hi jeroen, steve, if i understand correctly, there are two issues:
- offset of zero being treated as an error
- usage of ntpdate
for the former, i believe this issue is resolved in the latest version
of ntp. for the latter, is using ntpdate really a problem? jeroen:
you claim that ntpdate is deprecated, but is it? afaik it's still quite
useful in many situations, including this case.
On Sun, Oct 09, 2005 at 11:27:53PM +0200, Jeroen van Wolffelaar wrote:
> gathering from ntpdate output to gathering the same data from ntpq output. If
> ntpdate is really deprecated, something to pass along upstream, I'd say (true
> in general for these bugs, I might add).
you're talking to someone who's also on upstream :) in fact, i'd much
rather deal with bugs in the debian bts then the POS sourceforge
tracker, and then migrate them to upstream cvs in one big batch later on.
> That said, my patch in #268044 should still apply manually -- while the regex
> has changed, it became stricter rather than more lenient, and the point stands
> that the exact format of each column really doesn't matter and only makes
> things more fragile, and only the contents of those columns that are actually
> used matter (2, 4 and 11). So also the current regex can be replaced by the
> one I proposed in #268044.
i'll include this patch in the next update.
sean
--
signature.asc
Description: Digital signature
