ntp-4.0.91 is available (RELEASE CANDIDATE)
Harlan Stenn
ftp://ftp.udel.edu/pub/ntp/testing/
With the exception of folks who seem to have lousy motherboard
oscillators, we think this version is stable. I know of two outstanding
issues:
- Dave is looking at a situation where more than 3 refclocks are
apparently considered as "suitable"
- We are waiting to see if the burst mode stuff really works for analog
and ISDN modem-based refclocks
I've upgraded to autoconf-2.13, reordered several tests in configure.in,
and incorporated several portability patches. I've compiled this code
under both gcc and cc on a variety of platforms here.
Please try this out and let us know if you have any problems. Again,
for problems under Linux it's best to ask on the newsgroup.
From the NEWS file:
(4.0.91)
* New ONCORE driver
* New MX4200 driver
* Palisade improvements
* config file bugfixes and problem reporting
* autoconf upgrade and cleanup
* HP-UX, IRIX lint cleanup
* AIX portability patches
* NT cleanup
From the ChangeLog:
Sun Jan 10 15:15:07 1999 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.91
Sat Jan 9 00:11:34 1999 Harlan Stenn <st...@whimsy.udel.edu>
* include/ntp_stdlib.h: msyslog() is declared differently if we're
not __STDC__.
* include/ntp_types.h: It's HAVE_PROTOTYPES, not USE_PROTOTYPES.
* include/ntp_machine.h: Ditto.
Fri Jan 8 20:47:10 1999 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: Upgrade to autoconf-2.13
Do the prototypes check much earlier, as it might alter CFLAGS and
things which will affect other tests.
* ntpd/ntp_request.c (do_conf): The problem was with a template
for "version" on an IRIX C compiler...
From: Marc....@waii.com
* libntp/authkeys.c: #include config.h first.
Reported by: brian....@funb.com
Thu Jan 7 00:24:35 1999 Harlan Stenn <st...@whimsy.udel.edu>
* util/tickadj.c (main): return() instead of exit().
* ntpd/ntp_request.c (do_conf): Disambiguate ||.
* ntpd/ntp_proto.c (clock_select): Initialize variables.
From: Marc....@waii.com
* scripts/ntpver.in: Use PATH_SH
* configure.in (PATH_SH): Added.
Tue Jan 5 19:02:51 1999 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90h
* html/driver30.htm: Updated.
* html/refclock.htm: Refer to driver30
* ntpd/refclock_oncore.c: Vastly improve and make less FreeBSD centric,
From: Poul-Henning Kamp <p...@critter.freebsd.dk> and
Reg.Clemens <r...@dwf.com>
* include/ntp.h: Portability/lint patches
* libntp/binio.c: Ditto.
* libntp/caljulian.c: Ditto.
* libntp/caltontp.c: Ditto.
* libntp/ieee754io.c: Ditto.
* libntp/md5c.c: Ditto.
* libntp/mfp_mul.c: Ditto.
* libntp/msyslog.c: Ditto.
* libntp/statestr.c: Ditto.
* libntp/systime.c: Ditto.
* libparse/clk_trimtsip.c: Ditto.
* libparse/data_mbg.c: Ditto.
* libparse/parse.c: Ditto.
* ntpd/ntp_control.c: Ditto.
* ntpd/ntp_filegen.c: Ditto.
* ntpd/ntp_intres.c: Ditto.
* ntpd/ntp_io.c: Ditto.
* ntpd/ntp_peer.c: Ditto.
* ntpd/ntp_proto.c: Ditto.
* ntpd/ntp_util.c: Ditto.
* ntpd/ntpd.c: Ditto.
* ntpd/refclock_arc.c: Ditto.
* ntpd/refclock_chu.c: Ditto.
* ntpd/refclock_datum.c: Ditto.
* ntpd/refclock_leitch.c: Ditto.
* ntpd/refclock_parse.c: Ditto.
* ntpd/refclock_usno.c: Ditto.
* ntpq/ntpq.c: Ditto.
* util/tickadj.c: Ditto.
From: Marc....@waii.com
Mon Jan 4 00:56:55 1999 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90g
* ntpd/ntp_config.c (getconfig): MODE was setting ttl, not hmode.
Reported by: Carsten Emde <c...@ceag.ch>
Fri Dec 4 01:01:14 1998 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90f
* ntpd/refclock_mx4200.c: New version
From: Marc....@waii.com
1998-12-02 Harlan Stenn <st...@whimsy.udel.edu>
* ntpd/ntp_config.c (do_resolve_internal): If fork fails, say why.
Reported by: Jeff_D...@admin.tc.faa.gov
* ntpd/ntpd.c (ntpdmain): fork() can return a -1. Someday we'll
report this condition...
1998-12-02 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90e
* ntpd/refclock_palisade.c: Reformat code so ansi2knr will work
* ntpd/refclock_palisade.h: Ditto
From: Marc....@waii.com
Sun Nov 29 21:00:53 1998 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90d
* configure.in (CFLAGS): Use "-O2 -g3 -n32" by default for Irix6.2
and later.
Reported by: Jack Bryans <jbr...@csulb.edu>
1998-11-29 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90c
* ntpd/refclock_oncore.c (oncore_msg_En): Convert to nano
From: John Hay <jh...@mikom.csir.co.za>
* include/ntp_request.h (RM_VN_MODE): Add version parameter, so
xntpdc will work across v3 and v4.
* ntpd/ntp_request.c: Track requested version
(req_ack): Use requested version in RM_VN_MODE
(more_pkt): Ditto
(flush_pkt): Ditto
(process_private): Get requested version
* ntpd/ntp_intres.c (request): Use default version
* ntpdc/ntpdc.c (sendrequest): Ditto
From: John Hay <jh...@mikom.csir.co.za>
Fri Nov 27 14:27:21 1998 Harlan Stenn <st...@whimsy.udel.edu>
* ntpd/refclock_palisade.c: Lint cleanup
* ntpd/refclock_palisade.h: Ditto.
From: Marc Brett <mbr...@rgs0.london.waii.com>
Mon Nov 23 04:45:03 1998 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90b
* New code and cleanup for the NT stuff
From: Carl Byington <ca...@five-ten-sg.com>
Sat Nov 21 21:21:45 1998 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90a
* libntp/systime.c (step_systime): net_set_tod calls clock_settime.
* libntp/machines.c (ntp_set_tod): Take a 2nd arg for NT.
* include/ntp_machine.h: ntp_set_tod() has 2 args always.
* ports/winnt/bldrel.bat: Typo.
From: Carl Byington <ca...@five-ten-sg.com>
* ntpd/ntp_intres.c (findhostaddr): h_errno is a #define under AIX.
* configure.in: clock_settime is a stub in AIX4.
From: Perry Ross <pr...@platinum.com>
* libntp/Makefile.am (EXTRA_DIST): Lose libntp.mak
* ntpd/Makefile.am (EXTRA_DIST): Ditto.
* ntpdate/Makefile.am (EXTRA_DIST): Ditto.
* ntpdc/Makefile.am (EXTRA_DIST): Ditto.
* ntpq/Makefile.am (EXTRA_DIST): Ditto.
From: Greg Schueman <schu...@ix.netcom.com>
Sat Nov 21 12:33:16 1998 Harlan Stenn <st...@whimsy.udel.edu>
* configure.in: 4.0.90
Harlan Stenn
he says the initial report of more than 3 refclocks being considered as
"suitable" was probably due to a runtime error (ie, the code and
algorithms seem to be OK), and we got a confirmation that the existing
burst code works for modem and ISDN refclocks.
4.0.91 is still a release candidate, and will be released as ntp-4.1.0
in about a week or so barring any negative feedback.
H
Serge Maleyev
>
> It's at:
>
> ftp://ftp.udel.edu/pub/ntp/testing/
>
> With the exception of folks who seem to have lousy motherboard
> oscillators, we think this version is stable. I know of two outstanding
> issues:
>
> - Dave is looking at a situation where more than 3 refclocks are
> apparently considered as "suitable"
>
> - We are waiting to see if the burst mode stuff really works for analog
> and ISDN modem-based refclocks
>
> I've upgraded to autoconf-2.13, reordered several tests in configure.in,
> and incorporated several portability patches. I've compiled this code
> under both gcc and cc on a variety of platforms here.
>
> Please try this out and let us know if you have any problems.
I had a problem to compile it under Solaris 2.6.
Both gcc and cc could not compile parsesolaris.c
Has anyone successfully compiled it under Solaris 2.6?
Serge
--
smal...@cc.umanitoba.ca
Systems Analyst Academic Computing and Networking
624 Engineering Bldg., 15 Gillson Str., The U[niversity] of M[anitoba]
Winnipeg, MB R3T 5V6, Canada Tel:(204)474-8311 Fax:(204)474-7515
Syed Zaeem Hosain
>
> Harlan Stenn wrote:
> >
> > It's at:
> >
> > ftp://ftp.udel.edu/pub/ntp/testing/
> >
> > With the exception of folks who seem to have lousy motherboard
> > oscillators, we think this version is stable. I know of two outstanding
> > issues:
> >
> > - Dave is looking at a situation where more than 3 refclocks are
> > apparently considered as "suitable"
> >
> > - We are waiting to see if the burst mode stuff really works for analog
> > and ISDN modem-based refclocks
> >
> > I've upgraded to autoconf-2.13, reordered several tests in configure.in,
> > and incorporated several portability patches. I've compiled this code
> > under both gcc and cc on a variety of platforms here.
> >
> > Please try this out and let us know if you have any problems.
>
> Both gcc and cc could not compile parsesolaris.c
> Has anyone successfully compiled it under Solaris 2.6?
Yes, I have recently successfully built this version and deployed
on Solaris 2.5.1 and Solaris 2.6 using the gcc 2.8.1 compiler ...
things seems to be running fine.
Z
--------------------------------------------------------------------
| Syed Zaeem Hosain |
| Aeris Communications, Inc. tel: (408) 557-1905 |
| 1245 S. Winchester Blvd., Suite 216 fax: (408) 557-1925 |
| San Jose, CA 95128-3908. Syed....@aeris.net |
--------------------------------------------------------------------
Jon Baggott
around since at least Solaris 2.3 (1994) Sun still haven't got around to
fixing it!
You have to patch /usr/include/sys/systm.h to change the prototype for ffs
from int ffs(long) to int ffs(const int)
Serge Maleyev wrote in message <36A648...@cc.umanitoba.ca>...
>
>I had a problem to compile it under Solaris 2.6.
>Both gcc and cc could not compile parsesolaris.c
>Has anyone successfully compiled it under Solaris 2.6?
>
>Serge
Paul D. Smith
jb> This is because of a Sun bug (Bug ID 1131645). Although this bug
jb> has been around since at least Solaris 2.3 (1994) Sun still
jb> haven't got around to fixing it!
jb> You have to patch /usr/include/sys/systm.h to change the prototype
jb> for ffs from int ffs(long) to int ffs(const int)
This seems like _exactly_ the kind of thing autoconf is supposed to
handle for you...
--
-------------------------------------------------------------------------------
Paul D. Smith <psm...@baynetworks.com> Network Management Development
"Please remain calm...I may be mad, but I am a professional." --Mad Scientist
-------------------------------------------------------------------------------
These are my opinions---Nortel Networks takes no responsibility for them.