#include "internal.h"

static void mg_strlcpy(register char *dst, register const char *src, size_t n) {
  for (; *src != '\0' && n > 1; n--) {
    *dst++ = *src++;
  }
  *dst = '\0';
}

static int lowercase(const char *s) {
  return tolower(* (const unsigned char *) s);
}

static int mg_strncasecmp(const char *s1, const char *s2, size_t len) {
  int diff = 0;

  if (len > 0)
    do {
      diff = lowercase(s1++) - lowercase(s2++);
    } while (diff == 0 && s1[-1] != '\0' && --len > 0);

  return diff;
}

static int mg_strcasecmp(const char *s1, const char *s2) {
  int diff;

  do {
    diff = lowercase(s1++) - lowercase(s2++);
  } while (diff == 0 && s1[-1] != '\0');

  return diff;
}

static char * mg_strndup(const char *ptr, size_t len) {
  char *p;

  if ((p = (char *) malloc(len + 1)) != NULL) {
    mg_strlcpy(p, ptr, len + 1);
  }

  return p;
}

static char * mg_strdup(const char *str) {
  return mg_strndup(str, strlen(str));
}

static const char *mg_strcasestr(const char *big_str, const char *small_str) {
  int i, big_len = strlen(big_str), small_len = strlen(small_str);

  for (i = 0; i <= big_len - small_len; i++) {
    if (mg_strncasecmp(big_str + i, small_str, small_len) == 0) {
      return big_str + i;
    }
  }

  return NULL;
}

// Like snprintf(), but never returns negative value, or a value
// that is larger than a supplied buffer.
// Thanks to Adam Zeldis to pointing snprintf()-caused vulnerability
// in his audit report.
static int mg_vsnprintf(char *buf, size_t buflen, const char *fmt, va_list ap) {
  int n;

  if (buflen == 0) {
    return 0;
  }

  n = vsnprintf(buf, buflen, fmt, ap);

  if (n < 0) {
    n = 0;
  } else if (n >= (int) buflen) {
    n = (int) buflen - 1;
  }
  buf[n] = '\0';

  return n;
}

static int mg_snprintf(char *buf, size_t buflen,
    PRINTF_FORMAT_STRING(const char *fmt), ...) PRINTF_ARGS(3, 4);

static int mg_snprintf(char *buf, size_t buflen, const char *fmt, ...) {
  va_list ap;
  int n;

  va_start(ap, fmt);
  n = mg_vsnprintf(buf, buflen, fmt, ap);
  va_end(ap);

  return n;
}

// Skip the characters until one of the delimiters characters found.
// 0-terminate resulting word. Skip the delimiter and following whitespaces.
// Advance pointer to buffer to the next word. Return found 0-terminated word.
// Delimiters can be quoted with quotechar.
static char *skip_quoted(char **buf, const char *delimiters,
                         const char *whitespace, char quotechar) {
  char *p, *begin_word, *end_word, *end_whitespace;

  begin_word = *buf;
  end_word = begin_word + strcspn(begin_word, delimiters);

  // Check for quotechar
  if (end_word > begin_word) {
    p = end_word - 1;
    while (*p == quotechar) {
      // If there is anything beyond end_word, copy it
      if (*end_word == '\0') {
        *p = '\0';
        break;
      } else {
        size_t end_off = strcspn(end_word + 1, delimiters);
        memmove (p, end_word, end_off + 1);
        p += end_off; // p must correspond to end_word - 1
        end_word += end_off + 1;
      }
    }
    for (p++; p < end_word; p++) {
      *p = '\0';
    }
  }

  if (*end_word == '\0') {
    *buf = end_word;
  } else {
    end_whitespace = end_word + 1 + strspn(end_word + 1, whitespace);

    for (p = end_word; p < end_whitespace; p++) {
      *p = '\0';
    }

    *buf = end_whitespace;
  }

  return begin_word;
}

// Simplified version of skip_quoted without quote char
// and whitespace == delimiters
static char *skip(char **buf, const char *delimiters) {
  return skip_quoted(buf, delimiters, delimiters, 0);
}


// Return HTTP header value, or NULL if not found.
static const char *get_header(const struct mg_request_info *ri,
                              const char *name) {
  int i;

  for (i = 0; i < ri->num_headers; i++)
    if (!mg_strcasecmp(name, ri->http_headers[i].name))
      return ri->http_headers[i].value;

  return NULL;
}

const char *mg_get_header(const struct mg_connection *conn, const char *name) {
  return get_header(&conn->request_info, name);
}

// A helper function for traversing a comma separated list of values.
// It returns a list pointer shifted to the next value, or NULL if the end
// of the list found.
// Value is stored in val vector. If value has form "x=y", then eq_val
// vector is initialized to point to the "y" part, and val vector length
// is adjusted to point only to "x".
static const char *next_option(const char *list, struct vec *val,
                               struct vec *eq_val) {
  if (list == NULL || *list == '\0') {
    // End of the list
    list = NULL;
  } else {
    val->ptr = list;
    if ((list = strchr(val->ptr, ',')) != NULL) {
      // Comma found. Store length and shift the list ptr
      val->len = list - val->ptr;
      list++;
    } else {
      // This value is the last one
      list = val->ptr + strlen(val->ptr);
      val->len = list - val->ptr;
    }

    if (eq_val != NULL) {
      // Value has form "x=y", adjust pointers and lengths
      // so that val points to "x", and eq_val points to "y".
      eq_val->len = 0;
      eq_val->ptr = (const char *) memchr(val->ptr, '=', val->len);
      if (eq_val->ptr != NULL) {
        eq_val->ptr++;  // Skip over '=' character
        eq_val->len = val->ptr + val->len - eq_val->ptr;
        val->len = (eq_val->ptr - val->ptr) - 1;
      }
    }
  }

  return list;
}

// Perform case-insensitive match of string against pattern
static int match_prefix(const char *pattern, int pattern_len, const char *str) {
  const char *or_str;
  int i, j, len, res;

  if ((or_str = (const char *) memchr(pattern, '|', pattern_len)) != NULL) {
    res = match_prefix(pattern, or_str - pattern, str);
    return res > 0 ? res :
        match_prefix(or_str + 1, (pattern + pattern_len) - (or_str + 1), str);
  }

  i = j = 0;
  res = -1;
  for (; i < pattern_len; i++, j++) {
    if (pattern[i] == '?' && str[j] != '\0') {
      continue;
    } else if (pattern[i] == '$') {
      return str[j] == '\0' ? j : -1;
    } else if (pattern[i] == '*') {
      i++;
      if (pattern[i] == '*') {
        i++;
        len = (int) strlen(str + j);
      } else {
        len = (int) strcspn(str + j, "/");
      }
      if (i == pattern_len) {
        return j + len;
      }
      do {
        res = match_prefix(pattern + i, pattern_len - i, str + j + len);
      } while (res == -1 && len-- > 0);
      return res == -1 ? -1 : j + res + len;
    } else if (lowercase(&pattern[i]) != lowercase(&str[j])) {
      return -1;
    }
  }
  return j;
}

// Protect against directory disclosure attack by removing '..',
// excessive '/' and '\' characters
static void remove_double_dots_and_double_slashes(char *s) {
  char *p = s;

  while (*s != '\0') {
    *p++ = *s++;
    if (s[-1] == '/' || s[-1] == '\\') {
      // Skip all following slashes, backslashes and double-dots
      while (s[0] != '\0') {
        if (s[0] == '/' || s[0] == '\\') {
          s++;
        } else if (s[0] == '.' && s[1] == '.') {
          s += 2;
        } else {
          break;
        }
      }
    }
  }
  *p = '\0';
}

void mg_url_encode(const char *src, char *dst, size_t dst_len) {
  static const char *dont_escape = "._-$,;~()";
  static const char *hex = "0123456789abcdef";
  const char *end = dst + dst_len - 1;

  for (; *src != '\0' && dst < end; src++, dst++) {
    if (isalnum(*(const unsigned char *) src) ||
        strchr(dont_escape, * (const unsigned char *) src) != NULL) {
      *dst = *src;
    } else if (dst + 2 < end) {
      dst[0] = '%';
      dst[1] = hex[(* (const unsigned char *) src) >> 4];
      dst[2] = hex[(* (const unsigned char *) src) & 0xf];
      dst += 2;
    }
  }

  *dst = '\0';
}

int mg_url_decode(const char *src, int src_len, char *dst,
                  int dst_len, int is_form_url_encoded) {
  int i, j, a, b;
#define HEXTOI(x) (isdigit(x) ? x - '0' : x - 'W')

  for (i = j = 0; i < src_len && j < dst_len - 1; i++, j++) {
    if (src[i] == '%' && i < src_len - 2 &&
        isxdigit(* (const unsigned char *) (src + i + 1)) &&
        isxdigit(* (const unsigned char *) (src + i + 2))) {
      a = tolower(* (const unsigned char *) (src + i + 1));
      b = tolower(* (const unsigned char *) (src + i + 2));
      dst[j] = (char) ((HEXTOI(a) << 4) | HEXTOI(b));
      i += 2;
    } else if (is_form_url_encoded && src[i] == '+') {
      dst[j] = ' ';
    } else {
      dst[j] = src[i];
    }
  }

  dst[j] = '\0'; // Null-terminate the destination

  return i >= src_len ? j : -1;
}

// Check whether full request is buffered. Return:
//   -1  if request is malformed
//    0  if request is not yet fully buffered
//   >0  actual request length, including last \r\n\r\n
static int get_request_len(const char *buf, int buf_len) {
  int i;

  for (i = 0; i < buf_len; i++) {
    // Control characters are not allowed but >=128 is.
    // Abort scan as soon as one malformed character is found;
    // don't let subsequent \r\n\r\n win us over anyhow
    if (!isprint(* (const unsigned char *) &buf[i]) && buf[i] != '\r' &&
        buf[i] != '\n' && * (const unsigned char *) &buf[i] < 128) {
      return -1;
    } else if (buf[i] == '\n' && i + 1 < buf_len && buf[i + 1] == '\n') {
      return i + 2;
    } else if (buf[i] == '\n' && i + 2 < buf_len && buf[i + 1] == '\r' &&
               buf[i + 2] == '\n') {
      return i + 3;
    }
  }

  return 0;
}

int mg_get_cookie(const char *cookie_header, const char *var_name,
                  char *dst, size_t dst_size) {
  const char *s, *p, *end;
  int name_len, len = -1;

  if (dst == NULL || dst_size == 0) {
    len = -2;
  } else if (var_name == NULL || (s = cookie_header) == NULL) {
    len = -1;
    dst[0] = '\0';
  } else {
    name_len = (int) strlen(var_name);
    end = s + strlen(s);
    dst[0] = '\0';

    for (; (s = mg_strcasestr(s, var_name)) != NULL; s += name_len) {
      if (s[name_len] == '=') {
        s += name_len + 1;
        if ((p = strchr(s, ' ')) == NULL)
          p = end;
        if (p[-1] == ';')
          p--;
        if (*s == '"' && p[-1] == '"' && p > s + 1) {
          s++;
          p--;
        }
        if ((size_t) (p - s) < dst_size) {
          len = p - s;
          mg_strlcpy(dst, s, (size_t) len + 1);
        } else {
          len = -3;
        }
        break;
      }
    }
  }
  return len;
}

int mg_get_var(const char *data, size_t data_len, const char *name,
               char *dst, size_t dst_len) {
  const char *p, *e, *s;
  size_t name_len;
  int len;

  if (dst == NULL || dst_len == 0) {
    len = -2;
  } else if (data == NULL || name == NULL || data_len == 0) {
    len = -1;
    dst[0] = '\0';
  } else {
    name_len = strlen(name);
    e = data + data_len;
    len = -1;
    dst[0] = '\0';

    // data is "var1=val1&var2=val2...". Find variable first
    for (p = data; p + name_len < e; p++) {
      if ((p == data || p[-1] == '&') && p[name_len] == '=' &&
          !mg_strncasecmp(name, p, name_len)) {

        // Point p to variable value
        p += name_len + 1;

        // Point s to the end of the value
        s = (const char *) memchr(p, '&', (size_t)(e - p));
        if (s == NULL) {
          s = e;
        }
        assert(s >= p);

        // Decode variable into destination buffer
        len = mg_url_decode(p, (size_t)(s - p), dst, dst_len, 1);

        // Redirect error code from -1 to -2 (destination buffer too small).
        if (len == -1) {
          len = -2;
        }
        break;
      }
    }
  }

  return len;
}