Yukihiro Matsumoto wrote:
> Hi,
> 
> In message "Re: inspect question/request"
>     on 03/03/31, Julian Snitow <vangczung / yahoo.com> writes:
> 
> |> Who wants to undef "inspect"?
> |
> |If something, no matter how stupid, is possible, then someone, somewhere 
> |  will do it.
> |
> |Perhaps in the absence of an inspect method (for whatever reason), a 
> |fall-back to the original behavior is in order?
> 
> I know.  I did it once, and asked myself how often it occurs, and how
> much it cost to cover it, then removed fall back code.  Was I wrong?
> 
> 							matz.
> 

I doubt that the test would affect the performance of correct code if 
it's performed in the error-handling routine (unless there's an obscure 
implemetation detail that I'm overlooking :-), and it seems to me that 
being able to "fail correctly" is important in a language that allows 
graceful recovery from exceptions.

If, on the other hand, there were an actual performance hit for 
*non*-buggy code (i.e., code that doesn't call non-existent methods very 
often), then that would be another matter entirely. :-)