AP_RangeFinder: Notify different alerts

[muramura]

Multiple alert notifications.
Currently notifies the same alerts as last time.
It would be better to notify alerts that are not originally notified.

[peterbarker]

This looks like an important change.

How was this tested?

[peterbarker]

I've pushed up an additional patch here which saves a few bytes and makes things a little clearer (IMO).

I've tested this in SITL using the simulated LR-D1 and this patch:

diff --git a/libraries/AP_RangeFinder/AP_RangeFinder_Ainstein_LR_D1.cpp b/libraries/AP_RangeFinder/AP_RangeFinder_Ainstein_LR_D1.cpp
index 775a004268a..214cc3778ad 100644
--- a/libraries/AP_RangeFinder/AP_RangeFinder_Ainstein_LR_D1.cpp
+++ b/libraries/AP_RangeFinder/AP_RangeFinder_Ainstein_LR_D1.cpp
@@ -67,10 +67,17 @@ bool AP_RangeFinder_Ainstein_LR_D1::get_reading(float &reading_m)
             continue;
         }
 
-        const uint8_t malfunction_alert = buffer[1];
+        uint8_t malfunction_alert = buffer[1];
         reading_m = UINT16_VALUE(buffer[3], buffer[4]) * 0.01;
         const uint8_t snr = buffer[5];
 
+        const uint32_t nowx_ms = AP_HAL::millis();
+        static uint32_t last_error_ms = 0;
+        if (nowx_ms - last_error_ms > 5000) {
+            last_error_ms = nowx_ms;
+            malfunction_alert = 0x1;
+        }
+

[peterbarker]

The net effect of these patches is that if the device is currently reporting a temperature error and then you get a voltage error you will not be notified of the temperature error again. On master you will be warned of the temperature error twice and the voltage error once.