Handle RSS feeds without entry links

3 files changed, 69 insertions, 38 deletions

diff --git a/reader/feed/rss/parser.go b/reader/feed/rss/parser.go
index 52310e4..247c789 100644
--- a/reader/feed/rss/parser.go
+++ b/reader/feed/rss/parser.go
@@ -7,22 +7,23 @@ package rss
 import (
 	"encoding/xml"
 	"fmt"
-	"github.com/miniflux/miniflux2/model"
 	"io"
 
+	"github.com/miniflux/miniflux2/model"
+
 	"golang.org/x/net/html/charset"
 )
 
 // Parse returns a normalized feed struct.
 func Parse(data io.Reader) (*model.Feed, error) {
-	rssFeed := new(RssFeed)
+	feed := new(rssFeed)
 	decoder := xml.NewDecoder(data)
 	decoder.CharsetReader = charset.NewReaderLabel
 
-	err := decoder.Decode(rssFeed)
+	err := decoder.Decode(feed)
 	if err != nil {
-		return nil, fmt.Errorf("Unable to parse RSS feed: %v", err)
+		return nil, fmt.Errorf("unable to parse RSS feed: %v", err)
 	}
 
-	return rssFeed.Transform(), nil
+	return feed.Transform(), nil
 }
diff --git a/reader/feed/rss/parser_test.go b/reader/feed/rss/parser_test.go
index 42dc965..8265da4 100644
--- a/reader/feed/rss/parser_test.go
+++ b/reader/feed/rss/parser_test.go
@@ -136,6 +136,31 @@ func TestParseEntryWithoutTitle(t *testing.T) {
 	}
 }
 
+func TestParseEntryWithoutLink(t *testing.T) {
+	data := `<?xml version="1.0" encoding="utf-8"?>
+		<rss version="2.0">
+		<channel>
+			<link>https://example.org/</link>
+			<item>
+				<guid isPermaLink="false">1234</guid>
+			</item>
+		</channel>
+		</rss>`
+
+	feed, err := Parse(bytes.NewBufferString(data))
+	if err != nil {
+		t.Error(err)
+	}
+
+	if feed.Entries[0].URL != "https://example.org/" {
+		t.Errorf("Incorrect entry link, got: %s", feed.Entries[0].URL)
+	}
+
+	if feed.Entries[0].Hash != "03ac674216f3e15c761ee1a5e255f067953623c8b388b4459e13f978d7c846f4" {
+		t.Errorf("Incorrect entry hash, got: %s", feed.Entries[0].Hash)
+	}
+}
+
 func TestParseFeedURLWithAtomLink(t *testing.T) {
 	data := `<?xml version="1.0" encoding="utf-8"?>
 		<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0">
diff --git a/reader/feed/rss/rss.go b/reader/feed/rss/rss.go
index d88ed6b..fd28943 100644
--- a/reader/feed/rss/rss.go
+++ b/reader/feed/rss/rss.go
@@ -6,38 +6,39 @@ package rss
 
 import (
 	"encoding/xml"
-	"github.com/miniflux/miniflux2/helper"
-	"github.com/miniflux/miniflux2/model"
-	"github.com/miniflux/miniflux2/reader/feed/date"
-	"github.com/miniflux/miniflux2/reader/processor"
-	"github.com/miniflux/miniflux2/reader/sanitizer"
 	"log"
 	"path"
 	"strconv"
 	"strings"
 	"time"
-)
 
-type RssLink struct {
-	XMLName xml.Name
-	Data    string `xml:",chardata"`
-	Href    string `xml:"href,attr"`
-}
+	"github.com/miniflux/miniflux2/helper"
+	"github.com/miniflux/miniflux2/model"
+	"github.com/miniflux/miniflux2/reader/feed/date"
+	"github.com/miniflux/miniflux2/reader/processor"
+	"github.com/miniflux/miniflux2/reader/sanitizer"
+)
 
-type RssFeed struct {
+type rssFeed struct {
 	XMLName      xml.Name  `xml:"rss"`
 	Version      string    `xml:"version,attr"`
 	Title        string    `xml:"channel>title"`
-	Links        []RssLink `xml:"channel>link"`
+	Links        []rssLink `xml:"channel>link"`
 	Language     string    `xml:"channel>language"`
 	Description  string    `xml:"channel>description"`
 	PubDate      string    `xml:"channel>pubDate"`
 	ItunesAuthor string    `xml:"http://www.itunes.com/dtds/podcast-1.0.dtd channel>author"`
-	Items        []RssItem `xml:"channel>item"`
+	Items        []rssItem `xml:"channel>item"`
 }
 
-type RssItem struct {
-	Guid              string         `xml:"guid"`
+type rssLink struct {
+	XMLName xml.Name
+	Data    string `xml:",chardata"`
+	Href    string `xml:"href,attr"`
+}
+
+type rssItem struct {
+	GUID              string         `xml:"guid"`
 	Title             string         `xml:"title"`
 	Link              string         `xml:"link"`
 	OriginalLink      string         `xml:"http://rssnamespace.org/feedburner/ext/1.0 origLink"`
@@ -45,25 +46,25 @@ type RssItem struct {
 	Content           string         `xml:"http://purl.org/rss/1.0/modules/content/ encoded"`
 	PubDate           string         `xml:"pubDate"`
 	Date              string         `xml:"http://purl.org/dc/elements/1.1/ date"`
-	Authors           []RssAuthor    `xml:"author"`
+	Authors           []rssAuthor    `xml:"author"`
 	Creator           string         `xml:"http://purl.org/dc/elements/1.1/ creator"`
-	Enclosures        []RssEnclosure `xml:"enclosure"`
+	Enclosures        []rssEnclosure `xml:"enclosure"`
 	OrigEnclosureLink string         `xml:"http://rssnamespace.org/feedburner/ext/1.0 origEnclosureLink"`
 }
 
-type RssAuthor struct {
+type rssAuthor struct {
 	XMLName xml.Name
 	Data    string `xml:",chardata"`
 	Name    string `xml:"name"`
 }
 
-type RssEnclosure struct {
-	Url    string `xml:"url,attr"`
+type rssEnclosure struct {
+	URL    string `xml:"url,attr"`
 	Type   string `xml:"type,attr"`
 	Length string `xml:"length,attr"`
 }
 
-func (r *RssFeed) GetSiteURL() string {
+func (r *rssFeed) GetSiteURL() string {
 	for _, elem := range r.Links {
 		if elem.XMLName.Space == "" {
 			return elem.Data
@@ -73,7 +74,7 @@ func (r *RssFeed) GetSiteURL() string {
 	return ""
 }
 
-func (r *RssFeed) GetFeedURL() string {
+func (r *rssFeed) GetFeedURL() string {
 	for _, elem := range r.Links {
 		if elem.XMLName.Space == "http://www.w3.org/2005/Atom" {
 			return elem.Href
@@ -83,7 +84,7 @@ func (r *RssFeed) GetFeedURL() string {
 	return ""
 }
 
-func (r *RssFeed) Transform() *model.Feed {
+func (r *rssFeed) Transform() *model.Feed {
 	feed := new(model.Feed)
 	feed.SiteURL = r.GetSiteURL()
 	feed.FeedURL = r.GetFeedURL()
@@ -101,12 +102,16 @@ func (r *RssFeed) Transform() *model.Feed {
 		}
 		entry.Author = sanitizer.StripTags(entry.Author)
 
+		if entry.URL == "" {
+			entry.URL = feed.SiteURL
+		}
+
 		feed.Entries = append(feed.Entries, entry)
 	}
 
 	return feed
 }
-func (i *RssItem) GetDate() time.Time {
+func (i *rssItem) GetDate() time.Time {
 	value := i.PubDate
 	if i.Date != "" {
 		value = i.Date
@@ -125,7 +130,7 @@ func (i *RssItem) GetDate() time.Time {
 	return time.Now()
 }
 
-func (i *RssItem) GetAuthor() string {
+func (i *rssItem) GetAuthor() string {
 	for _, element := range i.Authors {
 		if element.Name != "" {
 			return element.Name
@@ -139,8 +144,8 @@ func (i *RssItem) GetAuthor() string {
 	return i.Creator
 }
 
-func (i *RssItem) GetHash() string {
-	for _, value := range []string{i.Guid, i.Link} {
+func (i *rssItem) GetHash() string {
+	for _, value := range []string{i.GUID, i.Link} {
 		if value != "" {
 			return helper.Hash(value)
 		}
@@ -149,7 +154,7 @@ func (i *RssItem) GetHash() string {
 	return ""
 }
 
-func (i *RssItem) GetContent() string {
+func (i *rssItem) GetContent() string {
 	if i.Content != "" {
 		return i.Content
 	}
@@ -157,7 +162,7 @@ func (i *RssItem) GetContent() string {
 	return i.Description
 }
 
-func (i *RssItem) GetURL() string {
+func (i *rssItem) GetURL() string {
 	if i.OriginalLink != "" {
 		return i.OriginalLink
 	}
@@ -165,12 +170,12 @@ func (i *RssItem) GetURL() string {
 	return i.Link
 }
 
-func (i *RssItem) GetEnclosures() model.EnclosureList {
+func (i *rssItem) GetEnclosures() model.EnclosureList {
 	enclosures := make(model.EnclosureList, 0)
 
 	for _, enclosure := range i.Enclosures {
 		length, _ := strconv.Atoi(enclosure.Length)
-		enclosureURL := enclosure.Url
+		enclosureURL := enclosure.URL
 
 		if i.OrigEnclosureLink != "" {
 			filename := path.Base(i.OrigEnclosureLink)
@@ -189,7 +194,7 @@ func (i *RssItem) GetEnclosures() model.EnclosureList {
 	return enclosures
 }
 
-func (i *RssItem) Transform() *model.Entry {
+func (i *rssItem) Transform() *model.Entry {
 	entry := new(model.Entry)
 	entry.URL = i.GetURL()
 	entry.Date = i.GetDate()