[ntp:hackers] hackers] NTP software numbering

Harlan Stenn stenn at ntp.org
Fri Dec 26 06:57:30 UTC 2014


Brian Utterback writes:
> >>>> stenn at ntp.org said:
> >>>>> We're talking about changing from:
> >>>>>    ntpd 4.2.8p1-beta1 at 1.3268-o Wed Dec 24 21:02:52 UTC 2014 (2)
> >>>>> to:
> >>>>>    ntpd 4 v2.8.1-beta1 at 1.3268-o Wed Dec 24 21:02:52 UTC 2014 (2)

The above examples are what the old numbers would translate to.  Think
SunOS4 and SunOS5.X, the latter being known as Solaris X.

I hope I'm not about to trim anything immediately useful...

>>>> The NTPv5 will likely contain several elements of properties which is
>>>> unlike todays NTPv4, so I suspect that it won't be a direct drop-in
>>>> replacement, which is why it's also a new major-release.

I would expect it to be pretty close to a drop-in replacement.  There
might very well be config file changes, and who knows what else.  But we
could also take steps to make sure an older config file worked, or
provide a conversion script.  So I'm still OK with the first release of
NTPv5 code being packaged as ntp-5.0.0-tar.gz .


>>>> Besides, moving from ntp-4.2.8p0 to ntp-2.8.1 is not working well
>>>> either, number revision-wise, while moving to ntp-4.2.8p0 to ntp4-2.8.1
>>>> works, but is indeed quirky, but sufficiently good in my mind.

Nobody is proposing that.

If we had the package name of ntp-4.2.8p0, that would become a package
name of ntp4-2.8.0 under the new numbering scheme.

We would not be doing that.  We understood this years ago.

That's way we were looking at the next -stable release being ntp4-5.0.0,
so the package name would become ntp4, and the first -stable release
would be 5.0.0, which collates properly with a 4.2.8 release.

Please don't put too much attention on this - I'm liking going from:

 ntp-4.2.8

to

 ntp-4.4.0

for the next release, which also properly addresses the version number
stuff.

> Right now we have the worst of all worlds where there is no functional
> meaning to any of the numbers except the major one, the one that we
> are discussing doing away with.

I don't see this - I see us going from PROTO.MAJOR.MINORpPOINT (protocol +
release numbers) to Proto.Release.Point.

We never had particularly clear rules about when to bump a Major number,
and with the new scheme we just won't worry about it.

> Anything can happen at any point, including daily patch releases.

We've always had frequent patch releases for -dev.  That's OK.

We've also been pretty "reserved" about -stable patch releases.

We can have as many patches to -stable as we want.  This is clearly
documented at https://support.ntp.org/bin/view/Dev/MaintainerIssues .

> In the last few months multiple new licenses crept in,

Where?

> likewise new dependencies, neither of which were
> reflected in the Changelog file.

Please let me know what you mean - I do my best to make sure no
incompatible licenses come in to the codebase.

> Every time the license changes I have to get the legal department's
> approval before we can release the new version, a process that is at
> best time consuming. If I knew that a new license would only creep in
> at certain revs, then I could get a blanket approval for lower level
> changes, but I can't make that assurance. The same thing goes for
> backwards incompatibilities. I don't mean to rant, but if you are
> going to have rev numbers, I think they should mean something.

Agreed, and I'm confused because I don't know what you're talking about.

libopts is good - it's triple-licensed (BSD, GPL, and something else).

If you are talking about -dev, I'm not following this one either.  If
there were any licensing issues with any of those changes, why didn't
you speak up sooner?

We note backward-incompatible changes in the ChangeLog:

(4.2.7p342) 2012/12/31 Released by Harlan Stenn <stenn at ntp.org>
* [Bug 2081 - Backward Incompatible] rawstats now logs everything.

I grant you that I did not list these in the notes when I released
4.2.8.  That was an oversight on my part, and I've just updated:

 https://support.ntp.org/bin/view/Dev/NewsFileMaintenance

so that shouldn't happen again.

H


More information about the hackers mailing list