如何用从scanf中获取的参数填充字符指针数组？

Question

我正在尝试用 C 编写一个非常基本的 shell 程序。我面临的问题是试图用从输入中获取的单词填充我的 argv 字符指针数组。当我尝试使用下面的 parse() 函数填充 argv 数组的内容后尝试打印它时，出现分段错误。我知道这意味着我可能正在尝试访问超出范围的 argv 数组的一部分。但是，即使只提供一个参数来填充数组，我仍然会遇到段错误。用于打印 argc 的 printf 调用根据输入返回正确的 argc 值，但带有 *argv[0] 的第二个 printf 调用是导致段错误的调用。我想知道我的错误是我尝试打印 argv 内容的方式，还是因为我试图错误地填充 argv。

编辑：我应该补充一点，getword() 函数接收一行文本并返回由空格分隔的第一个单词，以及许多其他分隔符。如有必要，我可以张贴所有将单词分开的分隔符，但我认为问题不是因为 getword()。

编辑2：在头文件中添加并在main中包含#include语句。

编辑3：在main()下添加getword函数，在p2.h下添加getword.h

这里是p2.h，main中包含的头文件：

#include <stdio.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include "getword.h"
#include <signal.h>

#define MAXITEM 100

getword.h:

#include <stdio.h>
#include <string.h>
#include <strings.h>

#define STORAGE 255

int getword(char *w);

int parse(char *, char *[]);

这里是主要功能：

#include "p2.h"
int main() {
    pid_t pid, child_pid;
    int argc, inputRedirect;
    char *devNull;
    devNull = (char *) malloc(10);
    strcpy(devNull, "/dev/null");
    char *argv[MAXITEM];
    char commandLine[STORAGE];


    for (;;) {
        printf("p2: ");
        scanf("%s", commandLine);
        argc = parse(commandLine, argv);
        printf("argc = %d\n", argc);

        if(argc == 0)
            continue;
        printf("*argv = %s\n", *argv[0]);
        child_pid = fork();
        if (child_pid < 0) {
            printf("Cannot fork! Terminating...");
            exit(1);
        } else if (child_pid == 0) {
            inputRedirect = open(devNull, O_RDONLY);
            dup2(inputRedirect, STDIN_FILENO);
            close(inputRedirect);
            execvp(*argv, argv);
        }
        else {
            for(;;) {
                pid = wait(NULL);
                if(pid == child_pid)
                   break;
            }
            printf("Child's pid is %d\n", child_pid);
        }
    }
    killpg(getpid(), SIGTERM);
    printf("p2 Terminated.\n");
    exit(0);
}

int parse(char *commandLine, char *argv[]) {
    int i, argc = 0;
    char *commandPointer = commandLine;
    while (*commandPointer != '\0') {
        *argv = commandPointer;
        argc++;
        getword(commandPointer);
    }
    *commandPointer = '\0';
    *argv = '\0';
    return argc;
}

getword.c:

#include "getword.h"
#include <stdlib.h>

/*Function Prototypes*/
int tilde(char *p, int i);
int BSFollowedByMetaCharacter(int c, char *w);


int getword(char *w) {

    int c;
    int index = 0;

    /*This while loop removes all leading blanks and whitespace characters
     * The if statement then tests if the first character is a new line or
     *  semicolon metacharacter*/
    while ((c = getchar()) == ' ' || c == '\t' || c == '\n' || c == ';') {
        if (c == '\n' || c == ';') {
            w[index] = '\0';
            return 0;
        }
    }

    /*This if statement calls ungetc() to push whatever character was taken
     * from the input stream in the previous while loop back to the input
     * stream. If EOF was taken from the input stream, ungetc() will return EOF,
     * which will then cause getword() to return -1, signalling that it reached
     * the End Of File. */
    if (ungetc(c, stdin) == EOF)
        return -1;

    /*This if statement deals with some of the "non-special" metacharacters.
     * If one of these metacharacters is pulled from the input stream by getchar(),
     * it is stored in w and null-terminated. getword() then returns the length of
     * the current string stored in w. If getchar() pulls anything besides one of the
     * specified metacharacters from the input stream, it is then returned using ungetc() after
     * the if statement.*/
    if ((c = getchar()) == '<' || c == '>' || c == '|' || c == '&') {
        w[index++] = c;
        int d = getchar();
        if (c == '>' && d == '>')
            w[index++] = d;
        else {
            ungetc(d, stdin);
        }
        w[index] = '\0';
        return index;
    }
    ungetc(c, stdin);

    /*This while statement handles plain text from the input stream, as well as a few 'special'
     * metacharacters. It also ensures that the word scanned is shorter than STORAGE-1 bytes.*/
    while ((c = getchar()) != ' ' && c != '<' && c != '>' && c != '|'
        && c != ';' && c != '&' && c != '\t' && c != '\n' && c != '\0'
        && index <= STORAGE - 1) {
        if (c == '~') {
            int *ip = &index;
            index = tilde(&w[index], *ip);
            continue;
        }/*END IF*/
        else if (c == '\\') {
            int d = c;
            c = getchar();
            if (BSFollowedByMetaCharacter(c, w)) {
                w[index++] = c;
                continue;
            } else {
                w[index++] = d;
            }

        }/*END ELSE IF*/
        w[index] = c;
        index++;
    }/*END WHILE*/

    ungetc(c, stdin);/*This final ungetc() call is used to push any meta characters*/
    w[index] = '\0'; /*used as delimiters back to the input stream, to be retrieved*/
    return index;    /*at the next call of getword().                                      */
}/*END getword()*/

int tilde(char *cp, int i) {
    int *ip;
    ip = &i;
    char *p = cp;
    char *o;
    o = (strcpy(p, getenv("HOME")));
    int offset = strlen(o);
    *ip = *ip + offset;
    return i;
}

int BSFollowedByMetaCharacter(int c, char *w) {
    if (c == '~' || c == '<' || c == '>' || c == '|' || c == ';' || c == '&'
        || c == ' ' || c == '\t' || c == '\\') {
        return 1;
    } else {
        return 0;
    }
}

Answer 1

getword.c 中的函数似乎是正确的。你的问题出在函数parse。

要使用execvp，argv 的内容应该如下（输入：“hello world”）：

argv[0] -> "hello"
argv[1] -> "world"
argv[2] -> NULL

这里，argv 是一个字符指针数组。但是，在parse 函数中，您在这里将argv 视为简单的字符指针：

*argv = commandPointer;

这里：

*argv = '\0';

把你的解析函数改成这样：

int parse(char *commandLine, char *argv[]) {
    int argc = 0;
    char *commandPointer;
    argv[argc++] = commandLine;

    do{
        commandPointer = (char*)malloc(sizeof(char) * STORAGE);
        argv[argc++] = commandPointer;
        getword(commandPointer);
    }while(*commandPointer != '\0');
    argv[argc] = NULL;
    return argc;
}

现在，您应该在 if-else 树之后释放分配的内存，例如：

for(int i = 0; i < argc; i++) free(argv[i]);