On Aug 8, 2007, at 12:22 PM, Yukihiro Matsumoto wrote:

> Hi,
>
> In message "Re: Is this really what we want?"
>     on Wed, 8 Aug 2007 05:08:43 +0900, James Edward Gray II  
> <james / grayproductions.net> writes:
>
> |I'm investigating some recent breakage in FasterCSV and have tracking
> |it down to a change in the Date standard library.  Is this really the
> |desired behavior now:
> |
> |   $ ruby -r date -e 'p Date.parse("junk")'
> |   #<Date: 4908505/2,0,2299161>
> |
> |?
>
> I asked the author, and he told us it's a side effect of parse
> accepting month name only.  The parse methods (both Date.parse and
> Time.parse) are not for validation.  Considering the complexity of
> date representation they accept, validation is nearly impossible.  The
> author does not think it is worth the cost of implementing more rigid
> check.

We definitely need to update the documentation then.  It says it  
throws ArgumentErrors which is no longer true.

James Edward Gray II