Am I missing something simple here?

Eric S. Raymond esr at thyrsus.com
Wed Jan 4 01:49:54 UTC 2017


Hal Murray <hmurray at megapathdsl.net>:
> 
> esr at thyrsus.com said:
> > The worst.  Runtime errors - the protocol machine never gets out of INIT,
> > even though the unit tests for the macros pass.
> 
> What does ntpq -p say?

Like I said, the pool host returns its usual number of peers.  Then they hang
in INIT state forever.

> If that is blank, look in rawstats and/or peerstats.  Peerstats has a status 
> word.  There are a bunch of status/flag bits that record why packets get 
> dropped, but I don't think they get included in rawstats.  If they don't get 
> to peerstats, I'd probably add a msyslog line to record why packets are 
> getting dropped.

I'll look at peerstats on my next try, I guess.

> Did your unit tests verify that the old and new interoperate or just that the 
> new is self consistent?

The unit tests verify that the macros can assemble l_fp instances from
32-bit halves and get the same 32-bit halves back out.  They also check that
incrementing a high half (the integer seconds part)  by 1 gives the
expected result.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>


More information about the devel mailing list