Re: [Evolution] Evolution not listening to quit

[Milan Crha <mcrha redhat com>]

        Hi,

On Sun, 2019-06-02 at 20:48 +0200, Ralf Mardorf via evolution-list
wrote:
> Ok, removing '.running' might be something done by 'evolution-
> backup', not by your script. My bad.

right, everything prefixed with "evolution-backup-Message" in the
original post is printed by the evolution-backup tool. Looking at it,
that means each pasted line in the original message from the
./evo-backup.sh invocation is printed by the evolution-backup tool.

On Sun, 2019-06-02 at 13:09 +0200, Ralf Mardorf via evolution-list
wrote:
> As a user we might not know what processes to SIGTERM and in which
> order, while the developers of an app do know this. IOW a SIGTERM
> isn't as evil as a SIGKILL, but do we know, in what order we need to
> SIGTERM evolution related processes?

a) the evolution-backup calls that "evolution -quit" command,
   not Michelle's script;
b) I'm not sure why you question SIGTERM and SIGKILL here, especially
   when the user cannot do anything about it;
c) it's not only about the correct order (which
   evolution --force-shutdown, aka killev, follows), but also about
   the desktop environment, because killing evolution(-data-server)
   processes can be tricky in some of them (I know of GNOME Shell,
   whose gnome-shell-calendar-server restarts
   evolution-calendar-factory process whenever it disappears, thus it's
   close to impossible to stop the calendar factory there).

I don't want to add more off-topic discussion here, I only wanted to
clarify certain things.

Michelle, by the way, the "--restart" argument of the evolution-backup
means to start evolution after the backup is done. I'm not sure whether
you do that on purpose. Also, the missing .running file indicates that
evolution is not running, thus, maybe, the other running instance is a
stale process, or it's running for a different user. I'm just guessing
here. I would try to `kill -9` it and then retest the
`evolution --quit`.
        Bye,
        Milan