Parsing bash command input in C

Question

I'm writing a simple shell in C. Part of this shell uses bash commands that need to be split up into an array of characters, so that I can properly execute the commands. As a simple proof of concept, I wanted to make a C program that would check to see if its a word, or a special character (|, &, >, and <).

Here is an example of how it should look:

input a command: cat hello.txt |wc -l
---output---
cat
hello.txt
|
wc
-l

So far, I've gotten every case to work, for example except for one, and that is when I do a special character like in |wc l where a letter is after the special character. Here is what my current output looks like:

input a command: cat hello.txt |wc -l
---output---
cat
hello.txt
|wc
-l

Below is my code, I've tried adding special cases, removing cases, and nothing seems to work correctly. How should I solve this?

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

#define ALPHABET 1000
#define LESS_THAN 60
#define GREATER_THAN 62
#define AMPERSAND 38
#define PIPE 124
#define WHITESPACE 32
#define NEWLINE 10
#define MYCOPOUT 9999

#define CAPITALS(i) ((i >= 65) && (i <= 90))
#define LOWERS(i) ((i >= 97) && (i <= 122))
#define BACKSLASH(i) i == 47

int classify(int i)
{
    if (CAPITALS(i) || LOWERS(i) || BACKSLASH(i))
        return ALPHABET;
    return i;
}

char isSpecial(int i)
{
    return (i == LESS_THAN) ||(i == GREATER_THAN) || (i == AMPERSAND) || (i == PIPE);
}

char isWhiteSpace(int i)
{
    return (i == WHITESPACE) || (i == NEWLINE);
}

void PRINT_N_NEW(int c, int last)
{
    if (isWhiteSpace(last)) {
        printf("%c",c);
    } else if(isSpecial(last)) {
        printf("\n%c\n",c);
    } else {
        printf("%c\n", c);
    }
    last = c;
}

int main(int argc, char** argv)
{

    int i;
    int last = MYCOPOUT;
    while ( (i = getchar()) != EOF){
        switch(classify(i)){
            case ALPHABET:
                printf("%c",i);
                last = i;
                break;
            case LESS_THAN:
                PRINT_N_NEW(i,last);
                break;
            case GREATER_THAN:
                PRINT_N_NEW(i,last);
                break;
            case AMPERSAND:
                PRINT_N_NEW(i,last);
                break;
            case PIPE:
                PRINT_N_NEW(i,last);
                break;
            case WHITESPACE:
                printf("\n");
                last = i;
                break;
            case NEWLINE:
                printf("\n");
                last = i;
                break;
            default:
                printf("%c",i);
                last = i;
                break;
        }
    }
    return 0;
}

Use ctype.h isspace() instead of isWhiteSpace(int i). And this, #define AMPERSAND 38 is not needed you just nee #define AMPERSAND '&' — Iharob Al Asimi
– Iharob Al Asimi, Commented Feb 20, 2015 at 10:39
Don't name functions with all uppercase. By convention, that's used for macros. — Barmar
– Barmar, Commented Feb 20, 2015 at 10:44
If you want to replicate bash behavior, better look at bash code... But you probably want just something resembling Unix shells? — hyde
– hyde, Commented Feb 20, 2015 at 10:57

Barmar · Accepted Answer · 2015-02-20 10:59:16Z

You're not setting last every time through the loop, only in certain cases of the switch. You should also use PRINT_N_NEW when printing ordinary characters, so that it will test the last character and put a newline before it. Finally, if the character you're printing is a special character, you need to print a newline before it.

There's no point in doing last = c at the end of PRINT_N_NEW. That doesn't set the variable in the caller, it just sets the local copy.

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

#define ALPHABET 1000
#define LESS_THAN 60
#define GREATER_THAN 62
#define AMPERSAND 38
#define PIPE 124
#define WHITESPACE 32
#define NEWLINE 10
#define MYCOPOUT 9999

#define CAPITALS(i) ((i >= 65) && (i <= 90))
#define LOWERS(i) ((i >= 97) && (i <= 122))
#define BACKSLASH(i) i == 47

int classify(int i)
{
    if (CAPITALS(i) || LOWERS(i) || BACKSLASH(i))
        return ALPHABET;
    return i;
}

char isSpecial(int i)
{
    return (i == LESS_THAN) ||(i == GREATER_THAN) || (i == AMPERSAND) || (i == PIPE);
}

char isWhiteSpace(int i)
{
    return (i == WHITESPACE) || (i == NEWLINE);
}

void PRINT_N_NEW(int c, int last)
{
    if (isWhiteSpace(last)) {
        printf("%c",c);
    } else if(isSpecial(last) || isSpecial(c)) {
        printf("\n%c",c);
    } else {
        printf("%c", c);
    }
}

int main(int argc, char** argv)
{

    int i;
    int last = MYCOPOUT;
    while ( (i = getchar()) != EOF){
        switch(classify(i)){
            case ALPHABET:
                PRINT_N_NEW(i, last);
                break;
            case LESS_THAN:
                PRINT_N_NEW(i,last);
                break;
            case GREATER_THAN:
                PRINT_N_NEW(i,last);
                break;
            case AMPERSAND:
                PRINT_N_NEW(i,last);
                break;
            case PIPE:
                PRINT_N_NEW(i,last);
                break;
            case WHITESPACE:
                printf("\n");
                break;
            case NEWLINE:
                printf("\n");
                break;
            default:
                PRINT_N_NEW(i, last);
                break;
        }
        last = i;
    }
    return 0;
}

Karthikeyan.R.S · Accepted Answer · 2015-02-20 10:48:42Z

0

You code is working fine,

if (isWhiteSpace(last)) {
    printf("%c\n",c);

Place the new line in this print statement. Or else place the new line after when pipe is came.

edited Feb 20, 2015 at 10:48

answered Feb 20, 2015 at 10:42

Karthikeyan.R.S

4,0512 gold badges21 silver badges31 bronze badges

1 Comment

Alex Over a year ago

This works somewhat.. however for some reason, if I do echo hello| world it now results in echo, hello|, world (each comma is a new line)

Borja · Accepted Answer · 2015-02-20 11:06:13Z

0

May you should use strtok?

http://www.tutorialspoint.com/c_standard_library/c_function_strtok.htm

try to separate substrings with " " symbol, it might be useful

To solve the problem, you could check if the first char of the token is a symbol

if ( token[0] == "|" ) aux_str = token+1;

It could be like:

enter code here
#define (CMD_SEPARATORS) ("|<>&")


int function interpretCMD(char *s){
    char **args = malloc (sizeof(char*) * MAX_ARGS);


    cmd = strtok (s, " ");

    args = strtok(s);
    while (cmd != NULL){
        args[i] = strtok(s);
    }
/*
    at this point you must have cmd the cmd
    and in args[i] the i argument
*/
}

int function parseCMD(char *s){
    cmd = strtok(s, CMD_SEPARATORS);
    while (cmd!= NULL){
        interpretCMD(token);
        cmd = strtok(s);
    }
}

main blablah

while ( (line = getline()) != EOF){
    parseCMD(line);


}

edited Feb 20, 2015 at 11:06

answered Feb 20, 2015 at 10:42

Borja

213 bronze badges

6 Comments

Barmar Over a year ago

What does this have to do with where he prints newlines while scanning?

Borja Over a year ago

well, the strtok function allows you to get every word in the line readed from input. So you'll be able to interpret as bash like: cmd = first_token, every next token are arguments unless token[0] its in ("|",";","&",">", "<") -> manage that options with a switch to redirect input or launch two commands at time, etc.

Barmar Over a year ago

What if he writes cat foo|wc -l. Checking just token[0] won't find that.

Borja Over a year ago

uuuhmm you're right, well, strtok allow to make the separation on one or more symbols... may be.... strtok (line, " |;&><"); i'm thinking on that please wait

Borja Over a year ago

uhum what if you make a function like ` for (i=0; i<strlen(token);i++) if (isSymbol(token[i])) then blablah()`

|

Collectives™ on Stack Overflow

Parsing bash command input in C

3 Answers 3

Comments

1 Comment

6 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

1 Comment

6 Comments

Your Answer

Sign up or log in

Post as a guest

Related