mirror of
https://github.com/mauiaaron/apple2.git
synced 2024-12-25 09:29:48 +00:00
600 lines
14 KiB
C
600 lines
14 KiB
C
#include "jsmn.h"
|
|
|
|
#include <assert.h>
|
|
|
|
typedef struct {
|
|
jsmntype_t type;
|
|
int start;
|
|
int end;
|
|
int size;
|
|
int skip;
|
|
int parent;
|
|
/* private data */
|
|
int _lasttype;
|
|
int _depth;
|
|
} jsmntok_priv_t;
|
|
|
|
/**
|
|
* Private parsing tokens.
|
|
*/
|
|
enum {
|
|
_JSMN_SEPARATOR_COLON = 201,
|
|
_JSMN_SEPARATOR_COMMA,
|
|
_JSMN_OBJECT_END,
|
|
_JSMN_ARRAY_END,
|
|
_JSMN_STRING_KEY,
|
|
};
|
|
|
|
/**
|
|
* Allocates a fresh unused token from the token pull.
|
|
*/
|
|
static jsmntok_t *jsmn_alloc_token(jsmn_parser *parser,
|
|
jsmntok_t *tokens, size_t num_tokens) {
|
|
jsmntok_t *tok;
|
|
if (parser->toknext >= num_tokens) {
|
|
return NULL;
|
|
}
|
|
tok = &tokens[parser->toknext++];
|
|
tok->start = tok->end = -1;
|
|
tok->size = 0;
|
|
tok->skip = 1;
|
|
tok->parent = -1;
|
|
((jsmntok_priv_t *)tok)->_lasttype = JSMN_UNDEFINED;
|
|
((jsmntok_priv_t *)tok)->_depth = -1;
|
|
return tok;
|
|
}
|
|
|
|
/**
|
|
* Fills token type and boundaries.
|
|
*/
|
|
static void jsmn_fill_token(jsmntok_t *token, jsmntype_t type, int start, int end) {
|
|
token->type = type;
|
|
token->start = start;
|
|
token->end = end;
|
|
token->size = 0;
|
|
}
|
|
|
|
/**
|
|
* Parse digit(s).
|
|
*/
|
|
static int jsmn_parse_primitive_digits(const char *js, jsmntok_t *token, int start, int *end) {
|
|
int i;
|
|
int ended = 0;
|
|
|
|
/* must be at least one digit */
|
|
i = start;
|
|
switch (js[i]) {
|
|
case '0': case '1': case '2': case '3': case '4':
|
|
case '5': case '6': case '7': case '8': case '9':
|
|
break;
|
|
default:
|
|
return JSMN_ERROR_PRIMITIVE_INVAL;
|
|
}
|
|
++i;
|
|
|
|
for (; i < token->end; i++) {
|
|
switch (js[i]) {
|
|
case '0': case '1': case '2': case '3': case '4':
|
|
case '5': case '6': case '7': case '8': case '9':
|
|
break;
|
|
default:
|
|
ended = 1;
|
|
break;
|
|
}
|
|
|
|
if (ended) {
|
|
break;
|
|
}
|
|
}
|
|
|
|
*end = i;
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* Parse exponent.
|
|
*/
|
|
static int jsmn_parse_primitive_exp(const char *js, jsmntok_t *token, int start, int *end) {
|
|
int i;
|
|
int ret;
|
|
|
|
i = start;
|
|
assert(i < token->end);
|
|
|
|
if (js[i] != 'e' && js[i] != 'E') {
|
|
return JSMN_ERROR_PRIMITIVE_INVAL;
|
|
}
|
|
|
|
++i;
|
|
if (i >= token->end) {
|
|
return JSMN_ERROR_PART;
|
|
}
|
|
|
|
if (js[i] == '-' || js[i] == '+') {
|
|
++i;
|
|
if (i >= token->end) {
|
|
return JSMN_ERROR_PART;
|
|
}
|
|
}
|
|
|
|
ret = jsmn_parse_primitive_digits(js, token, i, &i);
|
|
if (ret) {
|
|
return ret;
|
|
}
|
|
|
|
/* should be exactly at the end */
|
|
if (i != token->end) {
|
|
return JSMN_ERROR_PRIMITIVE_INVAL;
|
|
}
|
|
|
|
*end = i;
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* Parse fraction or exponent component.
|
|
*/
|
|
static int jsmn_parse_primitive_frac_or_exp(const char *js, jsmntok_t *token, int start, int *end) {
|
|
int i;
|
|
int ret;
|
|
|
|
i = start;
|
|
assert(i < token->end);
|
|
|
|
if (js[i] == '.') {
|
|
++i;
|
|
if (i >= token->end) {
|
|
return JSMN_ERROR_PART;
|
|
}
|
|
ret = jsmn_parse_primitive_digits(js, token, i, &i);
|
|
if (ret) {
|
|
return ret;
|
|
}
|
|
if (i < token->end) {
|
|
ret = jsmn_parse_primitive_exp(js, token, i, &i);
|
|
if (ret) {
|
|
return ret;
|
|
}
|
|
}
|
|
} else {
|
|
ret = jsmn_parse_primitive_exp(js, token, i, &i);
|
|
if (ret) {
|
|
return ret;
|
|
}
|
|
}
|
|
|
|
*end = i;
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* Parse primitive number.
|
|
*/
|
|
static int jsmn_parse_primitive_number(const char *js, jsmntok_t *token) {
|
|
int i;
|
|
int ret;
|
|
|
|
i = token->start;
|
|
|
|
if (token->end - token->start <= 0) {
|
|
return JSMN_ERROR_PRIMITIVE_INVAL;
|
|
}
|
|
|
|
/* parse negative */
|
|
if (js[i] == '-') {
|
|
++i;
|
|
if (i >= token->end) {
|
|
return JSMN_ERROR_PART;
|
|
}
|
|
}
|
|
|
|
/* parse beginning zero */
|
|
if (js[i] == '0') {
|
|
++i;
|
|
if (i < token->end) {
|
|
ret = jsmn_parse_primitive_frac_or_exp(js, token, i, &i);
|
|
if (ret) {
|
|
return ret;
|
|
}
|
|
if (i < token->end) {
|
|
return JSMN_ERROR_PRIMITIVE_INVAL;
|
|
}
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
/* parse main digits */
|
|
ret = jsmn_parse_primitive_digits(js, token, i, &i);
|
|
if (ret) {
|
|
return ret;
|
|
}
|
|
if (i == token->end) {
|
|
return 0;
|
|
}
|
|
|
|
/* parse remaining fraction or exponent */
|
|
assert(i < token->end);
|
|
ret = jsmn_parse_primitive_frac_or_exp(js, token, i, &i);
|
|
|
|
if (i < token->end) {
|
|
return JSMN_ERROR_PRIMITIVE_INVAL;
|
|
}
|
|
|
|
return ret;
|
|
}
|
|
|
|
/**
|
|
* Fills next available token with JSON primitive.
|
|
*/
|
|
static int jsmn_parse_primitive(jsmn_parser *parser, const char *js,
|
|
size_t len, jsmntok_t *tokens, size_t num_tokens) {
|
|
jsmntok_t *token;
|
|
int start;
|
|
|
|
start = parser->pos;
|
|
|
|
for (; parser->pos < len && js[parser->pos] != '\0'; parser->pos++) {
|
|
switch (js[parser->pos]) {
|
|
case '\t' : case '\r' : case '\n' : case ' ' :
|
|
case ',' : case ']' : case '}' :
|
|
goto found;
|
|
}
|
|
if (js[parser->pos] < 32 || js[parser->pos] >= 127) {
|
|
parser->pos = start;
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
}
|
|
if ( (parser->pos < len) && (js[parser->pos] != '\0') ) {
|
|
/* In strict mode primitive must be followed by a comma/object/array or end-of-string */
|
|
parser->pos = start;
|
|
return JSMN_ERROR_PART;
|
|
}
|
|
|
|
found:
|
|
token = jsmn_alloc_token(parser, tokens, num_tokens);
|
|
if (token == NULL) {
|
|
parser->pos = start;
|
|
return JSMN_ERROR_NOMEM;
|
|
}
|
|
jsmn_fill_token(token, JSMN_PRIMITIVE, start, parser->pos);
|
|
token->parent = parser->toksuper;
|
|
|
|
/* parse/validate primitives according */
|
|
do {
|
|
if (js[start] == 't') {
|
|
if ( (token->end - start == 4) && (js[start+1] == 'r') && (js[start+2] == 'u') && (js[start+3] == 'e') ) {
|
|
break;
|
|
}
|
|
} else if (js[start] == 'f') {
|
|
if ( (token->end - start == 5) && (js[start+1] == 'a') && (js[start+2] == 'l') && (js[start+3] == 's') && (js[start+4] == 'e') ) {
|
|
break;
|
|
}
|
|
} else if (js[start] == 'n') {
|
|
if ( (token->end - start == 4) && (js[start+1] == 'u') && (js[start+2] == 'l') && (js[start+3] == 'l') ) {
|
|
break;
|
|
}
|
|
} else {
|
|
if (jsmn_parse_primitive_number(js, token) == 0) {
|
|
break;
|
|
}
|
|
}
|
|
|
|
/* primitive validation failed */
|
|
parser->pos = token->start;
|
|
return JSMN_ERROR_PRIMITIVE_INVAL;
|
|
} while (0);
|
|
|
|
parser->pos--;
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* Fills next token with JSON string.
|
|
*/
|
|
static int jsmn_parse_string(jsmn_parser *parser, const char *js,
|
|
size_t len, jsmntok_t *tokens, size_t num_tokens) {
|
|
jsmntok_t *token;
|
|
|
|
int start = parser->pos;
|
|
|
|
parser->pos++;
|
|
|
|
/* Skip starting quote */
|
|
for (; parser->pos < len && js[parser->pos] != '\0'; parser->pos++) {
|
|
char c = js[parser->pos];
|
|
|
|
/* Quote: end of string */
|
|
if (c == '\"') {
|
|
token = jsmn_alloc_token(parser, tokens, num_tokens);
|
|
if (token == NULL) {
|
|
parser->pos = start;
|
|
return JSMN_ERROR_NOMEM;
|
|
}
|
|
jsmn_fill_token(token, JSMN_STRING, start+1, parser->pos);
|
|
token->parent = parser->toksuper;
|
|
return 0;
|
|
}
|
|
|
|
/* Backslash: Quoted symbol expected */
|
|
if (c == '\\' && parser->pos + 1 < len) {
|
|
int i;
|
|
parser->pos++;
|
|
switch (js[parser->pos]) {
|
|
/* Allowed escaped symbols */
|
|
case '\"': case '/' : case '\\' : case 'b' :
|
|
case 'f' : case 'r' : case 'n' : case 't' :
|
|
break;
|
|
/* Allows escaped symbol \uXXXX */
|
|
case 'u':
|
|
parser->pos++;
|
|
for(i = 0; i < 4 && parser->pos < len && js[parser->pos] != '\0'; i++) {
|
|
/* If it isn't a hex character we have an error */
|
|
if(!((js[parser->pos] >= 48 && js[parser->pos] <= 57) || /* 0-9 */
|
|
(js[parser->pos] >= 65 && js[parser->pos] <= 70) || /* A-F */
|
|
(js[parser->pos] >= 97 && js[parser->pos] <= 102))) { /* a-f */
|
|
parser->pos = start;
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
parser->pos++;
|
|
}
|
|
parser->pos--;
|
|
break;
|
|
/* Unexpected symbol */
|
|
default:
|
|
parser->pos = start;
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
}
|
|
}
|
|
parser->pos = start;
|
|
return JSMN_ERROR_PART;
|
|
}
|
|
|
|
static void jsmn_percolate_skip_counts(jsmntok_t *tokens, int parent) {
|
|
jsmntok_t *token;
|
|
|
|
token = &tokens[parent];
|
|
token->skip++;
|
|
for (;;) {
|
|
if (token->parent == -1) {
|
|
break;
|
|
}
|
|
token = &tokens[token->parent];
|
|
token->skip++;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Parse JSON string and fill tokens.
|
|
*/
|
|
int jsmn_parse(jsmn_parser *parser, const char *js, size_t len,
|
|
jsmntok_t *tokens, unsigned int num_tokens) {
|
|
int r;
|
|
int i;
|
|
jsmntok_priv_t *token;
|
|
int depth = 0;
|
|
int count = parser->toknext;
|
|
|
|
if (tokens == NULL) {
|
|
return JSMN_ERROR_NOMEM;
|
|
}
|
|
|
|
for (; parser->pos < len && js[parser->pos] != '\0'; parser->pos++) {
|
|
char c;
|
|
jsmntype_t type;
|
|
|
|
c = js[parser->pos];
|
|
switch (c) {
|
|
case '{': case '[':
|
|
/* check previous token is valid */
|
|
if (parser->toknext >= 1) {
|
|
token = (jsmntok_priv_t *)&tokens[parser->toknext - 1];
|
|
depth = token->_depth;
|
|
switch (token->_lasttype) {
|
|
case _JSMN_SEPARATOR_COMMA:
|
|
case _JSMN_SEPARATOR_COLON:
|
|
case JSMN_ARRAY:
|
|
break;
|
|
default:
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
}
|
|
/* check parent token is valid */
|
|
if (parser->toksuper != -1) {
|
|
token = (jsmntok_priv_t *)&tokens[parser->toksuper];
|
|
if (token->type == JSMN_OBJECT) {
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
}
|
|
count++;
|
|
token = (jsmntok_priv_t *)jsmn_alloc_token(parser, tokens, num_tokens);
|
|
if (token == NULL) return JSMN_ERROR_NOMEM;
|
|
if (parser->toksuper != -1) {
|
|
tokens[parser->toksuper].size++;
|
|
token->parent = parser->toksuper;
|
|
jsmn_percolate_skip_counts(tokens, parser->toksuper);
|
|
}
|
|
token->type = (c == '{' ? JSMN_OBJECT : JSMN_ARRAY);
|
|
token->_lasttype = token->type;
|
|
token->_depth = depth+1;
|
|
token->start = parser->pos;
|
|
parser->toksuper = parser->toknext - 1;
|
|
break;
|
|
case '}': case ']':
|
|
type = (c == '}' ? JSMN_OBJECT : JSMN_ARRAY);
|
|
if (parser->toksuper == -1) {
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
if (parser->toknext < 1) {
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
token = (jsmntok_priv_t *)&tokens[parser->toknext - 1];
|
|
switch (token->_lasttype) {
|
|
case JSMN_PRIMITIVE:
|
|
case JSMN_STRING:
|
|
case JSMN_OBJECT:
|
|
case JSMN_ARRAY:
|
|
case _JSMN_OBJECT_END:
|
|
case _JSMN_ARRAY_END:
|
|
break;
|
|
default:
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
for (;;) {
|
|
if (token->start != -1 && token->end == -1) {
|
|
if (token->type != type) {
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
token->end = parser->pos + 1;
|
|
parser->toksuper = token->parent;
|
|
token->_depth--;
|
|
break;
|
|
}
|
|
if (token->parent == -1) {
|
|
break;
|
|
}
|
|
token = (jsmntok_priv_t *)&tokens[token->parent];
|
|
}
|
|
token->_lasttype = token->type == JSMN_OBJECT ? _JSMN_OBJECT_END : _JSMN_ARRAY_END;
|
|
break;
|
|
case '\"':
|
|
if (parser->toknext >= 1) {
|
|
token = (jsmntok_priv_t *)&tokens[parser->toknext - 1];
|
|
depth = token->_depth;
|
|
switch (token->_lasttype) {
|
|
case JSMN_OBJECT:
|
|
case JSMN_ARRAY:
|
|
case _JSMN_SEPARATOR_COLON:
|
|
case _JSMN_SEPARATOR_COMMA:
|
|
break;
|
|
default:
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
}
|
|
r = jsmn_parse_string(parser, js, len, tokens, num_tokens);
|
|
if (r < 0) return r;
|
|
count++;
|
|
if (parser->toksuper != -1) {
|
|
tokens[parser->toksuper].size++;
|
|
jsmn_percolate_skip_counts(tokens, parser->toksuper);
|
|
token = (jsmntok_priv_t *)&tokens[parser->toksuper];
|
|
type = token->_lasttype == JSMN_OBJECT ? _JSMN_STRING_KEY : JSMN_STRING;
|
|
} else {
|
|
type = JSMN_STRING;
|
|
}
|
|
token = (jsmntok_priv_t *)&tokens[parser->toknext - 1];
|
|
token->_lasttype = type;
|
|
token->_depth = depth;
|
|
break;
|
|
case '\t' : case '\r' : case '\n' : case ' ':
|
|
break;
|
|
case ':':
|
|
if (parser->toksuper == -1) {
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
assert(parser->toknext >= 1);
|
|
token = (jsmntok_priv_t *)&tokens[parser->toknext - 1];
|
|
switch (token->_lasttype) {
|
|
case _JSMN_STRING_KEY:
|
|
break;
|
|
default:
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
parser->toksuper = parser->toknext - 1;
|
|
token->_lasttype = _JSMN_SEPARATOR_COLON;
|
|
break;
|
|
case ',':
|
|
if (parser->toksuper == -1) {
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
assert(parser->toknext >= 1);
|
|
token = (jsmntok_priv_t *)&tokens[parser->toknext - 1];
|
|
switch (token->_lasttype) {
|
|
case JSMN_PRIMITIVE:
|
|
case JSMN_STRING:
|
|
case _JSMN_OBJECT_END:
|
|
case _JSMN_ARRAY_END:
|
|
break;
|
|
default:
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
if (
|
|
tokens[parser->toksuper].type != JSMN_ARRAY &&
|
|
tokens[parser->toksuper].type != JSMN_OBJECT) {
|
|
parser->toksuper = tokens[parser->toksuper].parent;
|
|
}
|
|
token->_lasttype = _JSMN_SEPARATOR_COMMA;
|
|
break;
|
|
/* In strict mode primitives are: numbers and booleans */
|
|
case '-': case '0': case '1' : case '2': case '3' : case '4':
|
|
case '5': case '6': case '7' : case '8': case '9':
|
|
case 't': case 'f': case 'n' :
|
|
/* And they must not be keys of the object */
|
|
if (parser->toksuper != -1) {
|
|
token = (jsmntok_priv_t *)&tokens[parser->toksuper];
|
|
if (token->type == JSMN_OBJECT ||
|
|
(token->type == JSMN_STRING && token->size != 0)) {
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
depth = token->_depth;
|
|
}
|
|
if (parser->toknext >= 1) {
|
|
token = (jsmntok_priv_t *)&tokens[parser->toknext - 1];
|
|
switch (token->_lasttype) {
|
|
case JSMN_ARRAY:
|
|
case _JSMN_SEPARATOR_COLON:
|
|
case _JSMN_SEPARATOR_COMMA:
|
|
break;
|
|
default:
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
}
|
|
r = jsmn_parse_primitive(parser, js, len, tokens, num_tokens);
|
|
if (r < 0) return r;
|
|
count++;
|
|
if (parser->toksuper != -1) {
|
|
tokens[parser->toksuper].size++;
|
|
jsmn_percolate_skip_counts(tokens, parser->toksuper);
|
|
}
|
|
token = (jsmntok_priv_t *)&tokens[parser->toknext - 1];
|
|
token->_lasttype = JSMN_PRIMITIVE;
|
|
token->_depth = depth;
|
|
break;
|
|
|
|
/* Unexpected char in strict mode */
|
|
default:
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
}
|
|
|
|
if (tokens[0].type == JSMN_OBJECT || tokens[0].type == JSMN_ARRAY) {
|
|
/* Unmatched opened object or array */
|
|
if (((jsmntok_priv_t *)(&tokens[0]))->_depth != 0) {
|
|
return JSMN_ERROR_INVAL;
|
|
}
|
|
}
|
|
|
|
#ifndef NDEBUG
|
|
/* sanity checks */
|
|
for (i = 0; i < parser->toknext; i++) {
|
|
assert(tokens[i].start != -1);
|
|
assert(tokens[i].end != -1);
|
|
assert(((jsmntok_priv_t *)(&tokens[i]))->_depth != -1);
|
|
assert(((jsmntok_priv_t *)(&tokens[i]))->_lasttype != JSMN_UNDEFINED);
|
|
}
|
|
#endif
|
|
|
|
return count;
|
|
}
|
|
|
|
/**
|
|
* Creates a new parser based over a given buffer with an array of tokens
|
|
* available.
|
|
*/
|
|
void jsmn_init(jsmn_parser *parser) {
|
|
parser->pos = 0;
|
|
parser->toknext = 0;
|
|
parser->toksuper = -1;
|
|
}
|
|
|