From: Daniel Brockman <daniel@...>
Date: 2005-07-23T00:31:56+09:00
Subject: Re: ruby gnome question

Hal Fulton <hal9000@hypermetrics.com> writes:

> Joe Van Dyk wrote:
>
>> What the.  Is that the principle of Matz's least surprise
>> coming up again?  Cuz it sure didn't meet JVD's principle
>> of least surprise.
>
> Congratulations, you get to retrain your neural net. :)
>
> There are good reasons for the distinction here.
> Frequently (I might dare to say: most often) you want to
> pass the same args on to the parent's method.

You frequently need to do this to other methods as well.
So I think it would have been both more useful and less
confusing to invent a piece of syntax for passing along the
same arguments to another method.  Something like this:

   def foo(bar, baz, *rest, &block) super(...)
     # Do additional stuff.
   end

   def bar(baz, quux, *rest, &block)
     @moomin.bar(...).postprocess
   end

I'm not trying to stir up trouble; I just wanted to let Joe
know he's not the first person whose PLS this doesn't meet.

> But sometimes you need more control. In fact, there is no
> guarantee that the two methods will even have the
> same arity.
>
> So you get to pass whatever you want. If what you pass is
> "nothing" then this is a special case -- you have to
> surround the "nothing" with parentheses: super()

And that looks really strange if you're not familiar with
this language quirk.  One of the most basic facts you learn
about the Ruby syntax is that ‘foo()’ equals ‘foo’.

-- 
Daniel Brockman <daniel@brockman.se>

    So really, we all have to ask ourselves:
    Am I waiting for RMS to do this?   --TTN.