fix(gateway): delete partial message after fallback send on flood control

When Telegram flood control triggers 3+ consecutive edit failures, the
stream consumer enters fallback mode and sends the complete response as
a new message. This leaves the user seeing two messages: a frozen
partial (with cursor) and the full duplicate.

After the fallback chunks are sent successfully, delete the original
partial message so the user only sees one complete response. The delete
is best-effort — if it fails (e.g. flood still active, missing
permissions), the full answer is still delivered.

Fixes #16668

gateway/stream_consumer.py

@@ -640,6 +640,7 @@ class GatewayStreamConsumer:
         safe_limit = max(500, raw_limit - 100)
         chunks = self._split_text_chunks(continuation, safe_limit)
 
+        stale_message_id = self._message_id  # partial message to clean up
         last_message_id: Optional[str] = None
         last_successful_chunk = ""
         sent_any_chunk = False
@@ -687,6 +688,16 @@ class GatewayStreamConsumer:
             # so any stale tool-progress bubble gets closed off.
             self._notify_new_message()
 
+        # Remove the frozen partial message so the user only sees the
+        # complete fallback response.  Best-effort — if the delete fails
+        # (e.g. flood control still active, or bot lacks permission), the
+        # partial message remains but at least the full answer was delivered.
+        if stale_message_id and stale_message_id != last_message_id:
+            try:
+                await self.adapter.delete_message(self.chat_id, stale_message_id)
+            except Exception:
+                pass
+
         self._message_id = last_message_id
         self._already_sent = True
         self._final_response_sent = True