Merge branch 'strict-dump' of github.com:Shopify/json into Shopify-strict-dump

Author: hsbt

ext/json/ext/generator/generator.c

@@ -16,7 +16,7 @@ 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_script_safe, i_escape_slash;
+          i_buffer_initial_length, i_dup, i_script_safe, i_escape_slash, i_strict;
 
 /*
  * Copyright 2001-2004 Unicode, Inc.
@@ -749,6 +749,8 @@ static VALUE cState_configure(VALUE self, VALUE opts)
         tmp = rb_hash_aref(opts, ID2SYM(i_escape_slash));
         state->script_safe = RTEST(tmp);
     }
+    tmp = rb_hash_aref(opts, ID2SYM(i_strict));
+    state->strict = RTEST(tmp);
     return self;
 }
 
@@ -784,6 +786,7 @@ static VALUE cState_to_h(VALUE self)
     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_script_safe), state->script_safe ? Qtrue : Qfalse);
+    rb_hash_aset(result, ID2SYM(i_strict), state->strict ? 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;
@@ -1049,6 +1052,8 @@ static void generate_json(FBuffer *buffer, VALUE Vstate, JSON_Generator_State *s
         generate_json_bignum(buffer, Vstate, state, obj);
     } else if (klass == rb_cFloat) {
         generate_json_float(buffer, Vstate, state, obj);
+    } else if (state->strict) {
+        rb_raise(eGeneratorError, "%"PRIsVALUE" not allowed in JSON", RB_OBJ_STRING(CLASS_OF(obj)));
     } else if (rb_respond_to(obj, i_to_json)) {
         tmp = rb_funcall(obj, i_to_json, 1, Vstate);
         Check_Type(tmp, T_STRING);
@@ -1423,7 +1428,7 @@ static VALUE cState_script_safe(VALUE self)
 }
 
 /*
- * call-seq: script_safe=(depth)
+ * call-seq: script_safe=(enable)
  *
  * This sets whether or not the forward slashes will be escaped in
  * the json output.
@@ -1435,6 +1440,37 @@ static VALUE cState_script_safe_set(VALUE self, VALUE enable)
     return Qnil;
 }
 
+/*
+ * call-seq: strict
+ *
+ * If this boolean is false, types unsupported by the JSON format will
+ * be serialized as strings.
+ * If this boolean is true, types unsupported by the JSON format will
+ * raise a JSON::GeneratorError.
+ */
+static VALUE cState_strict(VALUE self)
+{
+    GET_STATE(self);
+    return state->strict ? Qtrue : Qfalse;
+}
+
+/*
+ * call-seq: strict=(enable)
+ *
+ * This sets whether or not to serialize types unsupported by the
+ * JSON format as strings.
+ * If this boolean is false, types unsupported by the JSON format will
+ * be serialized as strings.
+ * If this boolean is true, types unsupported by the JSON format will
+ * raise a JSON::GeneratorError.
+ */
+static VALUE cState_strict_set(VALUE self, VALUE enable)
+{
+    GET_STATE(self);
+    state->strict = RTEST(enable);
+    return Qnil;
+}
+
 /*
  * call-seq: allow_nan?
  *
@@ -1557,6 +1593,9 @@ void Init_generator(void)
     rb_define_alias(cState, "escape_slash", "script_safe");
     rb_define_alias(cState, "escape_slash?", "script_safe?");
     rb_define_alias(cState, "escape_slash=", "script_safe=");
+    rb_define_method(cState, "strict", cState_strict, 0);
+    rb_define_method(cState, "strict?", cState_strict, 0);
+    rb_define_method(cState, "strict=", cState_strict_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);
@@ -1615,6 +1654,7 @@ void Init_generator(void)
     i_max_nesting = rb_intern("max_nesting");
     i_script_safe = rb_intern("script_safe");
     i_escape_slash = rb_intern("escape_slash");
+    i_strict = rb_intern("strict");
     i_allow_nan = rb_intern("allow_nan");
     i_ascii_only = rb_intern("ascii_only");
     i_depth = rb_intern("depth");

ext/json/ext/generator/generator.h

@@ -73,6 +73,7 @@ typedef struct JSON_Generator_StateStruct {
     char allow_nan;
     char ascii_only;
     char script_safe;
+    char strict;
     long depth;
     long buffer_initial_length;
 } JSON_Generator_State;
@@ -153,6 +154,8 @@ static VALUE cState_depth(VALUE self);
 static VALUE cState_depth_set(VALUE self, VALUE depth);
 static VALUE cState_script_safe(VALUE self);
 static VALUE cState_script_safe_set(VALUE self, VALUE depth);
+static VALUE cState_strict(VALUE self);
+static VALUE cState_strict_set(VALUE self, VALUE strict);
 static FBuffer *cState_prepare_buffer(VALUE self);
 #ifndef ZALLOC
 #define ZALLOC(type) ((type *)ruby_zalloc(sizeof(type)))

java/src/json/ext/Generator.java

@@ -428,7 +428,11 @@ void generate(Session session, IRubyObject object, ByteList buffer) {
         new Handler<IRubyObject>() {
             @Override
             RubyString generateNew(Session session, IRubyObject object) {
-                if (object.respondsTo("to_json")) {
+                if (session.getState().strict()) {
+                    throw Utils.newException(session.getContext(),
+                            Utils.M_GENERATOR_ERROR,
+                            object + " not allowed in JSON");
+                } else if (object.respondsTo("to_json")) {
                     IRubyObject result = object.callMethod(session.getContext(), "to_json",
                               new IRubyObject[] {session.getState()});
                     if (result instanceof RubyString) return (RubyString)result;

java/src/json/ext/GeneratorState.java

@@ -88,6 +88,12 @@ public class GeneratorState extends RubyObject {
      */
     private boolean scriptSafe = DEFAULT_SCRIPT_SAFE;
     static final boolean DEFAULT_SCRIPT_SAFE = false;
+    /**
+     * If set to <code>true</code> types unsupported by the JSON format will
+     * raise a <code>JSON::GeneratorError</code>.
+     */
+    private boolean strict = DEFAULT_STRICT;
+    static final boolean DEFAULT_STRICT = false;
     /**
      * The initial buffer length of this state. (This isn't really used on all
      * non-C implementations.)
@@ -204,6 +210,7 @@ public IRubyObject initialize_copy(ThreadContext context, IRubyObject vOrig) {
         this.asciiOnly = orig.asciiOnly;
         this.quirksMode = orig.quirksMode;
         this.scriptSafe = orig.scriptSafe;
+        this.strict = orig.strict;
         this.bufferInitialLength = orig.bufferInitialLength;
         this.depth = orig.depth;
         return this;
@@ -379,6 +386,24 @@ public RubyBoolean script_safe_p(ThreadContext context) {
         return context.getRuntime().newBoolean(scriptSafe);
     }
 
+    /**
+     * Returns true if strict mode is enabled.
+     */
+    public boolean strict() {
+        return strict;
+    }
+
+    @JRubyMethod(name="strict")
+    public RubyBoolean strict_get(ThreadContext context) {
+        return context.getRuntime().newBoolean(strict);
+    }
+
+    @JRubyMethod(name="strict=")
+    public IRubyObject strict_set(IRubyObject isStrict) {
+        strict = isStrict.isTrue();
+        return isStrict.getRuntime().newBoolean(strict);
+    }
+
     public boolean allowNaN() {
         return allowNaN;
     }
@@ -467,6 +492,7 @@ public IRubyObject configure(ThreadContext context, IRubyObject vOpts) {
         if (!scriptSafe) {
             scriptSafe = opts.getBool("escape_slash", DEFAULT_SCRIPT_SAFE);
         }
+        strict = opts.getBool("strict", DEFAULT_STRICT);
         bufferInitialLength = opts.getInt("buffer_initial_length", DEFAULT_BUFFER_INITIAL_LENGTH);
 
         depth = opts.getInt("depth", 0);
@@ -495,6 +521,7 @@ public RubyHash to_h(ThreadContext context) {
         result.op_aset(context, runtime.newSymbol("ascii_only"), ascii_only_p(context));
         result.op_aset(context, runtime.newSymbol("max_nesting"), max_nesting_get(context));
         result.op_aset(context, runtime.newSymbol("script_safe"), script_safe_get(context));
+        result.op_aset(context, runtime.newSymbol("strict"), strict_get(context));
         result.op_aset(context, runtime.newSymbol("depth"), depth_get(context));
         result.op_aset(context, runtime.newSymbol("buffer_initial_length"), buffer_initial_length_get(context));
         for (String name: getInstanceVariableNameList()) {

lib/json/common.rb

@@ -3,6 +3,9 @@
 require 'json/generic_object'
 
 module JSON
+  NOT_SET = Object.new.freeze
+  private_constant :NOT_SET
+
   class << self
     # :call-seq:
     #   JSON[object] -> new_array or new_string
@@ -608,7 +611,7 @@ class << self
   #   puts File.read(path)
   # Output:
   #   {"foo":[0,1],"bar":{"baz":2,"bat":3},"bam":"bad"}
-  def dump(obj, anIO = nil, limit = nil)
+  def dump(obj, anIO = nil, limit = nil, strict: NOT_SET)
     if anIO and limit.nil?
       anIO = anIO.to_io if anIO.respond_to?(:to_io)
       unless anIO.respond_to?(:write)
@@ -618,6 +621,7 @@ def dump(obj, anIO = nil, limit = nil)
     end
     opts = JSON.dump_default_options
     opts = opts.merge(:max_nesting => limit) if limit
+    opts[:strict] = strict if NOT_SET != strict
     result = generate(obj, opts)
     if anIO
       anIO.write result

lib/json/pure/generator.rb

@@ -141,6 +141,7 @@ def initialize(opts = {})
           @allow_nan             = false
           @ascii_only            = false
           @script_safe          = false
+          @strict                = false
           @buffer_initial_length = 1024
           configure opts
         end
@@ -170,6 +171,10 @@ def initialize(opts = {})
         # all json strings.
         attr_accessor :script_safe
 
+        # If this attribute is set to true, attempting to serialize types not
+        # supported by the JSON spec will raise a JSON::GeneratorError
+        attr_accessor :strict
+
         # :stopdoc:
         attr_reader :buffer_initial_length
 
@@ -214,6 +219,11 @@ def script_safe?
           @script_safe
         end
 
+        # Returns true, if forward slashes are escaped. Otherwise returns false.
+        def strict?
+          @strict
+        end
+
         # Configure this State instance with the Hash _opts_, and return
         # itself.
         def configure(opts)
@@ -245,6 +255,8 @@ def configure(opts)
             false
           end
 
+          @strict                = !!opts[:strict] if opts.key?(:strict)
+
           if !opts.key?(:max_nesting) # defaults to 100
             @max_nesting = 100
           elsif opts[:max_nesting]
@@ -304,7 +316,13 @@ module Object
           # 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.
-          def to_json(*) to_s.to_json end
+          def to_json(generator_state)
+            if generator_state.strict?
+              raise GeneratorError, "#{self.class} not allowed in JSON"
+            else
+              to_s.to_json
+            end
+          end
         end
 
         module Hash
@@ -336,7 +354,9 @@ def json_transform(state)
               result << delim unless first
               result << state.indent * depth if indent
               result = "#{result}#{key.to_s.to_json(state)}#{state.space_before}:#{state.space}"
-              if value.respond_to?(:to_json)
+              if state.strict?
+                raise GeneratorError, "#{value.class} not allowed in JSON"
+              elsif value.respond_to?(:to_json)
                 result << value.to_json(state)
               else
                 result << %{"#{String(value)}"}
@@ -377,7 +397,9 @@ def json_transform(state)
             each { |value|
               result << delim unless first
               result << state.indent * depth if indent
-              if value.respond_to?(:to_json)
+              if state.strict?
+                raise GeneratorError, "#{value.class} not allowed in JSON"
+              elsif value.respond_to?(:to_json)
                 result << value.to_json(state)
               else
                 result << %{"#{String(value)}"}

tests/json_generator_test.rb

@@ -150,6 +150,7 @@ def test_pretty_state
       :buffer_initial_length => 1024,
       :depth                 => 0,
       :script_safe           => false,
+      :strict                => false,
       :indent                => "  ",
       :max_nesting           => 100,
       :object_nl             => "\n",
@@ -167,6 +168,7 @@ def test_safe_state
       :buffer_initial_length => 1024,
       :depth                 => 0,
       :script_safe           => false,
+      :strict                => false,
       :indent                => "",
       :max_nesting           => 100,
       :object_nl             => "",
@@ -184,6 +186,7 @@ def test_fast_state
       :buffer_initial_length => 1024,
       :depth                 => 0,
       :script_safe           => false,
+      :strict                => false,
       :indent                => "",
       :max_nesting           => 0,
       :object_nl             => "",
@@ -336,7 +339,13 @@ def test_hash_likeness_set_string
 
   def test_json_generate
     assert_raise JSON::GeneratorError do
-      assert_equal true, generate(["\xea"])
+      generate(["\xea"])
+    end
+  end
+
+  def test_json_generate_unsupported_types
+    assert_raise JSON::GeneratorError do
+      generate(Object.new, strict: true)
     end
   end