From: Gennady <bystr@...>
Date: 2003-05-06T15:53:11+09:00
Subject: Re: Variable/Method ambiguity

On Monday, May 5, 2003, at 06:31 PM, dblack@superlink.net wrote:
>
> I don't think it's inconsistent.  I think it would be inconsistent,
> and confusing, if this:
>
>   class A
>     def thing=(x); end
>     def other; thing = 100; end
>   end
>
> were different from this:
>
>   class A
>     def other; thing = 100; end
>     def thing=(x); end
>   end

These 2 cases would work the same, as actual execution happens after 
all methods are defined -- so it would call thing=() in both.

>
> or from this:
>
>   class A
>     def blah=(x); end
>     def other; thing = 100; end
>   end
>
> As I understand it, the consistency is that, in all cases, the
> receiver of a method must appear explicitly, unless both of these
> things are true:
>   (a) the receiver is self;
>   (b) the variable could not be an already-defined local variable,
>       or a local variable in the process of being defined.
>
> I think this is the only algorithm that won't cause shifting behavior
> in cases like the ones above.

I know that I can use self.name= as Gavin suggested (it does not look 
pretty for my task, though) and I understand all consequences presented 
by David, and yet ... it was so tempting to want to be able to use it 
my way. Nevermind, I use slightly different approach now that works ;-)

Gennady.