From: Rick DeNatale <rick.denatale@...>
Date: 2007-03-18T04:23:27+09:00
Subject: Re: attr_accessor question

On 3/17/07, Stefano Crocco <stefano.crocco@alice.it> wrote:
> Alle sabato 17 marzo 2007, Rick DeNatale ha scritto:
> > > def var=value
> > > @var=value
> > > end
> >
> > Actually the setter should be
> >
> > def var=(value)
> > @var=value
> > end
>
> Do you mean I should put value into brackets? If so, why? To me, it seems that
> my form works perfectly. Is there something I'm not aware of?

I guess it's a matter of style.

The pickaxe gives the syntax of a method definition as:

  def defname [( [arg [ =val], ...] [, *vararg ] [ , &blockarg ] ) ]
       body
  end

Which implies that if arguments are present they must be enclosed in
parentheses.

I've always done so, and I can't recall having seen method
*definitions* which omitted the parens around the args.  But trying
this with both ruby1.8 and ruby1.9 I see that it neither complains
about the missing parens, something I didn't know until just now.

But I greatly prefer using the parens.  In fact my preference is to
use them for most method invocations as well, although omitting  them
there is more common.  I tend only to omit them in idiomatic cases
like

   attr_accessor :foo, :bar

instead of
   attr_accessor(:foo, :bar)

or when ruby is being used as a domain specific language.

I also notice that the pickaxe syntax is a little loose because it
won't admit valid things like

   def foo(&block)
      ...
   end

So it's really a pedagogical syntax.


-- 
Rick DeNatale

My blog on Ruby
http://talklikeaduck.denhaven2.com/