[Pool] Leap Indicator bits

Harlan Stenn stenn at ntp.org
Mon Jul 6 10:21:06 UTC 2015


Martin Burnicki writes:
> Harlan,
> 
> Harlan Stenn wrote:
> > If a bug was fixed in 4.2.8 but was not fixed in 4.2.6, then there were
> > over 1,100 other bugs that were also fixed in 4.2.8 but not in 4.2.6.
> 
> And how many of those bugs were introduced after 4.2.6, in 4.2.7*, and 
> then fixed?

Not many.

But I also don't plan to count them.

> IMO there are different kinds of bug. Some of them can only be fixed in 
> the next major version, and others are worth to roll out a patch release.
> 
> IMO security fixes implemented around last christmas, and evil things 
> like the sticky leap bit are really worth to roll out a patch release. 
> This would save many users and admins of really much trouble.

As I have said, there were not sufficient resources to do this.

I gave plenty of warning about the release of 4.2.8.  Folks had lots of
opportunity to test it and report problems.  I did the best and most
complete coverage testing I could do with the time and resources available.

We have decades of history that we only maintain a single stable
release.  No secret, no surprise.

Folks have also long know that if they want support for older
versions of stable that's available.  Nobody has asked for it.

> I think we have already discussed this in detail by email back in 
> December or January.

Yes.

There are lots of folks who think NTP is trivial, or that "it's been
around for a long time so there must not be anything left to do".  They
think a few hours' week should be all that's needed.  The folks who
think this have never been near the codebase.

I've been putting in over 105 hours/week on this for over 4 years' time
now.  That's 15+ hours/day, 7 days a week.

I put in 16-18 hour days in June, except for the 2-3 all-nighters.

That's in excess of 3 full-time "jobs", and I estimate there is work for
at least 12 full-time people making all of this work.

Volunteers like you, Martin, and several others put in needed and
valuable time to NTP.  I don't know for sure how much time that is, but
I suspect that all of it might sum up to about a half-time person.

We all want this to be better.  But unless and until we get useful
funding and technical support from competent and qualified folks, we're
going to be getting less than what we want.

Reality bites.
-- 
Harlan Stenn <stenn at ntp.org>
http://networktimefoundation.org - be a member!


More information about the pool mailing list