C语言str系列库函数之strspn()、strcspn()和strpbrk()

在vc++8.0的函数库中，strtok()函数的实现采用了bit map，随后继续查阅strspn()、strcspn()和strpbrk()这三个函数源码，发现也是采用的bit map来减少存储空间，并公用了部分代码，因此写篇博客简单记录一下。

首先来看着三个函数的原型：

//return length of prefix of string consisting of characters in control
size_t __cdecl strspn (const char * string, const char * control);

//return length of prefix of string consisring of characters not in control
size_t __cdecl strcspn (const char * string, const char * control);

//return pointer to first occurrence in string of any charactrers of control,
//or NULL if nor present
char * __cdecl strpbrk (const char * string, const char * control);

XX\Microsoft Visual Studio 8\VC\crt\src这个文件夹下有这样三个文件：strcspn.c、strpbrk.c和strspn.c。

strcspn.c：

#define SSTRCSPN      //定义的宏
#include "strspn.c"

strpbrk.c：

#define SSTRPBRK      //定义的宏
#include "strspn.c"

关键的实现在strspn.c中：

/* Determine which routine we're compiling for (default to STRSPN) */
#define _STRSPN         1
#define _STRCSPN        2
#define _STRPBRK        3

#if defined (SSTRCSPN)
#define ROUTINE _STRCSPN
#elif defined (SSTRPBRK)
#define ROUTINE _STRPBRK
#else  
#define ROUTINE _STRSPN	   //默认
#endif  

/* Routine prototype */
#if ROUTINE == _STRSPN
size_t __cdecl strspn (
#elif ROUTINE == _STRCSPN
size_t __cdecl strcspn (
#else  /* ROUTINE == _STRCSPN */
char * __cdecl strpbrk (
#endif  /* ROUTINE == _STRCSPN */
				const char * string,
				const char * control
			)
{
	const unsigned char *str = string;
	const unsigned char *ctrl = control;

	unsigned char map[32];
	int count;

	/* Clear out bit map */
	for (count=0; count<32; count++)
		map[count] = 0;

	/* Set bits in control map */
	while (*ctrl)		  //跟strtok()函数中的使用方式是一样的
	{
		map[*ctrl >> 3] |= (1 << (*ctrl & 7));
		ctrl++;
	}

#if ROUTINE == _STRSPN	//	strspn()

	/* 1st char NOT in control map stops search */
	if (*str)
	{
		count=0;
		while (map[*str >> 3] & (1 << (*str & 7)))
		{
			count++;
			str++;
		}
		return(count);
	}
	return(0);

#elif ROUTINE == _STRCSPN	//strcspn()

	/* 1st char in control map stops search */
	count=0;
	map[0] |= 1;    /* null chars not considered */
	while (!(map[*str >> 3] & (1 << (*str & 7))))
	{
		count++;
		str++;
	}
	return(count);

#else  /* ROUTINE == _STRCSPN */  //strpbrk()

	/* 1st char in control map stops search */
	while (*str)
	{
		if (map[*str >> 3] & (1 << (*str & 7)))
			return((char *)str);
		str++;
	}
	return(NULL);

#endif  /* ROUTINE == _STRCSPN */

}

理解了strtok()的源码，这三个函数的理解就太容易了，主要是位图的使用，C++中提供了专门的标准类型bitset，巧妙使用，在很多算法中能节省空间，提高效率。