From: David Vallner <david@...>
Date: 2006-01-11T14:36:39+09:00
Subject: Re: use of "return"

Ryan Leavengood wrote:

>On 1/10/06, Jim Freeze <jim@freeze.org> wrote:
>  
>
>>Seems like every time I check, code with a 'return' runs slower
>>than code without. Less typing and faster. I say drop the return.
>>    
>>
>
>Confirmed:
>
>require 'quickbench'
>
>class Test
>  def initialize(test)
>    @test = test
>  end
>
>  def test
>    @test
>  end
>
>  def get_test
>    return @test
>  end
>end
>
>t = Test.new("something")
>QuickBench.go(3000000, 25) {}
>
>__END__
>t.test
>t.get_test
>
>Results:
>
>-------------------------------------------------------------------------
>|                      QuickBench Session Started                       |
>|                          3000000 Iterations                           |
>-------------------------------------------------------------------------
>                                  user     system      total        real
>1. t.test                     2.860000   0.000000   2.860000 (  2.859000)
>2. t.get_test                 3.078000   0.000000   3.078000 (  3.094000)
>-------------------------------------------------------------------------
>|                        Fastest was <1. t.test>                        |
>-------------------------------------------------------------------------
>
>Ryan
>
>  
>
Most weird. Let's hope someone notices and the compiler is modified to 
plain drop tail-returns. That is, if "return" is a reserved word and 
thus not viable to being overridden by someone who codes with a dead 
chicken in his left hand ;P (Gotta love it when flexibility bites you 
back. *daydreams about clean blocks*)