With the following patch:

--- a/ArduPlane/ArduPlane.cpp
+++ b/ArduPlane/ArduPlane.cpp
@@ -286,6 +286,9 @@ extern AP_IOMCU iomcu;
 
 void Plane::one_second_loop()
 {
+
+    gcs().send_text(MAV_SEVERITY_INFO, "SCR=%u MODE=%u CMDH=%u", LOG_SCRIPTING_MSG, LOG_MODE_MSG, LOG_CMDH_MSG);
+
     // make it possible to change control channel ordering at runtime
     set_control_channels();

I get the following output:

AP: SCR=197 MODE=112 CMDH=16

Which means we have 57 ids "after the break" - available for logger.Log_write(...) and logger.LogWriteStreaming(...) to use. We don't have that many calls - and if anyone's using more in scripting then they're almost certainly doing something very wrong...

This also introduces a gap before FMT - before this patch MODE was 127. That's very silly as anybody adding a logging message to e.g. NavEKF3 would get a compilation failure as it would push MODE past FMT!

Also adds a comment to warn people of the 32-id limit. Plane's the largest user of vehicle-specific IDs with 17 defined.

Also corrected a bad sanity check.

Tidied the vehicle definitions up to remove absolute setting of specific IDs (not needed), and to make enumerations in the two vehicles where we were using defines.

Tested only with the autotest suite and a very small amount in SITL.

Edit: I flashed Rover onto CubeBlack and had a look at a produced log. All looked in order.