oss/miniflux-v2
oss/miniflux-v2
1
0
Fork 0
mirror of https://github.com/miniflux/v2.git synced 2025-08-01 17:38:37 +00:00
Code Wiki Activity

Handle invalid feeds with relative URLs

Browse source
This commit is contained in:
Frédéric Guillot 2020-12-02 20:47:11 -08:00 • committed by fguillot
parent 49feb1958c
commit f722fd1208
18 changed files with 392 additions and 157 deletions
Download patch file Download diff file Expand all files Collapse all files

4
reader/rdf/parser.go
View file

@ -13,7 +13,7 @@ import (
)
// Parse returns a normalized feed struct from a RDF feed.
func Parse(data io.Reader) (*model.Feed, *errors.LocalizedError) {
func Parse(baseURL string, data io.Reader) (*model.Feed, *errors.LocalizedError) {
feed := new(rdfFeed)
decoder := xml.NewDecoder(data)
err := decoder.Decode(feed)
@ -21,5 +21,5 @@ func Parse(data io.Reader) (*model.Feed, *errors.LocalizedError) {
return nil, errors.NewLocalizedError("Unable to parse RDF feed: %q", err)
}
return feed.Transform(), nil
return feed.Transform(baseURL), nil
}

28
reader/rdf/parser_test.go
View file

@ -76,7 +76,7 @@ func TestParseRDFSample(t *testing.T) {
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://xml.com/pub/rdf.xml", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -85,7 +85,7 @@ func TestParseRDFSample(t *testing.T) {
t.Errorf("Incorrect title, got: %s", feed.Title)
}
if feed.FeedURL != "" {
if feed.FeedURL != "http://xml.com/pub/rdf.xml" {
t.Errorf("Incorrect feed URL, got: %s", feed.FeedURL)
}
@ -187,7 +187,7 @@ func TestParseRDFSampleWithDublinCore(t *testing.T) {
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://meerkat.oreillynet.com/feed.rdf", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -196,7 +196,7 @@ func TestParseRDFSampleWithDublinCore(t *testing.T) {
t.Errorf("Incorrect title, got: %s", feed.Title)
}
if feed.FeedURL != "" {
if feed.FeedURL != "http://meerkat.oreillynet.com/feed.rdf" {
t.Errorf("Incorrect feed URL, got: %s", feed.FeedURL)
}
@ -254,7 +254,7 @@ func TestParseItemWithOnlyFeedAuthor(t *testing.T) {
</item>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://meerkat.oreillynet.com", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -279,7 +279,7 @@ func TestParseItemRelativeURL(t *testing.T) {
</item>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://meerkat.oreillynet.com", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -308,7 +308,7 @@ func TestParseItemWithoutLink(t *testing.T) {
</item>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://meerkat.oreillynet.com", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -339,7 +339,7 @@ func TestParseItemWithDublicCoreDate(t *testing.T) {
</item>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://example.org", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -365,7 +365,7 @@ func TestParseItemWithoutDate(t *testing.T) {
</item>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://example.org", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -379,7 +379,7 @@ func TestParseItemWithoutDate(t *testing.T) {
func TestParseInvalidXml(t *testing.T) {
data := `garbage`
_, err := Parse(bytes.NewBufferString(data))
_, err := Parse("http://example.org", bytes.NewBufferString(data))
if err == nil {
t.Fatal("Parse should returns an error")
}
@ -394,7 +394,7 @@ func TestParseFeedWithHTMLEntity(t *testing.T) {
</channel>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://example.org", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -413,7 +413,7 @@ func TestParseFeedWithInvalidCharacterEntity(t *testing.T) {
</channel>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://example.org", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -469,7 +469,7 @@ func TestParseFeedWithURLWrappedInSpaces(t *testing.T) {
</item>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://biorxiv.org", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}
@ -504,7 +504,7 @@ func TestParseRDFWithContentEncoded(t *testing.T) {
</item>
</rdf:RDF>`
feed, err := Parse(bytes.NewBufferString(data))
feed, err := Parse("http://example.org/", bytes.NewBufferString(data))
if err != nil {
t.Fatal(err)
}

9
reader/rdf/rdf.go
View file

@ -25,10 +25,15 @@ type rdfFeed struct {
DublinCoreFeedElement
}
func (r *rdfFeed) Transform() *model.Feed {
func (r *rdfFeed) Transform(baseURL string) *model.Feed {
var err error
feed := new(model.Feed)
feed.Title = sanitizer.StripTags(r.Title)
feed.SiteURL = r.Link
feed.FeedURL = baseURL
feed.SiteURL, err = url.AbsoluteURL(baseURL, r.Link)
if err != nil {
feed.SiteURL = r.Link
}
for _, item := range r.Items {
entry := item.Transform()