Subclasses and constants

ruby-talk
1

Often, I find myself doing:

class RemoteCommand
  def submit
    sendto(URL, ...)
  end
end

class GetStockPriceCommand
  URL = 'http://...'
end

Which won't work.
I'm forced to do:

sendto(self.class::URL, ...)

which just seems akward. Same for using @@class_variables.

Why? More importantly, is there a better workaround?

2

List Recv wrote:

Often, I find myself doing:

class RemoteCommand
  def submit
    sendto(URL, ...)
  end
end

class GetStockPriceCommand
  URL = 'http://...'
end

Which won't work.
I'm forced to do:

sendto(self.class::URL, ...)

which just seems akward. Same for using @@class_variables.

Why? More importantly, is there a better workaround?

I assume you meant: class GetStockPriceCommand < RemoteCommand

If so then the following will work:

  class RemoteCommand
    def submit
      sendto(url, ...)
    end
    def url
      self.class::URL
    end
  end

  class GetStockPriceCommand < RemoteCommand
    URL = 'http://...'
  end

···

--
-- Jim Weirich

--
Posted via http://www.ruby-forum.com/\.

3

Jim,

Right, that's exactly what I meant, and that's in fact the work around
which I've been using.

My question is really why I need to do this, and if there is a better
way... It forces me to decide whether I'll ever subclass when I code
the parent, which seems unrubyish... I'd rather just code the parent,
and if the child changes the URL, so be it!

4

and that is exactly why you should not use a Constant.
Constants do not change, do they ;)?
You should use an instance variable and define accessor methods to it. (or
class variable for that matter)
These accessor methods are then used by the superclass.
This is a classic pattern, but I do not know its name in English, sorry.

Hope this helps
Robert

···

On 3/27/06, listrecv@gmail.com <listrecv@gmail.com> wrote:

the parent, which seems unrubyish... I'd rather just code the parent,
and if the child changes the URL, so be it!

--
Deux choses sont infinies : l'univers et la bêtise humaine ; en ce qui
concerne l'univers, je n'en ai pas acquis la certitude absolue.

- Albert Einstein

5

@@class_variables seemed to have the same behavior - if they were
changed in the subclass, and the parent defined a method, it wouldn't
have access to the subclass's definition.

(BTW - of course Constants don't change - but different subclasses
could have different ones - with them never changing - but I digress)

6

No do not worry you do not, I just looked at an implementation that might
fit your needs better, the conceptual discussion remains interesting anyhow.
I will check out the class variable behaviour you mentioned, I was surprised
about it.
Cheers
Robert

···

On 3/28/06, listrecv@gmail.com <listrecv@gmail.com> wrote:

@@class_variables seemed to have the same behavior - if they were
changed in the subclass, and the parent defined a method, it wouldn't
have access to the subclass's definition.

(BTW - of course Constants don't change - but different subclasses
could have different ones - with them never changing - but I digress)

--
Deux choses sont infinies : l'univers et la bêtise humaine ; en ce qui
concerne l'univers, je n'en ai pas acquis la certitude absolue.

- Albert Einstein