scan_urlencoded_query.c

   1 /* This software was written by Dirk Engling <erdgeist@erdgeist.org>
   2    It is considered beerware. Prost. Skol. Cheers or whatever.
   3
   4    $id$ */
   5
   6 /* Opentracker */
   7 #include "scan_urlencoded_query.h"
   8
   9 /* Libwofat */
  10 #include "scan.h"
  11
  12 /* System */
  13 #include <string.h>
  14
  15 /* Idea is to do a in place replacement or guarantee at least
  16    strlen( string ) bytes in deststring
  17    watch http://www.ietf.org/rfc/rfc2396.txt
  18          unreserved    = alphanum | mark
  19          mark          = "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")"
  20    we add '%' to the matrix to not stop at encoded chars.
  21    After losing too many requests to being too strict, add the following characters to reserved matrix
  22          relax         = "+" | "," | "/" | ";" | "<" | ">" | ":"
  23 */
  24
  25 /* This matrix holds for each ascii character the information,
  26    whether it is a non-terminating character for on of the three
  27    scan states we are in, that is 'path', 'param' and 'value' from
  28   /path?param=value&param=value, it is encoded in bit 0, 1 and 2
  29   respectively
  30
  31   The top bit of lower nibble indicates, whether this character is
  32   a hard terminator, ie. \0, \n or \s, where the whole scanning
  33   process should terminate
  34   */
  35 static const unsigned char is_unreserved[256] = {
  36   8,0,0,0,0,0,0,0,0,0,8,0,0,8,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  37   8,7,8,8,8,7,0,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,4,7,6,
  38   4,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,8,8,8,8,7,
  39   8,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,8,8,8,7,0,
  40   0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  41   0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  42   0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  43   0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
  44 };
  45
  46 /* Do a fast nibble to hex representation conversion */
  47 static unsigned char fromhex(unsigned char x) {
  48   x -= '0';
  49   if (x <= 9)
  50     return x;
  51   x &= ~0x20;
  52   x -= 'A' - '0';
  53   if (x < 6)
  54     return x + 10;
  55   return 0xff;
  56 }
  57
  58 /* Skip the value of a param=value pair */
  59 void scan_urlencoded_skipvalue(char **string) {
  60   const unsigned char *s = *(const unsigned char **)string;
  61   unsigned char        f;
  62
  63   /* Since we are asked to skip the 'value', we assume to stop at
  64      terminators for a 'value' string position */
  65   while ((f = is_unreserved[*s++]) & SCAN_SEARCHPATH_VALUE)
  66     ;
  67
  68   /* If we stopped at a hard terminator like \0 or \n, make the
  69      next scan_urlencoded_query encounter it again */
  70   if (f & SCAN_SEARCHPATH_TERMINATOR)
  71     --s;
  72
  73   *string = (char *)s;
  74 }
  75
  76 int scan_find_keywords(const ot_keywords *keywords, char **string, SCAN_SEARCHPATH_FLAG flags) {
  77   char   *deststring   = *string;
  78   ssize_t match_length = scan_urlencoded_query(string, deststring, flags);
  79
  80   if (match_length < 0)
  81     return match_length;
  82   if (match_length == 0)
  83     return -3;
  84
  85   while (keywords->key) {
  86     if (!strncmp(keywords->key, deststring, match_length) && !keywords->key[match_length])
  87       return keywords->value;
  88     keywords++;
  89   }
  90
  91   return -3;
  92 }
  93
  94 ssize_t scan_urlencoded_query(char **string, char *deststring, SCAN_SEARCHPATH_FLAG flags) {
  95   const unsigned char *s = *(const unsigned char **)string;
  96   unsigned char       *d = (unsigned char *)deststring;
  97   unsigned char        b, c;
  98
  99   /* This is the main decoding loop.
 100     'flag' determines, which characters are non-terminating in current context
 101     (ie. stop at '=' and '&' if scanning for a 'param'; stop at '?' if scanning for the path )
 102   */
 103   while (is_unreserved[c = *s++] & flags) {
 104
 105     /* When encountering an url escaped character, try to decode */
 106     if (c == '%') {
 107       if ((b = fromhex(*s++)) == 0xff)
 108         return -1;
 109       if ((c = fromhex(*s++)) == 0xff)
 110         return -1;
 111       c |= (b << 4);
 112     }
 113
 114     /* Write (possibly decoded) character to output */
 115     *d++ = c;
 116   }
 117
 118   switch (c) {
 119   case 0:
 120   case '\r':
 121   case '\n':
 122   case ' ':
 123     /* If we started scanning on a hard terminator, indicate we've finished */
 124     if (d == (unsigned char *)deststring)
 125       return -2;
 126
 127     /* Else make the next call to scan_urlencoded_param encounter it again */
 128     --s;
 129     break;
 130   case '?':
 131     if (flags != SCAN_PATH)
 132       return -1;
 133     break;
 134   case '=':
 135     if (flags != SCAN_SEARCHPATH_PARAM)
 136       return -1;
 137     break;
 138   case '&':
 139     if (flags == SCAN_PATH)
 140       return -1;
 141     if (flags == SCAN_SEARCHPATH_PARAM)
 142       --s;
 143     break;
 144   default:
 145     return -1;
 146   }
 147
 148   *string = (char *)s;
 149   return d - (unsigned char *)deststring;
 150 }
 151
 152 ssize_t scan_fixed_int(char *data, size_t len, int *tmp) {
 153   int minus = 0;
 154   *tmp      = 0;
 155   if (*data == '-')
 156     --len, ++data, ++minus;
 157   while ((len > 0) && (*data >= '0') && (*data <= '9')) {
 158     --len;
 159     *tmp = 10 * *tmp + *data++ - '0';
 160   }
 161   if (minus)
 162     *tmp = -*tmp;
 163   return len;
 164 }