From: Phrogz <phrogz@...>
Date: 2007-11-23T13:19:59+09:00
Subject: Re: syntax Float equality within epsilon

On Nov 22, 1:38 am, unbewusst.s...@weltanschauung.com.invalid (Une
Bévue) wrote:
> i need to test float numbers within an epsilon, then i've extended the
> Float clas like that :
>
> class Float
>   def ===( aFloat, eps = 1.0e-10)
>     begin
>       clazz = aFloat.class.to_s
>       raise "Argument \"#{aFloat}\" must be a Float (being of
> #{clazz})." if clazz != "Float"
>       ( self > aFloat - eps ) && ( self < aFloat + eps )
>     rescue
>       puts "An error occurred: #{$!}"
>       nil
>     end
>   end
> end
>
> this works as expected except when i want not to use the default value
> for eps where i couldn't find the correct syntax :
> a = 1.000000001
> b = 1.00000000012
> p ( a ===( b, 0.001) ).to_s
> gave me :
> [...] syntax error, unexpected ')', expecting tCOLON2 or '[' or '.'
> p ( a ===( b, 0.001) ).to_s
>                     ^
>
> why ???

Some operator methods have syntax sugar that prevents them from taking
multiple arguments using simple syntax. You can get around this,
albeit with a slightly less elegant syntax:

irb(main):013:0> class Float
irb(main):014:1>   def ===( a, b )
irb(main):015:2>     p a, b
irb(main):016:2>   end
irb(main):017:1> end
=> nil
irb(main):018:0> a = 5.4
=> 5.4
irb(main):019:0> a.send( :===, 42, 73 )
42
73
=> nil

If this is a case you'd use often, I suggest simply giving it a normal
method name. that you can call without using #send