[mbed-os-6.6.0] minimal-printf incorrect output

[caoddx]

Description of defect

Floating point number is not printed correctly.

test code

  printf("123.456: %f\n", 123.456);
  printf("0.456: %f\n", 0.456);
  printf("-0.456: %f\n", -0.456);
  printf("123.0: %f\n", 123.0);

  // output:
  // 123.456: 123.000123
  // 0.456: 0.000000
  // -0.456: -0.000000
  // 123.0: 123.000123

mbed_app.json

{
    "target_overrides": {
        "*": {
            "target.printf_lib": "minimal-printf",
            "platform.minimal-printf-enable-floating-point": true,
            "platform.minimal-printf-set-floating-point-max-decimals": 6,
            "platform.minimal-printf-enable-64-bit": true
        }
    }
}

It seems that the variable precision in the file mbed_printf_implementation.c is not used.

Target(s) affected by this defect ?

should be all

Toolchain(s) (name and version) displaying this defect ?

GCC_ARM

What version of Mbed-os are you using (tag or sha) ?

mbed-os-6.6.0

What version(s) of tools are you using. List all that apply (E.g. mbed-cli)

mbed-cli

How is this defect reproduced ?

use the above test code and mbed_app.json

[ciarmcom]

Thank you for raising this detailed GitHub issue. I am now notifying our internal issue triagers.
Internal Jira reference: https://jira.arm.com/browse/IOTOSM-3071

[0xc0170]

This PR might have caused this #13548. Have you tested 6.5 version if it works as expected?

cc @EmbedEngineer

[EmbedEngineer]

@0xc0170 yes this has been tested by the green tea tests and there it works. Just take a look at the appropriate file.

[helpfulchicken]

I have come across this problem myself and found the fix.
This line:

mbed-os/platform/source/minimal-printf/mbed_printf_implementation.c

Line 310 in e917282

MBED_UNSIGNED_STORAGE decimal = value;

should be
MBED_UNSIGNED_STORAGE decimal = value * precision;

Otherwise the assignment truncates all decimal information, which completely spoils the intended behaviour.

[caoddx]

@helpfulchicken
Have you tested my test case?

Here's my solution:

...
/* get decimal part */
double precision = 1.0;

for (size_t index = 0; index < dec_precision; index++) {
    precision *= 10;
}

value -= integer; // <---------------- add code
/* convert to positive number */
if (value < 0.0) {
    value *= -1.0;
}

value *= precision; // <---------------- add code
MBED_UNSIGNED_STORAGE decimal = value;

/* round up or down */
value -= decimal;

if (!((value > -0.5) && (value < 0.5))) {
    decimal++;
}
...

[0xc0170]

@caoddx please create a new PR with a fix (also a test case to have a failure fixed).

[helpfulchicken]

I have come across this problem myself and found the fix.
This line:

mbed-os/platform/source/minimal-printf/mbed_printf_implementation.c

Line 310 in e917282

MBED_UNSIGNED_STORAGE decimal = value;

should be
MBED_UNSIGNED_STORAGE decimal = value * precision;

Otherwise the assignment truncates all decimal information, which completely spoils the intended behaviour.

I have realised I was mistaken in my fix.
It should be
MBED_UNSIGNED_STORAGE decimal = (value - integer) * precision;

@caoddx It looks like your fix should achieve the same result.

[LDong-Arm]

@helpfulchicken @caoddx Please create a PR with your fix, and we'll review it and get it in. Thanks!

@evedon FYI