Hello,
we're discussing a release criterion which is tightly related to GNOME and KDE environments. Could your teams please provide some feedback on the proposed criterion update, see below? Is user switching something you believe should be working (and in which milestone?), or should we rather de-emphasize/remove that function?
Please respond to the test@ list, if possible, to avoid fragmenting the discussion.
Thanks!
----- Original Message -----
Hello,
yesterday we have discussed whether user switching should be included in our criteria. We agreed that Beta is a good target for it, and accepted this bug as a blocker:
https://bugzilla.redhat.com/show_bug.cgi?id=1184933 https://bugzilla.gnome.org/show_bug.cgi?id=719418
Now we need to adjust this criterion: https://fedoraproject.org/wiki/Fedora_22_Beta_Release_Criteria#Shutdown.2C_r... Shutting down, logging out and rebooting must work using standard console commands and the mechanisms offered (if any) by all release-blocking desktops. Work? [hide] Similar to the Alpha criterion for shutting down, shutdown and reboot mechanisms must take storage volumes down cleanly and correctly request a shutdown or reboot from the system firmware. Logging out must return the user to the environment from which they logged in, working as expected.
I propose this change: title: Shutdown, reboot, logout, switch Shutting down, rebooting, logging out and user switching must work using standard console commands and the mechanisms offered (if any) by all release-blocking desktops. Work? [hide] Similar to the Alpha criterion for shutting down, shutdown and reboot mechanisms must take storage volumes down cleanly and correctly request a shutdown or reboot from the system firmware. Logging out must return the user to the environment from which they logged in, working as expected. User switching must allow multiple users to perform live switching between their sessions, working as expected.
What do you think?
On Wed, 2015-01-28 at 07:51 -0500, Kamil Paral wrote:
Hello,
we're discussing a release criterion which is tightly related to GNOME and KDE environments. Could your teams please provide some feedback on the proposed criterion update, see below? Is user switching something you believe should be working (and in which milestone?), or should we rather de-emphasize/remove that function?
Please respond to the test@ list, if possible, to avoid fragmenting the discussion.
Thanks!
----- Original Message -----
Hello,
yesterday we have discussed whether user switching should be included in our criteria. We agreed that Beta is a good target for it, and accepted this bug as a blocker:
https://bugzilla.redhat.com/show_bug.cgi?id=1184933 https://bugzilla.gnome.org/show_bug.cgi?id=719418
Now we need to adjust this criterion: https://fedoraproject.org/wiki/Fedora_22_Beta_Release_Criteria#Shutdown.2C_r... Shutting down, logging out and rebooting must work using standard console commands and the mechanisms offered (if any) by all release-blocking desktops. Work? [hide] Similar to the Alpha criterion for shutting down, shutdown and reboot mechanisms must take storage volumes down cleanly and correctly request a shutdown or reboot from the system firmware. Logging out must return the user to the environment from which they logged in, working as expected.
I propose this change: title: Shutdown, reboot, logout, switch Shutting down, rebooting, logging out and user switching must work using standard console commands and the mechanisms offered (if any) by all release-blocking desktops. Work? [hide] Similar to the Alpha criterion for shutting down, shutdown and reboot mechanisms must take storage volumes down cleanly and correctly request a shutdown or reboot from the system firmware. Logging out must return the user to the environment from which they logged in, working as expected. User switching must allow multiple users to perform live switching between their sessions, working as expected.
What do you think?
I think user switching is much less important than the other things on your list (login, shutdown, reboot) - if user switching is broken, the majority of our users won't even notice because they are on single-user systems.
Testing it is fine of course, but I don't really think we should block on this. Also worth considering: user switching has never worked 100% reliably, so increased qa focus may just uncover old heisenbugs, and turn them into blocking issues.
On 01/28/2015 09:41 AM, Matthias Clasen wrote:
On Wed, 2015-01-28 at 07:51 -0500, Kamil Paral wrote:
Hello,
we're discussing a release criterion which is tightly related to GNOME and KDE environments. Could your teams please provide some feedback on the proposed criterion update, see below? Is user switching something you believe should be working (and in which milestone?), or should we rather de-emphasize/remove that function?
Please respond to the test@ list, if possible, to avoid fragmenting the discussion.
Thanks!
----- Original Message -----
Hello,
yesterday we have discussed whether user switching should be included in our criteria. We agreed that Beta is a good target for it, and accepted this bug as a blocker:
https://bugzilla.redhat.com/show_bug.cgi?id=1184933 https://bugzilla.gnome.org/show_bug.cgi?id=719418
Now we need to adjust this criterion: https://fedoraproject.org/wiki/Fedora_22_Beta_Release_Criteria#Shutdown.2C_r... Shutting down, logging out and rebooting must work using standard console commands and the mechanisms offered (if any) by all release-blocking desktops. Work? [hide] Similar to the Alpha criterion for shutting down, shutdown and reboot mechanisms must take storage volumes down cleanly and correctly request a shutdown or reboot from the system firmware. Logging out must return the user to the environment from which they logged in, working as expected.
I propose this change: title: Shutdown, reboot, logout, switch Shutting down, rebooting, logging out and user switching must work using standard console commands and the mechanisms offered (if any) by all release-blocking desktops. Work? [hide] Similar to the Alpha criterion for shutting down, shutdown and reboot mechanisms must take storage volumes down cleanly and correctly request a shutdown or reboot from the system firmware. Logging out must return the user to the environment from which they logged in, working as expected. User switching must allow multiple users to perform live switching between their sessions, working as expected.
What do you think?
I think user switching is much less important than the other things on your list (login, shutdown, reboot) - if user switching is broken, the majority of our users won't even notice because they are on single-user systems.
Testing it is fine of course, but I don't really think we should block on this. Also worth considering: user switching has never worked 100% reliably, so increased qa focus may just uncover old heisenbugs, and turn them into blocking issues.
I use user switching frequently on my in-law's laptop. What is not reliable about it?
kde mailing list kde@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/kde New to KDE4? - get help from http://userbase.kde.org
Matthias Clasen wrote: I think user switching is much less important than the other things on your list (login, shutdown, reboot)
I agree it's not as important as shutdown, but the question is whether it's important enough to be considered "basic functionality" and we block the release when it's broken heavily.
For example, by our current standards, gnome-calculator must work or we do not ship: https://fedoraproject.org/wiki/Fedora_22_Final_Release_Criteria#Default_appl... We also don't ship when we don't have high-contrast icons for all apps, or when we miss some artwork.
All of that is probably not as important as shutdown working, but still forms some kind of basis which needs to work.
Btw, I just realized that it could be argued that user switching is already covered by this criterion: https://fedoraproject.org/wiki/Fedora_22_Final_Release_Criteria#Default_pane... but it's good we're having this discussion, at least we can clarify this.
if user switching is broken, the majority of our users won't even notice because they are on single-user systems.
It's hard to make some estimates here, but judging by my surroundings, single user systems are very common for us IT geeks, but do not necessarily represent Fedora's larger audience. On my home laptop, I have user profiles for me and for my girlfriend (my account is mostly logged in when the laptop gets suspended), and I need her to be able to use it even when I'm not around to manually switch VTs. My dad's laptop has user profiles for him and my mom. And even I have a guest account on my work laptop when I need to lend it to someone for a minute, I guess that's also a very common use case.
It seems to me that family usage is the most commonly mentioned scenario when I see people complaining about this feature being broken.
Testing it is fine of course, but I don't really think we should block on this.
Do you mean not block on it even for Final, i.e. at all, or Beta?
What I see as a problem is that any serious bugs [1] cut off multi-user usage of that system almost completely. I feel that we're losing a really big part of our audience this way. And now that we have Fedora Workstation product (and the KDE spin) with its more targeted focus on a specific user base, it would be a real pity to sacrifice those people.
Also worth considering: user switching has never worked 100% reliably, so increased qa focus may just uncover old heisenbugs, and turn them into blocking issues.
It that a bad thing or a good thing? Could this be a worthy goal for F22? It's still pretty early in the cycle. Or if you think it's too late, does it make sense to target this for F23? Is it about technical complexity, or is there no will/manpower to maintain this? (Since this is one of my painful issues with GNOME, I can promise you some QA time from my side:)).
Thanks, Kamil
Matthias Clasen wrote:
On Wed, 2015-01-28 at 07:51 -0500, Kamil Paral wrote:
I propose this change: title: Shutdown, reboot, logout, switch Shutting down, rebooting, logging out and user switching must work
...
What do you think?
I think user switching is much less important than the other things on your list (login, shutdown, reboot) - if user switching is broken, the majority of our users won't even notice because they are on single-user systems.
Testing it is fine of course, but I don't really think we should block on this.
I agree with Matthias, user switching isn't worth considering a blocker.
Besides, even if broken, it can easily be fixed by subsequent updates.
-- Rex
Matthias Clasen wrote:
Testing it is fine of course, but I don't really think we should block on this. Also worth considering: user switching has never worked 100% reliably, so increased qa focus may just uncover old heisenbugs, and turn them into blocking issues.
Maybe for GNOME it hasn't. For KDE, using KDM or SDDM, user switching has always just worked. Nowadays, it should just work with GDM too, but of course it can be subject to bugs and undocumented behavior changes in GDM (which wouldn't be a first).
IMHO, adding user switching to the blocker criteria is a good idea. Bugs in user switching can easily lead to data loss, either directly (e.g., if user switching crashes the machine) or indirectly (e.g., if, due to user switching not working, somebody just logs the other account off without caring about running processes, or even hits the power or reset button if the session is locked).
Kevin Kofler
On 02/02/2015 09:48 PM, Kevin Kofler wrote:
Matthias Clasen wrote:
Testing it is fine of course, but I don't really think we should block on this. Also worth considering: user switching has never worked 100% reliably, so increased qa focus may just uncover old heisenbugs, and turn them into blocking issues.
Maybe for GNOME it hasn't. For KDE, using KDM or SDDM, user switching has always just worked. Nowadays, it should just work with GDM too, but of course it can be subject to bugs and undocumented behavior changes in GDM (which wouldn't be a first).
IMHO, adding user switching to the blocker criteria is a good idea. Bugs in user switching can easily lead to data loss, either directly (e.g., if user switching crashes the machine) or indirectly (e.g., if, due to user switching not working, somebody just logs the other account off without caring about running processes, or even hits the power or reset button if the session is locked).
Kevin Kofler
+1 - Agreed, or whatever. Since switching to KDE, I've not had _any_ issues with user switching in KDM or SDDM. You are correct in stating that it doesn't work in GNOME. This is one of the reasons I use KDE (among many other) - my family shares a computer, and we switch users all day; this feature was one thing we used heavily on Windows and was a requirement before I introduced Fedora to them.
With all that said, I feel adding user switching to the blocker criteria is a logical step - this is 2015. It has worked on Windows since Windows Vista, circa 2003. I use it at home, and I use it at work as an administrator to fix problems for my users without needing to log them out.
Just because it doesn't work on GNOME isn't a valid reason not to test it and block on it if it is a feature that is used and depended on.
On Tue, 03 Feb 2015 15:43:01 +0100, Dan Mossor danofsatx@gmail.com wrote:
On 02/02/2015 09:48 PM, Kevin Kofler wrote:
Matthias Clasen wrote:
Testing it is fine of course, but I don't really think we should block on this. Also worth considering: user switching has never worked 100% reliably, so increased qa focus may just uncover old heisenbugs, and turn them into blocking issues.
Maybe for GNOME it hasn't. For KDE, using KDM or SDDM, user switching has always just worked. Nowadays, it should just work with GDM too, but of course it can be subject to bugs and undocumented behavior changes in GDM (which wouldn't be a first).
IMHO, adding user switching to the blocker criteria is a good idea. Bugs in user switching can easily lead to data loss, either directly (e.g., if user switching crashes the machine) or indirectly (e.g., if, due to user switching not working, somebody just logs the other account off without caring about running processes, or even hits the power or reset button if the session is locked).
Kevin Kofler+1 - Agreed, or whatever. Since switching to KDE, I've not had _any_ issues with user switching in KDM or SDDM. You are correct in stating that it doesn't work in GNOME. This is one of the reasons I use KDE (among many other) - my family shares a computer, and we switch users all day; this feature was one thing we used heavily on Windows and was a requirement before I introduced Fedora to them.
With all that said, I feel adding user switching to the blocker criteria is a logical step - this is 2015. It has worked on Windows since Windows Vista, circa 2003. I use it at home, and I use it at work as an administrator to fix problems for my users without needing to log them out.
Just because it doesn't work on GNOME isn't a valid reason not to test it and block on it if it is a feature that is used and depended on.
Well, you don't get user switching (and the kickoff item) every time something's messed up with your PAM session. This is not exclusive to GDM though, SDDM suffered from this a lot too in past. Changing something in PAM, PAM configs or the DMs can bring the problem back. Anyway I agree this should be tested.
Dan Mossor wrote:
+1 - Agreed, or whatever. Since switching to KDE, I've not had _any_ issues with user switching in KDM or SDDM. You are correct in stating that it doesn't work in GNOME. This is one of the reasons I use KDE (among many other) - my family shares a computer, and we switch users all day; this feature was one thing we used heavily on Windows and was a requirement before I introduced Fedora to them.
GNOME these days supports user switching ONLY with GDM. :-( (Yet another example of how GNOME does not give a darn about interoperability with anything else.) Even with GDM, it could be broken, but with other display managers, it is not even expected to work anymore.
Kevin Kofler
I wrote:
GNOME these days supports user switching ONLY with GDM. :-( (Yet another example of how GNOME does not give a darn about interoperability with anything else.) Even with GDM, it could be broken, but with other display managers, it is not even expected to work anymore.
PS: Oh, and while for user switching, this has been the case already for a while, recently, GNOME has started requiring GDM even for SCREEN LOCKING! (!!!)
Kevin Kofler
-
Dan Mossor -
Kamil Paral -
Kevin Kofler -
Martin Bříza -
Matthias Clasen -
Patrick Boutilier -
Rex Dieter