From: Jim Freeze <jim@...>
Date: 2005-06-03T06:09:04+09:00
Subject: Re: Interesting discovery...

* Joel VanderWerf <vjoel@path.berkeley.edu> [2005-06-03 04:43:52 +0900]:

> Ara.T.Howard@noaa.gov wrote:
> >  rescue_test  2.940000   0.320000   3.260000 (  3.264758)
> >  respond_to_test  0.160000   0.050000   0.210000 (  0.208268)
> >
> >
> >looks like respond_to? is indeed faster.
> 
> But if the method *is* defined, rescue is faster:

I'm not sure I understand. If the method exists, what is the
need for rescue to be called?

Are the options being tested these:

  # with respond_to?
  if x.respond_to(:meth) x.meth # or x.call(:meth)
  # or with rescue
  begin
    x.meth
  rescue
    # no meth

  So, if meth exists, the only overhead should be the 'if x.respond_to'.
  If the method does nothing, then a check of this type may appear expensive,
  but if the method does something, then the 'if' should not be too much
  of a burden.
  I suppose that this could be a problem in a tight loop. Under such
  conditions, it may be better to just rescue.


-- 
Jim Freeze
Theory and practice are the same, in theory. -- Ryan Davis