Cleanup session/

core/src/session/session.cpp

@@ -148,7 +148,6 @@ void SessionModel::apply_channel_event(const voicecat::v1::ChannelEvent& ev) {
 
     } else if (ev.kind() == Kind::DELETED) {
         // deleted_id, not channel().id() — the proto leaves `channel` unset for deletes
-        // (docs/protocol.md, core/proto/voicecat.proto's ChannelEvent).
         uint32_t cid = ev.deleted_id();
         channels_.erase(std::remove_if(channels_.begin(), channels_.end(),
                                        [cid](const Channel& x) { return x.id == cid; }),

core/src/session/session.h

@@ -1,5 +1,5 @@
 /*
- * session/session.h — client-side mirror of the server's channel/user state.
+ * session/session.h: client-side mirror of the server's channel/user state.
  *
  * Populated from ServerStateSnapshot (full snapshot) and incremental UserEvent /
  * ChannelEvent messages. Not thread-safe — always called from io_thread_.
@@ -24,7 +24,7 @@ struct Channel {
     bool        password_protected{false};
     uint32_t    max_users{0};
     uint32_t    sort_order{0};
-    // Authoritative channel AudioConfig (docs/voice.md §3) — mirrors
+    // Authoritative channel AudioConfig  mirrors
     // voicecat::v1::AudioConfig field-for-field, same flat shape as Stream below. Populated
     // from ServerStateSnapshot / ChannelEvent so the channel-edit dialog can read back the
     // current config (the vc_channel read struct now carries these too).
@@ -45,7 +45,7 @@ struct Stream {
     uint32_t    ssrc{0};
     int         kind{0};
     std::string label;
-    // Full effective AudioConfig (docs/protocol.md §5), as broadcast by the server in
+    // Full effective AudioConfig as broadcast by the server in
     // StreamInfo.audio — mirrors voicecat::v1::AudioConfig field-for-field so per-channel
     // tuning (mono/stereo, bitrate, FEC/DTX, application) is observable client-side, not just
     // sample_rate/frame_ms.