firmware: clear panel to Loading screen on new stream connection

The panel used to hold its last framebuffer during the connect→first-frame gap,
flashing a stale old frame before video appeared — more visible now that the
Scrypted side no longer sends a bridging snapshot. The wake path already paints
the Loading screen, but state_machine_set(AWAKE) is a no-op when the device is
already awake (e.g. a stream restart), so the clear didn't happen there.

Tie the clear to the stream connection instead: recv-task paints
local_screens_show_loading() once per new conn_id, guarded by the decoder lock
so the RGB565 present can't race a trailing decode-task paint, and gated on
AWAKE so it never draws on a sleeping panel.

1 files changed, 15 insertions, 0 deletions

diff --git a/main/stream_server.c b/main/stream_server.c
index 24646cb..2bc58ac 100644
--- a/main/stream_server.c
+++ b/main/stream_server.c
@@ -14,6 +14,7 @@
 
 #include "display.h"
 #include "jpeg_decoder.h"
+#include "local_screens.h"
 #include "state_machine.h"
 #include "viewport_state.h"
 
@@ -216,6 +217,20 @@ static void handle_client_recv(int fd, const char *peer)
     conn_id = ++s_conn_id;
     portEXIT_CRITICAL(&s_slot_mux);
 
+    // Clear any stale prior frame to the "Loading..." screen at the start of
+    // every stream session, so the panel doesn't sit on an old image during
+    // the connect→first-frame gap (several seconds on a cold prebuffer; the
+    // Scrypted side no longer sends a bridging snapshot). The wake path also
+    // shows this, but a stream can start while already AWAKE (e.g. a restart)
+    // where the state-machine wake is a no-op — this covers that case too.
+    // Guard with the decoder lock so the RGB565 present can't race a trailing
+    // decode-task paint from the previous connection; skip if not awake.
+    if (state_machine_current() == VIEWPORT_STATE_AWAKE &&
+        jpeg_decoder_try_lock(500)) {
+        local_screens_show_loading();
+        jpeg_decoder_unlock();
+    }
+
     while (1) {
         uint8_t first4[4];
         if (read_n(fd, first4, 4) != ESP_OK) {