Having a hard time printing floats that are less than 1 with many digits behind the decimal point using NRF_LOG_FLOAT.

[Martin Sivertsen]

Here is the code I'm using to test it out:

static float number   = 0.5678905;

NRF_LOG_INFO("Float number: "NRF_LOG_FLOAT_MARKER"\n\r",  NRF_LOG_FLOAT(number));

And this is what is printed out:

:INFO:Float number: 0.56

The float is reduced to only two decimal points precision.

If I try changing the macro from

#define NRF_LOG_FLOAT_MARKER "%d.%02d"

to

#define NRF_LOG_FLOAT_MARKER "%10d.%08d"

It prints this:

:INFO:Float number:          0.00000056

So, changing the number behind the decimal point in the macro only moves the value, and fills the remaining space with 0's (or empty spaces if I leave out the zero from the macro definition)

Does anyone know how I can remedy this?

[Martin Sivertsen]

I figured it out.

I had to change the factor in this macro as well, from

#define NRF_LOG_FLOAT(val) (int32_t)(val),                                 \
                       (int32_t)(((val > 0) ? (val) - (int32_t)(val)       \
                                            : (int32_t)(val) - (val))*100)

to

#define NRF_LOG_FLOAT(val) (int32_t)(val),                                      \
                       (int32_t)(((val > 0) ? (val) - (int32_t)(val)            \
                                            : (int32_t)(val) - (val))*100000000)

to give the float the correct precision.

Now it prints:

:INFO:Float number:          0.56789052

[Turbo J]

Good find. Now try to print -0.5678905with these macros...

[Martin Sivertsen]

I saw the post about not being able to print negative floats between 0 and -1. Luckily I don't have to print negative floats between 0 and -1 right now, but it is an irritating bug for sure.