From: Robert Dober <robert.dober@...>
Date: 2006-08-04T05:35:22+09:00
Subject: Re: Creating Classes at runtime

------=_Part_94830_11810529.1154637319135
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

On 8/3/06, ara.t.howard@noaa.gov <ara.t.howard@noaa.gov> wrote:
>
> On Fri, 4 Aug 2006, David Stokar wrote:
>
> > David Stokar wrote:
> > it works like this (just copy and paste)
>
> <snip positive impl>
>
> i thought i'd chime in here and point out that the idea of this is flawed:
>
> if you know the name of the class you want to generate, let's say
> 'MyClass',
> and you also want to refer to it via the bareword constant MyClass then
> you
> can simply do
>
>    class MyClass; end
>
> and later
>
>    class MyClass
>      def an_added_method
>        # ...
>      end
>    end
>
> the point being that, unless you are going to __also__ going retrieve you
> classes via the string name, as in
>
>    c = const_get 'MyClass'
>
>    obj = c::new
>
>
> not
>
>    c = create_class 'MyClass'
>
>    obj = MyClass.new
>
> then you do not need to dynamically create classes by name since the name
> is
> know apriori


Hmm but that still leaves the OP's  question unanswered, how can I  create
a  named class dynamically.

eval "class #{x}; end"
myref2class = Object.const_get(x)


obviously does the job. ( I agree seems flawed to me too, but OP might have
his reasons )

I had a quick look into class.c and it seems that rb_define_class is the
only way to define the name of the class object, which would mean that eval
is the only way to do the trick.

Any opinnions?

Cheers
Robert

<snip>

>
> we can never obtain peace in the outer world until we make peace with
> ourselves.
> - h.h. the 14th dali lama
>
> Hmmm your quotes enlighten me (even more than your posts ;)

<cut my old quote>

------=_Part_94830_11810529.1154637319135--