oss/miniflux-v2
oss/miniflux-v2
1
0
Fork 0
mirror of https://github.com/miniflux/v2.git synced 2025-06-27 16:36:00 +00:00
Code Wiki Activity

fix(rss): handle item title with CDATA content correctly

Browse source 
Fix regression introduced in commit a3ce03cc
This commit is contained in:
Frédéric Guillot 2025-02-15 14:46:04 -08:00
parent a3ce03cc9d
commit 7f54b27079
3 changed files with 205 additions and 164 deletions
Download patch file Download diff file Expand all files Collapse all files

2
internal/reader/rss/adapter.go
View file

@ -173,7 +173,7 @@ func findFeedAuthor(rssChannel *RSSChannel) string {
} }
func findEntryTitle(rssItem *RSSItem) string { func findEntryTitle(rssItem *RSSItem) string {
title := sanitizer.StripTags(rssItem.Title.Inner) title := rssItem.Title.Content
if rssItem.DublinCoreTitle != "" { if rssItem.DublinCoreTitle != "" {
title = rssItem.DublinCoreTitle title = rssItem.DublinCoreTitle

335
internal/reader/rss/parser_test.go
View file

@ -311,6 +311,184 @@ func TestParseEntryWithDCTitleOnly(t *testing.T) {
} }
} }
func TestParseFeedTitleWithHTMLEntity(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>
<title>Example &nbsp; Feed</title>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Title != "Example \u00a0 Feed" {
t.Errorf(`Incorrect title, got: %q`, feed.Title)
}
}
func TestParseFeedTitleWithUnicodeEntityAndCdata(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>
<title><![CDATA[Jenny&#8217;s Newsletter]]></title>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Title != `Jenny’s Newsletter` {
t.Errorf(`Incorrect title, got: %q`, feed.Title)
}
}
func TestParseItemTitleWithHTMLEntity(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>
<title>Example</title>
<item>
<title>&lt;/example&gt;</title>
<link>http://www.example.org/entries/1</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "</example>" {
t.Errorf(`Incorrect title, got: %q`, feed.Entries[0].Title)
}
}
func TestParseItemTitleWithNumericCharacterReference(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>
<title>Example</title>
<item>
<title>&#931; &#xDF;</title>
<link>http://www.example.org/article.html</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "Σ ß" {
t.Errorf(`Incorrect title, got: %q`, feed.Entries[0].Title)
}
}
func TestParseItemTitleWithDoubleEncodedEntities(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>
<title>Example</title>
<item>
<title>&amp;#39;Text&amp;#39;</title>
<link>http://www.example.org/article.html</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "'Text'" {
t.Errorf(`Incorrect title, got: %q`, feed.Entries[0].Title)
}
}
func TestParseItemTitleWithWhitespaces(t *testing.T) {
data := `<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0">
<channel>
<title>Example</title>
<link>http://example.org</link>
<item>
<title>
Some Title
</title>
<link>http://www.example.org/entries/1</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "Some Title" {
t.Errorf("Incorrect entry title, got: %s", feed.Entries[0].Title)
}
}
func TestParseItemTitleWithCDATA(t *testing.T) {
data := `<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0">
<channel>
<title>Example</title>
<link>http://example.org</link>
<item>
<title><![CDATA[This is a title]]></title>
<link>http://www.example.org/entries/1</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "This is a title" {
t.Errorf("Incorrect entry title, got: %s", feed.Entries[0].Title)
}
}
func TestParseItemTitleWithInnerHTML(t *testing.T) {
data := `<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0">
<channel>
<title>Example</title>
<link>http://example.org</link>
<item>
<title>Test: <b>bold</b></title>
<link>http://www.example.org/entries/1</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "Test: bold" {
t.Errorf("Incorrect entry title, got: %s", feed.Entries[0].Title)
}
}
func TestParseEntryWithoutLink(t *testing.T) { func TestParseEntryWithoutLink(t *testing.T) {
data := `<?xml version="1.0" encoding="utf-8"?> data := `<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0"> <rss version="2.0">
@ -997,56 +1175,6 @@ func TestParseEntryWithFeedBurnerLink(t *testing.T) {
} }
} }
func TestParseEntryTitleWithWhitespaces(t *testing.T) {
data := `<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0">
<channel>
<title>Example</title>
<link>http://example.org</link>
<item>
<title>
Some Title
</title>
<link>http://www.example.org/entries/1</link>
<pubDate>Fri, 15 Jul 2005 00:00:00 -0500</pubDate>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "Some Title" {
t.Errorf("Incorrect entry title, got: %s", feed.Entries[0].Title)
}
}
func TestParseEntryTitleWithInnerHTML(t *testing.T) {
data := `<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0">
<channel>
<title>Example</title>
<link>http://example.org</link>
<item>
<title>Test: <b>bold</b></title>
<link>http://www.example.org/entries/1</link>
<pubDate>Fri, 15 Jul 2005 00:00:00 -0500</pubDate>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "Test: bold" {
t.Errorf("Incorrect entry title, got: %s", feed.Entries[0].Title)
}
}
func TestParseEntryWithEnclosures(t *testing.T) { func TestParseEntryWithEnclosures(t *testing.T) {
data := `<?xml version="1.0" encoding="utf-8"?> data := `<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0"> <rss version="2.0">
@ -1404,113 +1532,6 @@ func TestParseInvalidXml(t *testing.T) {
} }
} }
func TestParseFeedTitleWithHTMLEntity(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>
<title>Example &nbsp; Feed</title>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Title != "Example \u00a0 Feed" {
t.Errorf(`Incorrect title, got: %q`, feed.Title)
}
}
func TestParseFeedTitleWithUnicodeEntityAndCdata(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>
<title><![CDATA[Jenny&#8217;s Newsletter]]></title>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Title != `Jenny’s Newsletter` {
t.Errorf(`Incorrect title, got: %q`, feed.Title)
}
}
func TestParseItemTitleWithHTMLEntity(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>
<title>Example</title>
<item>
<title>&lt;/example&gt;</title>
<link>http://www.example.org/entries/1</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "</example>" {
t.Errorf(`Incorrect title, got: %q`, feed.Entries[0].Title)
}
}
func TestParseItemTitleWithNumericCharacterReference(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>
<title>Example</title>
<item>
<title>&#931; &#xDF;</title>
<link>http://www.example.org/article.html</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "Σ ß" {
t.Errorf(`Incorrect title, got: %q`, feed.Entries[0].Title)
}
}
func TestParseItemTitleWithDoubleEncodedEntities(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>
<title>Example</title>
<item>
<title>&amp;#39;Text&amp;#39;</title>
<link>http://www.example.org/article.html</link>
</item>
</channel>
</rss>`
feed, err := Parse("https://example.org/", bytes.NewReader([]byte(data)))
if err != nil {
t.Fatal(err)
}
if feed.Entries[0].Title != "'Text'" {
t.Errorf(`Incorrect title, got: %q`, feed.Entries[0].Title)
}
}
func TestParseFeedLinkWithInvalidCharacterEntity(t *testing.T) { func TestParseFeedLinkWithInvalidCharacterEntity(t *testing.T) {
data := `<?xml version="1.0" encoding="utf-8"?> data := `<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:slash="http://purl.org/rss/1.0/modules/slash/"> <rss version="2.0" xmlns:slash="http://purl.org/rss/1.0/modules/slash/">

32
internal/reader/rss/rss.go
View file

@ -111,7 +111,7 @@ type RSSImage struct {
type RSSItem struct { type RSSItem struct {
// Title is the title of the item. // Title is the title of the item.
Title RSSTitle `xml:"rss title"` Title InnerContent `xml:"rss title"`
// Link is the URL of the item. // Link is the URL of the item.
Link string `xml:"rss link"` Link string `xml:"rss link"`
@ -169,11 +169,6 @@ type RSSItem struct {
googleplay.GooglePlayItemElement googleplay.GooglePlayItemElement
} }
type RSSTitle struct {
Data string `xml:",chardata"`
Inner string `xml:",innerxml"`
}
type RSSAuthor struct { type RSSAuthor struct {
XMLName xml.Name XMLName xml.Name
Data string `xml:",chardata"` Data string `xml:",chardata"`
@ -203,3 +198,28 @@ type RSSSource struct {
URL string `xml:"url,attr"` URL string `xml:"url,attr"`
Name string `xml:",chardata"` Name string `xml:",chardata"`
} }
type InnerContent struct {
Content string
}
func (ic *InnerContent) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
var content strings.Builder
for {
token, err := d.Token()
if err != nil {
return err
}
switch t := token.(type) {
case xml.CharData:
content.Write(t)
case xml.EndElement:
if t == start.End() {
ic.Content = strings.TrimSpace(content.String())
return nil
}
}
}
}