From: "Hal E. Fulton" <hal9000@...>
Date: 2002-04-28T09:12:35+09:00
Subject: Re: $_ if no |var|

----- Original Message ----- 
From: "Lars Christensen" <larsch@cs.auc.dk>
To: "ruby-talk ML" <ruby-talk@ruby-lang.org>
Sent: Saturday, April 27, 2002 4:16 PM
Subject: Re: $_ if no |var|


> On Sun, 28 Apr 2002, David Alan Black wrote:
> 
> > I have to confess that to me it just doesn't look or feel like Ruby.
> > I know that's very vague... that's why I was trying to come up with a
> > case illustrating the importance (or at least Ruby-ness) of all the
> > assignments being made, independent of any branching within the code
> > of the block.
> 
> I think you are right--I think I would 'expect' all assignments to be
> made.  I.e.
> 
>    proc { if |a| then |b| else |c| end }
> 
> should only be syntatic sugar for
> 
>    proc { |a,b,c| if a then b else c end }
> 
> or, equivalently
> 
>   proc { |p| a,b,c=p; if a then b else c end }
> 
> The |a| would just be syntax to mark which symbols should be bound to the
> arguments when the block is evaluated.
> 
> I would also expect
> 
>    proc { if |a| then |b| else |c|+b end }
> 
> to work, except in more complex cases, i'd use the current syntax.

I agree with Lars, but I have misgivings about this construct
in general.

Would you allow:

    proc {|a| puts "b = #{|b|}"; do_something_with(c) }

and that sort of thing? Sounds like a can of worms to me, a
syntactic and semantic bad dream.

Hal Fulton