Add real podcast subsystem: RSS feed refresh, bounded background downloads, restart recovery

internal/podcasts/rss.go

@@ -0,0 +1,57 @@
+package podcasts
+
+import "encoding/xml"
+
+// rssFeed is a pragmatic subset of RSS 2.0 (with the de facto iTunes
+// podcast extensions most feeds use) — enough to get channel metadata
+// and each episode's title, description, publish date, and audio
+// enclosure. Real podcast feeds vary a lot in which optional fields they
+// include, so every field here is best-effort.
+type rssFeed struct {
+	Channel rssChannel `xml:"channel"`
+}
+
+type rssChannel struct {
+	Title       string    `xml:"title"`
+	Description string    `xml:"description"`
+	Image       rssImage  `xml:"image"`
+	Items       []rssItem `xml:"item"`
+}
+
+type rssImage struct {
+	URL string `xml:"url"`
+}
+
+type rssItem struct {
+	GUID        rssGUID      `xml:"guid"`
+	Title       string       `xml:"title"`
+	Description string       `xml:"description"`
+	PubDate     string       `xml:"pubDate"`
+	Enclosure   rssEnclosure `xml:"enclosure"`
+	Duration    string       `xml:"http://www.itunes.com/dtds/podcast-1.0.dtd duration"`
+}
+
+type rssGUID struct {
+	Value string `xml:",chardata"`
+}
+
+type rssEnclosure struct {
+	URL string `xml:"url,attr"`
+}
+
+func parseRSS(data []byte) (*rssFeed, error) {
+	var feed rssFeed
+	if err := xml.Unmarshal(data, &feed); err != nil {
+		return nil, err
+	}
+	return &feed, nil
+}
+
+// episodeGUID returns the item's GUID if present, falling back to its
+// enclosure URL — some minimal feeds omit <guid> entirely.
+func (i rssItem) episodeGUID() string {
+	if i.GUID.Value != "" {
+		return i.GUID.Value
+	}
+	return i.Enclosure.URL
+}