implemented utf sniffing, transcoding in parser

improved documentation
added to changes
bumped version

3 files changed, 67 insertions, 17 deletions

diff --git a/lib/json/common.rb b/lib/json/common.rb
index 467c7b2..39f6336 100644
--- a/lib/json/common.rb
+++ b/lib/json/common.rb
@@ -1,4 +1,5 @@
 require 'json/version'
+require 'iconv'
 
 module JSON
   class << self
@@ -105,7 +106,7 @@ module JSON
 
   module_function
 
-  # Parse the JSON string _source_ into a Ruby data structure and return it.
+  # Parse the JSON document _source_ into a Ruby data structure and return it.
   #
   # _opts_ can have the following
   # keys:
@@ -122,9 +123,9 @@ module JSON
     JSON.parser.new(source, opts).parse
   end
 
-  # Parse the JSON string _source_ into a Ruby data structure and return it.
+  # Parse the JSON document _source_ into a Ruby data structure and return it.
   # The bang version of the parse method, defaults to the more dangerous values
-  # for the _opts_ hash, so be sure only to parse trusted _source_ strings.
+  # for the _opts_ hash, so be sure only to parse trusted _source_ documents.
   #
   # _opts_ can have the following keys:
   # * *max_nesting*: The maximum depth of nesting allowed in the parsed data
@@ -145,9 +146,8 @@ module JSON
     JSON.parser.new(source, opts).parse
   end
 
-  # Unparse the Ruby data structure _obj_ into a single line JSON string and
-  # return it. _state_ is
-  # * a JSON::State object,
+  # Generate a JSON document from the Ruby data structure _obj_ and return
+  # it. _state_ is * a JSON::State object,
   # * or a Hash like object (responding to to_hash),
   # * an object convertible into a hash by a to_h method,
   # that is used as or to configure a State object.
@@ -180,7 +180,11 @@ module JSON
     else
       state = State.new
     end
-    obj.to_json(state)
+    result = obj.to_json(state)
+    if result !~ /\A\s*(?:\[.*\]|\{.*\})\s*\Z/m
+      raise GeneratorError, "only generation of JSON objects or arrays allowed"
+    end
+    result
   end
 
   # :stopdoc:
@@ -190,14 +194,17 @@ module JSON
   module_function :unparse
   # :startdoc:
 
-  # Unparse the Ruby data structure _obj_ into a single line JSON string and
-  # return it. This method disables the checks for circles in Ruby objects, and
-  # also generates NaN, Infinity, and, -Infinity float values.
+  # Generate a JSON document from the Ruby data structure _obj_ and return it.
+  # This method disables the checks for circles in Ruby objects.
   #
   # *WARNING*: Be careful not to pass any Ruby data structures with circles as
   # _obj_ argument, because this will cause JSON to go into an infinite loop.
   def fast_generate(obj)
-    obj.to_json(nil)
+    result = obj.to_json(nil)
+    if result !~ /\A(?:\[.*\]|\{.*\})\Z/
+      raise GeneratorError, "only generation of JSON objects or arrays allowed"
+    end
+    result
   end
 
   # :stopdoc:
@@ -206,8 +213,9 @@ module JSON
   module_function :fast_unparse
   # :startdoc:
 
-  # Unparse the Ruby data structure _obj_ into a JSON string and return it. The
-  # returned string is a prettier form of the string returned by #unparse.
+  # Generate a JSON document from the Ruby data structure _obj_ and return it.
+  # The returned document is a prettier form of the document returned by
+  # #unparse.
   #
   # The _opts_ argument can be used to configure the generator, see the
   # generate method for a more detailed explanation.
@@ -229,7 +237,11 @@ module JSON
       end
       state.configure(opts)
     end
-    obj.to_json(state)
+    result = obj.to_json(state)
+    if result !~ /\A\s*(?:\[.*\]|\{.*\})\s*\Z/m
+      raise GeneratorError, "only generation of JSON objects or arrays allowed"
+    end
+    result
   end
 
   # :stopdoc:
@@ -270,7 +282,6 @@ module JSON
       proc.call result
     end
   end
-  module_function :recurse_proc
 
   alias restore load
   module_function :restore
@@ -306,6 +317,11 @@ module JSON
   rescue JSON::NestingError
     raise ArgumentError, "exceed depth limit"
   end
+
+  # Shortuct for iconv.
+  def self.iconv(to, from, string)
+    Iconv.iconv(to, from, string).first
+  end
 end
 
 module ::Kernel
diff --git a/lib/json/pure/parser.rb b/lib/json/pure/parser.rb
index 7e8fe08..7a09f2f 100644
--- a/lib/json/pure/parser.rb
+++ b/lib/json/pure/parser.rb
@@ -66,7 +66,41 @@ module JSON
       # * *object_class*: Defaults to Hash
       # * *array_class*: Defaults to Array
       def initialize(source, opts = {})
-        super
+          if defined?(::Encoding)
+            if source.encoding == Encoding::ASCII_8BIT
+              b = source[0, 4].bytes.to_a
+              source = case
+                       when b.size >= 4 && b[0] == 0 && b[1] == 0 && b[2] == 0
+                         source.dup.force_encoding(Encoding::UTF_32BE).encode!(Encoding::UTF_8)
+                       when b.size >= 4 && b[0] == 0 && b[2] == 0
+                         source.dup.force_encoding(Encoding::UTF_16BE).encode!(Encoding::UTF_8)
+                       when b.size >= 4 && b[1] == 0 && b[2] == 0 && b[3] == 0
+                         source.dup.force_encoding(Encoding::UTF_32LE).encode!(Encoding::UTF_8)
+                       when b.size >= 4 && b[1] == 0 && b[3] == 0
+                         source.dup.force_encoding(Encoding::UTF_16LE).encode!(Encoding::UTF_8)
+                       else
+                         source.dup
+                       end
+            else
+              source = source.encode(Encoding::UTF_8)
+            end
+            source.force_encoding(Encoding::ASCII_8BIT)
+          else
+            b = source
+            source = case
+                     when b.size >= 4 && b[0] == 0 && b[1] == 0 && b[2] == 0
+                       JSON.iconv('utf-8', 'utf-32be', b)
+                     when b.size >= 4 && b[0] == 0 && b[2] == 0
+                       JSON.iconv('utf-8', 'utf-16be', b)
+                     when b.size >= 4 && b[1] == 0 && b[2] == 0 && b[3] == 0
+                       JSON.iconv('utf-8', 'utf-32le', b)
+                     when b.size >= 4 && b[1] == 0 && b[3] == 0
+                       JSON.iconv('utf-8', 'utf-16le', b)
+                     else
+                       b
+                     end
+          end
+        super source
         if !opts.key?(:max_nesting) # defaults to 19
           @max_nesting = 19
         elsif opts[:max_nesting]
diff --git a/lib/json/version.rb b/lib/json/version.rb
index 788ca2e..ff48b9b 100644
--- a/lib/json/version.rb
+++ b/lib/json/version.rb
@@ -1,6 +1,6 @@
 module JSON
   # JSON version
-  VERSION         = '1.1.9'
+  VERSION         = '1.2.0'
   VERSION_ARRAY   = VERSION.split(/\./).map { |x| x.to_i } # :nodoc:
   VERSION_MAJOR   = VERSION_ARRAY[0] # :nodoc:
   VERSION_MINOR   = VERSION_ARRAY[1] # :nodoc: