module Marshal

The marshaling library converts collections of Ruby objects into a byte stream, allowing them to be stored outside the currently active script. This data may subsequently be read and the original objects reconstituted.

Marshaled data has major and minor version numbers stored along with the object information. In normal use, marshaling can only load data written with the same major version number and an equal or lower minor version number. If Ruby’s “verbose” flag is set (normally using -d, -v, -w, or –verbose) the major and minor numbers must match exactly. Marshal versioning is independent of Ruby’s version numbers. You can extract the version by reading the first two bytes of marshaled data.

str = Marshal.dump("thing")
RUBY_VERSION   #=> "1.9.0"
str[0].ord     #=> 4
str[1].ord     #=> 8

Some objects cannot be dumped: if the objects to be dumped include bindings, procedure or method objects, instances of class IO, or singleton objects, a TypeError will be raised.

If your class has special serialization needs (for example, if you want to serialize in some specific format), or if it contains objects that would otherwise not be serializable, you can implement your own serialization strategy.

There are two methods of doing this, your object can define either marshal_dump and marshal_load or _dump and _load. marshal_dump will take precedence over _dump if both are defined. marshal_dump may result in smaller Marshal strings.

Security considerations¶ ↑

By design, Marshal.load can deserialize almost any class loaded into the Ruby process. In many cases this can lead to remote code execution if the Marshal data is loaded from an untrusted source.

As a result, Marshal.load is not suitable as a general purpose serialization format and you should never unmarshal user supplied input or other untrusted data.

If you need to deserialize untrusted data, use JSON or another serialization format that is only able to load simple, ‘primitive’ types such as String, Array, Hash, etc. Never allow user input to specify arbitrary types to deserialize into.

marshal_dump and marshal_load¶ ↑

When dumping an object the method marshal_dump will be called. marshal_dump must return a result containing the information necessary for marshal_load to reconstitute the object. The result can be any object.

When loading an object dumped using marshal_dump the object is first allocated then marshal_load is called with the result from marshal_dump. marshal_load must recreate the object from the information in the result.

Example:

class MyObj
  def initialize name, version, data
    @name    = name
    @version = version
    @data    = data
  end

  def marshal_dump
    [@name, @version]
  end

  def marshal_load array
    @name, @version = array
  end
end

_dump and _load¶ ↑

Use _dump and _load when you need to allocate the object you’re restoring yourself.

When dumping an object the instance method _dump is called with an Integer which indicates the maximum depth of objects to dump (a value of -1 implies that you should disable depth checking). _dump must return a String containing the information necessary to reconstitute the object.

The class method _load should take a String and use it to return an object of the same class.

Example:

class MyObj
  def initialize name, version, data
    @name    = name
    @version = version
    @data    = data
  end

  def _dump level
    [@name, @version].join ':'
  end

  def self._load args
    new(*args.split(':'))
  end
end

Since Marshal.dump outputs a string you can have _dump return a Marshal string which is Marshal.loaded in _load for complex objects.

class Klass
  def initialize(str)
    @str = str
  end
  def say_hello
    @str
  end
end

o = Klass.new("hello\n")
data = Marshal.dump(o)
obj = Marshal.load(data)
obj.say_hello  #=> "hello\n"

static VALUE
marshal_dump(int argc, VALUE *argv, VALUE _)
{
    VALUE obj, port, a1, a2;
    int limit = -1;

    port = Qnil;
    rb_scan_args(argc, argv, "12", &obj, &a1, &a2);
    if (argc == 3) {
        if (!NIL_P(a2)) limit = NUM2INT(a2);
        if (NIL_P(a1)) io_needed();
        port = a1;
    }
    else if (argc == 2) {
        if (FIXNUM_P(a1)) limit = FIX2INT(a1);
        else if (NIL_P(a1)) io_needed();
        else port = a1;
    }
    return rb_marshal_dump_limited(obj, port, limit);
}

serialized = Marshal.dump(['value1', 'value2', 'value1', 'value2'])

deserialized = Marshal.load(serialized)
deserialized.map(&:frozen?)
# => [false, false, false, false]
deserialized.map(&:object_id)
# => [1023900, 1023920, 1023940, 1023960] -- 4 different objects

deserialized = Marshal.load(serialized, freeze: true)
deserialized.map(&:frozen?)
# => [true, true, true, true]
deserialized.map(&:object_id)
# => [1039360, 1039380, 1039360, 1039380] -- only 2 different objects, object_ids repeating

# File marshal.rb, line 33
def self.load(source, proc = nil, freeze: false)
  Primitive.marshal_load(source, proc, freeze)
end