WIP: Fail or warn on duplicated key during generation

I think I found a way that has very minimal overhead as long
as the hash has only string or only symbols as keys.

Author: byroot

ext/json/ext/generator/generator.c

@@ -991,6 +991,7 @@ struct hash_foreach_arg {
     struct generate_json_data *data;
     int first_key_type;
     bool first;
+    bool mixed_keys_encountered;
 };
 
 static VALUE
@@ -1008,6 +1009,17 @@ convert_string_subclass(VALUE key)
     return key_to_s;
 }
 
+RBIMPL_ATTR_NOINLINE()
+static void
+json_inspect_hash_with_mixed_keys(struct hash_foreach_arg *arg)
+{
+    if (arg->mixed_keys_encountered) {
+        return;
+    }
+    arg->mixed_keys_encountered = true;
+    rb_funcall(mJSON, rb_intern("on_mixed_keys_hash"), 1, arg->hash);
+}
+
 static int
 json_object_i(VALUE key, VALUE val, VALUE _arg)
 {
@@ -1018,7 +1030,6 @@ json_object_i(VALUE key, VALUE val, VALUE _arg)
     JSON_Generator_State *state = data->state;
 
     long depth = state->depth;
-
     int key_type = rb_type(key);
 
     if (arg->first) {
@@ -1043,7 +1054,7 @@ json_object_i(VALUE key, VALUE val, VALUE _arg)
     switch (key_type) {
         case T_STRING:
             if (RB_UNLIKELY(arg->first_key_type != T_STRING)) {
-                fprintf(stderr, "mixed keys (str)\n");
+                json_inspect_hash_with_mixed_keys(arg);
             }
 
             if (RB_LIKELY(RBASIC_CLASS(key) == rb_cString)) {
@@ -1054,7 +1065,7 @@ json_object_i(VALUE key, VALUE val, VALUE _arg)
             break;
         case T_SYMBOL:
             if (RB_UNLIKELY(arg->first_key_type != T_SYMBOL)) {
-                fprintf(stderr, "mixed keys (sym)\n");
+                json_inspect_hash_with_mixed_keys(arg);
             }
 
             key_to_s = rb_sym2str(key);

lib/json/common.rb

@@ -186,6 +186,27 @@ def generator=(generator) # :nodoc:
 
     private
 
+    # Called from the extension when a hash has both string and symbol keys
+    def on_mixed_keys_hash(hash)
+      set = {}
+      hash.each_key do |key|
+        key_str = case key
+        when String
+          key
+        when Symbol
+          key.name
+        else
+          key.to_s
+        end
+
+        if set[key_str]
+          raise GeneratorError, "Duplicate key #{key_str.inspect} in #{hash.inspect}"
+        else
+          set[key_str] = true
+        end
+      end
+    end
+
     def deprecated_singleton_attr_accessor(*attrs)
       args = RUBY_VERSION >= "3.0" ? ", category: :deprecated" : ""
       attrs.each do |attr|

test/json/json_generator_test.rb

@@ -828,4 +828,11 @@ def test_numbers_of_various_sizes
       assert_equal "[#{number}]", JSON.generate([number])
     end
   end
+
+  def test_generate_duplicate_keys
+    error = assert_raise JSON::GeneratorError do
+      JSON.generate({ foo: 1, "foo" => 2})
+    end
+    assert_equal 'Duplicate key "foo" in {foo: 1, "foo" => 2}', error.message
+  end
 end