ntpdsim status and question

[Eric S. Raymond]

Hal Murray <hmurray at megapathdsl.net>:
> 
> esr at thyrsus.com said:
> > I somehow missed Tom's query.  But I note that the kind of problem Hal is
> > pointing at here is very similar to the one the intercept layer solves for
> > capture/replay.  Maybe that layer should implement a sim mode? 
> 
> That's making the decision at run time vs compile time.

With enough proprocessor hackery, runtime support implies the ability
to compile an ndpsim-only version.

> If/when we decide to build a minimal version of ntpd, I could easily imagine 
> that a compile time option to avoid the intercept layer would be a useful 
> idea.

The only reasonable way to do this would be to not compile in the
replay/capture support in the intercept layer, leaving the normal-mode
implementations in place.

> If we want to support the simulator, I think we should get the old code 
> running first so we have a reference point.  It doesn't have to be pretty, 
> just work well enough so that we can run it and understand what it does.  If 
> we decide we want to actively use it long term, then we can consider 
> repackaging things so it is easier to maintain in our build environment.

Reasonable.  I won't have any attention to spare for a while, though; I'm
fully occupied with pushing replay forward.  It's hard work.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>