the item has no `link`
+  and the `guid` starts with `http:` or `https:`.
+
 - `feedurl` - The url (string) of the feed. FeedParser is very good at
   resolving relative urls in feeds, including those embedded in HTML content
   fields. But some feeds use relative urls without declaring the `xml:base`

index.d.ts

@@ -36,6 +36,7 @@ declare namespace FeedParser {
         strict?: boolean;
         normalize?: boolean;
         addmeta?: boolean;
+        guidlink?: boolean;
         feedurl?: string;
         resume_saxerror?: boolean;
         MAX_BUFFER_LENGTH?: number;

lib/feedparser.js

@@ -77,6 +77,7 @@ function FeedParser (options) {
   if (!('strict' in this.options)) this.options.strict = false;
   if (!('normalize' in this.options)) this.options.normalize = true;
   if (!('addmeta' in this.options)) this.options.addmeta = true;
+  if (!('guidlink' in this.options)) this.options.guidlink = true;
   if (!('resume_saxerror' in this.options)) this.options.resume_saxerror = true;
   // MAX_BUFFER_LENGTH is not part of the public API of sax, but we need to be
   // able to handle nodes that are larger than the 64K default
@@ -1168,7 +1169,7 @@ FeedParser.prototype.handleItem = function handleItem (node, type, options) {
     if (item.categories.length) {
       item.categories = _.uniq(item.categories);
     }
-    if (!item.link) {
+    if (!item.link && (!options || options.guidlink)) {
       if (item.guid && /^https?:/.test(item.guid)) {
         item.link = item.guid;
       }

test/link.js

@@ -47,4 +47,64 @@ describe('links', function () {
       });
   });
 
+  it('should infer item link from http guid by default (issue #293)', function (done) {
+    var feedparser = new FeedParser();
+    var feed = '<?xml version="1.0" encoding="UTF-8"?>' +
+      '<rss version="2.0">' +
+      '<channel>' +
+      '<title>Linkless feed</title>' +
+      '<link>http://example.com/</link>' +
+      '<description>Feed with linkless items</description>' +
+      '<item>' +
+      '<title>One</title>' +
+      '<guid>http://example.com/posts/one</guid>' +
+      '</item>' +
+      '</channel>' +
+      '</rss>';
+
+    feedparser
+      .once('readable', function () {
+        var item = this.read();
+        assert.equal(item.guid, 'http://example.com/posts/one');
+        assert.equal(item.link, 'http://example.com/posts/one');
+        done();
+      })
+      .on('error', function (err) {
+        assert.ifError(err);
+        done(err);
+      });
+
+    feedparser.end(feed);
+  });
+
+  it('should not infer item link from guid when guidlink is false (issue #293)', function (done) {
+    var feedparser = new FeedParser({ guidlink: false });
+    var feed = '<?xml version="1.0" encoding="UTF-8"?>' +
+      '<rss version="2.0">' +
+      '<channel>' +
+      '<title>Linkless feed</title>' +
+      '<link>http://example.com/</link>' +
+      '<description>Feed with linkless items</description>' +
+      '<item>' +
+      '<title>One</title>' +
+      '<guid>http://example.com/posts/one</guid>' +
+      '</item>' +
+      '</channel>' +
+      '</rss>';
+
+    feedparser
+      .once('readable', function () {
+        var item = this.read();
+        assert.equal(item.guid, 'http://example.com/posts/one');
+        assert.equal(item.link, null);
+        done();
+      })
+      .on('error', function (err) {
+        assert.ifError(err);
+        done(err);
+      });
+
+    feedparser.end(feed);
+  });
+
 });

test/types.ts

@@ -8,6 +8,7 @@ const fp2 = new FeedParser({
   strict: false,
   normalize: true,
   addmeta: true,
+  guidlink: true,
   feedurl: 'https://example.com/feed',
   resume_saxerror: true,
   MAX_BUFFER_LENGTH: 1024 * 1024,