[ntp:hackers] The Distinction between Development and Stable

Dave Hart davehart at gmail.com
Sun Apr 10 03:27:31 UTC 2011


On Sun, Apr 10, 2011 at 02:42 UTC, Harlan Stenn <stenn at ntp.org> wrote:
> While I'd rather not start another RC cycle for -stable given that I am
> hoping to start the RC cycle on -dev within the next month or so, if you
> want to see if you can just backport refclock_oncore.c from -dev to
> -stable we can do that.

Harlan tried dropping ntp-dev's refclock_oncore.c into ntp-stable and
exposed a few issues.  Every one of the breaking changes is something
I recognize because I broke it, and none of them should be difficult
to overcome.

1)  You'll see incompatible pointer warnings because struct refclockio
member srcclock changed from a caddr_t (char *) to a struct peer * and
I removed the (struct peer *) cast in the drivers.  Ignore the
warning, or add the cast back.

2)  ntp-dev's refclock_oncore.c fails to compile in ntp-stable
complaining of no declaration of the months[] array.  Either copy and
paste the months[] declaration from ntp-dev's ntp_calendar.c into the
-stable refclock_oncore.c, or just backport all of
include/ntp_calendar.h and libntp/ntp_calendar.c.

3)  ntp-dev oncore_log() calls mprintf_clock_stats().  Either backport
the latter, which may also require backporting libntp/msyslog.c, or
decompose the call in refclock_oncore.c to [m]snprintf() followed by
record_clock_stats().

I am anxious to see the 4.2.8 -RC cycle start ASAP.

Cheers,
Dave Hart


More information about the hackers mailing list