json 2.8.1 → 2.9.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 39c161c94f7ae85545dc117ec7616aa65f69e9192f081dd1d0934fc57f7a1b9b
-  data.tar.gz: f980984efc78b2bf8b0aa81e67dc8115e503bcb03171ce7658bdc9ba653da0b7
+  metadata.gz: 583ef6b3c8caf7e7fc9ad5ce6e60b1336ca23438cbd2ed998bea27ffbe8fbf38
+  data.tar.gz: 4ee0a25a7f0854aa99c0c79f9cdbd4038ac1a00884a840e49d119ab21b229ca1
 SHA512:
-  metadata.gz: 02e0733a4e893e0c486bc3ea2785c15601c5e660d990c8c83543df3df6f93acdb41d29e656f951896de8854f9672d89bae462f7761746e27614e038b48f72296
-  data.tar.gz: 5b22f7e7aa15e8965293fb91ac5d45631f5ef16dbaf0911f5d0763175138dfde8bb104dcea2df73c5e99fda45aecde081c971ccd2bc7c2dae9b4d700376ef8e3
+  metadata.gz: 6ab5ed9d48b51602d00ace1657947c82bf3bb233eb163baa93b582c32c9eb1cc23dec69e88ed94cb26fcabf6dbb7c355f0b67696186345e911beb9736e504aa0
+  data.tar.gz: 7fb0db017a55c3598004d7e5f640959ff50dc37a7e1ebe9a9e04462ad76f806bd599f7468891b612eaa4648a0c33ca05c3c9183c24489204128f49b2ef149652

data/CHANGES.md

@@ -1,5 +1,17 @@
 # Changes
 
+### 2024-12-03 (2.9.0)
+
+* Fix C implementation of `script_safe` escaping to not confuse some other 3 wide characters with `\u2028` and `\u2029`.
+  e.g. `JSON.generate(["倩", "瀨"], script_safe: true)` would generate the wrong JSON.
+* `JSON.dump(object, some_io)` now write into the IO in chunks while previously it would buffer the entire JSON before writing.
+* `JSON::GeneratorError` now has a `#invalid_object` attribute, making it easier to understand why an object tree cannot be serialized.
+* Numerous improvements to the JRuby extension.
+
+### 2024-11-14 (2.8.2)
+
+* `JSON.load_file` explictly read the file as UTF-8.
+
 ### 2024-11-06 (2.8.1)
 
 * Fix the java packages to include the extension.
@@ -14,6 +26,7 @@
   pre-existing support for comments, make it suitable to parse `jsonc` documents.
 * Many performance improvements to `JSON.parse` and `JSON.load`, up to `1.7x` faster on real world documents.
 * Some minor performance improvements to `JSON.dump` and `JSON.generate`.
+* `JSON.pretty_generate` no longer include newline inside empty object and arrays. 
 
 ### 2024-11-04 (2.7.6)
 

data/ext/json/ext/fbuffer/fbuffer.h

@@ -46,9 +46,11 @@ typedef struct FBufferStruct {
     unsigned long len;
     unsigned long capa;
     char *ptr;
+    VALUE io;
 } FBuffer;
 
 #define FBUFFER_STACK_SIZE 512
+#define FBUFFER_IO_BUFFER_SIZE (16384 - 1)
 #define FBUFFER_INITIAL_LENGTH_DEFAULT 1024
 
 #define FBUFFER_PTR(fb) ((fb)->ptr)
@@ -66,7 +68,7 @@ static void fbuffer_append_long(FBuffer *fb, long number);
 #endif
 static inline void fbuffer_append_char(FBuffer *fb, char newchr);
 #ifdef JSON_GENERATOR
-static VALUE fbuffer_to_s(FBuffer *fb);
+static VALUE fbuffer_finalize(FBuffer *fb);
 #endif
 
 static void fbuffer_stack_init(FBuffer *fb, unsigned long initial_length, char *stack_buffer, long stack_buffer_size)
@@ -86,24 +88,19 @@ static void fbuffer_free(FBuffer *fb)
     }
 }
 
-#ifndef JSON_GENERATOR
 static void fbuffer_clear(FBuffer *fb)
 {
     fb->len = 0;
 }
-#endif
 
-static void fbuffer_do_inc_capa(FBuffer *fb, unsigned long requested)
+static void fbuffer_flush(FBuffer *fb)
 {
-    unsigned long required;
-
-    if (RB_UNLIKELY(!fb->ptr)) {
-        fb->ptr = ALLOC_N(char, fb->initial_length);
-        fb->capa = fb->initial_length;
-    }
-
-    for (required = fb->capa; requested > required - fb->len; required <<= 1);
+    rb_io_write(fb->io, rb_utf8_str_new(fb->ptr, fb->len));
+    fbuffer_clear(fb);
+}
 
+static void fbuffer_realloc(FBuffer *fb, unsigned long required)
+{
     if (required > fb->capa) {
         if (fb->type == FBUFFER_STACK_ALLOCATED) {
             const char *old_buffer = fb->ptr;
@@ -117,6 +114,32 @@ static void fbuffer_do_inc_capa(FBuffer *fb, unsigned long requested)
     }
 }
 
+static void fbuffer_do_inc_capa(FBuffer *fb, unsigned long requested)
+{
+    if (RB_UNLIKELY(fb->io)) {
+        if (fb->capa < FBUFFER_IO_BUFFER_SIZE) {
+            fbuffer_realloc(fb, FBUFFER_IO_BUFFER_SIZE);
+        } else {
+            fbuffer_flush(fb);
+        }
+
+        if (RB_LIKELY(requested < fb->capa)) {
+            return;
+        }
+    }
+
+    unsigned long required;
+
+    if (RB_UNLIKELY(!fb->ptr)) {
+        fb->ptr = ALLOC_N(char, fb->initial_length);
+        fb->capa = fb->initial_length;
+    }
+
+    for (required = fb->capa; requested > required - fb->len; required <<= 1);
+
+    fbuffer_realloc(fb, required);
+}
+
 static inline void fbuffer_inc_capa(FBuffer *fb, unsigned long requested)
 {
     if (RB_UNLIKELY(requested > fb->capa - fb->len)) {
@@ -174,11 +197,18 @@ static void fbuffer_append_long(FBuffer *fb, long number)
     fbuffer_append(fb, buffer_end - len, len);
 }
 
-static VALUE fbuffer_to_s(FBuffer *fb)
+static VALUE fbuffer_finalize(FBuffer *fb)
 {
-    VALUE result = rb_utf8_str_new(FBUFFER_PTR(fb), FBUFFER_LEN(fb));
-    fbuffer_free(fb);
-    return result;
+    if (fb->io) {
+        fbuffer_flush(fb);
+        fbuffer_free(fb);
+        rb_io_flush(fb->io);
+        return fb->io;
+    } else {
+        VALUE result = rb_utf8_str_new(FBUFFER_PTR(fb), FBUFFER_LEN(fb));
+        fbuffer_free(fb);
+        return result;
+    }
 }
 #endif
 #endif

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

@@ -54,7 +54,7 @@ struct generate_json_data {
 };
 
 static VALUE cState_from_state_s(VALUE self, VALUE opts);
-static VALUE cState_partial_generate(VALUE self, VALUE obj, generator_func);
+static VALUE cState_partial_generate(VALUE self, VALUE obj, generator_func, VALUE io);
 static void generate_json(FBuffer *buffer, struct generate_json_data *data, JSON_Generator_State *state, VALUE obj);
 static void generate_json_object(FBuffer *buffer, struct generate_json_data *data, JSON_Generator_State *state, VALUE obj);
 static void generate_json_array(FBuffer *buffer, struct generate_json_data *data, JSON_Generator_State *state, VALUE obj);
@@ -71,6 +71,31 @@ static void generate_json_float(FBuffer *buffer, struct generate_json_data *data
 
 static int usascii_encindex, utf8_encindex, binary_encindex;
 
+#ifdef RBIMPL_ATTR_NORETURN
+RBIMPL_ATTR_NORETURN()
+#endif
+static void raise_generator_error_str(VALUE invalid_object, VALUE str)
+{
+    VALUE exc = rb_exc_new_str(eGeneratorError, str);
+    rb_ivar_set(exc, rb_intern("@invalid_object"), invalid_object);
+    rb_exc_raise(exc);
+}
+
+#ifdef RBIMPL_ATTR_NORETURN
+RBIMPL_ATTR_NORETURN()
+#endif
+#ifdef RBIMPL_ATTR_FORMAT
+RBIMPL_ATTR_FORMAT(RBIMPL_PRINTF_FORMAT, 2, 3)
+#endif
+static void raise_generator_error(VALUE invalid_object, const char *fmt, ...)
+{
+    va_list args;
+    va_start(args, fmt);
+    VALUE str = rb_vsprintf(fmt, args);
+    va_end(args);
+    raise_generator_error_str(invalid_object, str);
+}
+
 /* Converts in_string to a JSON string (without the wrapping '"'
  * characters) in FBuffer out_buffer.
  *
@@ -130,7 +155,7 @@ static void convert_UTF8_to_JSON(FBuffer *out_buffer, VALUE str, const char esca
                 }
                 case 3: {
                     unsigned char b2 = ptr[pos + 1];
-                    if (RB_UNLIKELY(out_script_safe && b2 == 0x80)) {
+                    if (RB_UNLIKELY(out_script_safe && ch == 0xE2 && b2 == 0x80)) {
                         unsigned char b3 = ptr[pos + 2];
                         if (b3 == 0xA8) {
                             FLUSH_POS(3);
@@ -453,7 +478,7 @@ 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);
+    return cState_partial_generate(Vstate, self, generate_json_object, Qfalse);
 }
 
 /*
@@ -467,7 +492,7 @@ static VALUE mHash_to_json(int argc, VALUE *argv, VALUE self)
 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);
+    return cState_partial_generate(Vstate, self, generate_json_array, Qfalse);
 }
 
 #ifdef RUBY_INTEGER_UNIFICATION
@@ -480,7 +505,7 @@ 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);
+    return cState_partial_generate(Vstate, self, generate_json_integer, Qfalse);
 }
 
 #else
@@ -493,7 +518,7 @@ 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);
+    return cState_partial_generate(Vstate, self, generate_json_fixnum, Qfalse);
 }
 
 /*
@@ -505,7 +530,7 @@ 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);
+    return cState_partial_generate(Vstate, self, generate_json_bignum, Qfalse);
 }
 #endif
 
@@ -518,7 +543,7 @@ 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);
+    return cState_partial_generate(Vstate, self, generate_json_float, Qfalse);
 }
 
 /*
@@ -543,7 +568,7 @@ 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);
+    return cState_partial_generate(Vstate, self, generate_json_string, Qfalse);
 }
 
 /*
@@ -638,7 +663,7 @@ static VALUE mObject_to_json(int argc, VALUE *argv, VALUE self)
     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);
+    return cState_partial_generate(state, string, generate_json_string, Qfalse);
 }
 
 static void State_mark(void *ptr)
@@ -867,6 +892,17 @@ static inline int enc_utf8_compatible_p(int enc_idx)
     return 0;
 }
 
+static VALUE encode_json_string_try(VALUE str)
+{
+    return rb_funcall(str, i_encode, 1, Encoding_UTF_8);
+}
+
+static VALUE encode_json_string_rescue(VALUE str, VALUE exception)
+{
+    raise_generator_error_str(str, rb_funcall(exception, rb_intern("message"), 0));
+    return Qundef;
+}
+
 static inline VALUE ensure_valid_encoding(VALUE str)
 {
     int encindex = RB_ENCODING_GET(str);
@@ -886,7 +922,7 @@ static inline VALUE ensure_valid_encoding(VALUE str)
             }
         }
 
-        str = rb_funcall(str, i_encode, 1, Encoding_UTF_8);
+        str = rb_rescue(encode_json_string_try, str, encode_json_string_rescue, str);
     }
     return str;
 }
@@ -909,7 +945,7 @@ static void generate_json_string(FBuffer *buffer, struct generate_json_data *dat
             }
             break;
         default:
-            rb_raise(rb_path2class("JSON::GeneratorError"), "source sequence is illegal/malformed utf-8");
+            raise_generator_error(obj, "source sequence is illegal/malformed utf-8");
             break;
     }
     fbuffer_append_char(buffer, '"');
@@ -957,10 +993,8 @@ static void generate_json_float(FBuffer *buffer, struct generate_json_data *data
     char allow_nan = state->allow_nan;
     VALUE tmp = rb_funcall(obj, i_to_s, 0);
     if (!allow_nan) {
-        if (isinf(value)) {
-            rb_raise(eGeneratorError, "%"PRIsVALUE" not allowed in JSON", tmp);
-        } else if (isnan(value)) {
-            rb_raise(eGeneratorError, "%"PRIsVALUE" not allowed in JSON", tmp);
+        if (isinf(value) || isnan(value)) {
+            raise_generator_error(obj, "%"PRIsVALUE" not allowed in JSON", tmp);
         }
     }
     fbuffer_append_str(buffer, tmp);
@@ -1008,7 +1042,7 @@ static void generate_json(FBuffer *buffer, struct generate_json_data *data, JSON
             default:
             general:
                 if (state->strict) {
-                    rb_raise(eGeneratorError, "%"PRIsVALUE" not allowed in JSON", CLASS_OF(obj));
+                    raise_generator_error(obj, "%"PRIsVALUE" not allowed in JSON", CLASS_OF(obj));
                 } else if (rb_respond_to(obj, i_to_json)) {
                     tmp = rb_funcall(obj, i_to_json, 1, vstate_get(data));
                     Check_Type(tmp, T_STRING);
@@ -1036,21 +1070,19 @@ static VALUE generate_json_rescue(VALUE d, VALUE exc)
     struct generate_json_data *data = (struct generate_json_data *)d;
     fbuffer_free(data->buffer);
 
-    if (RBASIC_CLASS(exc) == rb_path2class("Encoding::UndefinedConversionError")) {
-        exc = rb_exc_new_str(eGeneratorError, rb_funcall(exc, rb_intern("message"), 0));
-    }
-
     rb_exc_raise(exc);
 
     return Qundef;
 }
 
-static VALUE cState_partial_generate(VALUE self, VALUE obj, generator_func func)
+static VALUE cState_partial_generate(VALUE self, VALUE obj, generator_func func, VALUE io)
 {
     GET_STATE(self);
 
     char stack_buffer[FBUFFER_STACK_SIZE];
-    FBuffer buffer = {0};
+    FBuffer buffer = {
+        .io = RTEST(io) ? io : Qfalse,
+    };
     fbuffer_stack_init(&buffer, state->buffer_initial_length, stack_buffer, FBUFFER_STACK_SIZE);
 
     struct generate_json_data data = {
@@ -1062,19 +1094,12 @@ static VALUE cState_partial_generate(VALUE self, VALUE obj, generator_func func)
     };
     rb_rescue(generate_json_try, (VALUE)&data, generate_json_rescue, (VALUE)&data);
 
-    return fbuffer_to_s(&buffer);
+    return fbuffer_finalize(&buffer);
 }
 
-/*
- * call-seq: generate(obj)
- *
- * Generates a valid JSON document from object +obj+ and returns the
- * result. If no valid JSON document can be created this method raises a
- * GeneratorError exception.
- */
-static VALUE cState_generate(VALUE self, VALUE obj)
+static VALUE cState_generate(VALUE self, VALUE obj, VALUE io)
 {
-    VALUE result = cState_partial_generate(self, obj, generate_json);
+    VALUE result = cState_partial_generate(self, obj, generate_json, io);
     GET_STATE(self);
     (void)state;
     return result;
@@ -1502,14 +1527,16 @@ static VALUE cState_configure(VALUE self, VALUE opts)
     return self;
 }
 
-static VALUE cState_m_generate(VALUE klass, VALUE obj, VALUE opts)
+static VALUE cState_m_generate(VALUE klass, VALUE obj, VALUE opts, VALUE io)
 {
     JSON_Generator_State state = {0};
     state_init(&state);
     configure_state(&state, opts);
 
     char stack_buffer[FBUFFER_STACK_SIZE];
-    FBuffer buffer = {0};
+    FBuffer buffer = {
+        .io = RTEST(io) ? io : Qfalse,
+    };
     fbuffer_stack_init(&buffer, state.buffer_initial_length, stack_buffer, FBUFFER_STACK_SIZE);
 
     struct generate_json_data data = {
@@ -1521,7 +1548,7 @@ static VALUE cState_m_generate(VALUE klass, VALUE obj, VALUE opts)
     };
     rb_rescue(generate_json_try, (VALUE)&data, generate_json_rescue, (VALUE)&data);
 
-    return fbuffer_to_s(&buffer);
+    return fbuffer_finalize(&buffer);
 }
 
 /*
@@ -1540,10 +1567,11 @@ void Init_generator(void)
     VALUE mExt = rb_define_module_under(mJSON, "Ext");
     VALUE mGenerator = rb_define_module_under(mExt, "Generator");
 
+    rb_global_variable(&eGeneratorError);
     eGeneratorError = rb_path2class("JSON::GeneratorError");
+
+    rb_global_variable(&eNestingError);
     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);
@@ -1583,9 +1611,9 @@ void Init_generator(void)
     rb_define_method(cState, "depth=", cState_depth_set, 1);
     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_private_method(cState, "_generate", cState_generate, 2);
 
-    rb_define_singleton_method(cState, "generate", cState_m_generate, 2);
+    rb_define_singleton_method(cState, "generate", cState_m_generate, 3);
 
     VALUE mGeneratorMethods = rb_define_module_under(mGenerator, "GeneratorMethods");
 

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

@@ -3,7 +3,6 @@ require 'mkmf'
 
 have_func("rb_enc_interned_str", "ruby.h") # RUBY_VERSION >= 3.0
 have_func("rb_hash_new_capa", "ruby.h") # RUBY_VERSION >= 3.2
-have_func("rb_gc_mark_locations", "ruby.h") # Missing on TruffleRuby
 have_func("rb_hash_bulk_insert", "ruby.h") # Missing on TruffleRuby
 have_func("rb_category_warn", "ruby.h") # Missing on TruffleRuby