[ntp:questions] ntpdate works, but ntpd doesn't (reach = 0)
stenn at ntp.org
Sun Feb 8 20:50:50 UTC 2009
>>> In article <498E6161.1040307 at ntp.isc.org>, mayer at ntp.isc.org (Danny Mayer) writes:
Harlan> I think you are dead wrong on this Danny, and next time we're
Harlan> trying it my way. In particular, we must have an easier way to
Harlan> import new releases from ISC, and I want to make sure that whenever
Harlan> possible, ISC takes back any patches we have.
Danny> No, I'm not wrong. The design was done by some of the sharpest
Danny> programmers in the business. These were not your average
Danny> programmers. I worked with them to get this done over several
Danny> years. You'd be hard put to improve over the layout. ntp is
Danny> structurally deficient in comparison. Fix the ntp structure before
Danny> you try and improve on ISC's.
We're talking about different things.
I'm saying we're going to import libisc/ using the directory structure they
use, and we're not going to perform surgery on that code to remove the bits
we don't want and take the remaining bits and distribute them in different
And I'll note a clear improvement below.
Harlan> This should get better when ISC releases libisc as a
Harlan> tearoff/standalone library.
Danny> It's standalone now. That was part of the original design.
Harlan> We have different definitions of "standalone" then.
Danny> Probably, but there is nothing you need to do to use the libisc
Danny> library for other applications.
Until libisc provides an autoconf macro, say ISC_LIBISC, that contains all
the needed configure.ac magic I'm gonna say that libisc may be the basis for
some excellent software, but it is not yet a product.
Harlan> My intent is to import libisc as a single unit, and this will affect
Harlan> the build scripts for Windows.
Danny> Not really. ... (Lots of other things)
I hear you, disagree with you, and after 4.2.6 is out we're going to find a
way to cleanly import libisc/ in one subdir of the NTP tree and the unix and
windows builds will have to deal with that.
If this turns out to be Difficult, I'll work with ISC to make it easier to
do, for at least an iteration or two. ISC has told me several times over
the years that they want libisc/ to be more generally useful. I'm gonna do
what I can to see that happen.
I find the current situation with what is needed to upgrade libisc/ in NTP
to be intolerable, and in this case my vote wins.
Harlan Stenn <stenn at ntp.org>
http://ntpforum.isc.org - be a member!
More information about the questions