json_value *json_parse(char *source, char **error_pos, int *error_line, block_allocator *allocator)
{
json_value *root = 0;
json_value *top = 0;
char *name = 0;
char *it = source;
int escaped_newlines = 0;
while (*it)
{
switch (*it)
{
case '{':
case '[':
{
// create new value
json_value *object = json_alloc(allocator);
// name
object->name = name;
name = 0;
// type
object->type = (*it == '{') ? JSON_OBJECT : JSON_ARRAY;
// skip open character
++it;
// set top and root
if (top)
{
json_append(top, object);
}
else if (!root)
{
root = object;
}
else
{
ERROR(it, "Second root. Only one root allowed");
}
top = object;
}
break;
case '}':
case ']':
{
if (!top || top->type != ((*it == '}') ? JSON_OBJECT : JSON_ARRAY))
{
ERROR(it, "Mismatch closing brace/bracket");
}
// skip close character
++it;
// set top
top = top->parent;
}
break;
case ':':
if (!top || top->type != JSON_OBJECT)
{
ERROR(it, "Unexpected character");
}
++it;
break;
case ',':
CHECK_TOP();
++it;
break;
case '"':
{
CHECK_TOP();
// skip '"' character
++it;
char *first = it;
char *last = it;
while (*it)
{
if ((unsigned char)*it < '\x20')
{
ERROR(first, "Control characters not allowed in strings");
}
else if (*it == '\\')
{
switch (it[1])
{
case '"':
*last = '"';
break;
case '\\':
*last = '\\';
break;
case '/':
*last = '/';
break;
case 'b':
*last = '\b';
break;
case 'f':
*last = '\f';
break;
case 'n':
*last = '\n';
++escaped_newlines;
break;
case 'r':
*last = '\r';
break;
case 't':
*last = '\t';
break;
case 'u':
{
unsigned int codepoint;
if (hatoui(it + 2, it + 6, &codepoint) != it + 6)
{
ERROR(it, "Bad unicode codepoint");
}
if (codepoint <= 0x7F)
{
*last = (char)codepoint;
}
else if (codepoint <= 0x7FF)
{
*last++ = (char)(0xC0 | (codepoint >> 6));
*last = (char)(0x80 | (codepoint & 0x3F));
}
else if (codepoint <= 0xFFFF)
{
*last++ = (char)(0xE0 | (codepoint >> 12));
*last++ = (char)(0x80 | ((codepoint >> 6) & 0x3F));
*last = (char)(0x80 | (codepoint & 0x3F));
}
}
it += 4;
break;
default:
ERROR(first, "Unrecognized escape sequence");
}
++last;
it += 2;
}
else if (*it == '"')
{
*last = 0;
++it;
break;
}
else
{
*last++ = *it++;
}
}
enum jtokentype getJsonToken(std::string& tokenVal, unsigned int& consumed,
const char *raw)
{
tokenVal.clear();
consumed = 0;
const char *rawStart = raw;
while ((*raw) && (isspace(*raw))) // skip whitespace
raw++;
switch (*raw) {
case 0:
return JTOK_NONE;
case '{':
raw++;
consumed = (raw - rawStart);
return JTOK_OBJ_OPEN;
case '}':
raw++;
consumed = (raw - rawStart);
return JTOK_OBJ_CLOSE;
case '[':
raw++;
consumed = (raw - rawStart);
return JTOK_ARR_OPEN;
case ']':
raw++;
consumed = (raw - rawStart);
return JTOK_ARR_CLOSE;
case ':':
raw++;
consumed = (raw - rawStart);
return JTOK_COLON;
case ',':
raw++;
consumed = (raw - rawStart);
return JTOK_COMMA;
case 'n':
case 't':
case 'f':
if (!strncmp(raw, "null", 4)) {
raw += 4;
consumed = (raw - rawStart);
return JTOK_KW_NULL;
} else if (!strncmp(raw, "true", 4)) {
raw += 4;
consumed = (raw - rawStart);
return JTOK_KW_TRUE;
} else if (!strncmp(raw, "false", 5)) {
raw += 5;
consumed = (raw - rawStart);
return JTOK_KW_FALSE;
} else
return JTOK_ERR;
case '-':
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
case '8':
case '9': {
// part 1: int
std::string numStr;
const char *first = raw;
const char *firstDigit = first;
if (!isdigit(*firstDigit))
firstDigit++;
if ((*firstDigit == '0') && isdigit(firstDigit[1]))
return JTOK_ERR;
numStr += *raw; // copy first char
raw++;
if ((*first == '-') && (!isdigit(*raw)))
return JTOK_ERR;
while ((*raw) && isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
// part 2: frac
if (*raw == '.') {
numStr += *raw; // copy .
raw++;
if (!isdigit(*raw))
return JTOK_ERR;
while ((*raw) && isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
}
// part 3: exp
if (*raw == 'e' || *raw == 'E') {
numStr += *raw; // copy E
raw++;
if (*raw == '-' || *raw == '+') { // copy +/-
numStr += *raw;
raw++;
}
if (!isdigit(*raw))
return JTOK_ERR;
while ((*raw) && isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
}
tokenVal = numStr;
consumed = (raw - rawStart);
return JTOK_NUMBER;
}
case '"': {
raw++; // skip "
std::string valStr;
while (*raw) {
if (*raw < 0x20)
return JTOK_ERR;
else if (*raw == '\\') {
raw++; // skip backslash
switch (*raw) {
case '"': valStr += "\""; break;
case '\\': valStr += "\\"; break;
case '/': valStr += "/"; break;
case 'b': valStr += "\b"; break;
case 'f': valStr += "\f"; break;
case 'n': valStr += "\n"; break;
case 'r': valStr += "\r"; break;
case 't': valStr += "\t"; break;
case 'u': {
char buf[4] = {0,0,0,0};
char *last = &buf[0];
unsigned int codepoint;
if (hatoui(raw + 1, raw + 1 + 4, codepoint) !=
raw + 1 + 4)
return JTOK_ERR;
if (codepoint <= 0x7f)
*last = (char)codepoint;
else if (codepoint <= 0x7FF) {
*last++ = (char)(0xC0 | (codepoint >> 6));
*last = (char)(0x80 | (codepoint & 0x3F));
} else if (codepoint <= 0xFFFF) {
*last++ = (char)(0xE0 | (codepoint >> 12));
*last++ = (char)(0x80 | ((codepoint >> 6) & 0x3F));
*last = (char)(0x80 | (codepoint & 0x3F));
}
valStr += buf;
raw += 4;
break;
}
default:
return JTOK_ERR;
}
raw++; // skip esc'd char
}
