Parsing a string (Unix Shell)

Question

I've tried thousand options to parse the set of strings I store inside the buffer, but can't find a good way to store them string by string in the *args[] array of strings. What I want to do afterwards is to run execvp inside the Child process.

Thanks!!

#include <unistd.h>
#include <stdio.h>
#include <sys/types.h>
#include <stdlib.h>


#define MAX_LINE 80

int main(){

char *args[MAX_LINE/2 + 1];
int should_run = 1;
int background = 1;
char buffer[1024];

    while (should_run){
            printf("jaime$ ");
            int i = 0;
            fflush(stdout);
            while (fgets(buffer, 1024, stdin)){
                    sscanf(buffer, "%s", args[i]);
                    i++;
            }

            pid_t pid;
            pid = fork();

            if (pid < 1){
                    fprintf(stderr, "fork() failed");
..........

I don't understand the question. Are you asking how to split and tokenize the input string to parse commands to run with arguments? — Etan Reisner
– Etan Reisner, Commented Dec 26, 2014 at 20:31
You need to read more about the fork system call, you're not checking for the correct return code. — Some programmer dude
– Some programmer dude, Commented Dec 26, 2014 at 20:31
Joachim, what do you suggest then for the fork system call? I'm following all the Operating Systems Concepts book instructions... thanks! — Jaime
– Jaime, Commented Dec 27, 2014 at 11:16
Etan, yes, that's right. I can't find a way to tokenize the input string. Going to follow some of the guys answers and see how it works... — Jaime
– Jaime, Commented Dec 27, 2014 at 11:19

chux · Accepted Answer · 2014-12-27 16:53:47Z

1

Code is not allocating memory to save the stirng

sscanf(buffer, "%s", args[i]);  args[i] is not initialized

Instead allocate needed memory.

Below uses "%n" to note the offset in the string at that point in the scan. " " consumes leading white-space. "%n" stores the scan offset in n1. "%*s" scans just like "%s", but does not store the result. Finally "%n" stores the scan offset in n2. Now code knows, n2-n1, the width in buffer to extract.

 while (fgets(buffer, 1024, stdin)) {
   int n1 = 0;
   int n2 = 0;
   sscanf(buffer, " %n%*s%n", &n1, &n2);
   int length = n2 - n1;
   args[i] = malloc(length + 1);
   if (args[i] == NULL) {
     Handle_OutOfMemory(); // TBD code to cope with OOM.
   } 
   memcpy(args[i], &buffer[n1], length);
   args[i][length] = '\0';
   i++;
 }

 // use args[0] to args[i-1]

 for (int j = 0  j < i; j++) {
   free(args[j]);
   j++;
 }

edited Dec 27, 2014 at 16:53

answered Dec 26, 2014 at 21:03

chux

158k17 gold badges160 silver badges311 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Jaime Over a year ago

Hi Chux, This is really helpful. I have two questions: 1. When I try to compile the program, it tells me that I have an "incompatible declaration of built-in function 'memcpy' [enabled by default]. It also tells me "undefined reference to Handle_OutOfMemory. Any ideas? 2. Could you add comments on lines 4 to 9 explaining exactly what they are doing? Thanks again!

chux Over a year ago

@Jaime Add <string.h> for memcpy(). Handle_OutOfMemory() is a place holder for code you write to cope with the rare event of running out of memory. Maybe print an error message and exit the program? Post updated

Jaime Over a year ago

Thanks Chux! Now I get it. (Can't vote your post since I still don't have 15 of reputation, but will do it as soon as I do :)

yossarian · Accepted Answer · 2014-12-26 20:36:53Z

0

Parsing complex strings using the standard C library isn't going to be easy, nor clean. This is especially true if you're trying to create a shell (which is what it looks like).

I would suggest taking a look at lex and yacc, which can be used to build tokenizers and parsers for complex syntaxes (including shell languages like bash).

Here's a decent introductory tutorial on lex and yacc. If you want to learn by example, you can check out bash/parse.y, although its grammar is fairly convoluted and complex.

answered Dec 26, 2014 at 20:36

yossarian

1,71717 silver badges21 bronze badges

Collectives™ on Stack Overflow

Parsing a string (Unix Shell)

2 Answers 2

3 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related