On Thu, Oct 26, 2006 at 09:47:39AM -0700, John W. Baxter wrote:
> On 10/26/06 9:30 AM, "gdub" <a507@???> wrote:
>
> > I was considering stripping comments,
> > replacing macros, and evaluating
> > conditional regions prior to distributing
> > config files. However, Mr. Hazel's
> > comments wrt forking and shared config
> > memory have me thinking that it would
> > not be effort well spent.
>
> That's the basic conclusion which has been reached over the years each time
> this has come up.
>
> I don't recall anyone making measurements, however, and an important rule of
> performance is that what one *thinks* about performance is wrong; what one
> *measures* about performance may be right--if the thinking that led to the
> instrumentation was reasonably close to right.
well, as a trivial example here (with a moderately complex
config with some perl),
echo QUIT | time exim -bhc 1.2.3.4
uses about 0.14s user + system time (and about the same
for exim -bt chris or exim -bP > /dev/null).
--
``Early to rise and early to bed,
makes a man healthy, wealthy and dead.'' (Thurber)
