https://bugs.kde.org/show_bug.cgi?id=343518
--- Comment #51 from Wolfgang Bauer <wba...@tmo.at> --- (In reply to Piotr Mierzwinski from comment #50) > > So this might just be your point of regression... > The regression is that none of GTK2 and GTK3 applications there are not > restored. This is regression comparing to previous Plasma, because In ver. > 5.5.x it was working. Yes, but this bug report was filed over a year before Plasma 5.6.0 was even released. So by definition this cannot be the same problem if you say it worked with 5.5.x and earlier. >> No idea about the others though. > But I'm sure that the last two were built using GTK2. My test was following. > I changed "Style of > GNOME application" in "System Settings" and just > restarted all tested applications. Ok, as I wrote I have no idea. It would have been a possible explanation for why Firefox is restored reliably here (where it is using GTK2), but not for you (using GTK3). > I tested Kubuntu 16.04. BTW in Plasma 5 provided by > this distribution applications based on KF are not restored. I'm not going > to evolve this topic. Anyway they plan update to Plasma 5.6.x. Yes, that's what this bug report originally was about, and why I marked it as duplicate of bug#354724 (which was about the same issue). It was a problem in Qt5, and should be fixed in Plasma 5.6 when built against Qt 5.6. > Subject is "Plasma doesn't restore everything after a reboot". So maybe is > too general. Yes, but still it's not a good idea to mix several unrelated problems into one bug report, even if the symptoms are similar, IMHO. > So where is placed the bug? In Plasma 5.5.4 GTK applications are restored > correctly and in 5.6.4 don't. I don't know. Please note that I am not the maintainer of ksmserver, I just closed this as duplicate of the other bug report which was about the same issue, the problem in Qt5 and KF5 applications (it wasn't even related to ksmserver actually). But as I already mentioned, there was exactly one change in ksmserver between 5.5.5 and 5.6 (unless I overlooked something). So if you notice a regression, this is the only possible cause in ksmserver for it, I'd say. If it isn't, the problem must lie somewhere else, and likely not even in KDE as those are GTK applications. Although XSM is outdated and shouldn't be used any more, at least according to the commit message. > From my perspective it looks like some informations necessary to restore GTK > based applications are not saved into ksmserverrc. ksmserver does not differentiate between KDE and GTK applications, obviously. It implements a protocol, and the applications (or the toolkit they use) have to implement that protocol correctly as well. The problem with Qt5 applications was caused by the fact that Qt5 closed the applications too soon on logout (when told to save the state), so they were not running any more when ksmserver actually recorded the running applications into ksmserverrc. Another possible problem in this regard might be that the application is crashing when told to save its state, this again would result in it not running any more when ksmserver records the running applications during logout. > Even after hardcoded them > in this file, they exist only till first relogin/restart. After that they > disappear. How it might be explained? If an app is restarted, the corresponding entry is removed from ksmserverrc I think. If the app doesn't start even though it was listed in ksmserverrc, the entry might have been wrong, or the app might crash on start or fail to start because of some other reason. > So I'm asking again. What happened? Developers of couple distributions made > some mistake? How should I know? It is even working correctly here AFAICT. (I don't have Thunderbird or Opera installed) It is a fact that this is the only change in ksmserver between 5.5.5 and 5.6 though. > > Or, as it might be a ksmserver bug after all (I never ruled that out), file > > a *new* bug report here. > Some time ago (May 4th) I just reported it here in bug: 362671 (no answer). Ok. But then I see even less reason to reopen this bug report, or even having this discussion here... -- You are receiving this mail because: You are watching all bug changes.
