在MCU上更改默认的printf

MCU中会经常使用printf来打印调试日志，但是编译器默认的printf会使用malloc函数，而且无法自己修改。malloc对于一般MCU平台来说，是有风险的，能不用就不用。

所以移植第三方的printf库函数就是很有必要的事情。

Github下载链接：

修改后的printf版本

本仓库中，查看branch：my-mcu-version。此pirntf库打印都是在栈上进行的，不会使用堆，所以非常安全。

所有修改项已经集中到cJson_Config.h中：

static inline void putchar_(char c)
{

}

/* Support for the decimal notation floating point conversion specifiers (%f, %F), 0 or 1 */
#define PRINTF_SUPPORT_DECIMAL_SPECIFIERS               1

/* Support for the exponential notation floating point conversion specifiers (%e, %g, %E, %G), 0 or 1 */
#define PRINTF_SUPPORT_EXPONENTIAL_SPECIFIERS           1

/* Support for the length write-back specifier (%n), 0 or 1 */
#define PRINTF_SUPPORT_WRITEBACK_SPECIFIER              1

/*
 * Support for the long long integral types (with the ll, z and t length modifiers for specifiers
 * %d,%i,%o,%x,%X,%u, and with the %p specifier).
 * Note: 'L' (long double) is not supported, 0 or 1.
 */
#define PRINTF_SUPPORT_LONG_LONG                        1

/*
 * 'ntoa' conversion buffer size, this must be big enough to hold one converted
 * numeric number including padded zeros (dynamically created on stack)
 */
#define PRINTF_INTEGER_BUFFER_SIZE                      32

/*
 * size of the fixed (on-stack) buffer for printing individual decimal numbers.
 * this must be big enough to hold one converted floating-point value including
 * padded zeros.
*/
#define PRINTF_DECIMAL_BUFFER_SIZE                      32

/* Default precision for the floating point conversion specifiers (the C standard sets this at 6) */
#define PRINTF_DEFAULT_FLOAT_PRECISION                  6

/*
 * According to the C languages standard, printf() and related functions must be able to print any
 * integral number in floating-point notation, regardless of length, when using the %f specifier -
 * possibly hundreds of characters, potentially overflowing your buffers. In this implementation,
 * all values beyond this threshold are switched to exponential notation.
 */
#define PRINTF_MAX_INTEGRAL_DIGITS_FOR_DECIMAL          9

/*
 * The number of terms in a Taylor series expansion of log_10(x) to
 * use for approximation - including the power-zero term (i.e. the
 * value at the point of expansion).
*/
#define PRINTF_LOG10_TAYLOR_TERMS                       4

/*
 * Be extra-safe, and don't assume format specifiers are completed correctly
 * before the format string end.
*/
#define PRINTF_CHECK_FOR_NUL_IN_FORMAT_SPECIFIER        1

/*
 * Support for msvc style integer specifiers (%I), comment it to disable.
*/
//#define PRINTF_SUPPORT_MSVC_STYLE_INTEGER_SPECIFIERS

移植仅需要修改该头文件中putchar_函数：

1.根据编译器修改其inline定义方式

2.将输出字节的方法添加到函数中

至此我们就可以愉快的使用printf_来进行打印了，函数名比默认的printf多了一个下划线。