[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [RFC] Allow cancelation of NTP synchronisation



On Sun, Jul 06, 2008 at 01:57:05PM -0400, Joey Hess wrote:
> Jérémy Bobbio wrote:
> > The first patch is against rdate, and adds, when -vv is specified on the
> > command-line, the output of single dot (".") for each NTP query sent.
> > 
> > The second patch modifies clock-setup to use these dots in order to
> > step the progress bar on each of these dots.  In case of a cancelation,
> > rdate is killed on SIGPIPE due to the standard output closure.
> 
> That's a beautiful approach. Especially if it works. :-)

It worked during my tests… :)  Except that I am working offline
and thus I never tested (yet) when a server is actually available.

> The clock-setup changelog could close #468565.

Will do. :)

> However, what about #436497? Until that's fixed, any use of PROGRESS
> commands while the clock could potentially be moved backwards can cause
> newt to hang. Hmm, looks like there's a new newt upstream that fixes
> this.

As PROGRESS is only sent when a '.' is received, and the later only
printed when a query is sent, no further PROGRESS should be issued after
the clock has been set by rdate.

But if you find the time to investigate the new newt upstream, please
do! :)

Cheers,
-- 
Jérémy Bobbio                        .''`. 
lunar@debian.org                    : :Ⓐ  :  # apt-get install anarchism
                                    `. `'` 
                                      `-   

Attachment: signature.asc
Description: Digital signature


Reply to: