c - char splitting and parsing -

so i'm attempting setup function correctly parse following type of input (note input jibberish try , illustrate example)

"./test script";ls -a -o;mkdir "te>st";ls > somefile.txt 

where each command separated ';' , each argument separated whitespace ' ' except in cases wrapped in "" in case should treated literal or whole. ie output want

cmd : "./test script"

cmd : ls args[2] {-a, -o}

cmd : mkdir args[1] { "te>st" }

cmd : ls args[2] {>, somefile.txt}

i've tried splitting via ; first, via ' ' first example fails (wrapped in "" should considered whole), i'm having trouble c i'm not familiar language, help? have far

    // commands split ;      char *cmdsplittoken = strtok(srcpointer, ";");      // store commands seperately can deal them 1 one     while(cmdsplittoken != null) {               cmds[cmdcount++] = cmdsplittoken;         cmdsplittoken = strtok(null, ";");     }      // loop on commands , gather arguments     for(int = 0; < cmdcount; i++) {              // args split ' '         char *argsplittoken = strtok(cmds[i], " ");          int argcount = 0;         while(argsplittoken != null) {                   printf("arg %s\n", argsplittoken);             argcount++;             argsplittoken = strtok(null, " ");         }     } 

roll own strtok , check quotes in there. (your example string doesn't contain ';' inside quotes, perhaps misunderstanding entire issue :)

anyway, here take on rough version of strtok works similar, except accepts 1 single token character instead of string (but if necessary, that's added) , meta-parsing on following:

• strings start " matched closing "
• strings start ' matched closing '
• any individual character can escaped prepending \

an unmatched " , ' match end of string.

#include <stdio.h>  char *get_token (char *input_str, char separator) {     static char *last_pos;      if (input_str)         last_pos = input_str;     else         input_str = last_pos;      if (last_pos && *last_pos)     {         while (*last_pos)         {             if (*last_pos == separator)             {                 *last_pos = 0;                 last_pos++;                 return input_str;             }             if (*last_pos == '\"')             {                 last_pos++;                 while (*last_pos && *last_pos != '\"')                     last_pos++;             } else             if (*last_pos == '\'')             {                 last_pos++;                 while (*last_pos && *last_pos != '\'')                     last_pos++;             } else             if (*last_pos == '\\' && last_pos[1])             {                 last_pos++;             }             last_pos++;         }         return input_str;     }     return null; }  void main (void) {     char str[] = "\"./test; script\";ls -a\\;b -o;mkdir \"te>st\";ls > 'some;file.txt'";     char *cmdsplittoken = get_token (str, ';');      while (cmdsplittoken != null)     {         printf("arg %s\n", cmdsplittoken);         cmdsplittoken = get_token (null, ';');     } } 

this fixes first half of command parsing. second part can handled same routine, or -- understand -- bog standard strtok.

by way, static char inside routine makes not re-entrant -- not use alternating strings. (possibly knew that, because avoid in own code.)