json 1.5.3 → 1.5.4

data/ext/json/ext/parser/parser.h

@@ -9,7 +9,7 @@
 
 #ifdef HAVE_RUBY_ENCODING_H
 #include "ruby/encoding.h"
-#define FORCE_UTF8(obj) rb_enc_associate((obj), rb_utf8_encoding())
+#define FORCE_UTF8(obj) ((obj) = rb_enc_associate(rb_str_dup(obj), rb_utf8_encoding()))
 #else
 #define FORCE_UTF8(obj)
 #endif
@@ -44,6 +44,7 @@ typedef struct JSON_ParserStruct {
     int allow_nan;
     int parsing_name;
     int symbolize_names;
+    int quirks_mode;
     VALUE object_class;
     VALUE array_class;
     int create_additions;
@@ -51,6 +52,9 @@ typedef struct JSON_ParserStruct {
 } JSON_Parser;
 
 #define GET_PARSER                          \
+    GET_PARSER_INIT;                        \
+    if (!json->Vsource) rb_raise(rb_eTypeError, "uninitialized instance")
+#define GET_PARSER_INIT                     \
     JSON_Parser *json;                      \
     Data_Get_Struct(self, JSON_Parser, json)
 

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

@@ -2,7 +2,7 @@
 
 /* unicode */
 
-static const char digit_values[256] = { 
+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,
@@ -38,7 +38,7 @@ static UTF32 unescape_unicode(const unsigned char *p)
     return result;
 }
 
-static int convert_UTF32_to_UTF8(char *buf, UTF32 ch) 
+static int convert_UTF32_to_UTF8(char *buf, UTF32 ch)
 {
     int len = 1;
     if (ch <= 0x7F) {
@@ -67,7 +67,7 @@ static int convert_UTF32_to_UTF8(char *buf, UTF32 ch)
 #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;
+static ID i_encoding, i_encode;
 #else
 static ID i_iconv;
 #endif
@@ -76,8 +76,9 @@ static VALUE mJSON, mExt, cParser, eParserError, eNestingError;
 static VALUE CNaN, CInfinity, CMinusInfinity;
 
 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, i_match, i_match_string, i_aset, i_leftshift;
+          i_chr, i_max_nesting, i_allow_nan, i_symbolize_names, i_quirks_mode,
+          i_object_class, i_array_class, i_key_p, i_deep_const_get, i_match,
+          i_match_string, i_aset, i_leftshift;
 
 %%{
     machine JSON_common;
@@ -97,7 +98,7 @@ static ID i_json_creatable_p, i_json_create, i_create_id, i_create_additions,
     VNaN                = 'NaN';
     VInfinity           = 'Infinity';
     VMinusInfinity      = '-Infinity';
-    begin_value         = [nft"\-[{NI] | digit;
+    begin_value         = [nft\"\-\[\{NI] | digit;
     begin_object        = '{';
     end_object          = '}';
     begin_array         = '[';
@@ -115,14 +116,14 @@ static ID i_json_creatable_p, i_json_create, i_create_id, i_create_additions,
 
     action parse_value {
         VALUE v = Qnil;
-        char *np = JSON_parse_value(json, fpc, pe, &v); 
+        char *np = JSON_parse_value(json, fpc, pe, &v);
         if (np == NULL) {
             fhold; fbreak;
         } else {
             if (NIL_P(json->object_class)) {
-              rb_hash_aset(*result, last_name, v);
+                rb_hash_aset(*result, last_name, v);
             } else {
-              rb_funcall(*result, i_aset, 2, last_name, v);
+                rb_funcall(*result, i_aset, 2, last_name, v);
             }
             fexec np;
         }
@@ -138,13 +139,14 @@ static ID i_json_creatable_p, i_json_create, i_create_id, i_create_additions,
 
     action exit { fhold; fbreak; }
 
-    a_pair  = ignore* begin_name >parse_name
-        ignore* name_separator ignore*
-        begin_value >parse_value;
+    pair  = ignore* begin_name >parse_name ignore* name_separator ignore* begin_value >parse_value;
+    next_pair   = ignore* value_separator pair;
 
-    main := begin_object
-          (a_pair (ignore* value_separator a_pair)*)?
-          ignore* end_object @exit;
+    main := (
+      begin_object
+      (pair (next_pair)*)? ignore*
+      end_object
+    ) @exit;
 }%%
 
 static char *JSON_parse_object(JSON_Parser *json, char *p, char *pe, VALUE *result)
@@ -178,6 +180,7 @@ static char *JSON_parse_object(JSON_Parser *json, char *p, char *pe, VALUE *resu
     }
 }
 
+
 %%{
     machine JSON_value;
     include JSON_common;
@@ -214,7 +217,7 @@ static char *JSON_parse_object(JSON_Parser *json, char *p, char *pe, VALUE *resu
 
     action parse_number {
         char *np;
-        if(pe > fpc + 9 && !strncmp(MinusInfinity, fpc, 9)) {
+        if(pe > fpc + 9 - json->quirks_mode && !strncmp(MinusInfinity, fpc, 9)) {
             if (json->allow_nan) {
                 *result = CMinusInfinity;
                 fexec p + 10;
@@ -230,7 +233,7 @@ static char *JSON_parse_object(JSON_Parser *json, char *p, char *pe, VALUE *resu
         fhold; fbreak;
     }
 
-    action parse_array { 
+    action parse_array {
         char *np;
         json->current_nesting++;
         np = JSON_parse_array(json, fpc, pe, result);
@@ -238,7 +241,7 @@ static char *JSON_parse_object(JSON_Parser *json, char *p, char *pe, VALUE *resu
         if (np == NULL) { fhold; fbreak; } else fexec np;
     }
 
-    action parse_object { 
+    action parse_object {
         char *np;
         json->current_nesting++;
         np =  JSON_parse_object(json, fpc, pe, result);
@@ -282,7 +285,7 @@ static char *JSON_parse_value(JSON_Parser *json, char *p, char *pe, VALUE *resul
 
     action exit { fhold; fbreak; }
 
-    main := '-'? ('0' | [1-9][0-9]*) (^[0-9] @exit);
+    main := '-'? ('0' | [1-9][0-9]*) (^[0-9]? @exit);
 }%%
 
 static char *JSON_parse_integer(JSON_Parser *json, char *p, char *pe, VALUE *result)
@@ -313,7 +316,7 @@ static char *JSON_parse_integer(JSON_Parser *json, char *p, char *pe, VALUE *res
     main := '-'? (
               (('0' | [1-9][0-9]*) '.' [0-9]+ ([Ee] [+\-]?[0-9]+)?)
               | (('0' | [1-9][0-9]*) ([Ee] [+\-]?[0-9]+))
-             )  (^[0-9Ee.\-] @exit );
+             )  (^[0-9Ee.\-]? @exit );
 }%%
 
 static char *JSON_parse_float(JSON_Parser *json, char *p, char *pe, VALUE *result)
@@ -342,7 +345,7 @@ static char *JSON_parse_float(JSON_Parser *json, char *p, char *pe, VALUE *resul
 
     action parse_value {
         VALUE v = Qnil;
-        char *np = JSON_parse_value(json, fpc, pe, &v); 
+        char *np = JSON_parse_value(json, fpc, pe, &v);
         if (np == NULL) {
             fhold; fbreak;
         } else {
@@ -419,7 +422,7 @@ static VALUE json_string_unescape(VALUE result, char *string, char *stringEnd)
                     unescape = (char *) "\f";
                     break;
                 case 'u':
-                    if (pe > stringEnd - 4) { 
+                    if (pe > stringEnd - 4) {
                         return Qnil;
                     } else {
                         char buf[4];
@@ -475,7 +478,7 @@ static VALUE json_string_unescape(VALUE result, char *string, char *stringEnd)
 
     action exit { fhold; fbreak; }
 
-    main := '"' ((^(["\\] | 0..0x1f) | '\\'["\\/bfnrt] | '\\u'[0-9a-fA-F]{4} | '\\'^(["\\/bfnrtu]|0..0x1f))* %parse_string) '"' @exit;
+    main := '"' ((^([\"\\] | 0..0x1f) | '\\'[\"\\/bfnrt] | '\\u'[0-9a-fA-F]{4} | '\\'^([\"\\/bfnrtu]|0..0x1f))* %parse_string) '"' @exit;
 }%%
 
 static int
@@ -521,35 +524,7 @@ static char *JSON_parse_string(JSON_Parser *json, char *p, char *pe, VALUE *resu
     }
 }
 
-
-%%{
-    machine JSON;
-
-    write data;
-
-    include JSON_common;
-
-    action parse_object {
-        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->current_nesting = 1;
-        np = JSON_parse_array(json, fpc, pe, &result);
-        if (np == NULL) { fhold; fbreak; } else fexec np;
-    }
-
-    main := ignore* (
-            begin_object >parse_object |
-            begin_array >parse_array
-            ) ignore*;
-}%%
-
-/* 
+/*
  * Document-class: JSON::Ext::Parser
  *
  * This is the JSON parser implemented as a C extension. It can be configured
@@ -573,22 +548,15 @@ static VALUE convert_encoding(VALUE source)
         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);
+                source = rb_funcall(source, i_encode, 2, CEncoding_UTF_8, CEncoding_UTF_32BE);
             } 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);
+                source = rb_funcall(source, i_encode, 2, CEncoding_UTF_8, CEncoding_UTF_16BE);
             } 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);
+                source = rb_funcall(source, i_encode, 2, CEncoding_UTF_8, CEncoding_UTF_32LE);
             } 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);
+                source = rb_funcall(source, i_encode, 2, CEncoding_UTF_8, CEncoding_UTF_16LE);
             } else {
+                source = rb_str_dup(source);
                 FORCE_UTF8(source);
             }
         } else {
@@ -637,14 +605,13 @@ static VALUE convert_encoding(VALUE source)
  */
 static VALUE cParser_initialize(int argc, VALUE *argv, VALUE self)
 {
-    char *ptr;
-    long len;
     VALUE source, opts;
-    GET_PARSER;
+    GET_PARSER_INIT;
+
+    if (json->Vsource) {
+        rb_raise(rb_eTypeError, "already initialized instance");
+    }
     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)) {
@@ -674,6 +641,13 @@ static VALUE cParser_initialize(int argc, VALUE *argv, VALUE self)
             } else {
                 json->symbolize_names = 0;
             }
+            tmp = ID2SYM(i_quirks_mode);
+            if (option_given_p(opts, tmp)) {
+                VALUE quirks_mode = rb_hash_aref(opts, tmp);
+                json->quirks_mode = RTEST(quirks_mode) ? 1 : 0;
+            } else {
+                json->quirks_mode = 0;
+            }
             tmp = ID2SYM(i_create_additions);
             if (option_given_p(opts, tmp)) {
                 json->create_additions = RTEST(rb_hash_aref(opts, tmp));
@@ -714,20 +688,44 @@ static VALUE cParser_initialize(int argc, VALUE *argv, VALUE self)
         json->object_class = Qnil;
         json->array_class = Qnil;
     }
+    if (!json->quirks_mode) {
+      source = convert_encoding(StringValue(source));
+    }
     json->current_nesting = 0;
-    json->len = len;
-    json->source = ptr;
+    json->len = RSTRING_LEN(source);
+    json->source = RSTRING_PTR(source);;
     json->Vsource = source;
     return self;
 }
 
-/*
- * call-seq: parse()
- *
- *  Parses the current JSON text _source_ and returns the complete data
- *  structure as a result.
- */
-static VALUE cParser_parse(VALUE self)
+%%{
+    machine JSON;
+
+    write data;
+
+    include JSON_common;
+
+    action parse_object {
+        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->current_nesting = 1;
+        np = JSON_parse_array(json, fpc, pe, &result);
+        if (np == NULL) { fhold; fbreak; } else fexec np;
+    }
+
+    main := ignore* (
+            begin_object >parse_object |
+            begin_array >parse_array
+            ) ignore*;
+}%%
+
+static VALUE cParser_parse_strict(VALUE self)
 {
     char *p, *pe;
     int cs = EVIL;
@@ -747,6 +745,62 @@ static VALUE cParser_parse(VALUE self)
     }
 }
 
+
+%%{
+    machine JSON_quirks_mode;
+
+    write data;
+
+    include JSON_common;
+
+    action parse_value {
+        char *np = JSON_parse_value(json, fpc, pe, &result);
+        if (np == NULL) { fhold; fbreak; } else fexec np;
+    }
+
+    main := ignore* (
+            begin_value >parse_value
+            ) ignore*;
+}%%
+
+static VALUE cParser_parse_quirks_mode(VALUE self)
+{
+    char *p, *pe;
+    int cs = EVIL;
+    VALUE result = Qnil;
+    GET_PARSER;
+
+    %% write init;
+    p = json->source;
+    pe = p + json->len;
+    %% write exec;
+
+    if (cs >= JSON_quirks_mode_first_final && p == pe) {
+        return result;
+    } else {
+        rb_raise(eParserError, "%u: unexpected token at '%s'", __LINE__, p);
+        return Qnil;
+    }
+}
+
+/*
+ * call-seq: parse()
+ *
+ *  Parses the current JSON text _source_ and returns the complete data
+ *  structure as a result.
+ */
+static VALUE cParser_parse(VALUE self)
+{
+  GET_PARSER;
+
+  if (json->quirks_mode) {
+    return cParser_parse_quirks_mode(self);
+  } else {
+    return cParser_parse_strict(self);
+  }
+}
+
+
 static JSON_Parser *JSON_allocate()
 {
     JSON_Parser *json = ALLOC(JSON_Parser);
@@ -786,6 +840,18 @@ static VALUE cParser_source(VALUE self)
     return rb_str_dup(json->Vsource);
 }
 
+/*
+ * call-seq: quirks_mode?()
+ *
+ * Returns a true, if this parser is in quirks_mode, false otherwise.
+ */
+static VALUE cParser_quirks_mode_p(VALUE self)
+{
+    GET_PARSER;
+    return json->quirks_mode ? Qtrue : Qfalse;
+}
+
+
 void Init_parser()
 {
     rb_require("json/common");
@@ -798,6 +864,7 @@ void Init_parser()
     rb_define_method(cParser, "initialize", cParser_initialize, -1);
     rb_define_method(cParser, "parse", cParser_parse, 0);
     rb_define_method(cParser, "source", cParser_source, 0);
+    rb_define_method(cParser, "quirks_mode?", cParser_quirks_mode_p, 0);
 
     CNaN = rb_const_get(mJSON, rb_intern("NaN"));
     CInfinity = rb_const_get(mJSON, rb_intern("Infinity"));
@@ -811,6 +878,7 @@ void Init_parser()
     i_max_nesting = rb_intern("max_nesting");
     i_allow_nan = rb_intern("allow_nan");
     i_symbolize_names = rb_intern("symbolize_names");
+    i_quirks_mode = rb_intern("quirks_mode");
     i_object_class = rb_intern("object_class");
     i_array_class = rb_intern("array_class");
     i_match = rb_intern("match");
@@ -828,9 +896,15 @@ void Init_parser()
     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
 }
+
+/*
+ * Local variables:
+ * mode: c
+ * c-file-style: ruby
+ * indent-tabs-mode: nil
+ * End:
+ */

data/java/src/json/ext/Generator.java

@@ -1,6 +1,6 @@
 /*
  * This code is copyrighted work by Daniel Luz <dev at mernen dot com>.
- * 
+ *
  * Distributed under the Ruby and GPLv2 licenses; see COPYING and GPL files
  * for details.
  */
@@ -85,11 +85,11 @@ public final class Generator {
     /**
      * A class that concentrates all the information that is shared by
      * generators working on a single session.
-     * 
+     *
      * <p>A session is defined as the process of serializing a single root
      * object; any handler directly called by container handlers (arrays and
      * hashes/objects) shares this object with its caller.
-     * 
+     *
      * <p>Note that anything called indirectly (via {@link GENERIC_HANDLER})
      * won't be part of the session.
      */
@@ -376,9 +376,9 @@ public final class Generator {
                 RubyString src;
 
                 if (info.encodingsSupported() &&
-                        object.encoding(session.getContext()) != info.utf8) {
+                        object.encoding(session.getContext()) != info.utf8.get()) {
                     src = (RubyString)object.encode(session.getContext(),
-                                                    info.utf8);
+                                                    info.utf8.get());
                 } else {
                     src = object;
                 }