…except that –debug isn’t one of gsd-color’s recognized parameters, and adding it caused GDM to fail to start. I changed it to –verbose which is available.

10.08 jani@saegusa:~$ LC_ALL=C /usr/lib/gnome-settings-daemon/gsd-color –help-all
Usage:
gsd-color [OPTION?] color

Help Options:
-h, –help Show help options
–help-all Show all help options
–help-gtk Show GTK+ Options

GTK+ Options
–class=CLASS Program class as used by the window manager
–name=NAME Program name as used by the window manager
–gdk-debug=FLAGS GDK debugging flags to set
–gdk-no-debug=FLAGS GDK debugging flags to unset
–gtk-module=MODULES Load additional GTK+ modules
–g-fatal-warnings Make all warnings fatal
–gtk-debug=FLAGS GTK+ debugging flags to set
–gtk-no-debug=FLAGS GTK+ debugging flags to unset

Application Options:
–exit-time Exit after n seconds time
–dummy-name Name when using the dummy daemon
-v, –verbose Verbose
–display=DISPLAY X display to use

There’s also –gtk-debug and –gdk-debug, but I could do with a little help for what to select for those, as just giving ’all’ seems a little counterproductive (for instance, ”touchscreen” and ”interactive” don’t seem very useful here).

10.03 jani@saegusa:~$ LC_ALL=C /usr/lib/gnome-settings-daemon/gsd-color –gtk-debug help
Supported debug values: misc plugsocket text tree updates keybindings multihead modules geometry icontheme printing builder size-request no-css-cache baselines pixel-cache no-pixel-cache interactive touchscreen actions resize layout all help

10.04 jani@saegusa:~$ LC_ALL=C /usr/lib/gnome-settings-daemon/gsd-color –gdk-debug help
Supported debug values: events misc dnd xim nograbs input cursor multihead xinerama draw eventloop frames settings opengl all help
Error parsing option –gdk-debug

(Looks like –gdk-debug reports an error with ’help’, despite having just listed it as one of the options.)

Sure, I’ve now set it up thus and will post the results once I hit this again (it seems pretty unpredictable, so could be anything from days to weeks).

Here you go. Had this occur twice in a row just now, so my original theory of the second login never failing was false.

A hint for anyone googling for this: make sure that the value of gcmSenderId in your AndroidManifest.xmlstill has the trailing backslash (\) after you’ve changed the project number to your own. Otherwise the value gets interpreted as int, when it should be a string, which is one way to trigger the issue of devices being assigned empty/null IDs.

(The building instructions do have this documented in bold, but it’s still very easy to miss, I know I did. Complicating the debugging was my Nexus 5X, which kept getting a proper device ID even when all other devices using the same broken build failed to.)

@juliushaertl Is backporting for docs also done as per the manual? It seems pretty straightforward, so I think I can produce those pull requests too, if needed.

Fix indefinite article before Nextcloud (an->a)

Sorry for necromancing, but because Google seems to like this thread, I just thought I’d add a mention that a check_data_directory_permissions option for config.php has since been implemented (and backported down to NC12), and it allows the admin to turn off the automatic permissions reset (by assigning it false).

This is discouraged however, and ACLs or other means of managing access are still the better option where available.

Thanks for the explanation, I was looking at North Europe thinking: Norway, what’s wrong with you?!

Fix typo: ”parameters and are” -> ”parameters are” in Using the occ command.