json_pure 1.0.0 → 1.4.6

data/ext/json/ext/parser/parser.rl

@@ -1,37 +1,83 @@
-/* vim: set cin et sw=4 ts=4: */
+#include "parser.h"
+
+/* unicode */
+
+static const char digit_values[256] = { 
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, -1,
+    -1, -1, -1, -1, -1, -1, 10, 11, 12, 13, 14, 15, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    10, 11, 12, 13, 14, 15, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
+    -1, -1, -1, -1, -1, -1, -1
+};
+
+static UTF32 unescape_unicode(const unsigned char *p)
+{
+    char b;
+    UTF32 result = 0;
+    b = digit_values[p[0]];
+    if (b < 0) return UNI_REPLACEMENT_CHAR;
+    result = (result << 4) | b;
+    b = digit_values[p[1]];
+    result = (result << 4) | b;
+    if (b < 0) return UNI_REPLACEMENT_CHAR;
+    b = digit_values[p[2]];
+    result = (result << 4) | b;
+    if (b < 0) return UNI_REPLACEMENT_CHAR;
+    b = digit_values[p[3]];
+    result = (result << 4) | b;
+    if (b < 0) return UNI_REPLACEMENT_CHAR;
+    return result;
+}
 
-#include "ruby.h"
-#include "re.h"
-#include "unicode.h"
+static int convert_UTF32_to_UTF8(char *buf, UTF32 ch) 
+{
+    int len = 1;
+    if (ch <= 0x7F) {
+        buf[0] = (char) ch;
+    } else if (ch <= 0x07FF) {
+        buf[0] = (char) ((ch >> 6) | 0xC0);
+        buf[1] = (char) ((ch & 0x3F) | 0x80);
+        len++;
+    } else if (ch <= 0xFFFF) {
+        buf[0] = (char) ((ch >> 12) | 0xE0);
+        buf[1] = (char) (((ch >> 6) & 0x3F) | 0x80);
+        buf[2] = (char) ((ch & 0x3F) | 0x80);
+        len += 2;
+    } else if (ch <= 0x1fffff) {
+        buf[0] =(char) ((ch >> 18) | 0xF0);
+        buf[1] =(char) (((ch >> 12) & 0x3F) | 0x80);
+        buf[2] =(char) (((ch >> 6) & 0x3F) | 0x80);
+        buf[3] =(char) ((ch & 0x3F) | 0x80);
+        len += 3;
+    } else {
+        buf[0] = '?';
+    }
+    return len;
+}
 
-#ifndef swap16
-#define swap16(x)	((((x)&0xFF)<<8) | (((x)>>8)&0xFF))
+#ifdef HAVE_RUBY_ENCODING_H
+static VALUE CEncoding_ASCII_8BIT, CEncoding_UTF_8, CEncoding_UTF_16BE,
+    CEncoding_UTF_16LE, CEncoding_UTF_32BE, CEncoding_UTF_32LE;
+static ID i_encoding, i_encode, i_encode_bang, i_force_encoding;
+#else
+static ID i_iconv;
 #endif
 
-#define EVIL 0x666
-
-static VALUE mJSON, mExt, cParser, eParserError;
-
-static ID i_json_creatable_p, i_json_create, i_create_id, i_chr; 
-
-typedef struct JSON_ParserStruct {
-    VALUE Vsource;
-    char *source;
-    long len;
-    char *memo;
-    VALUE create_id;
-} JSON_Parser;
-
-static char *JSON_parse_object(JSON_Parser *json, char *p, char *pe, VALUE *result);
-static char *JSON_parse_array(JSON_Parser *json, char *p, char *pe, VALUE *result);
-static char *JSON_parse_value(JSON_Parser *json, char *p, char *pe, VALUE *result);
-static char *JSON_parse_string(JSON_Parser *json, char *p, char *pe, VALUE *result);
-static char *JSON_parse_integer(JSON_Parser *json, char *p, char *pe, VALUE *result);
-static char *JSON_parse_float(JSON_Parser *json, char *p, char *pe, VALUE *result);
+static VALUE mJSON, mExt, cParser, eParserError, eNestingError;
+static VALUE CNaN, CInfinity, CMinusInfinity;
 
-#define GET_STRUCT                          \
-    JSON_Parser *json;                      \
-    Data_Get_Struct(self, JSON_Parser, json);
+static ID i_json_creatable_p, i_json_create, i_create_id, i_create_additions,
+          i_chr, i_max_nesting, i_allow_nan, i_symbolize_names, i_object_class,
+          i_array_class, i_key_p, i_deep_const_get;
 
 %%{
     machine JSON_common;
@@ -48,7 +94,10 @@ static char *JSON_parse_float(JSON_Parser *json, char *p, char *pe, VALUE *resul
     Vnull               = 'null';
     Vfalse              = 'false';
     Vtrue               = 'true';
-    begin_value         = [nft"\-[{] | digit;
+    VNaN                = 'NaN';
+    VInfinity           = 'Infinity';
+    VMinusInfinity      = '-Infinity';
+    begin_value         = [nft"\-[{NI] | digit;
     begin_object        = '{';
     end_object          = '}';
     begin_array         = '[';
@@ -68,7 +117,7 @@ static char *JSON_parse_float(JSON_Parser *json, char *p, char *pe, VALUE *resul
         VALUE v = Qnil;
         char *np = JSON_parse_value(json, fpc, pe, &v); 
         if (np == NULL) {
-            fbreak;
+            fhold; fbreak;
         } else {
             rb_hash_aset(*result, last_name, v);
             fexec np;
@@ -76,11 +125,14 @@ static char *JSON_parse_float(JSON_Parser *json, char *p, char *pe, VALUE *resul
     }
 
     action parse_name {
-        char *np = JSON_parse_string(json, fpc, pe, &last_name);
-        if (np == NULL) fbreak; else fexec np;
+        char *np;
+        json->parsing_name = 1;
+        np = JSON_parse_string(json, fpc, pe, &last_name);
+        json->parsing_name = 0;
+        if (np == NULL) { fhold; fbreak; } else fexec np;
     }
 
-    action exit { fbreak; }
+    action exit { fhold; fbreak; }
 
     a_pair  = ignore* begin_name >parse_name
         ignore* name_separator ignore*
@@ -95,17 +147,25 @@ static char *JSON_parse_object(JSON_Parser *json, char *p, char *pe, VALUE *resu
 {
     int cs = EVIL;
     VALUE last_name = Qnil;
-    *result = rb_hash_new();
+    VALUE object_class = json->object_class;
+
+    if (json->max_nesting && json->current_nesting > json->max_nesting) {
+        rb_raise(eNestingError, "nesting of %d is too deep", json->current_nesting);
+    }
+
+    *result = NIL_P(object_class) ? rb_hash_new() : rb_class_new_instance(0, 0, object_class);
 
     %% write init;
     %% write exec;
 
     if (cs >= JSON_object_first_final) {
-        VALUE klassname = rb_hash_aref(*result, json->create_id);
-        if (!NIL_P(klassname)) {
-            VALUE klass = rb_path2class(StringValueCStr(klassname));
-            if RTEST(rb_funcall(klass, i_json_creatable_p, 0)) {
-                *result = rb_funcall(klass, i_json_create, 1, *result);
+        if (RTEST(json->create_id)) {
+            VALUE klassname = rb_hash_aref(*result, json->create_id);
+            if (!NIL_P(klassname)) {
+                VALUE klass = rb_funcall(mJSON, i_deep_const_get, 1, klassname);
+                if RTEST(rb_funcall(klass, i_json_creatable_p, 0)) {
+                    *result = rb_funcall(klass, i_json_create, 1, *result);
+                }
             }
         }
         return p + 1;
@@ -129,36 +189,67 @@ static char *JSON_parse_object(JSON_Parser *json, char *p, char *pe, VALUE *resu
     action parse_true {
         *result = Qtrue;
     }
+    action parse_nan {
+        if (json->allow_nan) {
+            *result = CNaN;
+        } else {
+            rb_raise(eParserError, "%u: unexpected token at '%s'", __LINE__, p - 2);
+        }
+    }
+    action parse_infinity {
+        if (json->allow_nan) {
+            *result = CInfinity;
+        } else {
+            rb_raise(eParserError, "%u: unexpected token at '%s'", __LINE__, p - 8);
+        }
+    }
     action parse_string {
         char *np = JSON_parse_string(json, fpc, pe, result);
-        if (np == NULL) fbreak; else fexec np;
+        if (np == NULL) { fhold; fbreak; } else fexec np;
     }
 
     action parse_number {
         char *np;
+        if(pe > fpc + 9 && !strncmp(MinusInfinity, fpc, 9)) {
+            if (json->allow_nan) {
+                *result = CMinusInfinity;
+                fexec p + 10;
+                fhold; fbreak;
+            } else {
+                rb_raise(eParserError, "%u: unexpected token at '%s'", __LINE__, p);
+            }
+        }
         np = JSON_parse_float(json, fpc, pe, result);
         if (np != NULL) fexec np;
         np = JSON_parse_integer(json, fpc, pe, result);
         if (np != NULL) fexec np;
-        fbreak;
+        fhold; fbreak;
     }
 
     action parse_array { 
-        char *np = JSON_parse_array(json, fpc, pe, result);
-        if (np == NULL) fbreak; else fexec np;
+        char *np;
+        json->current_nesting++;
+        np = JSON_parse_array(json, fpc, pe, result);
+        json->current_nesting--;
+        if (np == NULL) { fhold; fbreak; } else fexec np;
     }
 
     action parse_object { 
-        char *np =  JSON_parse_object(json, fpc, pe, result);
-        if (np == NULL) fbreak; else fexec np;
+        char *np;
+        json->current_nesting++;
+        np =  JSON_parse_object(json, fpc, pe, result);
+        json->current_nesting--;
+        if (np == NULL) { fhold; fbreak; } else fexec np;
     }
 
-    action exit { fbreak; }
+    action exit { fhold; fbreak; }
 
 main := (
               Vnull @parse_null |
               Vfalse @parse_false |
               Vtrue @parse_true |
+              VNaN @parse_nan |
+              VInfinity @parse_infinity |
               begin_number >parse_number |
               begin_string >parse_string |
               begin_array >parse_array |
@@ -185,7 +276,7 @@ static char *JSON_parse_value(JSON_Parser *json, char *p, char *pe, VALUE *resul
 
     write data;
 
-    action exit { fbreak; }
+    action exit { fhold; fbreak; }
 
     main := '-'? ('0' | [1-9][0-9]*) (^[0-9] @exit);
 }%%
@@ -213,11 +304,11 @@ static char *JSON_parse_integer(JSON_Parser *json, char *p, char *pe, VALUE *res
 
     write data;
 
-    action exit { fbreak; }
+    action exit { fhold; fbreak; }
 
     main := '-'? (
               (('0' | [1-9][0-9]*) '.' [0-9]+ ([Ee] [+\-]?[0-9]+)?)
-              | ([1-9][0-9]* ([Ee] [+\-]?[0-9]+))
+              | (('0' | [1-9][0-9]*) ([Ee] [+\-]?[0-9]+))
              )  (^[0-9Ee.\-] @exit );
 }%%
 
@@ -249,14 +340,14 @@ static char *JSON_parse_float(JSON_Parser *json, char *p, char *pe, VALUE *resul
         VALUE v = Qnil;
         char *np = JSON_parse_value(json, fpc, pe, &v); 
         if (np == NULL) {
-            fbreak;
+            fhold; fbreak;
         } else {
             rb_ary_push(*result, v);
             fexec np;
         }
     }
 
-    action exit { fbreak; }
+    action exit { fhold; fbreak; }
 
     next_element  = value_separator ignore* begin_value >parse_value;
 
@@ -269,7 +360,12 @@ static char *JSON_parse_float(JSON_Parser *json, char *p, char *pe, VALUE *resul
 static char *JSON_parse_array(JSON_Parser *json, char *p, char *pe, VALUE *result)
 {
     int cs = EVIL;
-    *result = rb_ary_new();
+    VALUE array_class = json->array_class;
+
+    if (json->max_nesting && json->current_nesting > json->max_nesting) {
+        rb_raise(eNestingError, "nesting of %d is too deep", json->current_nesting);
+    }
+    *result = NIL_P(array_class) ? rb_ary_new() : rb_class_new_instance(0, 0, array_class);
 
     %% write init;
     %% write exec;
@@ -277,59 +373,78 @@ static char *JSON_parse_array(JSON_Parser *json, char *p, char *pe, VALUE *resul
     if(cs >= JSON_array_first_final) {
         return p + 1;
     } else {
-        rb_raise(eParserError, "unexpected token at '%s'", p);
+        rb_raise(eParserError, "%u: unexpected token at '%s'", __LINE__, p);
+        return NULL;
     }
 }
 
-static VALUE json_string_escape(char *p, char *pe)
+static VALUE json_string_unescape(VALUE result, char *string, char *stringEnd)
 {
-    VALUE result = rb_str_buf_new(pe - p + 1);
-
-    while (p < pe) {
-        if (*p == '\\') {
-            p++;
-            if (p >= pe) return Qnil; /* raise an exception later, \ at end */
-            switch (*p) {
+    char *p = string, *pe = string, *unescape;
+    int unescape_len;
+
+    while (pe < stringEnd) {
+        if (*pe == '\\') {
+            unescape = (char *) "?";
+            unescape_len = 1;
+            if (pe > p) rb_str_buf_cat(result, p, pe - p);
+            switch (*++pe) {
+                case 'n':
+                    unescape = (char *) "\n";
+                    break;
+                case 'r':
+                    unescape = (char *) "\r";
+                    break;
+                case 't':
+                    unescape = (char *) "\t";
+                    break;
                 case '"':
+                    unescape = (char *) "\"";
+                    break;
                 case '\\':
-                    rb_str_buf_cat(result, p, 1);
-                    p++;
+                    unescape = (char *) "\\";
                     break;
                 case 'b':
-                    rb_str_buf_cat2(result, "\b");
-                    p++;
+                    unescape = (char *) "\b";
                     break;
                 case 'f':
-                    rb_str_buf_cat2(result, "\f");
-                    p++;
-                    break;
-                case 'n':
-                    rb_str_buf_cat2(result, "\n");
-                    p++;
-                    break;
-                case 'r':
-                    rb_str_buf_cat2(result, "\r");
-                    p++;
-                    break;
-                case 't':
-                    rb_str_buf_cat2(result, "\t");
-                    p++;
+                    unescape = (char *) "\f";
                     break;
                 case 'u':
-                    if (p > pe - 4) { 
+                    if (pe > stringEnd - 4) { 
                         return Qnil;
                     } else {
-                        p = JSON_convert_UTF16_to_UTF8(result, p, pe, strictConversion);
+                        char buf[4];
+                        UTF32 ch = unescape_unicode((unsigned char *) ++pe);
+                        pe += 3;
+                        if (UNI_SUR_HIGH_START == (ch & 0xFC00)) {
+                            pe++;
+                            if (pe > stringEnd - 6) return Qnil;
+                            if (pe[0] == '\\' && pe[1] == 'u') {
+                                UTF32 sur = unescape_unicode((unsigned char *) pe + 2);
+                                ch = (((ch & 0x3F) << 10) | ((((ch >> 6) & 0xF) + 1) << 16)
+                                        | (sur & 0x3FF));
+                                pe += 5;
+                            } else {
+                                unescape = (char *) "?";
+                                break;
+                            }
+                        }
+                        unescape_len = convert_UTF32_to_UTF8(buf, ch);
+                        unescape = buf;
                     }
                     break;
+                default:
+                    p = pe;
+                    continue;
             }
+            rb_str_buf_cat(result, unescape, unescape_len);
+            p = ++pe;
         } else {
-            char *q = p;
-            while (*q != '\\' && q < pe) q++;
-            rb_str_buf_cat(result, p, q - p);
-            p = q;
+            pe++;
         }
     }
+    rb_str_buf_cat(result, p, pe - p);
     return result;
 }
 
@@ -340,24 +455,33 @@ static VALUE json_string_escape(char *p, char *pe)
     write data;
 
     action parse_string {
-        *result = json_string_escape(json->memo + 1, p);
-        if (NIL_P(*result)) fbreak; else fexec p + 1;
-    }
-
-    action exit { fbreak; }
-
-    main := '"' ((^(["\\] | 0..0x1f) | '\\'["\\/bfnrt] | '\\u'[0-9a-fA-F]{4})* %parse_string) '"' @exit;
+        *result = json_string_unescape(*result, json->memo + 1, p);
+        if (NIL_P(*result)) {
+			fhold;
+			fbreak;
+		} else {
+			FORCE_UTF8(*result);
+			fexec p + 1;
+		}
+	}
+
+    action exit { fhold; fbreak; }
+
+    main := '"' ((^(["\\] | 0..0x1f) | '\\'["\\/bfnrt] | '\\u'[0-9a-fA-F]{4} | '\\'^(["\\/bfnrtu]|0..0x1f))* %parse_string) '"' @exit;
 }%%
 
 static char *JSON_parse_string(JSON_Parser *json, char *p, char *pe, VALUE *result)
 {
     int cs = EVIL;
 
-    *result = rb_str_new("", 0);
+    *result = rb_str_buf_new(0);
     %% write init;
     json->memo = p;
     %% write exec;
 
+    if (json->symbolize_names && json->parsing_name) {
+      *result = rb_str_intern(*result);
+    }
     if (cs >= JSON_string_first_final) {
         return p + 1;
     } else {
@@ -374,13 +498,17 @@ static char *JSON_parse_string(JSON_Parser *json, char *p, char *pe, VALUE *resu
     include JSON_common;
 
     action parse_object {
-        char *np = JSON_parse_object(json, fpc, pe, &result);
-        if (np == NULL) fbreak; else fexec np;
+        char *np;
+        json->current_nesting = 1;
+        np = JSON_parse_object(json, fpc, pe, &result);
+        if (np == NULL) { fhold; fbreak; } else fexec np;
     }
 
     action parse_array {
-        char *np = JSON_parse_array(json, fpc, pe, &result);
-        if (np == NULL) fbreak; else fexec np;
+        char *np;
+        json->current_nesting = 1;
+        np = JSON_parse_array(json, fpc, pe, &result);
+        if (np == NULL) { fhold; fbreak; } else fexec np;
     }
 
     main := ignore* (
@@ -401,38 +529,156 @@ static char *JSON_parse_string(JSON_Parser *json, char *p, char *pe, VALUE *resu
  *
  */
 
-/*
- * call-seq: new(source)
- *
- * Creates a new JSON::Ext::Parser instance for the string _source_.
- */
-static VALUE cParser_initialize(VALUE self, VALUE source)
+static VALUE convert_encoding(VALUE source)
 {
-    char *ptr;
-    long len;
-    GET_STRUCT;
-    source = StringValue(source);
-    ptr = RSTRING(source)->ptr;
-    len = RSTRING(source)->len;
+    char *ptr = RSTRING_PTR(source);
+    long len = RSTRING_LEN(source);
     if (len < 2) {
         rb_raise(eParserError, "A JSON text must at least contain two octets!");
     }
-    /*
-       Convert these?
+#ifdef HAVE_RUBY_ENCODING_H
+    {
+        VALUE encoding = rb_funcall(source, i_encoding, 0);
+        if (encoding == CEncoding_ASCII_8BIT) {
+            if (len >= 4 &&  ptr[0] == 0 && ptr[1] == 0 && ptr[2] == 0) {
+                source = rb_str_dup(source);
+                rb_funcall(source, i_force_encoding, 1, CEncoding_UTF_32BE);
+                source = rb_funcall(source, i_encode_bang, 1, CEncoding_UTF_8);
+            } else if (len >= 4 && ptr[0] == 0 && ptr[2] == 0) {
+                source = rb_str_dup(source);
+                rb_funcall(source, i_force_encoding, 1, CEncoding_UTF_16BE);
+                source = rb_funcall(source, i_encode_bang, 1, CEncoding_UTF_8);
+            } else if (len >= 4 && ptr[1] == 0 && ptr[2] == 0 && ptr[3] == 0) {
+                source = rb_str_dup(source);
+                rb_funcall(source, i_force_encoding, 1, CEncoding_UTF_32LE);
+                source = rb_funcall(source, i_encode_bang, 1, CEncoding_UTF_8);
+            } else if (len >= 4 && ptr[1] == 0 && ptr[3] == 0) {
+                source = rb_str_dup(source);
+                rb_funcall(source, i_force_encoding, 1, CEncoding_UTF_16LE);
+                source = rb_funcall(source, i_encode_bang, 1, CEncoding_UTF_8);
+            } else {
+                FORCE_UTF8(source);
+            }
+        } else {
+            source = rb_funcall(source, i_encode, 1, CEncoding_UTF_8);
+        }
+    }
+#else
     if (len >= 4 &&  ptr[0] == 0 && ptr[1] == 0 && ptr[2] == 0) {
-        rb_raise(eParserError, "Only UTF8 octet streams are supported atm!");
+      source = rb_funcall(mJSON, i_iconv, 3, rb_str_new2("utf-8"), rb_str_new2("utf-32be"), source);
     } else if (len >= 4 && ptr[0] == 0 && ptr[2] == 0) {
-        rb_raise(eParserError, "Only UTF8 octet streams are supported atm!");
+      source = rb_funcall(mJSON, i_iconv, 3, rb_str_new2("utf-8"), rb_str_new2("utf-16be"), source);
     } else if (len >= 4 && ptr[1] == 0 && ptr[2] == 0 && ptr[3] == 0) {
-        rb_raise(eParserError, "Only UTF8 octet streams are supported atm!");
+      source = rb_funcall(mJSON, i_iconv, 3, rb_str_new2("utf-8"), rb_str_new2("utf-32le"), source);
     } else if (len >= 4 && ptr[1] == 0 && ptr[3] == 0) {
-        rb_raise(eParserError, "Only UTF8 octet streams are supported atm!");
+      source = rb_funcall(mJSON, i_iconv, 3, rb_str_new2("utf-8"), rb_str_new2("utf-16le"), source);
+    }
+#endif
+    return source;
+}
+
+/*
+ * call-seq: new(source, opts => {})
+ *
+ * Creates a new JSON::Ext::Parser instance for the string _source_.
+ *
+ * Creates a new JSON::Ext::Parser instance for the string _source_.
+ *
+ * It will be configured by the _opts_ hash. _opts_ can have the following
+ * keys:
+ *
+ * _opts_ can have the following keys:
+ * * *max_nesting*: The maximum depth of nesting allowed in the parsed data
+ *   structures. Disable depth checking with :max_nesting => false|nil|0, it
+ *   defaults to 19.
+ * * *allow_nan*: If set to true, allow NaN, Infinity and -Infinity in
+ *   defiance of RFC 4627 to be parsed by the Parser. This option defaults to
+ *   false.
+ * * *symbolize_names*: If set to true, returns symbols for the names
+ *   (keys) in a JSON object. Otherwise strings are returned, which is also
+ *   the default.
+ * * *create_additions*: If set to false, the Parser doesn't create
+ *   additions even if a matchin class and create_id was found. This option
+ *   defaults to true.
+ * * *object_class*: Defaults to Hash
+ * * *array_class*: Defaults to Array
+ */
+static VALUE cParser_initialize(int argc, VALUE *argv, VALUE self)
+{
+    char *ptr;
+    long len;
+    VALUE source, opts;
+    GET_PARSER;
+    rb_scan_args(argc, argv, "11", &source, &opts);
+    source = convert_encoding(StringValue(source));
+    ptr = RSTRING_PTR(source);
+    len = RSTRING_LEN(source);
+    if (!NIL_P(opts)) {
+        opts = rb_convert_type(opts, T_HASH, "Hash", "to_hash");
+        if (NIL_P(opts)) {
+            rb_raise(rb_eArgError, "opts needs to be like a hash");
+        } else {
+            VALUE tmp = ID2SYM(i_max_nesting);
+            if (option_given_p(opts, tmp)) {
+                VALUE max_nesting = rb_hash_aref(opts, tmp);
+                if (RTEST(max_nesting)) {
+                    Check_Type(max_nesting, T_FIXNUM);
+                    json->max_nesting = FIX2INT(max_nesting);
+                } else {
+                    json->max_nesting = 0;
+                }
+            } else {
+                json->max_nesting = 19;
+            }
+            tmp = ID2SYM(i_allow_nan);
+            if (option_given_p(opts, tmp)) {
+                VALUE allow_nan = rb_hash_aref(opts, tmp);
+                json->allow_nan = RTEST(allow_nan) ? 1 : 0;
+            } else {
+                json->allow_nan = 0;
+            }
+            tmp = ID2SYM(i_symbolize_names);
+            if (option_given_p(opts, tmp)) {
+                VALUE symbolize_names = rb_hash_aref(opts, tmp);
+                json->symbolize_names = RTEST(symbolize_names) ? 1 : 0;
+            } else {
+                json->symbolize_names = 0;
+            }
+            tmp = ID2SYM(i_create_additions);
+            if (option_given_p(opts, tmp)) {
+                VALUE create_additions = rb_hash_aref(opts, tmp);
+                if (RTEST(create_additions)) {
+                    json->create_id = rb_funcall(mJSON, i_create_id, 0);
+                } else {
+                    json->create_id = Qnil;
+                }
+            } else {
+                json->create_id = rb_funcall(mJSON, i_create_id, 0);
+            }
+            tmp = ID2SYM(i_object_class);
+            if (option_given_p(opts, tmp)) {
+                json->object_class = rb_hash_aref(opts, tmp);
+            } else {
+                json->object_class = Qnil;
+            }
+            tmp = ID2SYM(i_array_class);
+            if (option_given_p(opts, tmp)) {
+                json->array_class = rb_hash_aref(opts, tmp);
+            } else {
+                json->array_class = Qnil;
+            }
+        }
+    } else {
+        json->max_nesting = 19;
+        json->allow_nan = 0;
+        json->create_id = rb_funcall(mJSON, i_create_id, 0);
+        json->object_class = Qnil;
+        json->array_class = Qnil;
     }
-    */
+    json->current_nesting = 0;
     json->len = len;
     json->source = ptr;
     json->Vsource = source;
-    json->create_id = rb_funcall(mJSON, i_create_id, 0);
     return self;
 }
 
@@ -444,10 +690,10 @@ static VALUE cParser_initialize(VALUE self, VALUE source)
  */
 static VALUE cParser_parse(VALUE self)
 {
-    GET_STRUCT;
     char *p, *pe;
     int cs = EVIL;
     VALUE result = Qnil;
+    GET_PARSER;
 
     %% write init;
     p = json->source;
@@ -457,7 +703,8 @@ static VALUE cParser_parse(VALUE self)
     if (cs >= JSON_first_final && p == pe) {
         return result;
     } else {
-        rb_raise(eParserError, "unexpected token at '%s'", p);
+        rb_raise(eParserError, "%u: unexpected token at '%s'", __LINE__, p);
+        return Qnil;
     }
 }
 
@@ -472,11 +719,13 @@ static void JSON_mark(JSON_Parser *json)
 {
     rb_gc_mark_maybe(json->Vsource);
     rb_gc_mark_maybe(json->create_id);
+    rb_gc_mark_maybe(json->object_class);
+    rb_gc_mark_maybe(json->array_class);
 }
 
 static void JSON_free(JSON_Parser *json)
 {
-    free(json);
+    ruby_xfree(json);
 }
 
 static VALUE cJSON_parser_s_allocate(VALUE klass)
@@ -493,23 +742,51 @@ static VALUE cJSON_parser_s_allocate(VALUE klass)
  */
 static VALUE cParser_source(VALUE self)
 {
-    GET_STRUCT;
+    GET_PARSER;
     return rb_str_dup(json->Vsource);
 }
 
 void Init_parser()
 {
+    rb_require("json/common");
     mJSON = rb_define_module("JSON");
     mExt = rb_define_module_under(mJSON, "Ext");
     cParser = rb_define_class_under(mExt, "Parser", rb_cObject);
     eParserError = rb_path2class("JSON::ParserError");
+    eNestingError = rb_path2class("JSON::NestingError");
     rb_define_alloc_func(cParser, cJSON_parser_s_allocate);
-    rb_define_method(cParser, "initialize", cParser_initialize, 1);
+    rb_define_method(cParser, "initialize", cParser_initialize, -1);
     rb_define_method(cParser, "parse", cParser_parse, 0);
     rb_define_method(cParser, "source", cParser_source, 0);
 
+    CNaN = rb_const_get(mJSON, rb_intern("NaN"));
+    CInfinity = rb_const_get(mJSON, rb_intern("Infinity"));
+    CMinusInfinity = rb_const_get(mJSON, rb_intern("MinusInfinity"));
+
     i_json_creatable_p = rb_intern("json_creatable?");
     i_json_create = rb_intern("json_create");
     i_create_id = rb_intern("create_id");
+    i_create_additions = rb_intern("create_additions");
     i_chr = rb_intern("chr");
+    i_max_nesting = rb_intern("max_nesting");
+    i_allow_nan = rb_intern("allow_nan");
+    i_symbolize_names = rb_intern("symbolize_names");
+    i_object_class = rb_intern("object_class");
+    i_array_class = rb_intern("array_class");
+    i_key_p = rb_intern("key?");
+    i_deep_const_get = rb_intern("deep_const_get");
+#ifdef HAVE_RUBY_ENCODING_H
+    CEncoding_UTF_8 = rb_funcall(rb_path2class("Encoding"), rb_intern("find"), 1, rb_str_new2("utf-8"));
+    CEncoding_UTF_16BE = rb_funcall(rb_path2class("Encoding"), rb_intern("find"), 1, rb_str_new2("utf-16be"));
+    CEncoding_UTF_16LE = rb_funcall(rb_path2class("Encoding"), rb_intern("find"), 1, rb_str_new2("utf-16le"));
+    CEncoding_UTF_32BE = rb_funcall(rb_path2class("Encoding"), rb_intern("find"), 1, rb_str_new2("utf-32be"));
+    CEncoding_UTF_32LE = rb_funcall(rb_path2class("Encoding"), rb_intern("find"), 1, rb_str_new2("utf-32le"));
+    CEncoding_ASCII_8BIT = rb_funcall(rb_path2class("Encoding"), rb_intern("find"), 1, rb_str_new2("ascii-8bit"));
+    i_encoding = rb_intern("encoding");
+    i_encode = rb_intern("encode");
+    i_encode_bang = rb_intern("encode!");
+    i_force_encoding = rb_intern("force_encoding");
+#else
+    i_iconv = rb_intern("iconv");
+#endif
 }