Re: RFS: tenshi (updated)

[Ansgar Burchardt <ansgar@43-1.org>]

Hi,

sorry for the late reply, the mail got a bit lost at the bottom of my
mailbox.

Ignace Mouzannar <ghantoos@ghantoos.org> writes:
>> Upgrading from 0.11-1 fails:
>>
>>  Updating tenshi user's home directory ...
>>  usermod: user tenshi is currently logged in
>
> Oops, my daemon must have not been running when I tested my package.
> Sorry about that.
>
> The error was due to the -R flag that was passed to dh_installinit in
> the past release: it prevented the upgrade process from stopping the
> daemon before the new installation started.
>
> I have corrected this behavior by adding a preinst script that stop
> the daemon. But I'm not sure this is the best way to go. Is it?

I think it should be okay.

> Here are the relevant parts of the new changelog:
> /
> | tenshi (0.11-2) unstable; urgency=low
> |
> |   * debian/control:
> |    (...)
> |     - Set debhelper dependency back to (>= 7) as dh_overrides are not used by
> |       debian/rules anymore.

You still use the "--with quilt" feature which requires debhelper 7.0.8.

I noticed two other small things in the init script which might be nice
to have fixed in the next upload (I don't think it is important enough
to get it included in Squeeze):

 · In do_start, one time $DAEMONUSER is used and one time the expanded
   value for it ("tenshi").
 · In the stop action, the process with the pid from $PIDFILE is killed
   without making sure it is actually the right process.  A stale pid
   file might point to another process.

Regards,
Ansgar