From: George Ogata <george.ogata@...>
Date: 2006-12-14T13:51:00+09:00
Subject: Re: openstruct oddity

On 12/10/06, Vidar Hokstad <vidar.hokstad@gmail.com> wrote:
>
> George Ogata wrote:
> > Along a similar vein, ruby has the notion of the "current class".
> > This is like self in that it is defined at any point in your code, but
> > there's no short identifier for it.  In fact, I can't think of a
> > (non-destructive, thread-safe) expression that will return it.  But
> > this is the class that will gain a new method when you do "def foo".
>
> The current class (or module - I saw your correction) is in "self" when
> you are outside a "def" or . It is just another object (of class
> "Class" or "Module" respectively).
>
> Try this:
>
> class Foo
>    p self
>    p self.class
> end
>
> module Foo
>   p self
>   p self.class
> end
>
> If you are inside a "def" then it is "self.class", but in that case
> "self.class" will refer to the class of the object you call the method
> on, not the class or module the object was defined in.

You're correct (well, almost), but that's two expressions... ;-).  And
ones where you have to know the context up front (statically).  It falls short
on instance_eval.  Inside an instance_eval block, the current module is
self's singleton class.

But I was just making the point that there is this bit of state that
exists on the C-side that isn't so clearly visible on the ruby one.
And that was, I think, the source of my confusion until I looked into
it.