最近做cache lab 用到了getopt函数, 用man 3 getopt查看了下用法, 做个总结.

描述：getopt函数是用来解析命令行参数的, 以‘-’或‘--’开头的参数为选项元素，除去‘-’或‘--’的选项元素

为选项字符。如果getopt函数被重复调用，则它将会依次返回每个选项元素中的选项字符。

使用getopt函数需要包含以下头文件:

#include

#include

有几个全局变量与getopt函数解析参数有关:

optind: int型, 指示下一个要解析的参数位置，初始时为1.

optarg: char *, 必须接参数的选项元素的参数, 如上面的-nxzz, optarg 就指向"xzz"字符串.

opterr: int 型, 设为0将不打印错误信息.

函数原型为: int getopt(int argc, char * const argv[], const char *optstring);

参数说明: 前两个参数与main函数参数相同, argc保存参数个数，argv[]保存参数数组，第三个参数optstring是你定义

的选项字符组成的字符串, 如"abc",表示该命令有三个选项元素 -a, -b, -c, 选项字符后面如果有一个冒号说

明该选项元素一定有一个参数, 且该参数保存在optarg中, 如"n:t",

表示选项元素n后要接参数, 选项元素t后

不接参数，如 -n xzz -t 或 -nxzz t,有两个冒号说明该选项可接可选参数, 但可选参数不保存在optarg中.

返回值: 如果当前处理的参数为选项元素，且该选项字符在optstring字符串中, 即为你定义的选项, 则返回该

选项字符，如果该选项字符不是你定义的, 那么返回字符'?', 并更新全局变量optind, 指向argc数组中的下一

个参数. 如果当前处理的参数不是选项元素, 则optind偏移向下一个参数, 直到找到第一个选项元素为止, 然后

再按之前描述的操作,如果找不到选项元素, 说明解析结束, 则返回-1.

下面看例子:

#include

#include

int main(int argc, char * const argv[])

{

int opt;

while ((opt = getopt(argc, argv, "nb:o::t")) != -1) {

printf("opt = %c, optarg = %s, optind = %d, argv[%d] = %sn",

opt, optarg, optind, optind, argv[optind]);

}

return 0;

}

假设编译好的可执行文件名为test, test有3个有效参数-n, -b, -t, 其中-n, -t后不接参数, -b后一定要接参数, -o后接可选参数.

# ./test -x -y -z

输出:

./getopt: invalid option -- 'x'

opt = ?, optarg = (null), optind = 2, argv[2] = -y

./getopt: invalid option -- 'y'

opt = ?, optarg = (null), optind = 3, argv[3] = -z

./getopt: invalid option -- 'z'

opt = ?, optarg = (null), optind = 4, argv[4] = (null)

# ./test -n -b xzz -t

opt = n, optarg = (null), optind = 2, argv[2] = -b

opt = b, optarg = xzz, optind = 4, argv[4] = -t

opt = t, optarg = (null), optind = 5, argv[5] = (null)

# ./test -n -bxzz -t

opt = n, optarg = (null), optind = 2, argv[2] = -bxzz

opt = b, optarg = xzz, optind = 3, argv[3] = -t

opt = t, optarg = (null), optind = 4, argv[4] = (null)

# ./test -b -t

opt = b, optarg = -t, optind = 3, argv[3] = (null)

# ./test -t -b

opt = t, optarg = (null), optind = 2, argv[2] = -b

./getopt: option requires an argument -- 'b'

opt = ?, optarg = (null), optind = 3, argv[3] = (null)

# ./test -t a -b xzz

opt = t, optarg = (null), optind = 2, argv[2] = a

opt = b, optarg = xzz, optind = 5, argv[5] = (null)

# ./test -ntb xzz

opt = n, optarg = (null), optind = 1, argv[1] = -ntb

opt = t, optarg = (null), optind = 1, argv[1] = -ntb

opt = b, optarg = xzz, optind = 3, argv[3] = (null)

# ./test -t -o -b xzz

opt = t, optarg = (null), optind = 2, argv[2] = -o

opt = o, optarg = (null), optind = 3, argv[3] = -b

opt = b, optarg = xzz, optind = 5, argv[5] = (null)

# ./test -t -o 10 -b xzz

opt = t, optarg = (null), optind = 2, argv[2] = -o

opt = o, optarg = (null), optind = 3, argv[3] = 10

opt = b, optarg = xzz, optind = 6, argv[6] = (null)

常用用法:

一般是用while循环配合switch语句来使用getopt函数解析参数, 如下:

/*

* 得到参数信息

*/

int getinfo(int argc, char * const argv[], int *ps, int *pE, int *pb, char *trace_name, int *vflag)

{

int opt;

int count_arg = 0;

opterr = 0;

while ((opt = getopt(argc, argv, "vs:E:b:t:")) != -1) {

switch (opt) {

case 'v':

*vflag = 1;

break;

case 's':

++count_arg;

*ps = atoi(optarg);

break;

case 'E':

++count_arg;

*pE = atoi(optarg);

break;

case 'b':

++count_arg;

*pb = atoi(optarg);

break;

case 't':

++count_arg;

strcpy(trace_name, optarg);

break;

default: /* '?' */

Usage();

exit(-1);

break;

}

}

if (count_arg < 4) {

Usage();

exit(-1);

}

return 0;

}