QNX (was: Re: Raspberry Pi NTP config with fudge factors

[Eric S. Raymond]

Gary E. Miller <gem at rellim.com>:
> > We *will* assume POSIX-compliant shared memory on our target systems.
> 
> QNX has POSIX compliant shared memory, just the new one, not the old one.

Are we presently using the old one?  Where is documentation for the new one?

> What about Windows?

It too is not our problem yet.  At the rate Microsoft is now busily
Unixizing Windows, it might not be a problem at all by the time we get
there.  If we get there

Besides, while there's a case for supporting NTP client mode on Windows,
anybody who tries to use it as a primary time server is too stupid to live.

> > My intention is to officially deprecate refclocks 20 and 48 in favor
> > of 28 (SHM), explaining that the way 1PPS and in-band information is
> > mingled produces bad behavior on 1PPS dropouts.   Actual deprecation
> > will wait on confirmation from Gary's tests.
> 
> #48 does not have to work that way, it is just default configured that
> way.  If the #48 defaults are changed to sane one, and some testing, I
> would say it is would then be preferred over #28.
>
> For example, TAI to UTC offset is in the JSON stream (currently unused). 
> Adding TAI offset to SHM would be a major incompatible change.

OK, so deprecate #20 and fix #48. That's still clearer direction than we had.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 811 bytes
Desc: Digital signature
URL: <http://lists.ntpsec.org/pipermail/devel/attachments/20160506/7ff59cd5/attachment.bin>