`uBit.serial.printf(%d)` with float value triggers wrong output for other arguments

[microbit-carlos]

Floats are not supported in uBit.serial.printf(): https://github.com/lancaster-university/codal-core/blob/master/source/driver-models/Serial.cpp#L393-L473

If we accidentally try to print a float as an integer uBit.serial.printf("%d\n", 1.234); that serial output printed is an integer interpretation of the value, and that's expected 👍.

However, if there are more arguments, it affects their conversion as well and produces erroneous data. In this example, the second %d of the second printf produces 1072938614 instead of 1:

#include "MicroBit.h"

MicroBit uBit;

int main() {
    uBit.init();

    while (true) {
        uBit.serial.printf("%d, %d\n", 1, 1.234);    // <- the first value, an integer is converted correctly
        uBit.serial.printf("%d, %d\n\n", 1.234, 1);  // <- the second value, same integer, gets converted incorrectly
        uBit.sleep(10);
    }
}

The float-as-an-int conversion is also inconsistent, but it only happens sometimes, it's hard to catch. A small extract of the output of this example:

1, 0
-927712936, 1072938614

1, 16380
-927712936, 1072938614

1, 0
-927712936, 1072938614

1, 16380
-927712936, 1072938614

1, 0
-927712936, 1072938614

MICROBIT.hex.zip