json 2.1.0 → 2.6.0

data/ext/json/ext/generator/generator.c

@@ -1,11 +1,6 @@
 #include "../fbuffer/fbuffer.h"
 #include "generator.h"
 
-#ifdef HAVE_RUBY_ENCODING_H
-static VALUE CEncoding_UTF_8;
-static ID i_encoding, i_encode;
-#endif
-
 static VALUE mJSON, mExt, mGenerator, cState, mGeneratorMethods, mObject,
              mHash, mArray,
 #ifdef RUBY_INTEGER_UNIFICATION
@@ -15,14 +10,13 @@ static VALUE mJSON, mExt, mGenerator, cState, mGeneratorMethods, mObject,
 #endif
              mFloat, mString, mString_Extend,
              mTrueClass, mFalseClass, mNilClass, eGeneratorError,
-             eNestingError, CRegexp_MULTILINE, CJSON_SAFE_STATE_PROTOTYPE,
-             i_SAFE_STATE_PROTOTYPE;
+             eNestingError;
 
 static ID i_to_s, i_to_json, i_new, i_indent, i_space, i_space_before,
           i_object_nl, i_array_nl, i_max_nesting, i_allow_nan, i_ascii_only,
           i_pack, i_unpack, i_create_id, i_extend, i_key_p,
           i_aref, i_send, i_respond_to_p, i_match, i_keys, i_depth,
-          i_buffer_initial_length, i_dup;
+          i_buffer_initial_length, i_dup, i_escape_slash;
 
 /*
  * Copyright 2001-2004 Unicode, Inc.
@@ -130,7 +124,7 @@ static void unicode_escape_to_buffer(FBuffer *buffer, char buf[6], UTF16
 
 /* Converts string to a JSON string in FBuffer buffer, where all but the ASCII
  * and control characters are JSON escaped. */
-static void convert_UTF8_to_JSON_ASCII(FBuffer *buffer, VALUE string)
+static void convert_UTF8_to_JSON_ASCII(FBuffer *buffer, VALUE string, char escape_slash)
 {
     const UTF8 *source = (UTF8 *) RSTRING_PTR(string);
     const UTF8 *sourceEnd = source + RSTRING_LEN(string);
@@ -180,6 +174,11 @@ static void convert_UTF8_to_JSON_ASCII(FBuffer *buffer, VALUE string)
                         case '"':
                             fbuffer_append(buffer, "\\\"", 2);
                             break;
+                        case '/':
+                            if(escape_slash) {
+                                fbuffer_append(buffer, "\\/", 2);
+                                break;
+                            }
                         default:
                             fbuffer_append_char(buffer, (char)ch);
                             break;
@@ -229,7 +228,7 @@ static void convert_UTF8_to_JSON_ASCII(FBuffer *buffer, VALUE string)
  * characters required by the JSON standard are JSON escaped. The remaining
  * characters (should be UTF8) are just passed through and appended to the
  * result. */
-static void convert_UTF8_to_JSON(FBuffer *buffer, VALUE string)
+static void convert_UTF8_to_JSON(FBuffer *buffer, VALUE string, char escape_slash)
 {
     const char *ptr = RSTRING_PTR(string), *p;
     unsigned long len = RSTRING_LEN(string), start = 0, end = 0;
@@ -237,6 +236,7 @@ static void convert_UTF8_to_JSON(FBuffer *buffer, VALUE string)
     int escape_len;
     unsigned char c;
     char buf[6] = { '\\', 'u' };
+    int ascii_only = rb_enc_str_asciionly_p(string);
 
     for (start = 0, end = 0; end < len;) {
         p = ptr + end;
@@ -279,16 +279,25 @@ static void convert_UTF8_to_JSON(FBuffer *buffer, VALUE string)
                     escape =  "\\\"";
                     escape_len = 2;
                     break;
+                case '/':
+                    if(escape_slash) {
+                        escape = "\\/";
+                        escape_len = 2;
+                        break;
+                    }
                 default:
                     {
-                        unsigned short clen = trailingBytesForUTF8[c] + 1;
-                        if (end + clen > len) {
-                            rb_raise(rb_path2class("JSON::GeneratorError"),
-                                    "partial character in source, but hit end");
-                        }
-                        if (!isLegalUTF8((UTF8 *) p, clen)) {
-                            rb_raise(rb_path2class("JSON::GeneratorError"),
-                                    "source sequence is illegal/malformed utf-8");
+                        unsigned short clen = 1;
+                        if (!ascii_only) {
+                            clen += trailingBytesForUTF8[c];
+                            if (end + clen > len) {
+                                rb_raise(rb_path2class("JSON::GeneratorError"),
+                                        "partial character in source, but hit end");
+                            }
+                            if (!isLegalUTF8((UTF8 *) p, clen)) {
+                                rb_raise(rb_path2class("JSON::GeneratorError"),
+                                        "source sequence is illegal/malformed utf-8");
+                            }
                         }
                         end += clen;
                     }
@@ -324,6 +333,76 @@ static char *fstrndup(const char *ptr, unsigned long len) {
  *
  */
 
+/* Explanation of the following: that's the only way to not pollute
+ * standard library's docs with GeneratorMethods::<ClassName> which
+ * are uninformative and take a large place in a list of classes
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::Array
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::Bignum
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::FalseClass
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::Fixnum
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::Float
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::Hash
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::Integer
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::NilClass
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::Object
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::String
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::String::Extend
+ * :nodoc:
+ */
+
+/*
+ * Document-module: JSON::Ext::Generator::GeneratorMethods::TrueClass
+ * :nodoc:
+ */
+
 /*
  * call-seq: to_json(state = nil)
  *
@@ -535,13 +614,18 @@ static size_t State_memsize(const void *ptr)
     return size;
 }
 
+#ifndef HAVE_RB_EXT_RACTOR_SAFE
+#   undef RUBY_TYPED_FROZEN_SHAREABLE
+#   define RUBY_TYPED_FROZEN_SHAREABLE 0
+#endif
+
 #ifdef NEW_TYPEDDATA_WRAPPER
 static const rb_data_type_t JSON_Generator_State_type = {
     "JSON/Generator/State",
     {NULL, State_free, State_memsize,},
 #ifdef RUBY_TYPED_FREE_IMMEDIATELY
     0, 0,
-    RUBY_TYPED_FREE_IMMEDIATELY,
+    RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_FROZEN_SHAREABLE,
 #endif
 };
 #endif
@@ -642,6 +726,8 @@ static VALUE cState_configure(VALUE self, VALUE opts)
     state->allow_nan = RTEST(tmp);
     tmp = rb_hash_aref(opts, ID2SYM(i_ascii_only));
     state->ascii_only = RTEST(tmp);
+    tmp = rb_hash_aref(opts, ID2SYM(i_escape_slash));
+    state->escape_slash = RTEST(tmp);
     return self;
 }
 
@@ -676,6 +762,7 @@ static VALUE cState_to_h(VALUE self)
     rb_hash_aset(result, ID2SYM(i_allow_nan), state->allow_nan ? Qtrue : Qfalse);
     rb_hash_aset(result, ID2SYM(i_ascii_only), state->ascii_only ? Qtrue : Qfalse);
     rb_hash_aset(result, ID2SYM(i_max_nesting), LONG2FIX(state->max_nesting));
+    rb_hash_aset(result, ID2SYM(i_escape_slash), state->escape_slash ? Qtrue : Qfalse);
     rb_hash_aset(result, ID2SYM(i_depth), LONG2FIX(state->depth));
     rb_hash_aset(result, ID2SYM(i_buffer_initial_length), LONG2FIX(state->buffer_initial_length));
     return result;
@@ -692,7 +779,7 @@ static VALUE cState_aref(VALUE self, VALUE name)
     if (RTEST(rb_funcall(self, i_respond_to_p, 1, name))) {
         return rb_funcall(self, i_send, 1, name);
     } else {
-        return rb_ivar_get(self, rb_intern_str(rb_str_concat(rb_str_new2("@"), name)));
+        return rb_attr_get(self, rb_intern_str(rb_str_concat(rb_str_new2("@"), name)));
     }
 }
 
@@ -715,43 +802,83 @@ static VALUE cState_aset(VALUE self, VALUE name, VALUE value)
     return Qnil;
 }
 
-static void generate_json_object(FBuffer *buffer, VALUE Vstate, JSON_Generator_State *state, VALUE obj)
+struct hash_foreach_arg {
+    FBuffer *buffer;
+    JSON_Generator_State *state;
+    VALUE Vstate;
+    int iter;
+};
+
+static int
+json_object_i(VALUE key, VALUE val, VALUE _arg)
 {
+    struct hash_foreach_arg *arg = (struct hash_foreach_arg *)_arg;
+    FBuffer *buffer = arg->buffer;
+    JSON_Generator_State *state = arg->state;
+    VALUE Vstate = arg->Vstate;
+
     char *object_nl = state->object_nl;
     long object_nl_len = state->object_nl_len;
     char *indent = state->indent;
     long indent_len = state->indent_len;
-    long max_nesting = state->max_nesting;
     char *delim = FBUFFER_PTR(state->object_delim);
     long delim_len = FBUFFER_LEN(state->object_delim);
     char *delim2 = FBUFFER_PTR(state->object_delim2);
     long delim2_len = FBUFFER_LEN(state->object_delim2);
+    long depth = state->depth;
+    int j;
+    VALUE klass, key_to_s;
+
+    if (arg->iter > 0) fbuffer_append(buffer, delim, delim_len);
+    if (object_nl) {
+        fbuffer_append(buffer, object_nl, object_nl_len);
+    }
+    if (indent) {
+        for (j = 0; j < depth; j++) {
+            fbuffer_append(buffer, indent, indent_len);
+        }
+    }
+
+    klass = CLASS_OF(key);
+    if (klass == rb_cString) {
+        key_to_s = key;
+    } else if (klass == rb_cSymbol) {
+        key_to_s = rb_id2str(SYM2ID(key));
+    } else {
+        key_to_s = rb_funcall(key, i_to_s, 0);
+    }
+    Check_Type(key_to_s, T_STRING);
+    generate_json(buffer, Vstate, state, key_to_s);
+    fbuffer_append(buffer, delim2, delim2_len);
+    generate_json(buffer, Vstate, state, val);
+
+    arg->iter++;
+    return ST_CONTINUE;
+}
+
+static void generate_json_object(FBuffer *buffer, VALUE Vstate, JSON_Generator_State *state, VALUE obj)
+{
+    char *object_nl = state->object_nl;
+    long object_nl_len = state->object_nl_len;
+    char *indent = state->indent;
+    long indent_len = state->indent_len;
+    long max_nesting = state->max_nesting;
     long depth = ++state->depth;
-    int i, j;
-    VALUE key, key_to_s, keys;
+    int j;
+    struct hash_foreach_arg arg;
+
     if (max_nesting != 0 && depth > max_nesting) {
         fbuffer_free(buffer);
         rb_raise(eNestingError, "nesting of %ld is too deep", --state->depth);
     }
     fbuffer_append_char(buffer, '{');
-    keys = rb_funcall(obj, i_keys, 0);
-    for(i = 0; i < RARRAY_LEN(keys); i++) {
-        if (i > 0) fbuffer_append(buffer, delim, delim_len);
-        if (object_nl) {
-            fbuffer_append(buffer, object_nl, object_nl_len);
-        }
-        if (indent) {
-            for (j = 0; j < depth; j++) {
-                fbuffer_append(buffer, indent, indent_len);
-            }
-        }
-        key = rb_ary_entry(keys, i);
-        key_to_s = rb_funcall(key, i_to_s, 0);
-        Check_Type(key_to_s, T_STRING);
-        generate_json(buffer, Vstate, state, key_to_s);
-        fbuffer_append(buffer, delim2, delim2_len);
-        generate_json(buffer, Vstate, state, rb_hash_aref(obj, key));
-    }
+
+    arg.buffer = buffer;
+    arg.state = state;
+    arg.Vstate = Vstate;
+    arg.iter = 0;
+    rb_hash_foreach(obj, json_object_i, (VALUE)&arg);
+
     depth = --state->depth;
     if (object_nl) {
         fbuffer_append(buffer, object_nl, object_nl_len);
@@ -802,16 +929,27 @@ static void generate_json_array(FBuffer *buffer, VALUE Vstate, JSON_Generator_St
     fbuffer_append_char(buffer, ']');
 }
 
+#ifdef HAVE_RUBY_ENCODING_H
+static int enc_utf8_compatible_p(rb_encoding *enc)
+{
+    if (enc == rb_usascii_encoding()) return 1;
+    if (enc == rb_utf8_encoding()) return 1;
+    return 0;
+}
+#endif
+
 static void generate_json_string(FBuffer *buffer, VALUE Vstate, JSON_Generator_State *state, VALUE obj)
 {
     fbuffer_append_char(buffer, '"');
 #ifdef HAVE_RUBY_ENCODING_H
-    obj = rb_funcall(obj, i_encode, 1, CEncoding_UTF_8);
+    if (!enc_utf8_compatible_p(rb_enc_get(obj))) {
+        obj = rb_str_export_to_enc(obj, rb_utf8_encoding());
+    }
 #endif
     if (state->ascii_only) {
-        convert_UTF8_to_JSON_ASCII(buffer, obj);
+        convert_UTF8_to_JSON_ASCII(buffer, obj, state->escape_slash);
     } else {
-        convert_UTF8_to_JSON(buffer, obj);
+        convert_UTF8_to_JSON(buffer, obj, state->escape_slash);
     }
     fbuffer_append_char(buffer, '"');
 }
@@ -970,6 +1108,8 @@ static VALUE cState_generate(VALUE self, VALUE obj)
  * * *allow_nan*: true if NaN, Infinity, and -Infinity should be
  *   generated, otherwise an exception is thrown, if these values are
  *   encountered. This options defaults to false.
+ * * *ascii_only*: true if only ASCII characters should be generated. This
+ *   option defaults to false.
  * * *buffer_initial_length*: sets the initial length of the generator's
  *   internal buffer.
  */
@@ -1025,10 +1165,7 @@ static VALUE cState_from_state_s(VALUE self, VALUE opts)
     } else if (rb_obj_is_kind_of(opts, rb_cHash)) {
         return rb_funcall(self, i_new, 1, opts);
     } else {
-        if (NIL_P(CJSON_SAFE_STATE_PROTOTYPE)) {
-            CJSON_SAFE_STATE_PROTOTYPE = rb_const_get(mJSON, i_SAFE_STATE_PROTOTYPE);
-        }
-        return rb_funcall(CJSON_SAFE_STATE_PROTOTYPE, i_dup, 0);
+        return rb_class_new_instance(0, NULL, cState);
     }
 }
 
@@ -1252,6 +1389,31 @@ static VALUE cState_max_nesting_set(VALUE self, VALUE depth)
     return state->max_nesting = FIX2LONG(depth);
 }
 
+/*
+ * call-seq: escape_slash
+ *
+ * If this boolean is true, the forward slashes will be escaped in
+ * the json output.
+ */
+static VALUE cState_escape_slash(VALUE self)
+{
+    GET_STATE(self);
+    return state->escape_slash ? Qtrue : Qfalse;
+}
+
+/*
+ * call-seq: escape_slash=(depth)
+ *
+ * This sets whether or not the forward slashes will be escaped in
+ * the json output.
+ */
+static VALUE cState_escape_slash_set(VALUE self, VALUE enable)
+{
+    GET_STATE(self);
+    state->escape_slash = RTEST(enable);
+    return Qnil;
+}
+
 /*
  * call-seq: allow_nan?
  *
@@ -1267,7 +1429,7 @@ static VALUE cState_allow_nan_p(VALUE self)
 /*
  * call-seq: ascii_only?
  *
- * Returns true, if NaN, Infinity, and -Infinity should be generated, otherwise
+ * Returns true, if only ASCII characters should be generated. Otherwise
  * returns false.
  */
 static VALUE cState_ascii_only_p(VALUE self)
@@ -1335,6 +1497,11 @@ static VALUE cState_buffer_initial_length_set(VALUE self, VALUE buffer_initial_l
  */
 void Init_generator(void)
 {
+#ifdef HAVE_RB_EXT_RACTOR_SAFE
+    rb_ext_ractor_safe(true);
+#endif
+
+#undef rb_intern
     rb_require("json/common");
 
     mJSON = rb_define_module("JSON");
@@ -1343,6 +1510,8 @@ void Init_generator(void)
 
     eGeneratorError = rb_path2class("JSON::GeneratorError");
     eNestingError = rb_path2class("JSON::NestingError");
+    rb_gc_register_mark_object(eGeneratorError);
+    rb_gc_register_mark_object(eNestingError);
 
     cState = rb_define_class_under(mGenerator, "State", rb_cObject);
     rb_define_alloc_func(cState, cState_s_allocate);
@@ -1361,6 +1530,9 @@ void Init_generator(void)
     rb_define_method(cState, "array_nl=", cState_array_nl_set, 1);
     rb_define_method(cState, "max_nesting", cState_max_nesting, 0);
     rb_define_method(cState, "max_nesting=", cState_max_nesting_set, 1);
+    rb_define_method(cState, "escape_slash", cState_escape_slash, 0);
+    rb_define_method(cState, "escape_slash?", cState_escape_slash, 0);
+    rb_define_method(cState, "escape_slash=", cState_escape_slash_set, 1);
     rb_define_method(cState, "check_circular?", cState_check_circular_p, 0);
     rb_define_method(cState, "allow_nan?", cState_allow_nan_p, 0);
     rb_define_method(cState, "ascii_only?", cState_ascii_only_p, 0);
@@ -1408,7 +1580,6 @@ void Init_generator(void)
     mNilClass = rb_define_module_under(mGeneratorMethods, "NilClass");
     rb_define_method(mNilClass, "to_json", mNilClass_to_json, -1);
 
-    CRegexp_MULTILINE = rb_const_get(rb_cRegexp, rb_intern("MULTILINE"));
     i_to_s = rb_intern("to_s");
     i_to_json = rb_intern("to_json");
     i_new = rb_intern("new");
@@ -1418,6 +1589,7 @@ void Init_generator(void)
     i_object_nl = rb_intern("object_nl");
     i_array_nl = rb_intern("array_nl");
     i_max_nesting = rb_intern("max_nesting");
+    i_escape_slash = rb_intern("escape_slash");
     i_allow_nan = rb_intern("allow_nan");
     i_ascii_only = rb_intern("ascii_only");
     i_depth = rb_intern("depth");
@@ -1433,11 +1605,4 @@ void Init_generator(void)
     i_match = rb_intern("match");
     i_keys = rb_intern("keys");
     i_dup = rb_intern("dup");
-#ifdef HAVE_RUBY_ENCODING_H
-    CEncoding_UTF_8 = rb_funcall(rb_path2class("Encoding"), rb_intern("find"), 1, rb_str_new2("utf-8"));
-    i_encoding = rb_intern("encoding");
-    i_encode = rb_intern("encode");
-#endif
-    i_SAFE_STATE_PROTOTYPE = rb_intern("SAFE_STATE_PROTOTYPE");
-    CJSON_SAFE_STATE_PROTOTYPE = Qnil;
 }

data/ext/json/ext/generator/generator.h

@@ -49,8 +49,8 @@ static const UTF32 halfMask = 0x3FFUL;
 static unsigned char isLegalUTF8(const UTF8 *source, unsigned long length);
 static void unicode_escape(char *buf, UTF16 character);
 static void unicode_escape_to_buffer(FBuffer *buffer, char buf[6], UTF16 character);
-static void convert_UTF8_to_JSON_ASCII(FBuffer *buffer, VALUE string);
-static void convert_UTF8_to_JSON(FBuffer *buffer, VALUE string);
+static void convert_UTF8_to_JSON_ASCII(FBuffer *buffer, VALUE string, char escape_slash);
+static void convert_UTF8_to_JSON(FBuffer *buffer, VALUE string, char escape_slash);
 static char *fstrndup(const char *ptr, unsigned long len);
 
 /* ruby api and some helpers */
@@ -72,6 +72,7 @@ typedef struct JSON_Generator_StateStruct {
     long max_nesting;
     char allow_nan;
     char ascii_only;
+    char escape_slash;
     long depth;
     long buffer_initial_length;
 } JSON_Generator_State;
@@ -150,6 +151,8 @@ static VALUE cState_allow_nan_p(VALUE self);
 static VALUE cState_ascii_only_p(VALUE self);
 static VALUE cState_depth(VALUE self);
 static VALUE cState_depth_set(VALUE self, VALUE depth);
+static VALUE cState_escape_slash(VALUE self);
+static VALUE cState_escape_slash_set(VALUE self, VALUE depth);
 static FBuffer *cState_prepare_buffer(VALUE self);
 #ifndef ZALLOC
 #define ZALLOC(type) ((type *)ruby_zalloc(sizeof(type)))

data/ext/json/ext/parser/extconf.rb

@@ -2,5 +2,31 @@
 require 'mkmf'
 
 have_func("rb_enc_raise", "ruby.h")
+have_func("rb_enc_interned_str", "ruby.h")
+
+# checking if String#-@ (str_uminus) dedupes... '
+begin
+  a = -(%w(t e s t).join)
+  b = -(%w(t e s t).join)
+  if a.equal?(b)
+    $CFLAGS << ' -DSTR_UMINUS_DEDUPE=1 '
+  else
+    $CFLAGS << ' -DSTR_UMINUS_DEDUPE=0 '
+  end
+rescue NoMethodError
+  $CFLAGS << ' -DSTR_UMINUS_DEDUPE=0 '
+end
+
+# checking if String#-@ (str_uminus) directly interns frozen strings... '
+begin
+  s = rand.to_s.freeze
+  if (-s).equal?(s) && (-s.dup).equal?(s)
+    $CFLAGS << ' -DSTR_UMINUS_DEDUPE_FROZEN=1 '
+  else
+    $CFLAGS << ' -DSTR_UMINUS_DEDUPE_FROZEN=0 '
+  end
+rescue NoMethodError
+  $CFLAGS << ' -DSTR_UMINUS_DEDUPE_FROZEN=0 '
+end
 
 create_makefile 'json/ext/parser'