SIGHUP / SIGUSR2 not equivalent to restart

[Dave Abrahams]

on Fri Sep 30 2011, Sebastian Spaeth <Sebastian-AT-SSpaeth.de> wrote:

> On Thu, 29 Sep 2011 15:14:11 -0400, Dave Abrahams <dave at boostpro.com> wrote:
>> 
>
>> I just got back after being offline for a few hours.
>> I have something that pokes offlineimap with a SIGUSR2 when I
>> establish a network connection.  You can see it taking effect in the log
>> below (There are actually two of those; the first comes in when the
>> display wakes up at 1:57:05.  
>> 
>> However, I didn't receive any new mails until I did a SIGUSR1 and
>> relaunch, which you can also see at 3:00:21 where the pid changes.
>
> Actually, I think you want a SIGUSR1. Which will abort and do a new
> round of syncs, but continue to loop in general.
>
> Sending a SIGUSR2 should quit offlineimap as soon as all current syncs
> are finished.
>
> I just tested that on the current next branch, and it worked fine for
> me...

Sorry, that was just mis-reported.  It *is* sending a SIGUSR1, names in
the logs aside

>> It's my understanding that these ought to be equivalent operations for
>> most intents and purposes, and yet I've repeatedly seen situations where
>> I had to restart the process in order to get anywhere.
>
> SIGUSR1 and SIGUSR2 equivalent? No, SIGUSR2 will quit offlineimap and
> SIGUSR1 will continue to loop.

No, what I meant was that SIGUSR2 followed by starting offlineimap is
supposed to be equivalent to SIGUSR1

Sorry for the confusion.

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com