| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
| |
Add missing xml files to EXTRA_DIST and generated glue code to
BUILT_SOURCES.
Update .gitignore
|
| |
|
|
| |
One less dependency on dbus-glib, yay!
|
| |
|
|
|
|
| |
The daemon, all the slaves, and the session worker now all use
GDBus exclusively. Dependency on dbus-glib has been removed (but
it still remains for the GUI parts)
|
| |
|
|
|
|
|
| |
Use GDBusObjectManager for enumerating displays, instead of a
hand rolled interface. Also, fix a bunch of present but not
really working interfaces (slave bus name watching, display-added/removed
signals).
|
| |
|
|
|
|
| |
This is one big commit because it uses generated code both
in the daemon and in the slaves, so we need to port both at the
same time.
|
| |
|
|
|
| |
This finishes the GdmSession saga, bringing the last piece inline
with the generated code.
|
| |
|
|
| |
Port the CkConnector glue code to GLib and GDBus.
|
| |
|
|
|
|
|
|
| |
Since GdmSessionRelay went away in the last commit, GdmSessionDirect
was the only class implementing GdmSession, so drop the latter
and use GdmSessionDirect in slaves. Also, refactor it to be an
abstraction over multiple GdmDBusSession instead of talking DBus
directly.
|
| |
|
|
|
|
|
| |
This commit tackles the GdmSession infrastructure, trying to
consolidate GdmSessionRelay and GdmSessionDirect into a single
generated class. For now, only GdmSessionRelay is ported (both
client and server side).
|
| |
|
|
|
|
| |
Last commit moved global bus initialization to load GDBus instead
of dbus-glib, so this commit ensures that all traffic on the system
bus is through GIO.
|
| |
|
|
|
| |
Second patch in the GDBus port, this is primarily to ensure that
everything the slave expose is in what GIO thinks is the system bus.
|
| |
|
|
|
|
| |
GdmGreeterServer will soon see changes to allow multiple interactions
at the same time, and porting to GDBus is a prerequisite to make
the code more manageable.
|
| |
|
|
|
|
| |
Consolidates code that creates a private DBus server, including
generating the address and setting up the required message bus
interfaces.
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
If we've already logged a user in and the greeter is shutting
down and crashes, we should just ignore the crash and continue
with the login process.
Furthermore, if we explicitly kill the greeter and it dies with
a TERM signal, then we definitely don't want to treat that as
if it was a crash.
|
| |
|
|
|
|
|
|
| |
It's important we don't tell the slave the session
has exited/died until we've called pam_close_session,
since the slave reacts to it going way by closing
down which can result in the worker getting explicitly
killed before pam_close_session gets called.
|
| |
|
|
|
|
|
|
|
|
|
| |
This commit adds optional support for interacting with plymouth
from gdm. This feature can be enabled by passing --with-plymouth
to configure.
Hopefully, this will enable the various distributions that use
plymouth to drop a patch.
https://bugzilla.gnome.org/show_bug.cgi?id=572173
|
| |
|
|
|
|
|
|
|
| |
This combined with starting the X server with
-background none will give distros using plymouth
(or potentially other boot splashes) a a nice fade
transition when g-s-d starts.
https://bugzilla.gnome.org/show_bug.cgi?id=572173
|
| | |
|
| |
|
|
|
|
| |
Hopefully helpful later for whoever has to debug PAM stuff.
https://bugzilla.gnome.org/show_bug.cgi?id=675086
|
| |
|
|
|
|
|
|
| |
Documentation is the source of many cyclical build loops. Allow
operating system builders to manually cut these cycles initially
by disabling documentation for a first pass.
https://bugzilla.gnome.org/show_bug.cgi?id=675087
|
| |
|
|
|
|
|
|
| |
To make the restart and shutdown menu items in the fallback greeter
function without ConsoleKit, try the org.freedesktop.login1.Manager dbus
interface first. If that interface is not present, fallback to ConsoleKit.
Fixes: https://bugzilla.gnome.org/show_bug.cgi?id=674741
|
| |
|
|
|
|
|
|
|
|
| |
This moves the default location of the session log message
from ~/.xsession-errors to $XDG_CACHE_HOME/gdm/session.log
This helps avoid clobbering the log files when shared
home directories are used.
See https://live.gnome.org/GnomeGoals/XDGConfigFolders
https://bugzilla.gnome.org/show_bug.cgi?id=627939
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's convenient for debugging to "kill -9 $(pidof gdm-binary)", but
presently because gdm runs bare "dbus-launch" which daemonizes,
it won't terminate when the parent gdm welcome session dies.
Other approaches here are to use "dbus-daemon --no-fork --print-address"
directly as a child, and kill it when we get SIGTERM:
http://git.gnome.org/browse/at-spi2-core/tree/bus/at-spi-bus-launcher.c
But this is an easy fix, even if it only works on Linux.
https://bugzilla.gnome.org/show_bug.cgi?id=674982
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
| |
Two commits ago I tried to only strip the consolekit prefix
off the seat id, if there was a consolekit prefix. In the
process I ended up not initializing the seat_id variable
at all in the non-consolekit case.
This commit fixes that.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Now that we can use logind for seat management, there' may not be
/org/freedesktop/ConsoleKit in the seat identifier.
This commit makes sure the GDM_SEAT_ID environment variable doesn't
show gunk in the case the system is configured to use logind.
Related downstream report:
http://bugzilla.redhat.com/810451
|
| |
|
|
|
|
|
| |
Don't show locked (disabled) users in the user list.
Those are either "system" users with a shell or users disabled in the control-center.
See: https://bugzilla.gnome.org/show_bug.cgi?id=673784
|
| |
|
|
|
|
|
| |
Make session unlock work with logind.
This was not working due to a typo in the bus name of logind.
https://bugzilla.gnome.org/show_bug.cgi?id=673620
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
commit b436764275e513ebd5b0546ab047f0fce2f34b35 tried to prevent
autologin on auxilliary seats, but the conditional was inverted,
so it ended up preventing autologin on the primary seat.
https://bugzilla.gnome.org/show_bug.cgi?id=672773
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
