topic Re: PRINTF support for Scientific Notation, "e" specifier in KSDK 2.0 in Kinetis Software Development Kit

PRINTF support for Scientific Notation, "e" specifier in KSDK 2.0

marks — Fri, 26 Aug 2016 19:02:17 GMT

Hello,

How do you get the "e" specifier to work with PRINTF in KSDK 2.0? Using IAR.

For example, this code:

#include "fsl_debug_console.h"

#define PRINTF DbgConsole_Printf

float a_float_number = 123456.78;
PRINTF(" a_float_number = %e\n", a_float_number);
PRINTF(" a_float_number = %f\n", a_float_number);

generates this output:

a_float_number = e
a_float_number = 123456.781250

The same code works as expected in a KDSK 1.3 project with lower case printf. This is the output with KDSK 1.3:

a_float_number = 1.234568e+05
a_float_number = 123456.781250

I put #include <stdio.h> in the file but this made no difference.

Maybe it is just not supported. The e specifere is not listed in the Kinetis SDK v.2.0 API Reference Manual in section 12.1.2 Advanced Feature. However, the e specifier is mentioned when discussing the # flag "Used with e, E and f, it forces the written output to contain a decimal point even if no digits would follow."

If e is not supported maybe it could be added next rev.

Any help would be greatly appreciated!

Thanks,

Mark

Re: PRINTF support for Scientific Notation, "e" specifier in KSDK 2.0

egoodii — Fri, 26 Aug 2016 19:09:44 GMT

ASSUMING you are using the IAR-library (and not an explicit printf handler routine in your project) what size 'printf formatter' have you got selected on 'project-options-general_options-library_options' tab --- tiny, small, large, or full? Each 'bigger' version gives more runtime formatting options (at the expense of size and speed, of course).

I am not personally familiar with "fsl_debug_console.h", but if it were ME I would let the IAR tool-library support true 'printf', and do the proper tie from their output as __write() to 'whatever' console stream this is ending up on.

Re: PRINTF support for Scientific Notation, "e" specifier in KSDK 2.0

DavidS — Mon, 29 Aug 2016 16:16:04 GMT

Hi Mark,

Looking into the source code for PRINTF the "%e" format is not supported. The formats I see in fsl_debug_console.c are:

if ((c == 'd') || (c == 'i') || (c == 'f') || (c == 'F') || (c == 'x') || (c == 'X') || (c == 'o') || (c == 'b') || (c == 'p') || (c == 'u'))

I took KSDK_v2 hellow_world example for FRDM-K64F and made the following changes to create a workaround for you:

#include <stdio.h> //DES test for "e" format support in sprintf

#if 1 //DES 1=change output format, 0=default code
float a_float_number = 123456.78;
char my_string_buffer[60];

sprintf(my_string_buffer, "%e", a_float_number);
PRINTF(" a_float_number = %s\n\r", my_string_buffer);
PRINTF(" a_float_number = %e\n\r", a_float_number);
PRINTF(" a_float_number = %f\n\r", a_float_number);
#endif

Terminal Output:

a_float_number = 1.234568e+05 <-- This used sprintf then PRINTF %s to get desired result
a_float_number = e
a_float_number = 123456.781250

Source hellow_world.c attached.

Regards,

David

Re: PRINTF support for Scientific Notation, "e" specifier in KSDK 2.0

marks — Mon, 29 Aug 2016 21:29:42 GMT

Hi David, Tried your method first (before Earls above) and it did not work first pass. Then changed project-options-general_options-library_options' tab --- to full and your solution, as well as the lower case printf, now work!

Thank you for the help!!

Best regards,

Mark