fix(server+android): DIDL-Lite metadata for Sonos UPnP (error 1023)

After the X-Forwarded-Proto fix Sonos now gets a clean https:// URL
but returns vendor error 1023 - empty CurrentURIMetaData. Sonos
requires DIDL-Lite metadata with at minimum <res protocolInfo>
carrying the audio MIME type so it can validate the source before
playback. The original spec said 'Sonos accepts empty DIDL; recoverable
if a device rejects' - that was wrong for Sonos.

Server (cast_token.go):
- Look up the track and return mime (from tracks.file_format) +
  title in the cast-token response. mimeForFormat covers the common
  formats - mp3, flac, m4a/aac, ogg, opus, wav - falling through to
  audio/mpeg for unknowns.
- Missing track returns 404 (apierror.NotFound) instead of letting the
  caller mint a token for nothing.

Client (CastApi.kt, AVTransportClient.kt, OutputPickerController.kt):
- StreamTokenResponse gains mime + title (defaulted so old contracts
  stay parseable).
- AVTransportClient.setAVTransportURIWithMetadata builds minimal Sonos-
  acceptable DIDL-Lite around the URL + MIME + title. xml-escaped.
- selectUpnp calls the new overload; Timber.i now logs the MIME so the
  next on-device test shows it.

Generic UPnP renderers tolerate the DIDL shape too - no downside to
sending it everywhere.

android/app/src/main/java/com/fabledsword/minstrel/api/endpoints/CastApi.kt

@@ -39,11 +39,18 @@ data class StreamTokenRequest(
 /**
  * Response body. [url] is a fully-formed stream URL with [token] and
  * [exp] already embedded as query params — callers pass it verbatim
- * to `AVTransport.SetAVTransportURI`.
+ * to `AVTransport.SetAVTransportURI`. [mime] + [title] are the bits
+ * the client needs to build DIDL-Lite metadata for that call: Sonos
+ * rejects empty DIDL with vendor error 1023, so the server hands back
+ * the track's MIME (from `tracks.file_format`) and title so the
+ * client can populate `<res protocolInfo>` and `<dc:title>` without
+ * a follow-up round trip.
  */
 @Serializable
 data class StreamTokenResponse(
     val token: String,
     val exp: Long,
     val url: String,
+    val mime: String = "audio/mpeg",
+    val title: String = "",
 )

android/app/src/main/java/com/fabledsword/minstrel/player/output/OutputPickerController.kt

@@ -194,8 +194,12 @@ class OutputPickerController @Inject constructor(
         runCatching {
             Timber.i("UPnP select: mint token for track=$trackId, route=${route.name}")
             val token = castApi.streamToken(StreamTokenRequest(trackId = trackId))
-            Timber.i("UPnP select: SetAVTransportURI to ${token.url}")
-            transport.setAVTransportURI(token.url)
+            Timber.i("UPnP select: SetAVTransportURI to ${token.url} mime=${token.mime}")
+            transport.setAVTransportURIWithMetadata(
+                uri = token.url,
+                mime = token.mime,
+                title = token.title,
+            )
             Timber.i("UPnP select: Play")
             transport.play()
             playerController.pause()

android/app/src/main/java/com/fabledsword/minstrel/player/output/upnp/AVTransportClient.kt

@@ -26,6 +26,47 @@ class AVTransportClient(
         )
     }
 
+    /**
+     * Convenience overload that builds DIDL-Lite metadata around [uri]
+     * + [mime] + [title] and forwards to [setAVTransportURI]. Sonos
+     * rejects empty DIDL with vendor error 1023; this constructs the
+     * minimal-but-Sonos-acceptable shape:
+     *   <DIDL-Lite>
+     *     <item id="0" parentID="-1" restricted="1">
+     *       <dc:title>...</dc:title>
+     *       <upnp:class>object.item.audioItem.musicTrack</upnp:class>
+     *       <res protocolInfo="http-get:*:<mime>:*">...</res>
+     *     </item>
+     *   </DIDL-Lite>
+     * Generic UPnP renderers tolerate this shape too — there's no
+     * downside to always sending it. Title falls back to "Minstrel"
+     * when the caller doesn't supply one.
+     */
+    suspend fun setAVTransportURIWithMetadata(uri: String, mime: String, title: String) {
+        val safeTitle = title.ifBlank { "Minstrel" }
+        val didl = buildString {
+            append("<DIDL-Lite ")
+            append("xmlns=\"urn:schemas-upnp-org:metadata-1-0/DIDL-Lite/\" ")
+            append("xmlns:dc=\"http://purl.org/dc/elements/1.1/\" ")
+            append("xmlns:upnp=\"urn:schemas-upnp-org:metadata-1-0/upnp/\">")
+            append("<item id=\"0\" parentID=\"-1\" restricted=\"1\">")
+            append("<dc:title>").append(xmlEscape(safeTitle)).append("</dc:title>")
+            append("<upnp:class>object.item.audioItem.musicTrack</upnp:class>")
+            append("<res protocolInfo=\"http-get:*:").append(xmlEscape(mime))
+            append(":*\">").append(xmlEscape(uri)).append("</res>")
+            append("</item>")
+            append("</DIDL-Lite>")
+        }
+        setAVTransportURI(uri, didl)
+    }
+
+    private fun xmlEscape(v: String): String = v
+        .replace("&", "&amp;")
+        .replace("<", "&lt;")
+        .replace(">", "&gt;")
+        .replace("\"", "&quot;")
+        .replace("'", "&apos;")
+
     suspend fun play() {
         soap.call(
             controlUrl = controlUrl,

internal/api/cast_token.go

@@ -3,9 +3,11 @@ package api
 import (
 	"net/http"
 	"strconv"
+	"strings"
 	"time"
 
 	"git.fabledsword.com/bvandeusen/minstrel/internal/apierror"
+	"git.fabledsword.com/bvandeusen/minstrel/internal/db/dbq"
 )
 
 const (
@@ -23,6 +25,36 @@ type castTokenResponse struct {
 	Token string `json:"token"`
 	Exp   int64  `json:"exp"`
 	URL   string `json:"url"`
+	// MIME and Title let the client build proper DIDL-Lite metadata for
+	// SetAVTransportURI. Sonos rejects empty DIDL with vendor error 1023;
+	// passing back the track's MIME + title here lets the client populate
+	// `<res protocolInfo>` and `<dc:title>` without a follow-up round trip.
+	MIME  string `json:"mime"`
+	Title string `json:"title"`
+}
+
+// mimeForFormat maps the tracks.file_format column to an HTTP audio
+// MIME type. Sonos requires the protocolInfo MIME on DIDL-Lite to match
+// what the URL actually serves; "audio/*" wildcard is silently rejected.
+// Unknown formats fall back to audio/mpeg — most Sonos firmware probes
+// the URL anyway and recovers from a small MIME mismatch.
+func mimeForFormat(format string) string {
+	switch strings.ToLower(strings.TrimSpace(format)) {
+	case "mp3", "mpeg":
+		return "audio/mpeg"
+	case "flac":
+		return "audio/flac"
+	case "aac", "m4a", "mp4":
+		return "audio/mp4"
+	case "ogg", "vorbis":
+		return "audio/ogg"
+	case "opus":
+		return "audio/opus"
+	case "wav", "wave":
+		return "audio/wav"
+	default:
+		return "audio/mpeg"
+	}
 }
 
 // handleCastStreamToken issues a short-lived HMAC stream token for the
@@ -52,6 +84,14 @@ func (h *handlers) handleCastStreamToken(w http.ResponseWriter, r *http.Request)
 		writeErr(w, apierror.BadRequest("invalid_track_id", "trackId must be a UUID"))
 		return
 	}
+	// Track lookup for the DIDL-Lite metadata the client builds for
+	// SetAVTransportURI. A missing track is a 404 — there's nothing to
+	// cast in that case.
+	track, err := dbq.New(h.pool).GetTrackByID(r.Context(), trackUUID)
+	if err != nil {
+		writeErr(w, apierror.NotFound("track"))
+		return
+	}
 	expSec := clampExpSeconds(req.ExpSeconds)
 	exp := time.Now().Add(time.Duration(expSec) * time.Second).Unix()
 	token := SignStreamToken(h.streamSecret, req.TrackID, exp)
@@ -77,7 +117,13 @@ func (h *handlers) handleCastStreamToken(w http.ResponseWriter, r *http.Request)
 	url := scheme + "://" + host + "/api/tracks/" + req.TrackID +
 		"/stream?token=" + token + "&exp=" + strconv.FormatInt(exp, 10)
 
-	writeJSON(w, http.StatusOK, castTokenResponse{Token: token, Exp: exp, URL: url})
+	writeJSON(w, http.StatusOK, castTokenResponse{
+		Token: token,
+		Exp:   exp,
+		URL:   url,
+		MIME:  mimeForFormat(track.FileFormat),
+		Title: track.Title,
+	})
 }
 
 // clampExpSeconds applies the [60, 86400] window with a 6h default for