Allow only absolute URLs in comments URL

Some feeds are using invalid URLs (random text).

2 files changed, 31 insertions, 1 deletions

diff --git a/reader/rss/parser_test.go b/reader/rss/parser_test.go
index dd1d261..b7ebd0a 100644
--- a/reader/rss/parser_test.go
+++ b/reader/rss/parser_test.go
@@ -837,6 +837,31 @@ func TestParseEntryWithCommentsURL(t *testing.T) {
 	}
 }
 
+func TestParseEntryWithInvalidCommentsURL(t *testing.T) {
+	data := `<?xml version="1.0" encoding="utf-8"?>
+		<rss version="2.0" xmlns:slash="http://purl.org/rss/1.0/modules/slash/">
+		<channel>
+			<link>https://example.org/</link>
+			<item>
+				<title>Item 1</title>
+				<link>https://example.org/item1</link>
+				<comments>
+					Some text
+				</comments>
+			</item>
+		</channel>
+		</rss>`
+
+	feed, err := Parse(bytes.NewBufferString(data))
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	if feed.Entries[0].CommentsURL != "" {
+		t.Errorf("Incorrect entry comments URL, got: %q", feed.Entries[0].CommentsURL)
+	}
+}
+
 func TestParseInvalidXml(t *testing.T) {
 	data := `garbage`
 	_, err := Parse(bytes.NewBufferString(data))
diff --git a/reader/rss/rss.go b/reader/rss/rss.go
index fd120cb..172f6f8 100644
--- a/reader/rss/rss.go
+++ b/reader/rss/rss.go
@@ -317,7 +317,12 @@ func (r *rssItem) entryEnclosures() model.EnclosureList {
 func (r *rssItem) entryCommentsURL() string {
 	for _, commentLink := range r.CommentLinks {
 		if commentLink.XMLName.Space == "" {
-			return strings.TrimSpace(commentLink.Data)
+			commentsURL := strings.TrimSpace(commentLink.Data)
+			// The comments URL is supposed to be absolute (some feeds publishes incorrect comments URL)
+			// See https://cyber.harvard.edu/rss/rss.html#ltcommentsgtSubelementOfLtitemgt
+			if url.IsAbsoluteURL(commentsURL) {
+				return commentsURL
+			}
 		}
 	}