james/yarr
4
0
Fork 0
mirror of https://github.com/nkanaev/yarr.git synced 2025-05-24 00:33:14 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix atom links

Browse Source
This commit is contained in:
Nazar Kanaev 2023-09-07 18:19:17 +01:00
parent 479aebd023
commit 850ce195a0
9 changed files with 89 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
src/content/htmlutil/urlutils.go
View File

@ -2,6 +2,7 @@ package htmlutil
import (
"net/url"
"strings"
)
func Any(els []string, el string, match func(string, string) bool) bool {
@ -31,3 +32,7 @@ func URLDomain(val string) string {
}
return val
}
func IsAPossibleLink(val string) bool {
return strings.HasPrefix(val, "http://") || strings.HasPrefix(val, "https://")
}

11
src/parser/atom.go
View File

@ -81,9 +81,16 @@ func ParseAtom(r io.Reader) (*Feed, error) {
SiteURL: firstNonEmpty(srcfeed.Links.First("alternate"), srcfeed.Links.First("")),
}
for _, srcitem := range srcfeed.Entries {
link := firstNonEmpty(srcitem.OrigLink, srcitem.Links.First("alternate"), srcitem.Links.First(""))
linkFromID := ""
guidFromID := ""
if htmlutil.IsAPossibleLink(srcitem.ID) {
linkFromID = srcitem.ID
guidFromID = srcitem.ID + "::" + srcitem.Updated
}
link := firstNonEmpty(srcitem.OrigLink, srcitem.Links.First("alternate"), srcitem.Links.First(""), linkFromID)
dstfeed.Items = append(dstfeed.Items, Item{
GUID: firstNonEmpty(srcitem.ID, link),
GUID: firstNonEmpty(guidFromID, srcitem.ID, link),
Date: dateParse(firstNonEmpty(srcitem.Published, srcitem.Updated)),
URL: link,
Title: srcitem.Title.Text(),

45
src/parser/atom_test.go
View File

@ -131,3 +131,48 @@ func TestAtomImageLinkDuplicated(t *testing.T) {
t.Fatal("item.image_url must be unset if present in the content")
}
}
func TestAtomLinkInID(t *testing.T) {
feed, _ := Parse(strings.NewReader(`
<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xmlns:media="http://search.yahoo.com/mrss/">
<entry>
<title>one updated</title>
<id>https://example.com/posts/1</id>
<updated>2003-12-13T09:17:51</updated>
</entry>
<entry>
<title>two</title>
<id>urn:uuid:60a76c80-d399-11d9-b93C-0003939e0af6</id>
</entry>
<entry>
<title>one</title>
<id>https://example.com/posts/1</id>
</entry>
</feed>
`))
have := feed.Items
want := []Item{
Item{
GUID: "https://example.com/posts/1::2003-12-13T09:17:51",
Date: time.Date(2003, time.December, 13, 9, 17, 51, 0, time.UTC),
URL: "https://example.com/posts/1",
Title: "one updated",
},
Item{
GUID: "urn:uuid:60a76c80-d399-11d9-b93C-0003939e0af6",
Date: time.Date(1, time.January, 1, 0, 0, 0, 0, time.UTC), URL: "",
Title: "two",
},
Item{
GUID: "https://example.com/posts/1::",
Date: time.Date(1, time.January, 1, 0, 0, 0, 0, time.UTC),
URL: "https://example.com/posts/1",
Title: "one",
Content: "",
},
}
if !reflect.DeepEqual(want, have) {
t.Fatalf("\nwant: %#v\nhave: %#v\n", want, have)
}
}

10
src/parser/rss.go
View File

@ -20,7 +20,7 @@ type rssFeed struct {
}
type rssItem struct {
GUID rssGuid `xml:"guid"`
GUID rssGuid `xml:"guid"`
Title string `xml:"title"`
Link string `xml:"rss link"`
Description string `xml:"rss description"`
@ -86,10 +86,10 @@ func ParseRSS(r io.Reader) (*Feed, error) {
}
}
permalink := ""
if srcitem.GUID.IsPermaLink == "true" {
permalink = srcitem.GUID.GUID
}
permalink := ""
if srcitem.GUID.IsPermaLink == "true" {
permalink = srcitem.GUID.GUID
}
dstfeed.Items = append(dstfeed.Items, Item{
GUID: firstNonEmpty(srcitem.GUID.GUID, srcitem.Link),

10
src/parser/rss_test.go
View File

@ -217,11 +217,11 @@ func TestRSSIsPermalink(t *testing.T) {
`))
have := feed.Items
want := []Item{
{
GUID: "http://example.com/posts/1",
URL: "http://example.com/posts/1",
},
}
{
GUID: "http://example.com/posts/1",
URL: "http://example.com/posts/1",
},
}
for i := 0; i < len(want); i++ {
if want[i] != have[i] {
t.Errorf("Failed to handle isPermalink\nwant: %#v\nhave: %#v\n", want[i], have[i])

12
src/server/routes.go
View File

@ -314,12 +314,12 @@ func (s *Server) handleItem(c *router.Context) {
return
}
// runtime fix for relative links
if !strings.HasPrefix(item.Link, "http") {
if feed := s.db.GetFeed(item.FeedId); feed != nil {
item.Link = htmlutil.AbsoluteUrl(item.Link, feed.Link)
}
}
// runtime fix for relative links
if !htmlutil.IsAPossibleLink(item.Link) {
if feed := s.db.GetFeed(item.FeedId); feed != nil {
item.Link = htmlutil.AbsoluteUrl(item.Link, feed.Link)
}
}
item.Content = sanitizer.Sanitize(item.Link, item.Content)

12
src/storage/feed.go
View File

@ -29,12 +29,12 @@ func (s *Storage) CreateFeed(title, description, link, feedLink string, folderId
folderId,
)
var id int64
err := row.Scan(&id)
if err != nil {
log.Print(err)
return nil
}
var id int64
err := row.Scan(&id)
if err != nil {
log.Print(err)
return nil
}
return &Feed{
Id: id,
Title: title,

8
src/storage/feed_test.go
View File

@ -24,13 +24,13 @@ func TestCreateFeedSameLink(t *testing.T) {
t.Fatal("expected feed")
}
for i := 0; i < 10; i++ {
db.CreateFeed("title", "", "", "http://example2.com/feed.xml", nil)
}
for i := 0; i < 10; i++ {
db.CreateFeed("title", "", "", "http://example2.com/feed.xml", nil)
}
feed2 := db.CreateFeed("title", "", "http://example.com", "http://example1.com/feed.xml", nil)
if feed1.Id != feed2.Id {
t.Fatalf("expected the same feed.\nwant: %#v\nhave: %#v", feed1, feed2)
t.Fatalf("expected the same feed.\nwant: %#v\nhave: %#v", feed1, feed2)
}
}

8
src/storage/folder.go
View File

@ -17,11 +17,11 @@ func (s *Storage) CreateFolder(title string) *Folder {
on conflict (title) do update set title = ?
returning id`,
title, expanded,
// provide title again so that we can extract row id
title,
// provide title again so that we can extract row id
title,
)
var id int64
err := row.Scan(&id)
var id int64
err := row.Scan(&id)
if err != nil {
log.Print(err)