Asp Forum - Class inheriting from class within the namespace

Gavin Kistner

10/31/2006 4:41:00 PM

Tree structures are interesting. Often the root of the tree is used as a
handle for the tree itself. However, the tree often has additional
methods beyond that of the root node.

I couldn't figure out how to do this simply via inheritance due to
chicken-and-egg problems with namespacing:

class Tree < Tree::Node
class Node
end
end
#=> chicken-and-egg.rb:1: uninitialized constant Tree (NameError)

I came up with two alternative techniques to achieve the same goal. I'm
not wild about the way either are handled by RDoc (though the first is
substantially better):

# Technique 1: Using modules
class Tree
module NodeStuff
#...all the meat here...
end
include NodeStuff

class Node
include NodeStuff
end
end

# Technique 2: Conjur the chicken first, and kill it later
class TmpTreeNode
# ...all the meat here...
end
class Tree < TmpTreeNode
Node = TmpTreeNode
end
Object.class_eval{ remove_const( :TmpTreeNode ) }

Ideally, IMO, Ruby would let me do this:
# No "Tree" module or class has yet been defined;
# it is a pure namespace at this point
class Tree::Node
# ...all the meat here...
end

# Now Tree is a class
class Tree < Tree::Node
end

I don't have any question, and it's not important enough to me to make
an RCR for the above. I just thought I'd share these thoughts with the
world.

3 Answers

matt

10/31/2006 5:20:00 PM

Gavin Kistner <gavin.kistner@anark.com> wrote:

> Tree structures are interesting. Often the root of the tree is used as a
> handle for the tree itself. However, the tree often has additional
> methods beyond that of the root node.
>
> I couldn't figure out how to do this simply via inheritance due to
> chicken-and-egg problems with namespacing:
>
> class Tree < Tree::Node
> class Node
> end
> end
> #=> chicken-and-egg.rb:1: uninitialized constant Tree (NameError)

Why wouldn't you do this?

module TreeStuff
class Node
end
class Tree < Node
end
end

A tree does more than a node (you say), so a tree is a subclass of node
("a node and then some"). The pair are wrapped up in a module to isolate
their namespace.

Of course I could be missing some other desideratum. m.

--
matt neuburg, phd = matt@tidbits.com, http://www.tidbits...
Tiger - http://www.takecontrolbooks.com/tiger-custom...
AppleScript - http://www.amazon.com/gp/product/...
Read TidBITS! It's free and smart. http://www.t...

Jack Linthicum

11/21/2008 9:17:00 PM

On Nov 21, 3:51 pm, Jack Linthicum <jacklinthi...@earthlink.net>
wrote:
> On Nov 21, 3:11 pm, Les Cargill <lcarg...@cfl.rr.com> wrote:
>
>
>
> > Dan wrote:
> > > Les Cargill wrote:
> > >> J Antero wrote:
> > >>> I have a feeling that if the Detroit automakers wind up in Cahpter 7
> > >>> bankruptcy (doors closed, not just reorganization), that there will
> > >>> be a couple pages in Econ 101 textbooks 50 years from now, discussing
> > >>> what a very big mistake it was..
>
> > >>> The current financial crisis has a certain "superficial wounds"
> > >>> aspect to it.
>
> > >>> Losing Detroit automakers is more like a gut wound.
>
> > >> They've been putting lipstick on that corpse for twenty years. Maybe
> > >> longer.
>
> > >> I do have to wonder just how much CAFE standards are to blame, though
> > >> - not that Detroit handled them very well, but still....
>
> > > Lack of them in the last administration hurt Detroit enormously.
>
> > > Dan
>
> > But the end consumer simply purchased around CAFE standards by
> > buying SUVs. It's anything but clear that that "hurt" anybody
> > (discounting environmental damage).
>
> > --
> > Les Cargill

Correcting some gaffs.
>
> SUVs have more roll-overs, which are fatal 80% of the time when driver
> isn't wearing seat belt. There were 10k deaths in 2007 due to rollovers, which I think makes them something like 20-25% of all highway deaths. They also are high in back up
> accidents, and , because of their usually greater weight (and
> therefore mass) they do more damage than ordinary cars when a crash
> occurs. Also without air bags the occupants suffer more harm in side
> collisions, despite the higher seat. SUVs are classified as light
> trucks, some more of the Big 3 fictions to pass even the low CAFE
> standards. As such their bumpers do not have the "same height" rule as
> passenger cars. Try googling SUV accident rate.
>
> Personal observation says the driver of an SUV will be a small woman
> with children, usually not in restraints, who is talking on a phone.
>
> http://www.suv-rollovers.com...

Jack Linthicum

11/22/2008 12:31:00 AM

On Nov 21, 6:47 pm, "John Briggs" <john.brig...@ntlworld.com> wrote:
> "Jack Linthicum" wrote
>
>
>
> > Correcting some gaffs.
>
> gaffes.
>
> >> SUVs have more roll-overs, which are fatal 80% of the time when driver
> >> isn't wearing seat belt. There were 10k deaths in 2007 due to rollovers,
> >> which I think makes them something like 20-25% of all highway deaths.
> >> They also are high in back up accidents, and , because of their usually
> >> greater weight (and therefore mass) they do more damage than ordinary
> >> cars when a crash occurs.
>
> How (roughly) are you distinguishing between weight and mass?
> --
> John Briggs

If you print it in the manual it's weight, if you measure it in an
impact test it's mass.

comp.lang.ruby

Class inheriting from class within the namespace

Gavin Kistner

matt

Jack Linthicum

Jack Linthicum

x Login to ForumsZone