Issues #39 + #40: gate input until the scene presents; cap the wait-screen idle loop

#39 (field: 'I was able to input before the game screen actually came
up -- fired weapons, moving, could hear the audio'): the simulation
runs during load/wait and PadRIO fed it live input.  Poll now emits
only the analog heartbeat until gBTSceneHasPresented -- keyboard, pad
and joystick all stand down; edge state stays parked so keys held
across the boundary press cleanly at first frame.  BT_BTNTEST and the
panel-click seam stay live (event injection, same as before).

#40 (field: 'GPU usage during waiting screen is sometimes consuming up
to 45-50%'): ExecuteIdle Cleared+Presented uncapped for the whole join
wait.  Now capped at ~40 Hz with a 5 ms sleep on skipped passes (the
overlay's change gate is 12.5 Hz -- nothing visible changes faster).

Regression: 2-node relay round through the wait screen -> load ->
mission (135 in-mission ticks, autofire) -> clean.  Awaiting the
reporter's live confirm on both.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

engine/MUNGA_L4/L4PADRIO.cpp

@@ -403,6 +403,28 @@ void
 	}
 	lastPollMilliseconds = now;
 
+	//
+	// Issue #39 (field 2026-07-23): the simulation runs during load/wait,
+	// so a player could fire, walk and hear it all BEFORE the first frame
+	// presented ("I was able to input before the game screen came up").
+	// Until the scene is live, emit only the analog heartbeat -- keyboard,
+	// pad and joystick reads all stand down (edge state stays parked, so
+	// keys held across the boundary press cleanly at first frame).  The
+	// scripted BT_BTNTEST harness above stays live (it injects through the
+	// event seam, same as the on-screen panel clicks).
+	//
+	{
+		extern Logical gBTSceneHasPresented;
+		if (!gBTSceneHasPresented)
+		{
+			RIOEvent analog;
+			analog.Type = AnalogEvent;
+			analog.Data.Unit = 0;
+			PushEvent(analog);
+			return;
+		}
+	}
+
 	//
 	//-----------------------------------------------------------------
 	// XInput: hot-plug probe every ~3 s, then read the connected pad.

engine/MUNGA_L4/L4VIDEO.cpp

@@ -8917,6 +8917,22 @@ void DPLRenderer::ExecuteIdle()
 
 	if (gBTSceneHasPresented)
 		return;								// game graphics own the screen now
+
+	// Issue #40 (wait screen burned 45-50% GPU, field 2026-07-23): the idle
+	// loop Cleared+Presented at an UNCAPPED rate for the whole join wait.
+	// Cap it at ~40 Hz -- the overlay's own change gate runs at 12.5 Hz, so
+	// nothing visible changes faster than that anyway.  The 5 ms sleep also
+	// keeps the caller's spin off a CPU core.
+	{
+		static unsigned long s_lastIdlePresent = 0;
+		unsigned long now = GetTickCount();
+		if (now - s_lastIdlePresent < 25UL)
+		{
+			Sleep(5);
+			return;
+		}
+		s_lastIdlePresent = now;
+	}
 	{
 		static unsigned long s_lastIdleLog = 0;
 		unsigned long now = GetTickCount();