编译时的 sprintf 格式警告

Question

Variable freq is declared as follows:

void exciteFreqN(float freq, unsigned short N)

then I use the following instruction:

sprintf(debugstr, "Cntr ticks:%d freq:%1.1f\r\n", ctphp, freq);

Format specifier "%1.1f" for freq is clearly float (I think).
However, the compiler warns:

acquisitionXBEE.c: In function 'exciteFreqN':
***acquisitionXBEE.c:8519:5: warning: format '%1.1f' expects type 'double', but argument 4 has type 'float'

Why "%1.1f" expects double ? Shouldn't "f" stand for float?
How can I get rid of that warning?

Answer 1

Why %1.1f expects double? Shouldn't "f" stand for float? How can I get rid of that warning?

"%1.1f" expects a double as specified by the C standard.
In C, double is the default floating-point type for FP ... arguments and constants.

... arguments of type float are converted to double before being passed. sprintf(debugstr, "%1.1f %1.1f %1.1f ", 1.0, 2.0f, freq); should work.

Think of "%f" implying fixed-point format for floating-point, not float .

Compiler is buggy or simply designed that way and therefore non-compliant to C.

A cast may quiet the warning:

sprintf(debugstr, "Cntr ticks:%d freq:%1.1f\r\n", ctphp, (double) freq);

Report bug and/or move to another compiler.

Note: If the non-compliant compiler purposely does not promote float to double when a ... argument, sprintf() may support, as an extension, some flag, like "%$f" to indicate the argument is a float . Check your compiler documents. Be wary then of making such implementation specific code.

---

Warning

Even on a working machine sprintf(debugstr, "Cntr ticks:%d freq:%1.1f\r\n", ctphp, freq); is scary as the buffer may overrun with a large freq like FLT_MAX .

Control width, size and be more informative. What good is a debug message that causes a buffer overflow or is not so informative?

// snprintf,       v----size-----v                      %g
snprintf(debugstr, sizeof debugstr, "Cntr ticks:%d freq:%g\r\n", ctphp, freq);
// ... or pedantically to see all useful precision
snprintf(debugstr, sizeof debugstr, "Cntr ticks:%d freq:%.*g\r\n", 
    ctphp, FLT_DECIMAL_DIG, freq);