Re: Initialize not being called on objects created from literals

On Wed, May 27, 2009 at 2:10 AM, Robert Klemme
<[e-mail]> wrote:
> On 27.05.2009 02:26, Oliver Saunders wrote:
>>
>> Initialize doesn't appear to get called.
>>
>> class Object
>> ef initialize
>> foo = 'bar'
>> nd
>> ttr_reader :foo
>> end
>>
>> a = Object.new # => #<Object:0x54c964 @foo="bar">
>> a.foo # => "bar"
>> 'string'.foo # => nil
>> //.foo # => nil
>>
>> I want to have //.foo and 'string.foo give me "bar" as they should. How
>> can I do that?
>
> A class needs to explicitly invoke #initialize of the super class.  guess,
> since Object does not have any members by default String will not do the
> invocation.

This is true, see below
>
> class A
> ef initialize
> uts "A"
> nd
> end
>
> class B < A
> ef initialize
> uper
>  or super() in this case
> uts "B"
> nd
> end
>
> Having said that, it's generally not a too good idea to mess with built in
> classes - even though you can. ut you may produce unwanted side effects.

But that's not the whole story.  As the OP suspected, literals are
created by the parser without going through the normal initialization
route:

class Object
  alias_method :old_initialize, :initialize
  attr_reader :boo

  def initialize(*args, &b) # !> redefining Object#initialize may
cause infinite loop
    puts "Object#initialize"
    old_initialize(*args, &b)
    @boo = "who"
  end
end

puts "calling String.new"
s1 = String.new
puts "back"
puts "s1.boo is #{s1.boo.inspect}"
puts "making String#initialize call super"
class String
  alias_method :old_initialize, :initialize
  def initialize(*args, &b)
    super
    puts "String#initialize"
    old_initialize(*args, &b)
  end

end

puts "calling String.new"
s2 = String.new
puts "s2.boo is #{s2.boo.inspect}"

puts "\"abc\".boo is #{"abc".boo.inspect}"

Produces the following

calling String.new
back
s1.boo is nil
making String#initialize call super
calling String.new
Object#initialize
String#initialize
s2.boo is "who"
"abc".boo is nil
untitled:5: warning: redefining Object#initialize may cause infinite loop

Note that warning pointing out a potential unwanted side effect.

-- 
Rick DeNatale

Blog: http://talklikeaduck.denhaven2.com/
Twitter: http://twitter.com/RickDeNatale
WWR: http://www.workingwithrails.com/person/9021-rick-denatale
LinkedIn: http://www.linkedin.com/in/rickdenatale