From: Robert Dober <robert.dober@...>
Date: 2007-04-18T18:57:03+09:00
Subject: Re: nil in evaluations

On 4/18/07, progcat@comcast.net <progcat@comcast.net> wrote:
> Close but not really what I was looking for.
>
> example:
>    if g(x) > f(x)
>    then
>        put "rejoice because f(x) is greater than f(x)"
>    else
>        put "g(x) <= f(x) OR g(x) and/or f(x) returned a nil so I
> wanted to execute this"
>    end.
>
> In other words I want the function to work as it normally would for
> valid values, and if one or both are nil I want to make the test
> evaluate to false.
You ask nil to be greater than any other value and you ask nil to be
smaller than any other value at the same time.
I think you should rethink your modeling a bit.
Of course you can do it by defining Nil#> and redefine X#> for all X
that can be returned by g, but as Robert has pointed out correctly:
Think twice before doing this.

The quick hack if you really need it should look similar to
class Nil
  def >; false end ## I guess nil is a better choice than false
end

class Integer
   alias_method :old_gr, :>
   def > other
     return false if other.nil?
     old_gr other
   end
end

I you use it in production you will be bitten, be aware of it.

HTH
Robert
>
> Sorry I was not clear.
>
> Thanks,
> Tom
>

-- 
You see things; and you say Why?
But I dream things that never were; and I say Why not?
-- George Bernard Shaw