json 2.18.1 → 2.20.0

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

@@ -74,9 +74,6 @@ static void generate_json_string(FBuffer *buffer, struct generate_json_data *dat
 static void generate_json_null(FBuffer *buffer, struct generate_json_data *data, VALUE obj);
 static void generate_json_false(FBuffer *buffer, struct generate_json_data *data, VALUE obj);
 static void generate_json_true(FBuffer *buffer, struct generate_json_data *data, VALUE obj);
-#ifdef RUBY_INTEGER_UNIFICATION
-static void generate_json_integer(FBuffer *buffer, struct generate_json_data *data, VALUE obj);
-#endif
 static void generate_json_fixnum(FBuffer *buffer, struct generate_json_data *data, VALUE obj);
 static void generate_json_bignum(FBuffer *buffer, struct generate_json_data *data, VALUE obj);
 static void generate_json_float(FBuffer *buffer, struct generate_json_data *data, VALUE obj);
@@ -703,233 +700,6 @@ static void convert_UTF8_to_ASCII_only_JSON(search_state *search, const unsigned
     }
 }
 
-/*
- * Document-module: JSON::Ext::Generator
- *
- * This is the JSON generator implemented as a C extension. It can be
- * configured to be used by setting
- *
- *  JSON.generator = JSON::Ext::Generator
- *
- * with the method generator= in JSON.
- *
- */
-
-/* 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)
- *
- * Returns a JSON string containing a JSON object, that is generated from
- * this Hash instance.
- * _state_ is a JSON::State object, that can also be used to configure the
- * produced JSON string output further.
- */
-static VALUE mHash_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    VALUE Vstate = cState_from_state_s(cState, argc == 1 ? argv[0] : Qnil);
-    return cState_partial_generate(Vstate, self, generate_json_object, Qfalse);
-}
-
-/*
- * call-seq: to_json(state = nil)
- *
- * Returns a JSON string containing a JSON array, that is generated from
- * this Array instance.
- * _state_ is a JSON::State object, that can also be used to configure the
- * produced JSON string output further.
- */
-static VALUE mArray_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    VALUE Vstate = cState_from_state_s(cState, argc == 1 ? argv[0] : Qnil);
-    return cState_partial_generate(Vstate, self, generate_json_array, Qfalse);
-}
-
-#ifdef RUBY_INTEGER_UNIFICATION
-/*
- * call-seq: to_json(*)
- *
- * Returns a JSON string representation for this Integer number.
- */
-static VALUE mInteger_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    VALUE Vstate = cState_from_state_s(cState, argc == 1 ? argv[0] : Qnil);
-    return cState_partial_generate(Vstate, self, generate_json_integer, Qfalse);
-}
-
-#else
-/*
- * call-seq: to_json(*)
- *
- * Returns a JSON string representation for this Integer number.
- */
-static VALUE mFixnum_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    VALUE Vstate = cState_from_state_s(cState, argc == 1 ? argv[0] : Qnil);
-    return cState_partial_generate(Vstate, self, generate_json_fixnum, Qfalse);
-}
-
-/*
- * call-seq: to_json(*)
- *
- * Returns a JSON string representation for this Integer number.
- */
-static VALUE mBignum_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    VALUE Vstate = cState_from_state_s(cState, argc == 1 ? argv[0] : Qnil);
-    return cState_partial_generate(Vstate, self, generate_json_bignum, Qfalse);
-}
-#endif
-
-/*
- * call-seq: to_json(*)
- *
- * Returns a JSON string representation for this Float number.
- */
-static VALUE mFloat_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    VALUE Vstate = cState_from_state_s(cState, argc == 1 ? argv[0] : Qnil);
-    return cState_partial_generate(Vstate, self, generate_json_float, Qfalse);
-}
-
-/*
- * call-seq: to_json(*)
- *
- * This string should be encoded with UTF-8 A call to this method
- * returns a JSON string encoded with UTF16 big endian characters as
- * \u????.
- */
-static VALUE mString_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    VALUE Vstate = cState_from_state_s(cState, argc == 1 ? argv[0] : Qnil);
-    return cState_partial_generate(Vstate, self, generate_json_string, Qfalse);
-}
-
-/*
- * call-seq: to_json(*)
- *
- * Returns a JSON string for true: 'true'.
- */
-static VALUE mTrueClass_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    return rb_utf8_str_new("true", 4);
-}
-
-/*
- * call-seq: to_json(*)
- *
- * Returns a JSON string for false: 'false'.
- */
-static VALUE mFalseClass_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    return rb_utf8_str_new("false", 5);
-}
-
-/*
- * call-seq: to_json(*)
- *
- * Returns a JSON string for nil: 'null'.
- */
-static VALUE mNilClass_to_json(int argc, VALUE *argv, VALUE self)
-{
-    rb_check_arity(argc, 0, 1);
-    return rb_utf8_str_new("null", 4);
-}
-
-/*
- * call-seq: to_json(*)
- *
- * Converts this object to a string (calling #to_s), converts
- * it to a JSON string, and returns the result. This is a fallback, if no
- * special method #to_json was defined for some object.
- */
-static VALUE mObject_to_json(int argc, VALUE *argv, VALUE self)
-{
-    VALUE state;
-    VALUE string = rb_funcall(self, i_to_s, 0);
-    rb_scan_args(argc, argv, "01", &state);
-    Check_Type(string, T_STRING);
-    state = cState_from_state_s(cState, state);
-    return cState_partial_generate(state, string, generate_json_string, Qfalse);
-}
-
 static void State_mark(void *ptr)
 {
     JSON_Generator_State *state = ptr;
@@ -952,27 +722,24 @@ static void State_compact(void *ptr)
     state->as_json = rb_gc_location(state->as_json);
 }
 
-static void State_free(void *ptr)
-{
-    JSON_Generator_State *state = ptr;
-    ruby_xfree(state);
-}
-
 static size_t State_memsize(const void *ptr)
 {
+#ifdef HAVE_RUBY_TYPED_EMBEDDABLE
+    return 0;
+#else
     return sizeof(JSON_Generator_State);
+#endif
 }
 
 static const rb_data_type_t JSON_Generator_State_type = {
-    "JSON/Generator/State",
-    {
+    .wrap_struct_name = "JSON/Generator/State",
+    .function = {
         .dmark = State_mark,
-        .dfree = State_free,
+        .dfree = RUBY_DEFAULT_FREE,
         .dsize = State_memsize,
         .dcompact = State_compact,
     },
-    0, 0,
-    RUBY_TYPED_WB_PROTECTED | RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_FROZEN_SHAREABLE,
+    .flags = RUBY_TYPED_WB_PROTECTED | RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_FROZEN_SHAREABLE | RUBY_TYPED_EMBEDDABLE,
 };
 
 static void state_init(JSON_Generator_State *state)
@@ -1056,9 +823,17 @@ static VALUE encode_json_string_rescue(VALUE str, VALUE exception)
     return Qundef;
 }
 
+static inline int json_str_coderange(VALUE str) {
+    int coderange = RB_ENC_CODERANGE(str);
+    if (coderange == RUBY_ENC_CODERANGE_UNKNOWN) {
+        coderange = rb_enc_str_coderange(str);
+    }
+    return coderange;
+}
+
 static inline bool valid_json_string_p(VALUE str)
 {
-    int coderange = rb_enc_str_coderange(str);
+    int coderange = json_str_coderange(str);
 
     if (RB_LIKELY(coderange == ENC_CODERANGE_7BIT)) {
         return true;
@@ -1071,12 +846,8 @@ static inline bool valid_json_string_p(VALUE str)
     return false;
 }
 
-static inline VALUE ensure_valid_encoding(struct generate_json_data *data, VALUE str, bool as_json_called, bool is_key)
+NOINLINE(static) VALUE convert_invalid_encoding(struct generate_json_data *data, VALUE str, bool as_json_called, bool is_key)
 {
-    if (RB_LIKELY(valid_json_string_p(str))) {
-        return str;
-    }
-
     if (!as_json_called && data->state->strict && RTEST(data->state->as_json)) {
         VALUE coerced_str = json_call_as_json(data->state, str, Qfalse);
         if (coerced_str != str) {
@@ -1112,6 +883,16 @@ static inline VALUE ensure_valid_encoding(struct generate_json_data *data, VALUE
     return rb_rescue(encode_json_string_try, str, encode_json_string_rescue, str);
 }
 
+ALWAYS_INLINE(static) VALUE ensure_valid_encoding(struct generate_json_data *data, VALUE str, bool as_json_called, bool is_key)
+{
+    if (RB_LIKELY(valid_json_string_p(str))) {
+        return str;
+    }
+    else {
+        return convert_invalid_encoding(data, str, as_json_called, is_key);
+    }
+}
+
 static void raw_generate_json_string(FBuffer *buffer, struct generate_json_data *data, VALUE obj)
 {
     fbuffer_append_char(buffer, '"');
@@ -1130,7 +911,7 @@ static void raw_generate_json_string(FBuffer *buffer, struct generate_json_data
     search.chunk_end = NULL;
 #endif /* HAVE_SIMD */
 
-    switch (rb_enc_str_coderange(obj)) {
+    switch (json_str_coderange(obj)) {
         case ENC_CODERANGE_7BIT:
         case ENC_CODERANGE_VALID:
             if (RB_UNLIKELY(data->state->ascii_only)) {
@@ -1377,16 +1158,6 @@ static void generate_json_bignum(FBuffer *buffer, struct generate_json_data *dat
     fbuffer_append_str(buffer, StringValue(tmp));
 }
 
-#ifdef RUBY_INTEGER_UNIFICATION
-static void generate_json_integer(FBuffer *buffer, struct generate_json_data *data, VALUE obj)
-{
-    if (FIXNUM_P(obj))
-        generate_json_fixnum(buffer, data, obj);
-    else
-        generate_json_bignum(buffer, data, obj);
-}
-#endif
-
 static void generate_json_float(FBuffer *buffer, struct generate_json_data *data, VALUE obj)
 {
     double value = RFLOAT_VALUE(obj);
@@ -1430,7 +1201,7 @@ static void generate_json_fragment(FBuffer *buffer, struct generate_json_data *d
     fbuffer_append_str(buffer, fragment);
 }
 
-static void generate_json(FBuffer *buffer, struct generate_json_data *data, VALUE obj)
+static inline void generate_json_general(FBuffer *buffer, struct generate_json_data *data, VALUE obj, bool fallback)
 {
     bool as_json_called = false;
 start:
@@ -1457,15 +1228,15 @@ start:
                 generate_json_bignum(buffer, data, obj);
                 break;
             case T_HASH:
-                if (klass != rb_cHash) goto general;
+                if (fallback && klass != rb_cHash) goto general;
                 generate_json_object(buffer, data, obj);
                 break;
             case T_ARRAY:
-                if (klass != rb_cArray) goto general;
+                if (fallback && klass != rb_cArray) goto general;
                 generate_json_array(buffer, data, obj);
                 break;
             case T_STRING:
-                if (klass != rb_cString) goto general;
+                if (fallback && klass != rb_cString) goto general;
 
                 if (RB_LIKELY(valid_json_string_p(obj))) {
                     raw_generate_json_string(buffer, data, obj);
@@ -1481,7 +1252,7 @@ start:
                 generate_json_symbol(buffer, data, obj);
                 break;
             case T_FLOAT:
-                if (klass != rb_cFloat) goto general;
+                if (fallback && klass != rb_cFloat) goto general;
                 generate_json_float(buffer, data, obj);
                 break;
             case T_STRUCT:
@@ -1505,6 +1276,16 @@ start:
     }
 }
 
+static void generate_json(FBuffer *buffer, struct generate_json_data *data, VALUE obj)
+{
+    generate_json_general(buffer, data, obj, true);
+}
+
+static void generate_json_no_fallback(FBuffer *buffer, struct generate_json_data *data, VALUE obj)
+{
+    generate_json_general(buffer, data, obj, false);
+}
+
 static VALUE generate_json_try(VALUE d)
 {
     struct generate_json_data *data = (struct generate_json_data *)d;
@@ -1522,15 +1303,13 @@ static VALUE generate_json_ensure(VALUE d)
     return Qundef;
 }
 
-static VALUE cState_partial_generate(VALUE self, VALUE obj, generator_func func, VALUE io)
+static inline VALUE cState_partial_generate(VALUE self, VALUE obj, generator_func func, VALUE io)
 {
     GET_STATE(self);
 
     char stack_buffer[FBUFFER_STACK_SIZE];
-    FBuffer buffer = {
-        .io = RTEST(io) ? io : Qfalse,
-    };
-    fbuffer_stack_init(&buffer, state->buffer_initial_length, stack_buffer, FBUFFER_STACK_SIZE);
+    FBuffer buffer = { 0 };
+    fbuffer_init(&buffer, state->buffer_initial_length, io, stack_buffer, FBUFFER_STACK_SIZE);
 
     struct generate_json_data data = {
         .buffer = &buffer,
@@ -1540,9 +1319,7 @@ static VALUE cState_partial_generate(VALUE self, VALUE obj, generator_func func,
         .obj = obj,
         .func = func
     };
-    VALUE result = rb_ensure(generate_json_try, (VALUE)&data, generate_json_ensure, (VALUE)&data);
-    RB_GC_GUARD(self);
-    return result;
+    return rb_ensure(generate_json_try, (VALUE)&data, generate_json_ensure, (VALUE)&data);
 }
 
 /* call-seq:
@@ -1561,6 +1338,15 @@ static VALUE cState_generate(int argc, VALUE *argv, VALUE self)
     return cState_partial_generate(self, obj, generate_json, io);
 }
 
+/* :nodoc: */
+static VALUE cState_generate_no_fallback(int argc, VALUE *argv, VALUE self)
+{
+    rb_check_arity(argc, 1, 2);
+    VALUE obj = argv[0];
+    VALUE io = argc > 1 ? argv[1] : Qnil;
+    return cState_partial_generate(self, obj, generate_json_no_fallback, io);
+}
+
 static VALUE cState_initialize(int argc, VALUE *argv, VALUE self)
 {
     rb_warn("The json gem extension was loaded with the stdlib ruby code. You should upgrade rubygems with `gem update --system`");
@@ -1583,12 +1369,14 @@ static VALUE cState_init_copy(VALUE obj, VALUE orig)
     if (!objState) rb_raise(rb_eArgError, "unallocated JSON::State");
 
     MEMCPY(objState, origState, JSON_Generator_State, 1);
-    objState->indent = origState->indent;
-    objState->space = origState->space;
-    objState->space_before = origState->space_before;
-    objState->object_nl = origState->object_nl;
-    objState->array_nl = origState->array_nl;
-    objState->as_json = origState->as_json;
+
+    RB_OBJ_WRITTEN(obj, Qundef, objState->indent);
+    RB_OBJ_WRITTEN(obj, Qundef, objState->space);
+    RB_OBJ_WRITTEN(obj, Qundef, objState->space_before);
+    RB_OBJ_WRITTEN(obj, Qundef, objState->object_nl);
+    RB_OBJ_WRITTEN(obj, Qundef, objState->array_nl);
+    RB_OBJ_WRITTEN(obj, Qundef, objState->as_json);
+
     return obj;
 }
 
@@ -1795,7 +1583,21 @@ static VALUE cState_max_nesting(VALUE self)
 
 static long long_config(VALUE num)
 {
-    return RTEST(num) ? FIX2LONG(num) : 0;
+    return RTEST(num) ? NUM2LONG(num) : 0;
+}
+
+// depth must never be negative; reject early with a clear error.
+static long depth_config(VALUE num)
+{
+    if (!RTEST(num)) return 0;
+    long d = NUM2LONG(num);
+    if (RB_UNLIKELY(d < 0)) {
+        rb_raise(rb_eArgError, "depth must be >= 0 (got %ld)", d);
+    }
+    if (RB_UNLIKELY(d > INT_MAX)) {
+        rb_raise(rb_eArgError, "depth is too large (got %ld)", d);
+    }
+    return d;
 }
 
 /*
@@ -1954,7 +1756,7 @@ static VALUE cState_depth_set(VALUE self, VALUE depth)
 {
     rb_check_frozen(self);
     GET_STATE(self);
-    state->depth = long_config(depth);
+    state->depth = depth_config(depth);
     return Qnil;
 }
 
@@ -2019,7 +1821,7 @@ static int configure_state_i(VALUE key, VALUE val, VALUE _arg)
     else if (key == sym_max_nesting)           { state->max_nesting = long_config(val); }
     else if (key == sym_allow_nan)             { state->allow_nan = RTEST(val); }
     else if (key == sym_ascii_only)            { state->ascii_only = RTEST(val); }
-    else if (key == sym_depth)                 { state->depth = long_config(val); }
+    else if (key == sym_depth)                 { state->depth = depth_config(val); }
     else if (key == sym_buffer_initial_length) { buffer_initial_length_set(state, val); }
     else if (key == sym_script_safe)           { state->script_safe = RTEST(val); }
     else if (key == sym_escape_slash)          { state->script_safe = RTEST(val); }
@@ -2059,17 +1861,15 @@ static VALUE cState_configure(VALUE self, VALUE opts)
     return self;
 }
 
-static VALUE cState_m_generate(VALUE klass, VALUE obj, VALUE opts, VALUE io)
+static VALUE cState_m_do_generate(VALUE klass, VALUE obj, VALUE opts, VALUE io, generator_func func)
 {
     JSON_Generator_State state = {0};
     state_init(&state);
     configure_state(&state, Qfalse, opts);
 
     char stack_buffer[FBUFFER_STACK_SIZE];
-    FBuffer buffer = {
-        .io = RTEST(io) ? io : Qfalse,
-    };
-    fbuffer_stack_init(&buffer, state.buffer_initial_length, stack_buffer, FBUFFER_STACK_SIZE);
+    FBuffer buffer = { 0 };
+    fbuffer_init(&buffer, state.buffer_initial_length, io, stack_buffer, FBUFFER_STACK_SIZE);
 
     struct generate_json_data data = {
         .buffer = &buffer,
@@ -2077,14 +1877,21 @@ static VALUE cState_m_generate(VALUE klass, VALUE obj, VALUE opts, VALUE io)
         .state = &state,
         .depth = state.depth,
         .obj = obj,
-        .func = generate_json,
+        .func = func,
     };
     return rb_ensure(generate_json_try, (VALUE)&data, generate_json_ensure, (VALUE)&data);
 }
 
-/*
- *
- */
+static VALUE cState_m_generate(VALUE klass, VALUE obj, VALUE opts, VALUE io)
+{
+    return cState_m_do_generate(klass, obj, opts, io, generate_json);
+}
+
+static VALUE cState_m_generate_no_fallback(VALUE klass, VALUE obj, VALUE opts, VALUE io)
+{
+    return cState_m_do_generate(klass, obj, opts, io, generate_json_no_fallback);
+}
+
 void Init_generator(void)
 {
 #ifdef HAVE_RB_EXT_RACTOR_SAFE
@@ -2149,46 +1956,12 @@ void Init_generator(void)
     rb_define_method(cState, "buffer_initial_length", cState_buffer_initial_length, 0);
     rb_define_method(cState, "buffer_initial_length=", cState_buffer_initial_length_set, 1);
     rb_define_method(cState, "generate", cState_generate, -1);
+    rb_define_method(cState, "_generate_no_fallback", cState_generate_no_fallback, -1);
 
     rb_define_private_method(cState, "allow_duplicate_key?", cState_allow_duplicate_key_p, 0);
 
     rb_define_singleton_method(cState, "generate", cState_m_generate, 3);
-
-    VALUE mGeneratorMethods = rb_define_module_under(mGenerator, "GeneratorMethods");
-
-    VALUE mObject = rb_define_module_under(mGeneratorMethods, "Object");
-    rb_define_method(mObject, "to_json", mObject_to_json, -1);
-
-    VALUE mHash = rb_define_module_under(mGeneratorMethods, "Hash");
-    rb_define_method(mHash, "to_json", mHash_to_json, -1);
-
-    VALUE mArray = rb_define_module_under(mGeneratorMethods, "Array");
-    rb_define_method(mArray, "to_json", mArray_to_json, -1);
-
-#ifdef RUBY_INTEGER_UNIFICATION
-    VALUE mInteger = rb_define_module_under(mGeneratorMethods, "Integer");
-    rb_define_method(mInteger, "to_json", mInteger_to_json, -1);
-#else
-    VALUE mFixnum = rb_define_module_under(mGeneratorMethods, "Fixnum");
-    rb_define_method(mFixnum, "to_json", mFixnum_to_json, -1);
-
-    VALUE mBignum = rb_define_module_under(mGeneratorMethods, "Bignum");
-    rb_define_method(mBignum, "to_json", mBignum_to_json, -1);
-#endif
-    VALUE mFloat = rb_define_module_under(mGeneratorMethods, "Float");
-    rb_define_method(mFloat, "to_json", mFloat_to_json, -1);
-
-    VALUE mString = rb_define_module_under(mGeneratorMethods, "String");
-    rb_define_method(mString, "to_json", mString_to_json, -1);
-
-    VALUE mTrueClass = rb_define_module_under(mGeneratorMethods, "TrueClass");
-    rb_define_method(mTrueClass, "to_json", mTrueClass_to_json, -1);
-
-    VALUE mFalseClass = rb_define_module_under(mGeneratorMethods, "FalseClass");
-    rb_define_method(mFalseClass, "to_json", mFalseClass_to_json, -1);
-
-    VALUE mNilClass = rb_define_module_under(mGeneratorMethods, "NilClass");
-    rb_define_method(mNilClass, "to_json", mNilClass_to_json, -1);
+    rb_define_singleton_method(cState, "_generate_no_fallback", cState_m_generate_no_fallback, 3);
 
     rb_global_variable(&Encoding_UTF_8);
     Encoding_UTF_8 = rb_const_get(rb_path2class("Encoding"), rb_intern("UTF_8"));