On Thu, 21 Feb 2008 15:32:51 -0800, Mike Schilling
<mscottschilling@hotmail.com> wrote:
[...] So I'm
willing, *in this case*, to act as if that all implementations do
it
until I'm shown one that doesn't.
I would as well. Frankly, it's such a no-brainer to cache the
result
that even if I did run into an implementation that didn't, and as a
result caused some sort of serious performance problem, my first
inclination would be to tell whatever user ran across the problem
"go
use a decent Java implementation". :)
But I still think it's reasonable to question whether this is
documented as guaranteed behavior, or is simply
implementation-dependent, even if any sensible implementation would
in fact do it.
But that's just me. I've been known to be extra-particular about
such
things, sometimes for no apparent reason except for the sake of
being
particular. :)
to propose, and I'm already married, so I'll stop here.