Red Hat Bugzilla – Bug 869595
No log out / reboot notification after 'critical' updates (eg kernel or qt-x11)
Last modified: 2013-02-14 17:09:50 EST
Description of problem:
Some updates require a log out or reboot to take effect (eg. kernel update). At the end of the update apper doesn't show a notification in the system tray. It only says that the update was succesful and last check was x hours en y minutes ago.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. remove latest kernel or qt-x11
2. run Software Management -> updates
3. update the removed package
No log out / reboot notification
log out / remove notification
With apper-0.8.0-0.11.20121126git.fc18.x86_64 some progress has been made. Now I get a notification. In this message there is a "Restart" button - for both reboot and log out - that is doing nothing.
Besides this issue, the notification itself is IMHO not a very good choice. Take the following situation: I let apper install updates and because of the amount of them I decide to leave my computer. The need to reboot my system after the update (a kernel was updated or so), I naturally misses. When I come back everything looks fine, nothing has to be done. The problem is that I won't get the need to reboot my system, because the current notification system has already timed out the message. So I don't think that the 'permanent' icon (as in previous versions of apper) in the system tray can be replaced. It's an important notifier.
This is why I wouldn't suppose to close this report yet.
it is possible to make persistent notifications, I believe.
Is it a wish for upstream?
The popup is persistent,
so when you get back, everything doesn't just looks fine,
it has an unread notification in the tray.
The ugly truth is that most users (including me - the Apper developer)
just don't care about that, even the updater icon is just ignored
by those who don't want to install updates.
So it's just easier to close a notification an keep on
working, after all most tasks that do require a restart
are a kernel* update or some more important system package
but most things (except Firefox) works fine, KDE for instance
you will be forced to logout on some situations where everything
just stops working.
I've been analyzing this tray icon for quite some time,
it just doesn't work, the only way it would work
is if the icon keep jumping on you face, and you'd just
kill it's process id :P
About it not working well that's precisely the reason I didn't
make a release yet, some stuff is not well tested yet, I really
hope I can make it ready this week :D and thanks for the tip
so I test it more carefully to see if it's working.
*And they are often installed side by side so no need to rush.
Thanks for your clear explanation. I'll try to better keep an eye on the notifications :-)
Bwt. The ktp-project uses a respond button in their notification. When a message comes in and I immidiatly click on this button the conversation window opens. When I'm to late - more than two seconds - the button does nothing. May be this can be helpful.
If the notification is persistent they are doing
the wrong thing to handle it. If not I think plasma
should hide the buttons (but it doesn't currently).
> The popup is persistent,
> so when you get back, everything doesn't just looks fine,
> it has an unread notification in the tray.
What if the user still uses the old-school non-Plasma notifications? (I do…)
you mean from knotify? if so it should just work too.
Yes, I mean KNotify ones.
BTW, thanks for the report I just saw
that I forgot to make them persistent,
so the bug is properly fixed now :)
Confirmed fixed in apper-0.8.0-2 builds for me.
Yes, 0.8.0-2 now suffers from https://bugs.kde.org/show_bug.cgi?id=314816 instead. (We already had that one once, it seems that that longstanding bug was just masked by this bug.)
That longstanding bug was fixed in 0.7.x, but PackageKit 0.8.x introduced it again, so a change in PackageKit is now needed in order to fix it, tho it's not related to the original bug report.
And it's tracked in a new bug, so everything's fine. :-) This one is closed (as it should be), as is the original bug against PackageKit 0.7.x.