[ntp:hackers] Re: ntp 4.1.2

David L. Mills mills at udel.edu
Fri Aug 15 21:56:28 PDT 2003


Harlan,

You weren't listening to me. I took an embarassing hit from the Navy
last week. I'm trying to interest them in funding a NTPv4 specification.

IPv6 and multicast and autokey worked okay once upon a time. I don't
even mind waiting a short (week?) time to fix what likely is a small
problem so these things can work again. I've spent a lot of time testing
things and made no substantive changes except to fix very old bugs.
Except for IPv6, the stuff is ready for prime time. There will be no
ntp-dev-ng; there will be a 4.2 release version when IPv6 is fixed.

I have told you several times the release procedure is broken to the
max. I really and truly believe it. Releases as old as 16 months are not
releases and attempts to track and fix bugs in 16-month old fossils are
broken to the max. Old releases should be abandoned, not propped up,
especially if major incompatibilities like crypto continue to spread and
cause problems later. I already have a couple of those.

I give it a week or two. If not resolved by then, the R&D department at
UDel will divorce from the sales, marketing and maintenance division and
the latter will move elsewhere.

Dave

Harlan Stenn wrote:
> 
> Dave,
> 
> You have *only* been working on ntp-dev, and people who choose to use
> ntp-stable have sent me patches which I continue to integrate.  There
> have been enough changes since 4.1.1 (ntp-stable) to warrant a 4.1.2.
> 
> This is a feature.
> 
> It is documented.
> 
> You sometimes tell folks that their problems are due to them not reading
> the docs.  This is simply another case of that.
> 
> You and I are responsible for the release schedule, I submit that we
> have been optimistic about the base features needed for 4.2.0.
> 
> You said you wanted IPv6 in 4.2.0, and when other functionality came
> along that seemed "significant" I got your approval before slating it
> for inclusion in ntp-dev, which we all agreed would be the 4.2.0 release.
> 
> If you want, I am happy to have 2 ntp-dev repos if you wish.  I will
> take the current work and call it ntp-dev-ng and then you can focus on
> making ntp-dev become ready for release.
> 
> When ntp-dev-ng is ready we can release that as the next 4.2.x version.
> 
> After 4.2.0 is release I plan on releasing 4.2.1 (and any subsequent
> 4.2.x release) as often as are warranted, probably several times per year.
> 
> If you would prefer the "divorce" route I can focus on that instead.  If
> that is your choice I believe it will be a complete separation, at least
> on my part.  I cannot speak for any of the other webdudes, committers,
> or hackers.
> 
> My preference is that you and I get along and work better together.
> 
> Harlan
> --
> > Harlan,
> >
> > No. I'm just getting started.
> >
> > The "new" 4.1.2 release version is dated 17 July 2003. When I saw that
> > date I sighed relief and assumed that version was somewhere near modern
> > times. I checked and found it had Autokey Versio 1 support dated 19
> > February 2002. A lot of water has flowed under the bridge since then.
> > This explains why I have been getting complaints about broken
> > documentation and public key cryptography. Yes, the documentation in
> > that version does go with the version and explains why I have been
> > getting strange typo reports about *.htm pages.
> >
> > The dating of the release and development versions gave me some pause,
> > since the date of the former is later than the latter, and led to me
> > giving bad and confusing advice to several folks who contacted me
> > directly. I have asked, pleaded cajoled and sometimes been quite rude
> > about a modern release update. I just had to defend to the Navy why the
> > things I have been pitching them are not in an official release and I
> > had to tell them I have no control over the release process. There have
> > been several epoches in recent time when that could have been done with
> > generally good results, like just before Danny's recent I/O changes.
> > Those changes are to the very best good, more things kinda did work
> > before that adventure began.
> >
> > Now, if Danny is close to fixing the last bug in the IPv6 code, that
> > becomes the new releae version, not a release candidate, a release
> > version. If that doesn't happen reasonably soon, I will do that myself
> > and we should start talking divorce from Delaware.
> >
> > Dave
> > _______________________________________________
> > hackers mailing list
> > hackers at ntp.org
> > http://mailman.ntp.org/mailman/listinfo/hackers
> >



More information about the hackers mailing list