utils.cpp

   1 /***************************************************************************
   2                      utils.cpp  -  sytem dependent utilities
   3                              -------------------
   4     begin                : lun ott 24 2005
   5     copyright            : (C) 2005-2007 by Maurizio Monge
   6     email                : monge@linuz.sns.it
   7  ***************************************************************************/
   8
   9 /***************************************************************************
  10  *                                                                         *
  11  *   This program is free software; you can redistribute it and/or modify  *
  12  *   it under the terms of the GNU General Public License as published by  *
  13  *   the Free Software Foundation; either version 2 of the License, or     *
  14  *   (at your option) any later version.                                   *
  15  *                                                                         *
  16  ***************************************************************************/
  17
  18 #include "utils.h"
  19 #include <stdlib.h>
  20 #include <string.h>
  21
  22 char **get_tokenized(int* numtokens, bool accept_quotes)
  23 {
  24     return fget_tokenized(stdin, numtokens, accept_quotes);
  25 }
  26
  27 char **get_tokenized(const char *whitespaces, int* numtokens, bool accept_quotes)
  28 {
  29     return fget_tokenized(stdin, whitespaces, numtokens, accept_quotes);
  30 }
  31
  32 char **fget_tokenized(FILE* f, int* numtokens, bool accept_quotes)
  33 {
  34     return fget_tokenized(f, " \v\t\r\n", numtokens, accept_quotes);
  35 }
  36
  37 char **fget_tokenized(FILE* f, const char *whitespaces, int* numtokens, bool accept_quotes)
  38 {
  39     char str[4096];
  40     if(!fgets(str, 4096, f))
  41         return NULL;
  42     return tokenize(str, whitespaces, numtokens, accept_quotes);
  43 }
  44
  45 char **tokenize(const char *str, int* numtokens, bool accept_quotes)
  46 {
  47     return tokenize(str, " \v\t\r\n", numtokens, accept_quotes);
  48 }
  49
  50 char **tokenize(const char *str, const char *whitespaces, int* numtokens, bool accept_quotes)
  51 {
  52     int len = strlen(str);
  53     const char *ptr = str;
  54     int ntoks = 0;
  55
  56     while(1)
  57     {
  58         ptr += strspn(ptr, whitespaces);
  59         int s = strcspn(ptr, whitespaces);
  60
  61         if(!s)
  62             break;
  63         ptr += s;
  64         ntoks++;
  65     }
  66
  67     //printf("ntoks = %d\n", ntoks);
  68
  69     void* block = malloc((ntoks+1)*sizeof(char*) + len+1);
  70     char** retv = (char**)block;
  71     char* buf = (char*)block + (ntoks+1)*sizeof(char*);
  72
  73     ptr = str;
  74     ntoks = 0;
  75
  76     while(1)
  77     {
  78         ptr += strspn(ptr, whitespaces);
  79         //printf("-> \"%s\"\n", ptr);
  80         if(!*ptr)
  81             break;
  82         retv[ntoks++] = buf;
  83
  84         bool quoted = false;
  85         while( *ptr && (quoted || (strchr(whitespaces, *ptr) == NULL) ) )
  86         {
  87             if(accept_quotes && *ptr == '"')
  88             {
  89                 quoted = !quoted;
  90                 ptr++;
  91                 continue;
  92             }
  93
  94             if(quoted && *ptr == '\\')
  95             {
  96                 ptr++;
  97                 switch(*ptr)
  98                 {
  99                     case '"':
 100                         *buf++ = '"';
 101                         ptr++;
 102                         break;
 103                     case '\\':
 104                         *buf++ = '\\';
 105                         ptr++;
 106                         break;
 107                     case '\0':
 108                         *buf++ = '\\';
 109                         break;
 110                     default:
 111                         *buf++ = *ptr++;
 112                         break;
 113                 }
 114                 continue;
 115             }
 116
 117             *buf++ = *ptr++;
 118         }
 119         *buf++ = '\0';
 120         if(quoted)
 121             printf("Error! Missing terminating '\"' while parsing!\n");
 122         if(!*ptr)
 123             break;
 124     }
 125     if(numtokens)
 126         *numtokens = ntoks;
 127     retv[ntoks++] = NULL;
 128     return retv;
 129 }