1 files changed, 170 insertions, 0 deletions

diff --git a/ot_urlencoded.c b/ot_urlencoded.c
new file mode 100644
index 0000000..17f12a4
--- /dev/null
+++ b/ot_urlencoded.c
@@ -0,0 +1,170 @@
+/* This software was written by Dirk Engling <erdgeist@erdgeist.org>
+   It is considered beerware. Prost. Skol. Cheers or whatever.
+
+   $id$ */
+
+/* Opentracker */
+#include "ot_urlencoded.h"
+
+/* System */
+#include <string.h>
+
+/* Idea is to do a in place replacement or guarantee at least
+   strlen( string ) bytes in deststring
+   watch http://www.ietf.org/rfc/rfc2396.txt
+         unreserved    = alphanum | mark
+         mark          = "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")"
+   we add '%' to the matrix to not stop at encoded chars.
+   After losing too many requests to being too strict, add the following characters to reserved matrix
+         relax         = "+" | "," | "/" | ";" | "<" | ">" | ":"
+*/
+
+/* This matrix holds for each ascii character the information,
+   whether it is a non-terminating character for one of the three
+   scan states we are in, that is 'path', 'param' and 'value' from
+  /path?param=value&param=value, it is encoded in bit 0, 1 and 2
+  respectively
+
+  The top bit of lower nibble indicates, whether this character is
+  a hard terminator, ie. \0, \n or \s, where the whole scanning
+  process should terminate
+  */
+static const unsigned char is_unreserved[256] = {
+  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,
+  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,
+  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,
+  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,
+  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,
+  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,
+  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,
+  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
+};
+
+/* Do a fast nibble to hex representation conversion */
+static unsigned char fromhex(unsigned char x) {
+  x -= '0';
+  if (x <= 9)
+    return x;
+  x &= ~0x20;
+  x -= 'A' - '0';
+  if (x < 6)
+    return x + 10;
+  return 0xff;
+}
+
+/* Skip the value of a param=value pair */
+void urlencoded_skip_value(char **string) {
+  const unsigned char *s = *(const unsigned char **)string;
+  unsigned char        f;
+
+  /* Since we are asked to skip the 'value', we assume to stop at
+     terminators for a 'value' string position */
+  while ((f = is_unreserved[*s++]) & SCAN_SEARCHPATH_VALUE)
+    ;
+
+  /* If we stopped at a hard terminator like \0 or \n, make the
+     next urlencoded_parse encounter it again */
+  if (f & SCAN_SEARCHPATH_TERMINATOR)
+    --s;
+
+  *string = (char *)s;
+}
+
+int urlencoded_match_keyword(const ot_keywords *keywords, char **string, SCAN_SEARCHPATH_FLAG flags) {
+  char   *deststring   = *string;
+  ssize_t match_length = urlencoded_parse(string, deststring, flags);
+
+  if (match_length < 0)
+    return match_length;
+  if (match_length == 0)
+    return -3;
+
+  while (keywords->key) {
+    if (!strncmp(keywords->key, deststring, match_length) && !keywords->key[match_length])
+      return keywords->value;
+    keywords++;
+  }
+
+  return -3;
+}
+
+ssize_t urlencoded_parse(char **string, char *deststring, SCAN_SEARCHPATH_FLAG flags) {
+  const unsigned char *s = *(const unsigned char **)string;
+  unsigned char       *d = (unsigned char *)deststring;
+  unsigned char        b, c;
+
+  /* This is the main decoding loop.
+    'flag' determines, which characters are non-terminating in current context
+    (ie. stop at '=' and '&' if scanning for a 'param'; stop at '?' if scanning for the path )
+  */
+  while (is_unreserved[c = *s++] & flags) {
+
+    /* When encountering an url escaped character, try to decode */
+    if (c == '%') {
+      if ((b = fromhex(*s++)) == 0xff)
+        return -1;
+      if ((c = fromhex(*s++)) == 0xff)
+        return -1;
+      c |= (b << 4);
+    }
+
+    /* Write (possibly decoded) character to output */
+    *d++ = c;
+  }
+
+  switch (c) {
+  case 0:
+  case '\r':
+  case '\n':
+  case ' ':
+    /* If we started scanning on a hard terminator, indicate we've finished */
+    if (d == (unsigned char *)deststring)
+      return -2;
+
+    /* Else make the next call to urlencoded_parse encounter it again */
+    --s;
+    break;
+  case '?':
+    if (flags != SCAN_PATH)
+      return -1;
+    break;
+  case '=':
+    if (flags != SCAN_SEARCHPATH_PARAM)
+      return -1;
+    break;
+  case '&':
+    if (flags == SCAN_PATH)
+      return -1;
+    if (flags == SCAN_SEARCHPATH_PARAM)
+      --s;
+    break;
+  default:
+    return -1;
+  }
+
+  *string = (char *)s;
+  return d - (unsigned char *)deststring;
+}
+
+
+size_t urlencoded_format(char *dest, const uint8_t *src, size_t len) {
+  static const char safe[]     = "$/.=~-_";
+  static const char hexdigit[] = "0123456789abcdef";
+  size_t            written    = 0, i;
+
+  for (i = 0; i < len; ++i) {
+    uint8_t c = src[i];
+
+    /* strchr would find the terminator, so a zero byte never counts as safe */
+    if ((c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z') || (c >= '0' && c <= '9') || (c && strchr(safe, c))) {
+      dest[written++] = (char)c;
+    } else {
+      dest[written]     = '%';
+      dest[written + 1] = hexdigit[c >> 4];
+      dest[written + 2] = hexdigit[c & 15];
+      written          += 3;
+    }
+  }
+
+  return written;
+}