Just for reference, ============================== this is the service file we produce now: ==============================
[Unit] Description=Simple Desktop Display Manager Documentation=man:sddm(1) man:sddm.conf(5) [email protected] After=systemd-user-sessions.service systemd-logind.service [Service] ExecStart=/usr/bin/sddm Restart=always StandardOutput=syslog BusName=org.freedesktop.DisplayManager [Install] Alias=display-manager.service ==================================== This is the service file as shipped on Archlinux: ==================================== [Unit] Description=Simple Desktop Display Manager Documentation=man:sddm(1) man:sddm.conf(5) [email protected] After=systemd-user-sessions.service [email protected] plymouth-quit.service [Service] ExecStart=/usr/bin/sddm Restart=always [Install] Alias=display-manager.service ==================================== So what is the real difference and let's assume the archlinux file is correct. On Arch they also instruct to start after [email protected] (I would assume setting it in Conflicts would be enough) On Arch they don't define BusName=org.freedesktop.DisplayManager. On Arch doesn't define StandardOutput. (souldn't affect our issue) On Tue, Mar 20, 2018 at 1:34 AM, Jerrod Frost <[email protected]> wrote: > 3.19.18 KDE-dev.iso > > SDDM is still broken, now in a different way. SDDM doesn't even load on > the live disc on its own anymore. > > dubs.socket: Failed at step EXEC Spawning /bin/systemctl: no such file or > directory > > what did we change to get this ^^^^^ > > after starting sddm manually, everything appears fine, but thats because > the race condition didn't occur. > and konqueror also appears to be having some strange delimma. open it up, > do a google search, press alt+F4 to close. try to open it back up... NOPE. > konqueror[3534]: QCommandLineParser: argument list cannot be empty, it > should contain at least the executable name > > to test again, I performed a > killall konqueror > konqueror > ALT+F4 > (terminal still in use. Konqueror is still running!) > Open another terminal > konqueror > _IceTransSocketUNIXConnect: Cannot connect to non-local host sabayon > _IceTransSocketUNIXConnect: Cannot connect to non-local host sabayon > Qt: Session management error: Could not open network socket > > OK, what happens if I just close using the window's close button? > Same thing. > What about Ctrl+Q (quit) > Same issue! > What if its not on the live disc? > Same issue. > > My advice? ditch konqueror. its a terrible web browser. we have chrome, > then we also have firefox available for install. Nothing I can think of > depends on konqueror. > > On Sun, Mar 11, 2018 at 8:18 PM Jerrod Frost <[email protected]> > wrote: > >> Its testing in general. I don't just test raw images. If I see something >> or hear complaints, I test, and file a bug where necessary. >> the idea is we're testing all the time, not just ISOs during/after freeze. >> >> On Sun, Mar 11, 2018 at 5:56 PM Joost Ruis <[email protected]> >> wrote: >> >>> obs-studio isn't part of the images. What does this have to do with >>> image testing? >>> >>> On Sun, Mar 11, 2018 at 8:11 PM, Jerrod Frost <[email protected] >>> > wrote: >>> >>>> As far as testing is concerned, we're testing all the time. We aren't >>>> just testing right before or after a freeze :) . I get complaints in single >>>> packages every once in a while and we submit bugzilla bugs to have packages >>>> recompiled or upgraded. I've even gone upstream before if the ebuild needed >>>> updated in portage. obs-studio is a perfect example of this. I requested >>>> the 21.0.2 revision which fixed some audio glitches and tested just bumping >>>> the build name/version of the ebuild which worked fine. when confirmed they >>>> quickly updated the ebuild in portage and we now have 21.0.2 in entropy as >>>> well :) >>>> >>>> Right now, we're fairly stable. I wanted to perform a final test after >>>> freeze to confirm there were no strange quirks that were missed that would >>>> impact the OOTB experience for our users. we have 1 week (till end of day >>>> saturday) to finish testing and stamp approval for release. >>>> >>>> On Fri, Mar 9, 2018 at 5:36 PM Sławomir Nizio < >>>> [email protected]> wrote: >>>> >>>>> Hello, >>>>> >>>>> We've been talking on IRC and figured some pieces of information >>>>> related >>>>> to actions might be missing here. >>>>> >>>>> So, >>>>> >>>>> Has testing started? >>>>> How is it going? >>>>> Is it blocked by anything? Are you waiting for someone or something? >>>>> Is there help needed with something? >>>>> >>>>> Let's have it straight where's the ball now. :) >>>>> >>>>> >>>> >>>> >>>> >>> >>> > > >
